HP XP RAID Manager Software User Manual

HP XP P9000 Performance for Open and Mainframe Systems User Guide

Abstract

This guide describes and provides instructions for using HP XP P9000 Performance Monitor Software and HP XP P9000 Cache Residency Manager Software to configure and perform operations on HP XP P9000 disk arrays. The intended audience is a storage system administrator or authorized service provider with independent knowledge of HP XP P9000 disk arrays and the HP StorageWorks Remote Web Console.

HP Part Number: AV400-96634 Published: March 2014 Edition: Twelfth

Confidential computer software. Valid license from HP required for possession, use or copying. Consistent with FAR 12.211 and 12.212, Commercial Computer Software, Computer Software Documentation, and Technical Data for Commercial Items are licensed to the U.S. Government under vendor's standard commercial license.

The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein.

Acknowledgements

Microsoft®, Windows®, Windows® XP, and Windows NT® are U.S. registered trademarks of Microsoft Corporation.

UNIX® is a registered trademark of The Open Group.

Export Requirements

You may not export or re-export this document or any copy or adaptation in violation of export laws or regulations.

Without limiting the foregoing, this document may not be exported, re-exported, transferred or downloaded to or within (or to a national resident of) countries under U.S. economic embargo, including Cuba, Iran, North Korea, Sudan, and Syria. This list is subject to change.

This document may not be exported, re-exported, transferred, or downloaded to persons or entities listed on the U.S. Department of Commerce Denied Persons List, Entity List of proliferation concern or on any U.S. Treasury Department Designated Nationals exclusion list, or to parties directly or indirectly involved in the development or production of nuclear, chemical, biological weapons, or in missile technology programs as specified in the U.S. Export Administration Regulations (15 CFR 744).

Revision History

DescriptionDateEdition

Applies to microcode version 70-01-01-00/00 or later.October 2010First

Applies to microcode version 70-01-24-00/00 or later.November 2010Second

Applies to microcode version 70-01-62-00/00 or later.January 2011Third

Applies to microcode version 70-02-01-00/00 or later.May 2011Fourth

Applies to microcode version 70-02-5x-00/00 or later.September 2011Fifth

Applies to microcode version 70-03-00-00/00 or later.November 2011Sixth

Applies to microcode version 70-03-30-00/00 or later.April 2012Seventh

Applies to microcode version 70-04-00-00/00 or later.August 2012Eighth

Applies to microcode version 70-04-00-00/00 or later.November 2012Ninth

Applies to microcode version 70-06-04-00/00 or later.July 2013Tenth

Applies to microcode version 70-06-11-00/00 or later.January 2014Eleventh

Applies to microcode version 70-06-15-00/00 or later.March 2014Twelfth

1 Performance overview.................................................................................8

Performance Monitor overview...................................................................................................8

Performance Control overview....................................................................................................8

Performance of high-priority hosts...........................................................................................8

Upper-limit control................................................................................................................9

Threshold control.................................................................................................................9

Cache Residency overview.........................................................................................................9

Prestaging data in cache......................................................................................................9

Priority mode (read data only).............................................................................................10

Bind mode (read and write data).........................................................................................11

2 About performance monitoring...................................................................13

Monitoring storage system resources.........................................................................................13

Placing data in cache.............................................................................................................13

Prestaging specific data...........................................................................................................13

Priority mode.........................................................................................................................14

Bind mode.............................................................................................................................14

3 Interoperability of Performance Monitor and other products............................17

Cautions and restrictions for monitoring.....................................................................................17

Cautions and restrictions for usage statistics...............................................................................18

Using Performance Control.......................................................................................................18

WWN monitoring data...........................................................................................................20

Microprogram replacement......................................................................................................20

Changing the time setting of SVP..............................................................................................21

Using Performance Control.......................................................................................................21

4 Monitoring WWNs..................................................................................23

Viewing the WWNs that are being monitored............................................................................23

Adding new WWNs to monitor...............................................................................................23

Removing WWNs to monitor...................................................................................................23

Adding WWNs to ports..........................................................................................................23

Editing the WWN nickname....................................................................................................24

Connecting WWNs to ports....................................................................................................24

Deleting unused WWNs from monitoring targets........................................................................25

5 Monitoring CUs.......................................................................................26

Displaying CUs to monitor.......................................................................................................26

Adding or removing CUs to monitor..........................................................................................26

Confirming the status of CUs to monitor ....................................................................................26

6 Monitoring operation................................................................................28

Performing monitoring operations.............................................................................................28

Starting monitoring.................................................................................................................28

Stopping monitoring...............................................................................................................28

7 Setting statistical storage ranges.................................................................30

About statistical storage ranges................................................................................................30

Viewing statistics................................................................................................................30

Setting the storing period of statistics.........................................................................................30

8 Working with graphs................................................................................32

Basic operation .....................................................................................................................32

Objects that can be displayed in graphs ...................................................................................33

Usage rates of MPs ................................................................................................................34

Contents 3

Usage rate of a data recovery and reconstruction processor.........................................................35

Usage rate of cache memory...................................................................................................35

Write pending statistics...........................................................................................................35

Access paths usage statistics....................................................................................................36

Throughput of storage system...................................................................................................37

Size of data transferred...........................................................................................................38

Response times.......................................................................................................................39

Cache hit rates.......................................................................................................................40

Back-end performance.............................................................................................................42

Hard disk drive usage statistics.................................................................................................42

Hard disk drive access rates.....................................................................................................43

Business Copy usage statistics..................................................................................................43

Detailed information of resources on top 20 usage rates..............................................................44

9 Changing display of graphs......................................................................45

Graph operation....................................................................................................................45

Changing displayed items.......................................................................................................45

Changing a display period......................................................................................................45

Adding a new graph .............................................................................................................45

Deleting graph panel..............................................................................................................46

10 Performance Control operations................................................................47

Overview of Performance Control operations..............................................................................47

If one-to-one connections link HBAs and ports............................................................................47

If many-to-many connections link HBAs and ports........................................................................49

Port tab operations..................................................................................................................53

Analyzing traffic statistics....................................................................................................54

Setting priority for ports on the storage system.......................................................................54

Setting upper-limit values to traffic at non-prioritized ports........................................................55

Setting a threshold ............................................................................................................56

WWN tab operations.............................................................................................................57

Monitoring all traffic between HBAs and ports.......................................................................57

Excluding traffic between a host bus adapter and a port from the monitoring target...............59

Analyzing traffic statistics....................................................................................................59

Setting priority for host bus adapters....................................................................................60

Setting upper-limit values for non-prioritized WWNs...............................................................61

Setting a threshold.............................................................................................................62

Changing the PFC name of a host bus adapter......................................................................63

Registering a replacement host bus adapter...........................................................................64

Grouping host bus adapters................................................................................................65

Containing multiple HBAs in an PFC group.......................................................................65

Deleting an HBA from an PFC group...............................................................................65

Switching priority of an PFC group..................................................................................66

Setting an upper-limit value to HBAs in an PFC group........................................................66

Renaming an PFC group................................................................................................67

Deleting an PFC group...................................................................................................68

11 Estimating cache size...............................................................................69

About cache size....................................................................................................................69

Calculating cache size for open systems....................................................................................70

Calculating cache size for mainframe systems.............................................................................73

Cache Residency cache areas..................................................................................................73

Cache Residency system specifications......................................................................................74

12 Managing resident cache........................................................................75

Cache Residency rules, restrictions, and guidelines......................................................................75

Launching Cache Residency.....................................................................................................77

4 Contents

Viewing Cache Residency information.......................................................................................77

Placing specific data into Cache Residency cache.......................................................................77

Placing LDEVs into Cache Residency cache................................................................................79

Releasing specific data from Cache Residency cache..................................................................80

Releasing LDEVs from Cache Residency cache............................................................................80

Changing mode after Cache Residency is registered in cache......................................................81

13 Troubleshooting......................................................................................82

Troubleshooting resources........................................................................................................82

14 Support and other resources.....................................................................83

Contacting HP........................................................................................................................83

Subscription service............................................................................................................83

Documentation feedback....................................................................................................83

Related information.................................................................................................................83

HP websites......................................................................................................................83

Conventions for storage capacity values....................................................................................84

Typographic conventions.........................................................................................................84

A Export Tool..............................................................................................86

About the Export Tool..............................................................................................................86

Installing the Export Tool..........................................................................................................86

System requirements...........................................................................................................86

Installing the Export Tool on a Windows system.....................................................................87

Installing the Export Tool on a UNIX system...........................................................................87

Using the Export Tool..............................................................................................................88

Preparing a command file...................................................................................................88

Preparing a batch file.........................................................................................................91

Running the Export Tool......................................................................................................92

File formats...................................................................................................................92

Processing time.............................................................................................................93

Termination code..........................................................................................................93

Log files.......................................................................................................................94

Error handling..............................................................................................................94

Export Tool command reference................................................................................................95

Export Tool command syntax...............................................................................................95

Conventions.................................................................................................................95

Syntax descriptions........................................................................................................95

Writing a script in the command file................................................................................96

Viewing the online Help for subcommands.......................................................................96

Subcommand list...............................................................................................................96

svpip................................................................................................................................96

retry.................................................................................................................................97

login................................................................................................................................98

show................................................................................................................................98

group...............................................................................................................................99

Short-range.....................................................................................................................109

long-range......................................................................................................................111

outpath...........................................................................................................................113

option............................................................................................................................113

apply.............................................................................................................................114

set.................................................................................................................................114

help...............................................................................................................................115

Java...............................................................................................................................116

Exported files.......................................................................................................................117

Monitoring data exported by the Export Tool.......................................................................117

Contents 5

Resource usage and write-pending rate statistics..................................................................118

Parity groups, external volume groups, or V-VOL groups statistics............................................121

Volumes in parity/external volume groups or V-VOL groups statistics.......................................122

Volumes in parity groups, external volume groups, or V-VOL groups (at volumes controlled by a

particular CU).................................................................................................................123

Port statistics....................................................................................................................124

Host bus adapters connected to ports statistics.....................................................................125

Volumes (LU) statistics.......................................................................................................125

All host bus adapters connected to ports.............................................................................126

MP blades......................................................................................................................126

Remote copy operations by Cnt Ac-S/Cnt Ac-S Z (whole volumes)...........................................127

Remote copy operations by Cnt Ac-S and Cnt Ac-S Z (for each volume (LU)).............................127

Remote copy by Cnt Ac-S and Cnt Ac-S Z (volumes controlled by a particular CU)....................128

Remote copy by Cnt Ac-J and Cnt Ac-J Z (whole volumes)......................................................129

Remote copy by Cnt Ac-J and Cnt Ac-J Z (at journals)............................................................130

Remote copy by Cnt Ac-J and Cnt Ac-J Z (for each volume (LU))..............................................131

Remote copy by Cnt Ac-J and Cnt Ac-J Z (at volumes controlled by a particular CU)..................131

Causes of Invalid Monitoring Data..........................................................................................132

Troubleshooting the Export Tool..............................................................................................133

Messages issued by the Export tool....................................................................................134

B Performance Monitor GUI reference .........................................................137

Performance Monitor main window.........................................................................................137

Edit Monitoring Switch wizard................................................................................................139

Edit Monitoring Switch window..........................................................................................139

Confirm window..............................................................................................................140

Monitor Performance window.................................................................................................140

Edit CU Monitor Mode wizard...............................................................................................148

Edit CU Monitor Mode window.........................................................................................148

Confirm window..............................................................................................................150

View CU Matrix window.......................................................................................................152

Select by Parity Groups window.............................................................................................153

Parity Group Properties window..............................................................................................154

Edit WWN wizard...............................................................................................................155

Edit WWN window.........................................................................................................155

Confirm window..............................................................................................................155

Edit WWN Monitor Mode wizard..........................................................................................156

Edit WWN Monitor Mode window....................................................................................156

Confirm window..............................................................................................................158

Delete Unused WWNs window..............................................................................................160

Add New Monitored WWNs wizard......................................................................................160

Add New Monitored WWNs window................................................................................160

Confirm window..............................................................................................................162

Add to Ports wizard..............................................................................................................163

Add to Ports window........................................................................................................163

Confirm window..............................................................................................................164

Monitor window...................................................................................................................165

MP Properties window...........................................................................................................167

Edit Time Range window.......................................................................................................168

Edit Performance Objects window...........................................................................................169

Add Graph window..............................................................................................................177

Wizard buttons....................................................................................................................185

Navigation buttons...............................................................................................................185

C Performance Control GUI reference...........................................................187

Performance Control window..................................................................................................187

6 Contents

Port tab of the Performance Control main window.....................................................................188

WWN tab of the Performance Control main window................................................................190

D Cache Residency GUI reference...............................................................195

Cache Residency window......................................................................................................195

Multi Set dialog box ............................................................................................................199

Multi Release dialog box ......................................................................................................200

Glossary..................................................................................................202

Index.......................................................................................................204

Contents 7

1 Performance overview

Unless specified otherwise, the term HP XP P9000 refers to the HP XP P9500 Disk Array. This chapter provides an overview of the Remote Web Console software products that enable you

to monitor and manage the performance of the HP XP P9500 storage system.

Performance Monitor overview

Performance Monitor enables you to monitor your HP XP P9500 storage system and collect detailed usage and performance statistics. You can view the storage system data on graphs to identify changes in usage rates, workloads, and traffic, analyze trends in disk I/O, and detect peak I/O times. If there is a decrease in storage system performance (for example, delayed host response times), Performance Monitor can help you detect the cause of the problem and resolve it.

Performance Monitor provides data about storage system resources such as drives, volumes, and microprocessors as well as statistics about front-end (host I/O) and back-end (disk I/O) workloads. Using the Performance Monitor data you can configure Performance Control, Cache Residency, and Cache Partition operations to manage and fine-tune the performance of your storage system.

NOTE:

• To correctly display the performance statistics of a parity group, all volumes belonging to the

parity group must be specified as monitoring targets.

• To correctly display the performance statistics of a LUSE volume, all volumes making up the

LUSE volume must be specified as monitoring targets.

• The volumes to be monitored by Performance Monitor are specified by control unit (CU). If

the range of used CUs does not match the range of CUs monitored by Performance Monitor, performance statistics may not be collected for some volumes.

Performance Control overview

Performance Control allows you to designate prioritized ports (for example, for production servers) and non-prioritized ports (for example, for development servers) and set upper limits and thresholds for the I/O activity of these ports to prevent low-priority activities from negatively impacting high-prority activities. Performance Control operations can be performed only for ports connected to open-systems hosts.

Performance of high-priority hosts

In a storage area network (SAN) environment, the storage system is usually connected with many host servers. Some types of host servers often require higher performance than others. For example, production servers such as database and application servers that are used to perform daily tasks of business organizations usually require high performance. If production servers experience decreased performance, productivity in business activities can be negatively impacted. To prevent this from happening, the system administrator needs to maintain the performance of production servers at a relatively high level.

Computer systems in business organizations often include development servers, which are used for developing, testing, and debugging business applications, as well as production servers. If development servers experience decreased performance, development activities can be negatively impacted, but a drop in development server performance does not have as much negative impact to the entire organization as a drop in production server performance. In this case, you can use Performance Control to give higher priority to I/O activity from production servers than I/O activity from development servers to manage and control the impact of development activities.

8 Performance overview

Upper-limit control

Using Performance Control you can limit the number of I/O requests from servers to the storage system as well as the amount of data that can be transferred between the servers and the storage system to maintain production server performance at the required levels. This practice of limiting the performance of low-priority host servers is called upper-limit control.

Threshold control

While upper-limit control can help production servers to perform at higher levels during periods of heavy use, it may not be useful when production servers are not busy. For example, if the I/O activity from production servers is high between 09:00 and 15:00 hours and decreases significantly after 15:00, upper-limit control for development servers may not be required after 15:00.

To address this situation Performance Control provides threshold control, which automatically disables upper-limit control when I/O traffic between production servers and the storage system decreases to a user-specified level. This user-specified level at which upper-limit control is disabled is called the threshold. You can specify the threshold as an I/O rate (number of I/Os per second) and a data transfer rate (amount of data transferred per second).

For example, if you set a threshold of 500 I/Os per second to the storage system, the upper-limit controls for development servers are disabled when the I/O rate of the production servers drops below 500 I/Os per second. If the I/O rate of the production servers goes up and exceeds 500 I/Os per second, upper-limit control is restored to the development servers.

If you also set a threshold of 20 MB per second to the storage system, the data transfer rate limit for the development servers is not reached when the amount of data transferred between the storage system and the production servers is less than 20 MB per second.

Cache Residency overview

Cache Residency enables you to store frequently accessed data in the storage system's cache memory so that it is immediately available to hosts. Using Cache Residency you can increase the data access speed for specific data by enabling read and write I/Os to be performed at the higher front-end access speeds. You can use Cache Residency for both open-systems and mainframe data.

When Cache Residency is used, total storage system cache capacity must be increased to avoid data access performance degradation for non-cache-resident data. The maximum allowable Cache Residency cache area is configured when the cache is installed, so you must plan carefully for Cache Residency operations and work with your HP representative to calculate the required amount of cache memory for your configuration and requirements.

Cache Residency provides the following functions:

• Prestaging data in cache

• Priority cache mode

• Bind cache mode

Once data has been placed in cache, the cache mode cannot be changed without cache extension. If you need to change the cache mode without cache extension, you must release the data from cache, and then place the data back in cache with the desired mode.

Prestaging data in cache

Using Cache Residency you can place specific data into user-defined Cache Residency cache areas, also called cache extents, before it is accessed by the host. This is called prestaging data in cache. When prestaging is used, the host locates the prestaged data in the Cache Residency cache during the first access, thereby improving data access performance. Prestaging can be used for both priority mode and bind mode operations.

Cache Residency overview 9

Prestaging occurs under any of the following circumstances:

• When prestaging is performed using Cache Residency.

• When the storage system is powered on.

• When cache maintenance is performed.

Figure 1 Cache Residency cache area

NOTE:

• If the Cache Residency cache area is accessed for I/O before the prestaging operation is

complete, the data may not be available in cache at the first I/O access.

• To prevent slow response times for host I/Os, the storage system may interrupt the prestaging

operation when the cache load is heavy.

• Do not use the prestaging function if you specify the Cache Residency setting on a volume

during the quick formatting operation. To use the prestaging function after the quick formatting operation completes, first release the Cache Residency setting and then specify the setting again with the prestaging setting enabled. For information about quick formatting, see the HP

XP P9000 Provisioning for Open Systems User Guide or HP XP P9000 Provisioning for Mainframe Systems User Guide.

• When external volumes are configured in the storage system, you need to disconnect the

external storage system before powering off the storage system. If you power off the storage system without performing the disconnect external storage system operation and then turn on the power supply again, the prestaging process is aborted. If the prestaging process is aborted, you need to perform the prestaging operation again.

The prestaging process is aborted if a volume is created, deleted, or restored during the prestaging operation. If the prestaging process is aborted, you need to perform the prestaging operation again after the create, delete, or restore volume operation is complete.

Priority mode (read data only)

In priority mode the Cache Residency extents are used to hold read data for specific extents on volumes. Write data is write duplexed in cache other than Cache Residency cache, and the data is destaged to the drives when disk utilization is low.

The required total cache capacity for priority mode (normal mode) is: standard cache + Cache Residency cache + additional cache The next table specifies the standard cache capacity requirements for priority mode operations.

Meeting these requirements is important for preventing performance degradation. For more information about calculating cache size for priority mode, see “Estimating cache size” (page 69).

10 Performance overview

Table 1 Cache capacity requirements for CR priority mode

capacity is 128 GB or less

capacity exceeds 128 GB

1 GB = 1,073,741,824 bytes

Bind mode (read and write data)

In bind mode the Cache Residency extents are used to hold read and write data for specific extents on volumes. Data written to the Cache Residency bind area is not destaged to the drives. To ensure data integrity, write data is duplexed in the Cache Residency cache area, which consumes a significant amount of the Cache Residency cache.

Bind mode provides the following advantages over priority mode:

• The accessibility of read data is the same as Cache Residency priority mode.

• Write operations do not have to wait for available cache segments.

• There is no back-end contention caused by destaging data.

The required total cache capacity for bind mode is: standard cache + Cache Residency cache Cache Residency bind data that has write attributes is normally not destaged. However, the data

is destaged to disk in the following cases:

Standard cache capacitySettings of priority mode

16 GBSpecified number of cache areas is 8,192 or less and the specified

32 GBSpecified number of cache areas exceeds 8,192 or the specified

• During cache blockage that is caused by certain maintenance operations (for example, cache

upgrades) or by cache failure.

• When the storage system is powered off.

• When the volume is deleted from Cache Residency bind mode.

The next table specifies the cache requirements for bind mode operations. Meeting these requirements is important for preventing performance degradation. For more information about calculating cache size for bind mode, see “Estimating cache size” (page 69).

Table 2 Bind mode cache requirements

System Type

Open systems

Type

RAID 5 (3390) or RAID 6

RAID 1, or external volumes

Capacity SpecificationsRAID Level or Volume

Slot capacity: 264 KB Cache segment capacity: 16.5 KB Cache segments needed per slot: 48

(slot capacity / cache segment capacity)

Slot capacity: 264 KB Cache segment capacity: 16.5 KB Cache segments needed per slot: 32

(slot capacity / cache segment capacity)

Cache Residency Cache Requirement

3 times the space required for user data: 1 slot = 3 × 264 KB = 792 KB = 48 cache segments

2 times the space required for user data: 1 slot = 2 × 264 KB = 528 KB = 32 cache segments

Mainframe (for example, 3390-3, 3390-9)

RAID 5 mainframe or RAID 6

Slot capacity: 66 KB Cache segment capacity: 16.5 KB

3 times the space required for user data: 1 slot = 3 × 66 KB

Cache Residency overview 11

Table 2 Bind mode cache requirements (continued)

System Type

Type

RAID 1 mainframe, or external volumes

Capacity SpecificationsRAID Level or Volume

Cache segments needed per slot: 12 (slot capacity / cache segment capacity)

Note: Even though a mainframe track is 56 KB, because cache is divided into 16.5 KB segments, it requires 4 segments.

Slot capacity: 66 KB Cache segment capacity: 16.5 KB Cache segments needed per slot: 8

(slot capacity / cache segment capacity)

Cache Residency Cache Requirement

= 198 KB = 12 cache segments

2 times the space required for user data: 1 slot = 2 × 66 KB = 132 KB = 8 cache segments

12 Performance overview

2 About performance monitoring

This topic presents an overview of Performance Monitor and Cache Residency.

Monitoring storage system resources

Performance Monitor enables you to monitor your storage system and collect detailed usage and performance statistics. If there is a decrease in storage system performance (for example, delayed host response times), Performance Monitor can help you detect the cause of the problem and resolve it.

Performance Control uses the data collected by Performance Monitor to identify and resolve bottlenecks of activity on the ports.

Placing data in cache

Cache Residency enables you to place specific data into user-defined Cache Residency cache areas, also called cache extents, so that the data is available to hosts at front-end data transfer speeds. This is called prestaging data in cache.

Prestaging specific data

Cache Residency supports prestaging of data in which specific data is placed into the Cache Residency cache before it is accessed by the host. When prestaging is used, the host locates the prestaged data in the Cache Residency cache during the first access, thereby improving performance.

Figure 2 Cache Residency cache area

CAUTION: Total storage system cache capacity must be increased when using Cache Residency

to avoid data access performance degradation for non-cache-resident data. Cache Residency is available only on HP XP P9500 storage systems configured with at least 512 MB of cache. The HP representative configures the maximum allowable Cache Residency area when the cache is installed.

Both priority mode and bind mode permit prestaging. Prestaging occurs under any of the following circumstances:

• When prestaging is performed from the SVP or from Remote Web Console.

• When the power is turned on.

• When cache maintenance is performed.

Monitoring storage system resources 13

NOTE:

• If you access the Cache Residency area for input and output before the prestaging operation

is performed from the SVP or Remote Web Console, the host may not be able to find data in the cache at the first I/O access after Cache Residency is configured.

• To prevent the slow response time of the host I/O, the prestaging operation may be interrupted

when the cache load is heavy.

• If you specify the Cache Residency setting on the volume during the quick formatting, do not

use the prestaging function. If you want to use the prestaging function after the quick formatting processing completes, first release the setting and then specify the Cache Residency setting again, with the prestaging setting enabled. For information about the quick formatting, see

HP XP P9000 Provisioning for Open Systems User Guide or HP XP P9000 Provisioning for Mainframe Systems User Guide.

• When external volumes are set in the storage system, execute the disconnect external storage

system operation to the external storage system before turning off the power supply of the storage system. If you turn off the power supply of the storage system without executing the disconnect external storage system operation to the external storage system and then turn on the power supply again, the prestaging processing is aborted. If the prestaging processing is aborted, perform the prestaging operation from SVP or Remote Web Console.

The prestaging processing is aborted if a volume is created, deleted, or restored during the prestaging operation. If the prestaging processing is aborted, perform the prestaging operation from SVP or Remote Web Console after finishing a volume create, delete, or restore volume operation.

Priority mode

In priority mode the Cache Residency extents are used to hold read data for specific extents on volumes. Write data is write duplexed in cache other than that of Cache Residency, and the data is destaged to the disk drive when disk utilization is low.

In priority mode (normal mode), the total capacity of cache required is: standard cache + Cache Residency cache + additional cache The following table specifies the standard cache capacity requirements for Cache Residency priority

mode operations. Meeting these requirements is important for preventing performance degradation. For more information about calculating cache size for priority mode, see “Estimating cache size”

(page 69).

Table 3 Cache capacity requirements for CR priority mode

and the specified capacity is 128 GB or less

or the specified capacity exceeds 128 GB

1 GB = 1,073,741,824 bytes

Bind mode

Standard cache capacitySettings of priority mode

16 GBSpecified number of cache areas is 8,192 or less

32 GBSpecified number of cache areas exceeds 8,192

Cache Residency uses extents to hold read and write data for specific extents on volumes. Data written to the Cache Residency bind area is not destaged to the disk. To ensure data integrity, write data is duplexed in the Cache Residency area, which consumes a significant amount of the Cache Residency cache.

14 About performance monitoring

Bind mode has the following advantages:

• The accessibility of read data is the same as Cache Residency priority mode.

• Write operations do not have to wait for available cache segments.

• There is no back-end contention caused by destaging data.

In Bind Mode, the total capacity of cache required is: standard cache + Cache Residency cache For more information about calculating case size for bind mode, see “Estimating cache size”

(page 69).

The next table specifies the cache requirements for bind mode operations:

Table 4 Bind mode cache requirements

System Type

OPEN Systems

Mainframe (for example, 3390-3, 3390-9)

Type

RAID 5 (3390) or RAID 6

RAID 1, or external volumes

RAID 5 mainframe or RAID 6

Capacity SpecificationsRAID Level or Volume

Slot capacity: 264 KB Cache segment capacity: 16.5

KB Cache segments needed per

slot: 48 (slot capacity / cache segment capacity)

Slot capacity: 264 KB Cache segment capacity: 16.5

KB Cache segments needed per

slot: 32 (slot capacity / cache segment capacity)

Slot capacity: 66 KB Cache segment capacity: 16.5

KB Cache segments needed per

slot: 12 (slot capacity / cache segment capacity)

Note: Even though a mainframe track is 56 KB, because cache is divided into 16.5 KB segments, it requires 4 segments.

Cache Residency Cache Requirement

3 times the space required for user data: 1 slot = 3 × 264 KB = 792 KB = 48 cache segments

2 times the space required for user data: 1 slot = 2 × 264 KB = 528 KB = 32 cache segments

3 times the space required for user data: 1 slot = 3 × 66 KB = 198 KB = 12 cache segments

RAID 1 mainframe, or external volumes

Slot capacity: 66 KB Cache segment capacity: 16.5

KB Cache segments needed per

slot: 8 (slot capacity / cache segment capacity)

2 times the space required for user data: 1 slot = 2 × 66 KB = 132 KB = 8 cache segments

Cache Residency bind data that has write attributes is normally not destaged. However, the data is destaged in the following cases:

• During cache blockage that is caused by certain maintenance operations (for example, cache

upgrades) or by cache failure.

• If the storage system is powered off.

• If the volume is deleted from Cache Residency bind mode.

Bind mode 15

Changing the mode without cache extension requires reconfiguring Cache Residency (for example, release the data from cache, and then place the data back in cache with the desired mode).

16 About performance monitoring

3 Interoperability of Performance Monitor and other products

This chapter describes the interoperability of Performance Monitor and other products.

Cautions and restrictions for monitoring

• Storage system maintenance

If the storage system is undergoing the following maintenance operations during monitoring, the monitoring data might contain extremely large values.

◦ Adding, replacing, or removing cache memory.

◦ Adding, replacing, or removing data drives.

◦ Changing the storage system configuration.

◦ Replacing the microprogram.

◦ Formatting or quick formatting logical devices

◦ Adding on, replacing, or removing MP blades

• Storage system power-off

If the storage system is powered off during monitoring, monitoring stops while the storage system is powered off. When the storage system is powered up again, monitoring continues. However, Performance Monitor cannot display information about the period while the storage system is powered off. Therefore, the monitoring data immediately after powering on again might contain extremely large values.

• Microprogram replacement

After the microprogram is replaced, monitoring data is not stored until the service engineer releases the SVP from Modify mode. While the SVP is in Modify mode, inaccurate data is displayed.

• Changing the SVP time setting

If the SVP time setting is changed while the monitoring switch is enabled, the following monitoring errors can occur:

◦ Invalid monitoring data appears.

◦ No monitoring data is collected.

To change the SVP time setting, first disable the monitoring switch, change the SVP time setting, and then re-enable the monitoring switch. After that, obtain the monitoring data. For details about the monitoring switch, see “Starting monitoring” (page 28).

• WWN monitoring

You must configure some settings before the traffic between host bus adapters and storage system ports can be monitored. For details, see “Adding new WWNs to monitor” (page 23),

“Adding WWNs to ports” (page 23), and “Connecting WWNs to ports” (page 24).

Cautions and restrictions for monitoring 17

Cautions and restrictions for usage statistics

• Usage statistics for the last three months (93 days) are displayed in long-range monitoring,

and usage statistics for up to the last 15 days are displayed in short-range monitoring. Usage statistics outside of these ranges are deleted from the storage system.

• In the short range, monitoring results are retained for the last 8 hours to 15 days depending

on the specified gathering interval. If the retention period has passed since a monitoring result was obtained, the previous result has been deleted from the storage system and cannot be displayed.

• When the monitoring switch is set to disabled, no monitoring data is collected. This applies

to both long-range and short-range data.

• For short-range monitoring, if the host I/O workload is high, the storage system gives higher

priority to I/O processing than to monitoring. If this occurs, some monitoring data might be missing. If monitoring data is missing frequently, use the Edit Time Range option to lengthen the collection interval. For details, see “Starting monitoring” (page 28).

• The monitoring data (short-range and long-range) may have a margin of error.

• If the SVP is overloaded, the system might require more time than the gathering interval allows

to update the display of monitoring data. If this occurs, some portion of monitoring data is not displayed. For example, suppose that the gathering interval is 1 minute. In this case, if the display in the Performance Management window is updated at 9:00 and the next update occurs at 9:02, the window (including the graph) does not display the monitoring result for the period of 9:00 to 9:01. This situation can occur when the following maintenance operations are performed:

◦ Adding, replacing, or removing cache memory.

◦ Adding, replacing, or removing data drives.

◦ Changing the storage system configuration.

◦ Replacing the microprogram.

• Pool-VOLs of Fast Snap, Snapshot, Thin Provisioning, and Thin Provisioning Z are not monitored.

NOTE: When you run the RAID Manager horctakeover or pairresync-swaps command for a Cnt

Ac-J pair or the BCM YKRESYNC REVERSE command for a Cnt Ac-J Z pair, the primary and secondary volumes are swapped. You can collect the before-swapped information immediately after you run any of the commands. Incorrect monitoring data will be generated for a short time but will be corrected automatically when the monitoring data gets updated. The incorrect data will also be generated when the volume used for a secondary volume is used as a primary volume after a Cnt Ac-J or Cnt Ac-J Z pair is deleted.

Using Performance Control

• Starting Performance Control. Ensure that the Time Range in the Monitor Performance window

is not set to Use Real Time. You cannot start Performance Control in real-time mode.

• I/O rates and transfer rates. Performance Control runs based on I/O rates and transfer rates

measured by Performance Monitor. Performance Monitor measures I/O rates and transfer rates every second, and calculates the average I/O rate and the average transfer rate for every gathering interval (specified between 1 and 15 minutes) regularly.

Suppose that 1 minute is specified as the gathering interval and the I/O rate at the port 1-A changes as illustrated in Graph 1. When you use Performance Monitor to display the I/O rate graph for 1A, the line in the graph indicates changes in the average I/O rate calculated every minute (refer to Graph 2). If you select the Detail check box in the Performance Monitor

18 Interoperability of Performance Monitor and other products

windows, the graph displays changes in the maximum, average, and minimum I/O rates in one minute.

Performance Control applies upper limits and thresholds to the average I/O rate or the average transfer rate calculated every gathering interval. For example, in the following figures in which the gathering interval is 1 minute, if you set an upper limit of 150 I/Os to the port 1A, the highest data point in the line CL1-A in Graph 2 and the line Ave.(1 min) in Graph 3 is somewhere around 150 I/Os. It is possible that the lines Max (1 min.) and Min (1 min.) in Graph 3 might exceed the upper limit.

Figure 3 Graph 1: actual I/O rate (measured every second)

Figure 4 Graph 2: I/O rate displayed in Performance Monitor (the Detail check box is not selected)

Figure 5 Graph 3: I/O rate displayed in Performance Monitor (the Detail check box is selected)

Using Performance Control 19

NOTE:

• Continuous Access Synchronous: Performance Control monitors write I/O requests issued from

initiator ports of your storage system.

• When the remote copy functions: (Continuous Access Synchronous, Continuous Access

Synchronous Z, Continuous Access Journal, and Continuous Access Journal Z) are used in your environment, Performance Control monitors write I/O requests issued from initiator ports of your storage system.

If you give a priority attribute to the RCU target port, all I/Os received on the port will be controlled as the threshold control and its performance data will be added to the total number of I/Os (or the transfer rate) of all prioritized ports. I/Os on the port will not be limited.

If you give a non priority attribute to the RCU target port, I/O requests from the initiator port will not be controlled as threshold control and I/Os on the port will not be limited. On the other hand, I/O requests from a host will be controlled as the upper limit control and I/Os on the port will be limited.

• Statistics of Initiator/External ports: The initiator ports and external ports of your storage system

are not controlled by Performance Control. Although you can set Prioritize or Non-Prioritize to initiator ports and external ports by using Performance Control, the initiator ports and the external ports become the prioritized ports that are not under threshold control, regardless of whether the setting of the ports are Prioritize or Non-Prioritize. If the port attributes are changed from Initiator/External into Target/RCU Target, the settings by Performance Control take effect instantly and the ports are subject to threshold or upper limit control.

The statistics of the Monitor Performance window are the sum of the statistics on Target/RCU Target ports that are controlled by Performance Control. The statistics do not include the statistics of Initiator/External ports. Because the statistics of Initiator/External ports and Target/RCU Target ports are based on different calculation methods, it is impossible to sum up the statistics of Initiator/External ports and Target/RCU Target ports.

• Settings of Performance Control main window: The Performance Control main window has

two tabs: the Port tab and the WWN tab. The settings on only one tab at a time can be applied to the storage system. If you make settings on both tabs, the settings cannot be applied at the same time. When you select Apply, the settings on the last tab on which you made settings are applied, and the settings on the other tab are discarded.

NOTE:

• Settings for Performance Control from RAID Manager: You cannot operate Performance Control

from RAID Manager and Remote Web Console simultaneously.If you change some settings for Performance Control from RAID Manager, you cannot change those settings from Remote Web Console. If you do, some settings might not appear. Before you change features that use Performance Control, delete all Performance Control settings from the currently used features.

WWN monitoring data

You must configure some settings before you start monitoring traffic between host bus adapters and storage system ports. For details, see topics between “Viewing the WWNs that are being

monitored” (page 23) and “Deleting unused WWNs from monitoring targets” (page 25).

Microprogram replacement

After the microprogram is replaced, the monitoring data is not stored until a service engineer releases the SVP from Modify mode. While the SVP is in Modify mode, inaccurate data is displayed.

20 Interoperability of Performance Monitor and other products

Changing the time setting of SVP

If the monitoring switch function is performed, do not change the time setting of the SVP. If the time setting is changed, the following problems can happen:

• Invalid monitoring data appears.

• No monitoring data is collected.

If the time setting of the SVP is changed, turn off the monitoring switch, and then turn on the switch. After that, obtain the monitoring data. For details about the monitoring switch, see “Starting

monitoring” (page 28).

Using Performance Control

• When starting Performance Control. Ensure that the Time Range in the Monitor Performance

window is not set to Use Real Time. You cannot start Performance Control in real-time mode.

• Performance Control runs based on I/O rates and transfer rates measured by Performance

Monitor. Performance Monitor measures I/O rates and transfer rates every second, and calculates the average I/O rate and the average transfer rate for each gathering interval (specified between 1 and 15 minutes).

Figure 6 Graph 1: actual I/O rate (measured every second)

Figure 7 Graph 2: I/O rate displayed in Performance Monitor (the Detail check box is not selected)

Changing the time setting of SVP 21

Figure 8 Graph 3: I/O rate displayed in Performance Monitor (the Detail check box is selected)

NOTE:

• Continuous Access Synchronous: Performance Control monitors write I/O requests issued from

initiator ports of your storage system.

• When the remote copy functions: (Continuous Access Synchronous, Continuous Access

• Statistics of Initiator/External ports: The initiator ports and external ports of your storage system

The statistics of the Monitor Performance window are sum total of statistics on Target/RCU Target ports that are controlled by Performance Control. The statistics does not include the statistics of Initiator/External ports. Because the statistics of Initiator/External ports and Target/RCU Target ports are based on different calculation methods, it is impossible to sum up the statistics of Initiator/External ports and Target/RCU Target ports.

• Settings of Performance Control main window: The settings in the Port tab or the WWN tab

are valid as the current settings in the Performance Control main window. All port settings in the invalid tab are invalid regardless of the resources that are assigned to you. You can confirm the current status in Current Control Status which appears in the upper right of the Performance Control main window.

NOTE:

• Settings for Performance Control from RAID Manager: You cannot operate Performance Control

22 Interoperability of Performance Monitor and other products

4 Monitoring WWNs

This topic describes how to set up WWNs to be monitored.

Viewing the WWNs that are being monitored

To view the WWNs that are being monitored:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Select the Monitored WWNs tab to see the list of WWNs that are currently being monitored.

Adding new WWNs to monitor

To add new WWNs to monitor:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Select the Monitored WWNs tab.

4. Click Edit WWN Monitor Mode. The Edit WWN Monitor Mode window opens.

5. Select the WWNs in the Unmonitored WWNs list, and click Add.

6. Click Finish to display the Confirm window.

7. Click Apply in the Confirm window to apply the settings to the storage system.

Removing WWNs to monitor

To remove WWNs to monitor:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click the Monitored WWNs tab.

4. Click Edit WWN Monitor Mode. The Edit WWN Monitor Mode window opens.

5. Select the WWNs in the Monitored WWNs list that you want to remove, and click Remove.

6. Click Finish to display the Confirm window.

7. Click Apply in the Confirm window.

8. When the warning message appears, click OK to close the message. The settings are applied to the storage system.

Adding WWNs to ports

If you want to monitor WWNs that are not connected to the storage system, you can add them to ports and set them up for monitoring with Performance Monitor.

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click the Monitored WWNs tab.

Viewing the WWNs that are being monitored 23

4. Click Add New Monitored WWNs. The Add New Monitored WWNs window opens.

5. Specify the following information for each new WWN:

• HBA WWN (required)

Enter the 16-digit hexadecimal number.

• WWN Name (optional)

Enter the unique name to distinguish the host bus adapter from others. The WWN Name must be less than 64 characters and must consist of alphanumeric characters and at least one symbol.

• Port (In Available Ports)

In the Available Ports list select the port connected to the WWN. Ports connected to mainframe hosts are not displayed, because they are not supported for Performance Monitor.

6. Click Add. The added WWN is displayed in Selected WWNs.

7. If you need to remove a WWN from the Selected WWNs list, select the WWN and click Remove.

8. When you are done adding new WWNs, click Finish.

9. Click Apply in the Confirm window to apply the settings to the storage system.

Editing the WWN nickname

To edit the nickname of a WWN being monitored:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click the Monitored WWNs tab to see the list of WWNs being monitored.

4. Select the WWN to edit, and click Edit WWN. You can edit only one WWN at a time. If you select multiple WWNs, an error will occur.

The Edit WWN window opens.

5. Edit the HBA WWN and WWN Name fields as needed.

• HBA WWN

A 16-digit hexadecimal number. The value of HBA WWN must be unique in the storage system.

• WWN Name

The nickname distinguishes the host bus adapter from others. The WWN Name must be less than 64 digits and must consist of alphanumeric characters and at least one symbol.

6. Click Finish to display the Confirm window.

7. Click Apply in the Confirm window to apply the settings to the storage system.

Connecting WWNs to ports

To connect the WWNs to monitor to ports:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click the Monitored WWNs tab.

24 Monitoring WWNs

4. Select the WWN to connect to the port, and click Add to Ports. The Add to Ports window opens. If you select a WWN to connect, select one WWN in the

list. If you select multiple WWNs and click Add to Ports, an error occurs.

5. Select a port to connect in Available Ports, and then click Add. However, the ports of the mainframe system are not displayed in the list because they are not supported for Performance Monitor.

The added WWN and the port are specified for the Selected WWNs.

6. If necessary, select unnecessary row of a WWN and port in Selected WWNs, and then click Remove.

WWNs are deleted.

7. Click Finish to display the Confirm window.

8. Click Apply in the Confirm window to apply the settings to the storage system.

Deleting unused WWNs from monitoring targets

To delete WWNs that are being monitored:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click the Monitored WWNs tab.

4. Click Delete Unused WWNs to display the Confirm window.

5. Click Apply in the Confirm window.

6. When the warning message appears, click OK to close the message. The settings are applied to the storage system.

Deleting unused WWNs from monitoring targets 25

5 Monitoring CUs

This topic describes how to set up CUs to be monitored.

Displaying CUs to monitor

To display the list of CUs to monitor:

1. Open the Remote Web Console main window.

2. Select Performance Monitor in Explorer and select Performance Monitor from the tree.

3. Open the Monitored CUs tab. View the list of CUs.

Adding or removing CUs to monitor

NOTE: When a CU is removed from monitoring, the monitor data for that CU is deleted. If you

want to save the data, export it first using the Export Tool (see “Export Tool” (page 86)), and then remove the CU.

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Open the Monitored CUs tab.

4. Click Edit CU Monitor Mode. The Edit CU Monitor Mode window opens.

5. To add CUs as monitoring target objects, select the CUs in the Unmonitored CUs list, and click Add to move the selected CUs into the Monitored CUs list.

To add all CUs in a parity group as monitoring target objects:

1. Click Select by Parity Groups in the Unmonitored CUs area.

The Select by Parity Groups window opens. The available parity group IDs and number of CUs are displayed.

2. Select the parity group ID from the list and click Detail.

The Parity Group Properties window opens. CUs and the number of LDEVs are displayed.

3. Confirm the properties of the parity group and click Close.

The Select by Parity Groups window opens.

4. Select the parity group to be the monitoring target in the Select by Parity Groups window,

and click OK. The CUs in the parity group are selected in the Unmonitored CUs list.

5. Click Add to move the selected CUs into the Monitored CUs list.

6. To remove CUs as monitoring target objects, select the CUs in the Monitored CUs list, and click Remove to move the selected CUs into the Unmonitored CUs list.

7. When you are done adding and/or deleting CUs, click Finish.

8. When the confirmation dialog box opens, click Apply. If you are removing CUs, a warning message appears asking whether you want to continue

this operation even though monitor data will be deleted.

9. To add and remove the CUs, click OK. The new settings are registered in the system.

Confirming the status of CUs to monitor

To view the monitoring status of CUs:

26 Monitoring CUs

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Open the Monitored CUs tab.

4. Click Edit CU Monitor Mode. The Edit CU Monitor Mode window opens.

5. Click View CU Matrix in the Edit CU Monitor Mode window. The View CU Matrix window opens. The following CUs are displayed in the Monitored CUs

window:

• Monitored CUs

• Set monitored CUs

• Release monitored CUs

6. Click Close. The Edit CU Monitor Mode window opens.

Confirming the status of CUs to monitor 27

6 Monitoring operation

This topic describes how to start and stop the monitoring operation.

Performing monitoring operations

This topic describes how to start or stop the monitoring operation.

• To start the monitoring operation, see “Starting monitoring” (page 28).

• To stop the monitoring operation, see “Stopping monitoring” (page 28).

Starting monitoring

To start monitoring the storage system, start Performance Monitor and open the Edit Monitoring Switch window. If this operation is performed, the monitoring result will be deleted.

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click Edit Monitoring Switch in the Performance Monitor window. The Edit Monitoring Switch window opens.

4. Click Enable in the Monitoring Switch field.

5. Select the collecting interval in the Sample Interval. Specify the interval to obtain usage statistics about the storage system for short range

monitoring. This option is activated when you specify Enable for Current Status. If CUs to be monitored are 64 or less, you can specify the value between 1 and 15 minutes by minutes, and the default setting is 1 minute. For example, if you specify 1 minute for the gathering interval, Performance Monitor collect statistics (for example, I/O rates and transfer rates) every one minute.

If CUs to be monitored are 65 or more, the gathering interval can be specified to the value 5, 10 or 15 minutes (in the 5 minuted interval), and default is 5 minutes. For example, if you specify the gathering interval to 5 minutes, Performance Monitor gathers statistics (for example, I/O rate and transfer rate) every 5 minutes.

6. Click Finish to display the Confirm window.

7. Click Apply in the Confirm window. A warning message appears, asking whether you continue this operation although graph

data is deleted.

8. Click OK to start monitoring. When statistics are collected, a heavy workload is likely to be placed on servers. Therefore,

the client processing might slow down.

Stopping monitoring

To stop monitoring the storage system:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click Edit Monitoring Switch in the Performance Monitor window. The Edit Monitoring Switch window opens.

28 Monitoring operation

4. Click Disable in the Monitoring Switch field. The Sample Interval list is grayed out and becomes ineffective.

5. Click Finish to display the Confirm window.

6. Click Apply in the Confirm window to stop monitoring.

Stopping monitoring 29

7 Setting statistical storage ranges

This topic describes setting statistical storage ranges.

About statistical storage ranges

Performance Monitor collects and stores statistics for two time periods (ranges): short range and long range. The difference between the two ranges and the statistics they target are as follows:

• Short range

If the number of CUs to be monitored is 64 or less, statistics are collected at a user-specified interval between 1 and 15 minutes, and stored between 1 and 15 days.

If the number of CUs to be monitored is 65 or more, statistics are collected at a user-specified intervals of 5, 10, or 15 minutes, and stored for 8 hours, 16 hours, or 1 day, respectively.

• Long range

Statistics are collected at fixed 15-minute (0, 15, 30, and 45 minutes of every hour), and stored for 93 days (for example, 3 months).

Usage statistics about storage system resources are collected and stored in long range, in parallel with in short range. However, some of usage statistics about resources cannot be collected in long range.

Viewing statistics

Use the Monitor Performance window to view statistics within short and long storage ranges. All statistics, except some information related to Auto LUN V2, can be viewed in short range (for the storing period corresponding to the collecting interval setting). In addition, usage statistics about storage system resources can be viewed in both short range and long range because they are monitored in both ranges. When viewing usage statistics about resources, you can specify the range to view and which part of the storing period to depict on lists and graphics.

Setting the storing period of statistics

To set the storing period of statistics:

1. Display the Remote Web Console main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click Monitor Performance in the Performance Monitor window. The Monitor Performance window opens.

4. Select Long-Range or Short-Range in the Data Range as the periods (ranges) for collecting and storing statistics.

30 Setting statistical storage ranges

5. Select Set Range or Use Real Time in the Time Range as the periods (ranges) for displaying

statistics. If Long-Range is selected, you can specify only Set Range. If Short-Range is selected, you can

select Set Range or Use Real Time. Performance Monitor saves the statistics obtained up to 1440 times in SVP. Therefore, you

can estimate the storing period of statistics with “gathering interval multiplied by 1440.” For example, if you specify one minute for the gathering interval, the statistics for one day can be stored at the maximum from the following formula:

1 minute x 1440 = 1440 minutes = 24 hours = 1 day

This storing period is the range of display in the Monitor Performance windows. When you specify one minute for the gathering interval like the example above, Performance Monitor can display the statistics for one day (for example, 24 hours) in the list and graph at the maximum. Also, when you specify 15 minutes for the gathering interval, Performance Monitor can display the statistics for 15 days in the list and graph at the maximum.

Setting the storing period of statistics 31

8 Working with graphs

This topic describes how to display statistical graphs.

Basic operation

Use Monitor Performance window to display graphs.

1. Display the Performance Monitor main window.

2. Select Performance Monitor in Explorer, and select Performance Monitor in the tree. The Performance Monitor window opens.

3. Click Monitor Performance in the Performance Monitor window. The Monitor Performance window opens.

4. Select Long-Range or Short-Range as the storing period of statistics in the Data Range field.

5. Select Set Range or Use Real Time as the displaying period of statistics in the Time Range field. However, Use Real Time can be specified when Short-Range is selected.

Specify items to display graphs in the Performance Objects field.

6. Select items in the Object field. Select items in the left field and then select detailed items in the right field. Detailed item

changes by the items selected in the left field.

7. Select items in the Monitor Data field. Select items in the left field and then select detailed items in the right field.

8. Select the item to display graph in the Performance Object Selection field. Select the object in the Available Objects field.

9. Click Add. The added object is displayed in the Selected Objects field.

10. To delete the unnecessary object, select the object and click Remove.

11. Click Apply. The line graph is displayed on the graph panel in the Monitor window.

• Graphs appear on the left side and notes appear on the right of panels.

• You can change the size of the panel by clicking the icon in the upper right of the panel.

• You can view up to 8 lines in one panel.

• You can view up to 16 graphs across a total of four panels.

• In the graph panel, the unit of scale on vertical axis can be changed.

By using the list on the upper left of the graph panel, adjust the scale to display the maximum value of the graph. If the graph is too big, the display may not be able to displaying properly. For example, the line of the graph is too thick, or the graph panel is painted out in the color of the graph.

• If you locate the mouse cursor to each point of the graph, a detailed value is displayed

with the tool tip.

• When you click the explanatory note on the right of the graph panel, you can display or

hide points on the graph panel. However, If the graph is displayed only the one point on X axis, the graph is always displayed. Therefore you cannot switch the display of the point to non-display by clicking the explanatory note.

• If Time Range is set to Use Real Time and the MP blades are displayed in explanatory

notes on the right in the graph panel, the MP blade names are displayed as text links. If

32 Working with graphs

you click the text link, the resources assigned to an MP blade of top 20 in usage rates are displayed on the detailed window.

12. To close the graph, click Delete Graph.

Objects that can be displayed in graphs

Set items to display graph in the Performance Objects field of the Monitor Performance window. The outline of target objects and monitoring data that can be displayed in graphs is shown in the following table. The monitoring data shows the average value of sampling interval. The sampling intervals are 1 to 15 minutes and 15 minutes for Short Range and Long Range, respectively, that can be set in the Edit Monitoring Switch window.

Monitoring dataMonitoring target object

Usage rates of MPs (%).Controller

Usage rates of DRR (%).

Usage rates of cache (%).Cache

Write pending rates (%).

Usage rates of access path between CHA and ESW (%).Access Path

Usage rates of access path between DKA and ESW (%).

Usage rates of access path between MP Blade and ESW (%).

Usage rates of access path between cache and ESW (%).

Throughput (IOPS).Port

Data transfer (MB/s).

Response time (ms).

Throughput of WWN (IOPS).WWN

Data transfer of WWN (MB/s).

Response time of WWN (ms).

Throughput of port (IOPS).

Data transfer of port (MB/s).

Response time of port (ms).

Total throughput (IOPS).Logical Device

Read throughput (IOPS).

Write throughput (IOPS).

Cache hit (%).

Data transfer (MB/s).

Response time (ms).

Back transfer (count/sec).

Drive usage rate (%). *

Drive access rate (%). *

Usage rates of Business Copy (%). *

Total throughput (IOPS).Parity Group

Read throughput (IOPS).

Objects that can be displayed in graphs 33

Monitoring dataMonitoring target object

Write throughput (IOPS).

Cache hit (%).

Data transfer (MB/s).

Response time (ms).

Back transfer (count/sec).

Drive usage rate (%). *

Total throughput (IOPS).LUN

Read throughput (IOPS).

Write throughput (IOPS).

Cache hit (%).

Data transfer (MB/s).

Response time (ms).

Back transfer (count/sec).

Data transfer of logical devices (MB/s).External Storage

Response time of logical devices (ms).

1. Only information on internal volumes is displayed. Information on external volumes and FICON DM volumes is not

displayed.

2. The same value is output to all LUNs mapped to the LDEV.

Usage rates of MPs

Function

The usage rate of the MP shows the usage rate of an MP assigned to a logical device. If a usage rate of an MP is high, I/Os concentrate to an MP. Examine the distribution of I/Os to other MP blades.

Storing period

Short-Range or Long-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Data transfer of parity groups (MB/s).

Response time of parity groups (ms).

Item on left side of Object field

34 Working with graphs

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneUsage Rate (%)MPController

Usage rate of a data recovery and reconstruction processor

Function

A data recovery and reconstruction processor (DRR) is a microprocessor (located on the DKAs and channel adapters) that is used to generate parity data for RAID 5 or RAID 6 parity groups. The DRR uses the formula "old data + new data + old parity" to generate new parity.

If the monitor data shows high DRR usage overall, this can indicate high write penalty condition. Please consult your HP representative about high write penalty conditions.

Storing period

Short-Range or Long-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Usage rate of cache memory

Function

When you display monitoring results in a short range, the window displays the usage rates about the cache memory for the specified period of time.

Storing period

Short-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneUsage Rate (%)DRRController

Item on right side of Monitor Data field

NoneUsage Rate (%)NoneCache

Write pending statistics

Function

The write pending rate indicates the ratio of write pending data to the cache memory capacity. The Monitor Performance window displays the average and the maximum write pending rate for the specified period of time.

Storing period

Short-Range or Long-Range can be specified.

Usage rate of a data recovery and reconstruction processor 35

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Access paths usage statistics

Function

An access path is a path through which data and commands are transferred within a storage system.

In a storage system, channel adapters control data transfer between hosts and the cache memory. Disk adapters control data transfer between the cache memory and hard disk drives. Data transfer does not occur between channel adapters and disk adapters. Data is transferred via the ESW (PCI Express Switch adapter) to the cache memory.

When hosts issue commands, the commands are transferred via channel adapters to the shared memory (SM). The content of the shared memory is checked by disk adapters.

Performance Monitor tracks and displays the usage rate for the following access paths.

• Access paths between channel adapters and the cache switch (CHA ESW)

• Access paths between disk adapters and the cache switch (DKA ESW)

• Access paths between the cache switch and the cache memory (Cache ESW)

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneWrite Pending Rate (%)NoneCache

• Access paths between the MP blade and the cache switch (MP Blade ESW)

Figure 9 Access paths

Storing period

Short-Range or Long-Range can be specified.

36 Working with graphs

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Throughput of storage system

Function

Total throughput is the sum of I/Os per second. The read throughput is I/Os to the disk per second when the file read processing is performed. The write throughput is I/Os to the disk per second when the file write processing is performed.

Throughput in the following modes can be displayed.

• Sequential access mode

• Random access mode

• Cache fast write (CFW) mode

• Total value in the above-mentioned mode

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneUsage Rate (%)CHA ESWAccess Path

NoneUsage Rate (%)DKA ESW

NoneUsage Rate (%)MP Blade ESW

NoneUsage Rate (%)Cache ESW

Storing period

Short-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneThroughput (IOPS)NonePort*

NoneThroughput (IOPS)WWNWWN*

NoneThroughput (IOPS)Port

TotalTotal Throughput (IOPS)NoneLogical Device*

Sequential

Random

CFW

TotalRead Throughput (IOPS)

Sequential

Random

CFW

Throughput of storage system 37

Item on left side of Object field

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

TotalWrite Throughput (IOPS)

Sequential

Random

CFW

TotalTotal Throughput (IOPS)NoneParity Group*

Sequential

Random

CFW

TotalRead Throughput (IOPS)

Sequential

Random

CFW

TotalWrite Throughput (IOPS)

Sequential

* Volumes that do not accept I/O from the host, such as pool-VOLs, are not monitored.

Random

CFW

TotalTotal Throughput (IOPS)NoneLUN*

Sequential

Random

CFW

TotalRead Throughput (IOPS)

Sequential

Random

CFW

TotalWrite Throughput (IOPS)

Sequential

Random

CFW

Size of data transferred

Function

The amount of data per second transferred from the host server. The transferred data of reading or writing process can be monitored.

Storing period

Short-Range can be specified.

38 Working with graphs

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneData Trans. (MB/s)NonePort*

NoneData Trans. (MB/s)WWNWWN*

NoneData Trans. (MB/s)Port

TotalData Trans. (MB/s)NoneLogical Device*

Read

Write

TotalData Trans. (MB/s)NoneParity Group*

Read

Write

TotalData Trans. (MB/s)NoneLUN*

Read

Write

TotalData Trans. (MB/s)Parity GroupExternal Storage*

Read

* Volumes that do not accept I/O from the host, such as pool-VOLs, are not monitored.

Response times

Function

Time (in milliseconds) for replying from an external volume group when I/O accesses are made from the HP XP P9500 storage system to the external volume group. The average response time in the period specified at Monitoring Term is displayed.

Items that can be monitored response times are ports, WWNs, LDEVs, parity groups, LUNs, and external storages (parity groups and LDEVs).

Storing period

Short-Range can be specified.

Write

TotalData Trans. (MB/s)Logical Device

Read

Write

Response times 39

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneResponse Time (ms)NonePort*

NoneResponse Time (ms)WWNWWN*

NoneResponse Time (ms)Port

TotalResponse Time (ms)NoneLogical Device*

Read

Write

TotalResponse Time (ms)NoneParity Group*

Read

Write

TotalResponse Time (ms)NoneLUN*

Read

Write

TotalResponse Time (ms)Parity GroupExternal Storage*

Read

* Volumes that do not accept I/O from the host, such as pool-VOLs, are not monitored.

Cache hit rates

Function

The cash hit rate is a rate that the input or output data of the disk exists in the cache. The cash hit rate is displayed for the sequential access mode, the random access mode, the cache fast write (CFW) mode, and the entire these modes.

• Read hit ratio

For a read I/O, when the requested data is already in cache, the operation is classified as a read hit. For example, if ten read requests have been made from hosts to devices in a given time period and the read data was already on the cache memory three times out of ten, the read hit ratio for that time period is 30 percent. A higher read hit ratio implies higher processing speed because fewer data transfers are made between devices and the cache memory.

• Write hit ratio

For a write I/O, when the requested data is already in cache, the operation is classified as a write hit. For example, if ten write requests were made from hosts to devices in a given time period and the write data was already on the cache memory three cases out of ten, the write

Write

TotalResponse Time (ms)Logical Device

Read

Write

40 Working with graphs

hit ratio for that time period is 30 percent. A higher write hit ratio implies higher processing speed because fewer data transfers are made between devices and the cache memory.

Storing period

Short-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

Read (Total)Cache Hit (%)NoneLogical Device*

Read (Sequential)

Read (Random)

Read (CFW)

Write (Total)

Write (Sequential)

Write (Random)

Write (CFW)

Read (Total)Cache Hit (%)NoneParity Group*

Read (Sequential)

Read (Random)

Read (CFW)

Write (Total)

Write (Sequential)

* Volumes that do not accept I/O from the host, such as pool-VOLs, are not monitored.

Write (Random)

Write (CFW)

Read (Total)Cache Hit (%)NoneLUN*

Read (Sequential)

Read (Random)

Read (CFW)

Write (Total)

Write (Sequential)

Write (Random)

Write (CFW)

Cache hit rates 41

Back-end performance

Function

The back-end transfer can be monitored. The back-end transfer is the number of data transfers between the cache memory and the hard disk drive. The graph contains following information.

• Cache to Drive

The number of data transfers from the cache memory to hard disk drives.

• Drive to Cache Sequential

The number of data transfers from hard disk drives to the cache memory in sequential access mode

• Drive to Cache Random

The number of data transfers from hard disk drives to the cache memory in random access mode

Storing period

Short-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

* Volumes that do not accept I/O from the host, such as pool-VOLs, are not monitored.

Item on right side of Object field

Hard disk drive usage statistics

Item on left side of Monitor Data field

Item on right side of Monitor Data field

TotalBack Trans. (count/sec)NoneLogical Device*

Cache to Drive

Drive to Cache (Sequential)

Drive to Cache (Random)

TotalBack Trans. (count/sec)NoneParity Group*

Cache to Drive

Drive to Cache (Sequential)

Drive to Cache (Random)

TotalBack Trans. (count/sec)NoneLUN*

Cache to Drive

Drive to Cache (Sequential)

Drive to Cache (Random)

Function

The usage rates of the hard disk drive of each LDEV or parity group can be displayed.

42 Working with graphs

Storing period

Short-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

* Only information on internal volumes is displayed. Information about external volumes, FICON DM volumes, and virtual volumes such as DP-VOLs, Fast Snap V-VOLs, and Snapshot V-VOLs is not displayed.

Item on right side of Object field

Hard disk drive access rates

Function

The hard disk drive access rate shows the access rate of each hard disk drive (HDD). The rate of the file reading Read (Sequential) or the file writing Write (Sequential) processing of

HDD in the sequential access mode is displayed. The rate of file reading Read (Random) or file writing Write (Random) processing of HDD in the

random access mode is displayed.

Storing period

Long-Range or Short-Range can be specified.

Selection of monitoring objects

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneDrive Usage Rate (%)NoneLogical Device*

NoneDrive Usage Rate (%)NoneParity Group*

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

* Only information on internal volumes is displayed. Information about external volumes, FICON DM volumes, and virtual volumes such as DP-VOLs, Fast Snap V-VOLs, and Snapshot V-VOLs is not displayed.

Item on right side of Object field

Business Copy usage statistics

Function

The access rate of volume by Business Copy can be displayed the percentage of the processing of the program to all the processing of the physical drives, for each volume. This value is found by dividing access time to physical drives by the program by all the access time to physical drives.

Item on left side of Monitor Data field

Item on right side of Monitor Data field

Read (Sequential)Drive Access Rate (%)NoneLogical Device*

Read (Random)

Write (Sequential)

Write (Random)

Hard disk drive access rates 43

Storing period

Short-Range can be specified.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

* Only information on internal volumes is displayed. Information about external volumes, FICON DM volumes, and virtual volumes such as DP-VOLs, Fast Snap V-VOLs, and Snapshot V-VOLs is not displayed.

Item on right side of Object field

Item on left side of Monitor Data field

Detailed information of resources on top 20 usage rates

Function

While the capability to view the top 20 most used MP blades is supported, the maximum number of MP blades in a P9500 is 8. The blade information is displayed in the order of use.

Storing period

Only the Short-Range real time monitoring data can be supported.

Selection of monitoring objects

Select monitoring objects in Performance Objects field. The combination of items is shown as follows.

Item on left side of Object field

Item on right side of Object field

Item on left side of Monitor Data field

Item on right side of Monitor Data field

NoneBusiness Copy (%)NoneLogical Device*

Item on right side of Monitor Data field

NoneUsage Rate (%)MPController

Viewing MP blade resource details

To view the resources assigned to an individual MP blade, click the link to the name of the MP blade in the right panel of the Monitor window. The MP Properties window lists the 20 most-used resources by blade name.

44 Working with graphs

9 Changing display of graphs

This topic describes how to change displaying of graphs.

Graph operation

Information displayed in the graph can be changed. The following operations can be performed:

• Displayed items in the graph can be changed.

For details, see “Changing displayed items” (page 45).

• Displayed periods in the graph can be changed.

For details, see “Changing a display period” (page 45)

• New graphs can be added.

For details, see “Adding a new graph ” (page 45).

• Graph panels can be deleted.

For detail, see “Deleting graph panel” (page 46).

Changing displayed items

Use the Edit Performance Objects window to change displayed items in the graph.

1. Display a graph in the Monitor Performance window.

For details, see “Basic operation ” (page 32).

2. Click Edit Performance Objects. The Edit Performance Objects window opens.

3. Change displayed items in the information setting field at the left of the window. For details, see “Basic operation ” (page 32) .

4. Click Add. Items are added in the Selected Objects field.

5. If you want to delete the item, select the item and then click Remove.

6. Click OK. The graph is displayed.

Changing a display period

To change a display period in the graph.

1. Display graph in the Monitor Performance window. For details, see “Basic operation ” (page 32).

2. Click Edit Time Range. The Edit Time Range window opens.

3. Input the date when the display of the graph begins in the From field. Input the date when the display of the graph is ended in the To field.

4. Click OK. The graph is displayed.

Adding a new graph

To add a new graph:

Graph operation 45

1. Display graph in the Monitor Performance window.

For details, see “Basic operation ” (page 32).

2. Click Add Graph. The Add Graph window opens.

3. Change displayed items in the information setting field at the left of the window. For details, see “Basic operation ” (page 32).

4. Click Add. Items are added in the Selected Objects field.

5. If you want to delete the item, select the item and then click Remove.

6. Click OK. The graph is added.

Deleting graph panel

A graph panel can be deleted from the Monitor Performance window.

1. Display graph in the Monitor Performance window. For details, see “Basic operation ” (page 32).

2. Click Delete Graph or the icon to close the window displayed in the upper right of the graph panel.

A warning message appears, asking whether you want to delete the graph panel.

3. Click OK to close the message. The graph panel is deleted.

46 Changing display of graphs

10 Performance Control operations

This topic provides information and instructions for using Performance Control software to perform upper-limit control.

Overview of Performance Control operations

Procedures for using Performance Control depend on the connection between host bus adapters (HBAs) and storage system ports. HBAs are adapters contained in hosts and serve as host ports for connecting the hosts and the storage system.

If one-to-one connections are established between host bus adapters and ports, you specify the priority of I/O operations, upper limit value, and threshold value on each port. Because one port connects to one HBA, you can define the server priority by the port.

However, if many-to-many connections are established between host bus adapters and ports, you cannot define the server priority by the port, because one port can connect to multiple host bus adapters, and also multiple ports can connect to one host bus adapter. Therefore, in the many-to-many connection environment, you specify the priority of I/O operations and upper limit value on each host bus adapter. In this case, you specify one threshold value for the entire storage system.

If one-to-one connections are established between host bus adapters and ports, you use the Port tab of the Performance Control main window. If many-to-many connections are established between host bus adapters and ports, you use the WWN tab of the Performance Control main window. This topic explains the operation procedures in each tab.

If one-to-one connections link HBAs and ports

The next figure shows an example of a network in which each host bus adapter is connected to only one port on the storage system (henceforth, this network is referred to as network A). Host bus adapters and the storage system ports are directly connected and are not connected via hubs and switches.

Figure 10 Network A (1-to-1 connections between HBAs and ports)

If one-to-one connections are established between HBAs and ports, take the following major steps:

Overview of Performance Control operations 47

1. Set priority to ports on the storage system using the Port tab of the Performance Control main window.

In network A, the ports 1A and 1C are connected to high-priority production servers. The port 2A is connected to a low-priority development server. Therefore, the ports 1A and 1C should be given high priority, and the port 2A should be given low priority.

The next figure shows a portion of the Performance Control main window where the abbreviation Prio. indicates that the associated port is given high priority, and the abbreviation Non-Prio. indicates that the port is given low priority.

NOTE: The term prioritized port is used to refer to a high-priority port, and the term

non-prioritized port is used to refer to a low-priority port.

Figure 11 Priority specified in the Performance Control main window

2. Monitor traffic at ports. You must obtain statistics about traffic at each port on the storage system.

There are two types of traffic statistics: the I/O rate and the transfer rate. The I/O rate is the number of I/Os per second. The transfer rate is the size of data transferred between a host and the storage system. When you view traffic statistics in the window, you select either the I/O rate or the transfer rate. Use the Performance Monitor window of Performance Monitor to view a line graph illustrating changes in traffic.

The next figure shows the changes in the I/O rate for the three ports (1A, 1C, and 2A). According to the graph, the I/O rate for 1A and 1C was approximately 400 IO/s at first. The I/O rate for 2A was approximately 100 IO/s at first. However, as the I/O rate for 2A gradually increased from 100 IO/s to 200 IO/s, the I/O rate for 1A and 1C decreased from 400 IO/s to 200 IO/s. This fact indicates that the high-priority production servers have suffered lowered performance. If you were the network administrator, you probably would like to maintain the I/O rate for prioritized ports (1A and 1C) at 400 IO/s. To maintain the I/O rate at 400 IO/s, you must set an upper limit to the I/O rate for the port 2A.

For detailed information about monitoring traffic, see “Setting priority for ports on the storage

system” (page 54) and “Analyzing traffic statistics” (page 54).

Figure 12 Traffic at ports

48 Performance Control operations

3. Set an upper limit to traffic at the non-prioritized port. To prevent decline in I/O rates at prioritized ports, you set upper limit values to the I/O rate for non-prioritized ports.

When you set an upper limit for the first time, it is recommended that the upper limit be approximately 90 percent of the peak traffic. In network A, the peak I/O rate for the non-prioritized port (2A) is 200 IO/s. So, the recommended upper limit for 2A is 180 IO/s.

For details on how to set an upper limit, see “Setting upper-limit values for non-prioritized

WWNs” (page 61).

4. Check the result of applying upper limit values. After applying upper limit values, you must measure traffic at ports. You must view traffic statistics for prioritized ports 1A and 1C to check whether the host performance is improved to a desirable level.

In network A, the desirable I/O rate for ports 1A and 1C is 400 IO/s. If the I/O rate reaches 400 IO/s, production server performance has reached to a desirable level. If production server performance is not improved to a desirable level, you can change the upper limit to a smaller value and then apply the new upper limit to the storage system. In network A, if the upper limit is set to 180 IO/s but the I/O rate for 1A and 1C is still below 400 IO/s, the administrator needs to change the upper limit until the I/O rate reaches 400 IO/s.

5. If necessary, apply a threshold. If you want to use threshold control, set threshold values in the Port tab in the Performance Control main window. You can set threshold values in either of the following ways:

• Set one threshold to each prioritized port

In network A, if you set a threshold of 200 IO/s to the port 1A and set a threshold of 100 IO/s to the port 1C, the upper limit on the non-prioritized port (2A) is disabled when either of the following conditions is satisfied:

The I/O rate for the port 1A is 200 IO/s or lower. The I/O rate for the port 1C is 100 IO/s or lower.

• Set only one threshold to the entire storage system

In network A, if you set a threshold of 500 IO/s to the storage system, the upper limit on the non-prioritized port (2A) is disabled when the sum of the I/O rates for all prioritized ports (1A and 1C) goes below 500 IO/s.

For details on how to set a threshold, see “Setting a threshold ” (page 56).

If many-to-many connections link HBAs and ports

The next figure shows an example of a network in which a production server and a development server are connected to the storage system (Henceforth, this network is referred to as network B). The host bus adapter (wwn01) in the production server is connected to four ports (1A, 1C, 2A and 2C). The host bus adapters (wwn02 and wwn03) in the development server are also connected to the four ports.

If many-to-many connections link HBAs and ports 49

Figure 13 Network B (many-to-many connections are established between HBAs and ports)

If many-to-many connections are established between HBAs and ports, take the next steps:

1. Find WWNs of host bus adapters. Before using Performance Control, you must find the WWN (Worldwide Name) of each host bus adapter in host servers. WWNs are 16-digit hexadecimal numbers used to identify host bus adapters. For details on how to find WWNs, see the HP XP P9000 Provisioning for Open Systems User Guide.

2. Ensure that all host bus adapters connected to ports in the storage system are monitored. Use the WWN tab of the Performance Control main window to define which port is connected to which host bus adapter. Place host bus adapters connected to each port below the Monitor icons.

In network B, each of the four ports is connected to three host bus adapters (wwn01, wwn02, and wwn03). Place the host bus adapter icons of wwn01, wwn02, and wwn03 below the Monitor icons for all the four port icons.

The resulting definitions on the window are as follows:

Figure 14 Specifying host bus adapters to be monitored

For more detailed instruction, see “Setting priority for ports on the storage system” (page 54). Performance Control is unable to monitor and control the performance of hosts whose host

bus adapters are placed below the Non-Monitor icon.

50 Performance Control operations

3. Set priority to host bus adapters using the WWN tab of the Performance Control main window. In network B, the production server is given high priority and the development server is given

low priority. If your network is configured as in Figure 13 (page 50), you must give high priority to wwn01 and also give low priority to wwn02 and wwn03.

To give priority to the three host bus adapters, take the following steps:

• In the WWN tab, select one of the four ports that the HBAs are connected to (i.e., ports

1A, 1C, 2A, and 2C).

• Set Prio. to wwn01. Also, set Non-Prio. to wwn02 and wwn03.

Figure 15 Priority specified in the Performance Control main window

NOTE: The term prioritized WWN to refers to a high-priority host bus adapter (for example,

wwn01). The term non-prioritized port refers to a low-priority host bus adapter (for example, wwn02 and wwn03).

4. Monitor traffic between host bus adapter and ports. You must obtain statistics about traffic between host bus adapter and ports.

If your network is configured as network B, you must do the following:

• Measure traffic between the port 1A and the three host bus adapters (wwn01, wwn02

and wwn03.

• Measure traffic between the port 1C and the three host bus adapters (wwn01, wwn02

and wwn03.

• Measure traffic between the port 2A and the three host bus adapters (wwn01, wwn02

and wwn03.

• Measure traffic between the port 2C and the three host bus adapters (wwn01, wwn02

and wwn03. The following graph illustrates the I/O rate at the paths between each port and the host

bus adapters. According to the graph, the I/O rate at the path between 1A and the prioritized WWN (wwn01) was approximately 400 IO/s at first. The I/O rate at the path between 1A and the non-prioritized WWNs (wwn02 and wwn03) was approximately 100 IO/s at first. However, as the I/O rate for non-prioritized WWNs (wwn02 and wwn03) gradually increased from 100 IO/s to 200 IO/s, the I/O rate for the prioritized WWN (wwn01) decreased from 400 IO/s to 200 IO/s. This indicates that the high-priority production server has degraded. If you were the network administrator, you probably would like to maintain the I/O rate for the prioritized WWN (wwn01) at 400 IO/s.

For more information about monitoring traffic, see “Setting priority for host bus adapters”

(page 60) and “Analyzing traffic statistics” (page 59).

If many-to-many connections link HBAs and ports 51

Figure 16 Traffic at ports

5. Set an upper limit to traffic between ports and the non-prioritized WWN to prevent decline in I/O rates at prioritized WWNs.

When you set an upper limit for the first time, the upper limit should be approximately 90 percent of the peak traffic level.

In network B, the peak I/O rate at the paths between port 1A and the non-prioritized WWNs (wwn02 and wwn03) is 200 IO/s. The peak I/O rate at the paths between port 1C and the non-prioritized WWNs is also 200 IO/s. The peak I/O rate at the paths between port 2A and the non-prioritized WWNs is also 200 IO/s. The peak I/O rate at the paths between port 2C and the non-prioritized WWNs is also 200 IO/s. So, the recommended upper limit for the non-prioritized WWNs is 720 IO/s (= 200 ´ 4 ´ 0.90).

If your network is configured as in Figure 13 (page 50), you must do the following in this order:

• In the WWN tab, select one of the four ports that the HBAs are connected to (i.e., ports

1A, 1C, 2A, and 2C).

• Set an upper limit to the non-prioritized WWNs (wwn02 and wwn03).

The following figure is the result of setting the upper limit of 720 IO/s to the paths between 1A and the non-prioritized WWNs. For details on how to set an upper limit, see “Setting

upper-limit values for non-prioritized WWNs” (page 61).

Figure 17 Setting upper limits

6. Check the result of applying upper limit values. After applying upper limit values, you must measure traffic at ports. View traffic statistics for the prioritized WWN to check whether the host performance is improved to a desirable level.

In network B, the desirable I/O rate for the prioritized WWN is 400 IO/s. If the I/O rate reaches 400 IO/s, production server performance has reached to a desirable level. If production server performance is not improved to a desirable level, you can change the upper limit to a smaller value and then apply the new upper limit to the storage system. In network B, if the upper limit is set to 720 IO/s but the I/O rate for wwn01 is still below 400 IO/s, the administrator needs to change the upper limit until the I/O rate reaches 400 IO/s.

52 Performance Control operations

If an upper limit of the non-prioritized WWN is set to zero or nearly zero, I/O performance might be lowered. If I/O performance is lowered, the host cannot be connected to the storage system in some cases.

7. If necessary, apply a threshold. If you want to use threshold control, set a threshold in the WWN tab in the Performance Control main window.

In the WWN tab, you can specify only one threshold for the entire storage system, regardless of the number of prioritized WWNs. For example, if there are three prioritized WWNs in the network and the threshold is 100 IO/s, the upper limit on the non-prioritized WWNs is disabled when the sum of the I/O rates for all prioritized WWNs goes below 100 IO/s.

For details on how to set a threshold, see “Setting a threshold” (page 62).

CAUTION: If you enter zero (0) in a cell to disable the upper limit, the cell displays a hyphen

(-) and the threshold for the prioritized port becomes ineffective. If the thresholds of all the prioritized ports are ineffective, threshold control will not be performed but upper limit control will be performed.

The following table shows the relationship between thresholds of a prioritized WWN and the upper limits of a non-prioritized WWN.

Table 5 Prioritized WWN threshold setting relationships

Threshold settings

Threshold Is Set to The Prioritized WWN

The Prioritized WWN

Port tab operations

If one-to-one connections are established between host bus adapters (HBAs) and storage system ports, use the Port tab in the Performance Control main window to do the following:

• Analyze traffic statistics

• Measure traffic between host bus adapters and storage system ports

A number other than zero is set to the upper limit of the non-prioritized WWN

When thresholds are set to prioritized WWNs, depending on the I/O rate or the transfer rate, the following controls are executed.

• If the total value of I/O rate or transfer rate

exceeds the threshold in all prioritized WWNs, upper limits of all the non-prioritized WWNs take effect.

• If the total value of I/O rate or transfer rate

goes below the threshold in all prioritized WWNs, upper limits of all the non-prioritized WWNs do not take effect.

The specified upper limit always takes effect.Threshold Is Not Set to

Zero is set to the upper limit of the non-prioritized WWN

The threshold control of the prioritized WWN is not executed.

• Set priority to ports on the storage system

• Set an upper limit to traffic at each non-prioritized port

• Set a threshold to the storage system or to each prioritized port, if necessary

If one-to-one connections are established between host bus adapters and ports, you should specify the priority of I/O operations on each port. You can specify the upper limit values on the non-prioritized ports, and if necessary, you can also specify the threshold values on the prioritized ports. Moreover, you can use one threshold value applied for the entire storage system.

Port tab operations 53

For details on the system configuration of one-to-one connections between host bus adapters and ports, see “If one-to-one connections link HBAs and ports” (page 47). This topic explains operation procedures you can perform for ports and the entire storage system.

Analyzing traffic statistics

The traffic statistics reveal the number of I/Os that have been made via ports. The traffic statistics also reveal the amount of data that have been transferred via ports. You must analyze the traffic statistics to determine upper limit values that should be applied to I/O rates or transfer rates for non-prioritized ports.

The following is the procedure for using the Performance Control main window to analyze traffic statistics. You can also use the Performance Monitor window to analyze traffic statistics. Performance Monitor can display a line graph that indicates changes in traffic.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Select the Port tab.

5. Select All from the list at the top right corner of the window.

6. Do one of the following:

• To analyze I/O rates, select IOPS from the list at the upper left corner of the list.

• To analyze transfer rates, select MB/s from the list at the upper left corner of the list.

The list displays traffic statistics (i.e., the average and peak I/O rates or transfer rates) of the ports.

7. Analyze the information in the list and then determine upper limit values that should be applied to non-prioritized ports. If necessary, determine threshold values that should be applied to prioritized ports. For details on the upper limit values and threshold values, see “If one-to-one

connections link HBAs and ports” (page 47).

Setting priority for ports on the storage system

If one-to-one connection is established between HBAs and ports, you need to measure traffic between high-priority HBAs and prioritized ports. You also need to measure traffic between low-priority HBAs and non-prioritized ports.

Prioritized ports are ports on which the processing has high priority and non-prioritized ports are ports on which the processing has low priority. Specify a port that connects to a high-priority host bus adapter as a prioritized port. Specify a port that connects to a low-priority host bus adapter as a non-prioritized port.

1. Click Reports >Performance Monitor> Performance Control to open the Performance Control window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main window appears.

4. Ensure that the Port tab is displayed.

5. Select All from the list at the top right corner of the window.

6. Right-click a high-priority port and then select Non-Prio ->> Prio from the pop-up menu. If there is more than one high-priority port, repeat this operation. The Attribute column displays Prio.

54 Performance Control operations

7. Right-click a low-priority port and then select Prio ->> Non-Prio from the pop-up menu. If there

is more than one low-priority port, repeat this operation. The Attribute column displays Non-Prio. You must set upper limit values for the Non-prio. specified ports. For detail about the setting

of upper limit values, see “Setting upper-limit values to traffic at non-prioritized ports” (page 55).

8. Click Apply. The settings on the window are applied to the storage system. After priority has been set, you can implement the procedure for measuring traffic (I/O rates and

transfer rates). See “Monitoring operation” (page 28).

Setting upper-limit values to traffic at non-prioritized ports

After you analyze traffic statistics, you must set upper limit values to I/O rates or transfer rates for non-prioritized ports. Upper limit values for I/O rates are used to suppress the number of I/Os from the low priority host servers and thus provide better performance for high-priority host servers. Upper limit values for transfer rates are used to suppress the amount of data that should be transferred between the storage system and the low priority ports, and thus provide better performance for high-priority host servers.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Select the Port tab.

5. Do one of the following:

• To limit the I/O rate for the non-prioritized port, select IOPS from the list at the upper left

corner of the list.

• To limit the transfer rate for the non-prioritized port, select MB/s from the list at the upper

left corner of the list.

6. Locate the non-prioritized port in the list.

NOTE: The Attribute column of the list indicates whether ports are prioritized or non

prioritized. If you cannot find any non prioritized port in the list, check the list at the top right corner of the window. If the list displays Prioritize, select All or Non Prioritize from the list.

7. Do one of the following:

• To limit the I/O rate for the non-prioritized port, double-click the desired cell in the IOPS

column in Upper. Next, enter the upper limit value in the cell.

• To limit the transfer rate for the non-prioritized port, double-click the desired cell in the

MB/s column in Upper. Next, enter the upper limit value in the cell. In the list, either of IOPS or MB/s column is activated depending on the rate selected at

step 3 above. You can use either of them to specify the upper limit value for one port. You can specify different types of rates (IOPS or MB/s) for the upper limit values of different non-prioritized ports.

The upper limit value that you entered is displayed in blue.

8. Click Apply. The settings in the window are applied to the storage system. The upper limit

value that you entered turns black.

Port tab operations 55

Setting a threshold

If threshold control is used, upper limit control is automatically disabled when traffic between production servers and the storage system is reduced to a specified level. For details, see “Threshold

control” (page 9) and “If one-to-one connections link HBAs and ports” (page 47).

If one-to-one connections are established between HBAs and ports, you can set the threshold value by the following two ways:

• Set a threshold value for each prioritized port

• Set one threshold value for the entire storage system

The procedures for these operations are explained below.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main window appears.

4. Select the Port tab.

5. To set a threshold value for each prioritized port, select the type of rates for the threshold value from the list at the upper lest corner of the list.

• To use the I/O rates for the threshold value, select IOPS.

• To use the transfer rates for the threshold value, select MB/s.

NOTE: If you want to set one threshold value for the entire storage system, this step is

unnecessary.

6. Do one of the following:

• To set a threshold to each prioritized port, locate the desired prioritized port, which is

indicated by Prio. in the Attribute column. Next, double-click the cell in the IOPS or MB/s column in Threshold, and then enter the threshold value. In the list, either of IOPS or MB/s column is activated depending on the rate selected at step 5 above. Repeat this operation to set the thresholds for all the prioritized ports. You can use different types of rates (IOPS or MB/s) for thresholds of different prioritized ports.

CAUTION: If you enter zero (0) in a cell to disable the upper limit, the cell displays a

hyphen (-) and the threshold for the prioritized port becomes ineffective. If the thresholds of all the prioritized ports are ineffective, threshold control will not be performed but upper limit control will be performed. If you set thresholds for multiple prioritized ports and the I/O rate or transfer rate becomes below the threshold at one of the prioritized port, threshold control works in the entire storage system and the upper limits of the non-prioritized ports are disabled. The following table shows the relationship between the thresholds and the upper limits.

56 Performance Control operations

Table 6 Relationship between the thresholds of the prioritized port and the upper limits of the non-prioritized port

Thresholds settings

Threshold Is Set to The Prioritized port

A number other than zero is set to the upper limit of the non-prioritized port

When thresholds are set to multiple prioritized ports, depending on the transfer rate, following controls are executed.

Zero is set to the upper limit of the non-prioritized port

The threshold control of the prioritized port is not executed.

◦ If I/O rate or transfer rate exceeds the

threshold in any prioritized port, upper limits of all the non-prioritized ports take effect.

◦ If I/O rate or transfer rate goes below the

threshold in any prioritized port, upper limits of all the non-prioritized ports do not take effect.

The specified upper limit always takes effect.Threshold Is Not Set to The Prioritized port

• To set one threshold to the entire storage system, select the All Thresholds check box.

Next, select IOPS or MB/s from the list of right side in All Thresholds and enter the threshold value in the text box. Even if the types of rates for upper limit values and the threshold are different, the threshold control can work for all the non-prioritized ports.

7. Click Apply. The settings in the window are applied to the storage system.

WWN tab operations

If many-to-many connections are established between host bus adapters (HBAs) and storage system ports, you use the WWN tab in the Performance Control main window to do the following:

• Make all the traffics between host bus adapters and ports monitored

• Analyze traffic statistics

• Measure traffic between host bus adapters and storage system ports

• Set priority to host bus adapters

• Set an upper limit on traffic at non-prioritized WWNs

• Set a threshold, if necessary

If many-to-many connections are established between host bus adapters and ports, you should specify the priority of I/O operations on each host bus adapter. You can specify the upper limit values on the non-prioritized WWNs. If necessary, you can set one threshold value applied for the entire storage system. When many-to-many connections are established between host bus adapters and ports, you cannot set individual thresholds for prioritized WWNs.

For the system configuration of many-to-many connections between host bus adapters and ports, see “If many-to-many connections link HBAs and ports” (page 49).

For details on the system configuration of many-to-many connections between host bus adapters and ports, see “If many-to-many connections link HBAs and ports” (page 49). This topic explains operation procedures you can perform for host bus adapters and the entire storage system.

Monitoring all traffic between HBAs and ports

When many-to-many connections are established between host bus adapters (HBAs) and ports, you should make sure that all the traffic between HBAs and ports is monitored.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

WWN tab operations 57

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Ensure that the WWN tab is visible. The two trees are displayed in the left side of the WWN tab. The upper-left tree lists ports in

the storage system.

5. Select All from the list at the top right corner of the window.

6. In the upper-left tree, double-click a port.

7. Double-click Non-Monitor below the specified port. If there are any host bus adapters whose traffics with the specified port are not monitored,

those host bus adapters are displayed below Non-Monitor.

8. Right-click Monitor and then select Add WWN. The Add WWN window opens where you can add a WWN of a host bus adapter to Monitor.

9. In the Add WWN window, specify the WWN and the PFC name. Expand the WWN list to show the WWNs of the host bus adapters that are connected to the

port but are not monitored. These host bus adapters are the same as that displayed in step 7. From that list, select a WWN and specify the PFC name (up to 64 characters).

We recommend that you specify the same names for the PFC names and the nicknames of the host bus adapters for convenience of host bus adapter management. Nicknames are aliases of host bus adapters defined by LUN Manager. In the Performance Monitor window, not only PFC names but also nicknames are displayed as the aliases of host bus adapters (WWNs) in the list. Therefore, if you specify both the same aliases, the management of the host bus adapters is easier.

10. Click OK. The selected WWN (of the host bus adapter) is moved from Non-Monitor to Monitor. If the specified host bus adapter is connected to other ports, after clicking OK, a message

appears asking whether to change the settings of that host bus adapter for other ports, too. Make the same setting for all the ports.

11. Repeat step 8 to 10 to move all the host bus adapters displayed below Non-Monitor to below Monitor.

If you disconnect a host that has been connected via a cable to your storage system or change the port to the another port of the host, the WWN for the host will remain in the WWN list of the WWN tab. If you want to delete the WWN from the WWN list, you can delete the WWN by using LUN Manager. For detail information of the deleting old WWNs from the WWN list, see the HP XP P9000 Provisioning for Open Systems User Guide.

12. Click Apply in the Performance Control main window. The settings on the window are applied to the storage system.

58 Performance Control operations

If you add a port or host bus adapter to the storage system after the settings above, the traffics about connections to the newly added port or host bus adapter will not be monitored. In this case, follow the procedure above again to make all the traffics between host bus adapters and ports monitored.

Up to 32 host bus adapters (WWNs) can be monitored for one port. If more than 32 host bus adapters are connected to one port, the traffics about some host bus adapters will be obliged to be excluded from the monitoring target. Consider the intended use of each host, and move the host bus adapters that are not necessary to monitor to Non-Monitor by completing the steps in the following section.

Excluding traffic between a host bus adapter and a port from the monitoring target

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Ensure that the WWN tab is displayed.

5. Select All from the list at the top right corner of the window.

6. In the upper-left tree, double-click a port to which more than 32 host bus adapters are connected.

7. Double-click Monitor below the specified port.

8. Right-click the WWN of a host bus adapter you want to exclude from the monitoring target and then select Delete WWN from the pop-up menu.

NOTE:

• If the selected host bus adapter is connected to multiple ports, when you select the host

bus adapter and select the Delete WWN pop-up menu, a message will appear that asks you whether to move the host bus adapter from Monitor to Non-Monitor below all other ports, too.

• If the selected host bus adapter is contained in an PFC group, a message will appear

that tell you to delete the host bus adapter from the PFC group on ahead. You cannot move a host bus adapter which is contained in an PFC group from Monitor to Non-Monitor. For details on how to delete a host bus adapter from an PFC group, see “Deleting an

HBA from an PFC group” (page 65).

9. Click OK for the confirmation message that asks you whether to delete the host bus adapter. The deleted host bus adapter (WWN) is moved from Monitor to Non-Monitor.

10. Click Apply in the Performance Control main window. The settings on the window are applied to the storage system.

Analyzing traffic statistics

The traffic statistics reveal the number of I/Os that have been made via ports from HBAs. They also reveal the amount of data that have been transferred between ports and HBAs. You must analyze the traffic statistics to determine upper limit values that should be applied to I/O rates or transfer rates for low-priority HBAs.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control window.

WWN tab operations 59

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Select the WWN tab.

5. Select All from the list at the top right corner of the window.

6. Do one of the following:

• To analyze I/O rates, select IOPS from the list at the upper left corner.

• To analyze transfer rates, select MB/s from the list at the upper left corner of the list.

7. Below the Storage System folder in the upper-left tree, click the icon of the port whose traffic

statistics you want to collect. The list displays traffic statistics (I/O rates or transfer rates) about the host bus adapters that

connect to the selected port. The following two types of traffic are shown. The traffic has attributes including the average

and maximum values.

• Traffic between the host bus adapter and the selected port (shown in Per Port)

• Sum of the traffic between the host bus adapter and all the ports connected to the host

bus adapter (shown in WWN Total)

NOTE: The traffic statistics only about the host bus adapters below Monitor appear in

the list. The WWN Total traffic statistics will also be displayed in the list when you click an icon

in the lower left tree. If you click the Storage System folder in the lower left tree, the sum of the traffic of the host bus adapters registered on each PFC group is displayed. For details on PFC groups, see “Grouping host bus adapters” (page 65).

8. Analyze the information in the list and then determine upper limit values that should be applied to non-prioritized WWNs. If necessary, determine threshold values that should be applied to prioritized WWNs. For details, see “If many-to-many connections link HBAs and ports”

(page 49).

Setting priority for host bus adapters

If many-to-many connection is established between host bus adapters (HBAs) and ports, you need to define the priority of WWNs, measure traffic between each HBA and the port that the HBA is connected to, and analyze the traffics.

The host bus adapters (HBAs) are divided into two types: Prioritized WWNs and non-prioritized WWNs. Prioritized WWNs are the host bus adapters that are used for the high-priority processing, and non-prioritized WWNs are the host bus adapters that are used for the low-priority processing. Specify a host bus adapter existed in a server, on which the high-priority processing is performed, as a prioritized WWNs. Specify a host bus adapter existed in a server, on which the low-priority processing is performed, as a non-prioritized WWNs.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main window appears.

4. Select the WWN tab.

5. Select All from the list at the top right corner of the window.

6. In the upper-left tree, double-click a port.

60 Performance Control operations

7. Double-click Monitor, which is displayed below the specified port.

8. Check to see if all the WWNs of the host bus adapters to be controlled by using Performance

Control appear below Monitor. If some of the WWNs are missing, use the procedure in “Monitoring all traffic between HBAs

and ports” (page 57) to move all WWNs to below Monitor.

9. Click Monitor to display the information of the host bus adapters that are monitored in the list

on the right of the tree.

10. Right-click a host bus adapter (WWN) in the list and then select Non-Prio ->> Prio from the

pop-up menu. The Attribute column of the selected WWN in the list displays Prio. If you want to specify

more than one prioritized WWN, repeat this operation.

NOTE: You cannot change the priority of a WWN that is contained in an PFC group. For

details on how to change the attribute of a WWN contained in an PFC group, see “Switching

priority of an PFC group” (page 66).

11. Right-click a host bus adapter (WWN) in the list and then select Prio ->> Non-Prio from the

pop-up menu. The Attribute column of the selected WWN in the list displays Non-Prio. If you want to specify

more than one non-prioritized WWN, repeat this operation.

NOTE: You cannot change the priority of a WWN which is contained in an PFC group. For

details on how to change the attribute of a WWN contained in an PFC group, see “Switching

priority of an PFC group” (page 66).

You must set upper limit values for the Non-prio. specified ports. For details, see “Setting

upper-limit values for non-prioritized WWNs” (page 61).

12. Repeat steps 6 to 11 for ports (except for the port selected in step 6). If one host bus adapter is connected to multiple ports and you specify the priority of the host

bus adapter for one port, the specified priority will be also applied to the host bus adapter settings for other connected ports automatically.

13. Click Apply in the Performance Control main window. The settings on the window are applied to the storage system.

Follow the instructions in “Starting monitoring” (page 28) to measure traffic (that is, I/O rates and transfer rates).

Setting upper-limit values for non-prioritized WWNs

After you analyze traffic statistics about prioritized WWNs and non-prioritized WWNs, you must set upper limit values to I/O rates or transfer rates for non-prioritized WWNs. Upper limit values for I/O rates are used to suppress the number of I/Os from the low priority host servers and thus provide better performance for high-priority host servers. Upper limit values for transfer rates are used to suppress the amount of data that should be transferred between the storage system and the low priority ports, thus providing better performance for high-priority host servers.

TIP: To set the same upper limit value to more than one non-prioritized WWN, use an PFC group.

For details on PFC groups, see “Grouping host bus adapters” (page 65).

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control window.

2. Click to change to Modify mode.

WWN tab operations 61

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Ensure that the WWN tab is displayed.

5. Do one of the following:

• To limit the I/O rate of the non-prioritized WWN, select IOPS from the list at the upper

left corner.

• To limit the transfer rate of the non-prioritized WWN, select MB/s from the list at the

upper left corner.

6. In the upper-left tree, click the icon of the port whose traffic you want to limit below the Storage System folder.

The information about the host bus adapters which connect to the selected port is displayed in the list.

7. Locate the non-prioritized WWN in the list.

NOTE:

• The Attribute column of the list indicates whether WWNs are prioritized or non-prioritized.

The Attribute column of a non-prioritized WWN displays Non-Prio.

• If you cannot find any non-prioritized WWN in the list, check the list at the top right

corner of the window. If the list displays Prioritize, select All or Non-Prioritize.

8. Do one of the following:

• To limit the I/O rate of the non-prioritized WWN, double-click the desired cell in the IOPS

column in Upper. Next, enter the upper limit value in the cell.

• To limit the transfer rate of the non-prioritized WWN, double-click the desired cell in the

MB/s column in Upper. Next, enter the upper limit value in the cell. In the list, either of the IOPS cells or MB/s cells are activated depending on the rate you

specified in step 5. You can specify the limit value by using either of the I/O rate or transfer rate for each host bus adapter. The upper limit value that you entered is displayed in blue. It is allowed that you specify upper limit values by using the I/O rate for some host bus adapters and specify them by using the transfer rate for the other host bus adapters.

NOTE:

◦ You cannot specify or change the upper limit value of a host bus adapter which is

contained in an PFC group. The upper limit value of such a host bus adapter is defined by the PFC group settings. For details on how to specify an upper limit value for an PFC group, see “Setting an upper-limit value to HBAs in an PFC group”

(page 66).

◦ If one host bus adapter is connected to multiple ports and you specify an upper limit

value of the host bus adapter for one port, the specified upper limit value will be applied to the host bus adapter settings for other connected ports automatically.

9. Click Apply. The settings in the window are applied to the storage system. The upper limit value that you

entered turns black.

Setting a threshold

control” (page 9) and “If many-to-many connections link HBAs and ports” (page 49).

62 Performance Control operations

If many-to-many connections are established between host bus adapters and storage system ports, you can set one threshold value for the entire storage system. In this environment, you cannot set individual threshold values for each prioritized WWN.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Select the WWN tab.

5. Select the All Thresholds check box.

6. Select IOPS or MB/s from the All Thresholds list, and do one of the following:

• To specify the threshold value by using the I/O rate, select IOPS from the list below the

check box.

• To specify the threshold value by using the transfer rate, select MB/s from the list below

the check box. Even if the types of rates differ between the upper limit values and the threshold value,

the threshold control is effective for all the non-prioritized WWNs.

7. Enter the threshold in the text box of All Thresholds.

8. Click Apply.

The settings in the window are applied to the storage system.

Changing the PFC name of a host bus adapter

Use the Performance Control main window to assign an PFC name to a host bus adapter (HBA). Although you can identify HBAs by WWNs (Worldwide Names), you will be able to identify HBAs more easily if you assign PFC names. WWNs are 16-digit hexadecimal numbers and cannot be changed. However, PFC names should not necessarily be 16-digit hexadecimal numbers and can be changed.

The following is the procedure for changing an already assigned PFC name. For details on how to assign an PFC name, see “Monitoring all traffic between HBAs and ports” (page 57).

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Ensure that the WWN tab is displayed.

5. In the upper-left tree, select a host bus adapter ( ) from below Monitor and then right-click

the selection.

6. From the pop-up menu, select Change WWN and PFC Name. The Change WWN and PFC Name window opens.

WWN tab operations 63

7. Enter a new PFC name in the PFC Name box and then select OK. You can use up to 64

characters for an PFC name.

8. In the Performance Control main window, click Apply. The settings in the window are applied

to the storage system.

Registering a replacement host bus adapter

If a host bus adapter fails, replace the adapter with a new one. After you finish replacement, you will need to delete the old host bus adapter from the Performance Control main window and then register the new host bus adapter.

When you add a new host bus adapter rather than replacing an old one, the WWN of the added host bus adapter is automatically displayed below Non-Monitor for the connected port in the list.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Select the WWN tab.

5. In the upper-left tree, select the old host bus adapter ( ) from below Monitor and then right-click

the selection.

6. From the pop-up menu, select Change WWN and PFC Name. The Change WWN and PFC Name window opens.

7. Enter the WWN of the new host bus adapter in the WWN combo box. You can select the WWN of the newly connected host bus adapter in the WWN combo box.

8. If necessary, enter a new PFC name in the PFC Name box. You can use up to 64 characters for an PFC name.

9. Select OK to close the Change WWN and PFC Name window.

10. In the Performance Control main window, click Apply. The settings in the window are applied to the storage system.

64 Performance Control operations

Grouping host bus adapters

Use Performance Control to create an PFC group to contain multiple host bus adapters. You can include a maximum of 32 host bus adapters in a PFC group. You can create up to 512 PFC groups in the storage system. All the host bus adapters (HBAs) in one PFC group must be of the same priority. Prioritized WWNs and non-prioritized WWNs cannot be mixed in the same group.

You can use an PFC group to switch priority of multiple HBAs from prioritized to non-prioritized, or vice versa. You can also use an PFC group to set the same upper limit value to all the HBAs in the group.

Containing multiple HBAs in an PFC group

A host bus adapter can be contained in only one PFC group. To create an PFC group and contain multiple host bus adapters in the group:

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Select the WWN tab.

5. In the lower-left tree, select and right-click the Storage System folder.

6. From the pop-up menu, select Add New PFC Group.

7. In the Add New PFC Group window, enter the name of the PFC group and then select OK. An

PFC group is created and an PFC group icon ( ) is added to the lower-left tree.

8. Select an HBA from the upper-left tree and select an PFC group from the lower-left tree.

Next, click Add WWN. Repeat this operation until all desired HBAs are added to the PFC group.

NOTE:

• Select a host bus adapter from below Monitor. You cannot add HBAs from below

Non-Monitor to PFC groups.

• When selecting a host bus adapter that is already contained in some PFC group from

the upper-left tree, the Add WWN button is not activated. Select a host bus adapter that is not contained in any PFC groups.

9. Click Apply. The settings in the window are applied to the storage system.

Deleting an HBA from an PFC group

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

WWN tab operations 65

4. Select the WWN tab.

5. In the lower-left tree, double-click the PFC group ( ) that contains the host bus adapter to be deleted.

6. Below the PFC icon, right-click the icon the host bus adapter ( ) you want to delete.

7. Select Delete WWN from the pop-up menu. The selected host bus adapter icon is deleted from the tree.

8. Click Apply. The settings on the window are applied to the storage system.

Switching priority of an PFC group

All the host bus adapters (HBAs) in one PFC group must be of the same priority. Prioritized WWNs and non-prioritized WWNs cannot be mixed in one PFC group.

You can use an PFC group to switch priority of multiple HBAs from prioritized to non-prioritized, or vice versa.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main window appears.

4. Select the WWN tab.

5. In the lower-left tree, select and right-click an PFC group ( ).

6. Do one of the following:

• To switch priority from prioritized to non-prioritized, select Prio ->> Non-Prio from the

pop-up menu.

• To switch priority from non-prioritized to prioritized, select Non-Prio ->> Prio from the

pop-up menu.

7. Click Apply. The settings in the window are applied to the storage system.

Setting an upper-limit value to HBAs in an PFC group

If all the host bus adapters in an PFC group are non-prioritized WWNs, you can set an upper limit value to HBA performance (i.e., I/O rate or transfer rate). You can assign one upper limit value for one PFC group.

For example, suppose that the upper limit value 100 IOPS is assigned to an PFC group consisting of four host bus adapters. If the sum of the I/O rate of the four HBAs reaches 100 IOPS, Performance Control controls the system so that the sum of the I/O rates will not exceed 100 IOPS.

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main window appears.

4. Select the WWN tab.

5. In the lower-left tree, select and right-click the Storage System folder or an PFC group ( ).

66 Performance Control operations

6. If you selected the Storage System folder, take the following steps:

• Select IOPS or MB/s from the list at the upper-left corner of the list. Select IOPS if you

want to assign an upper limit to the I/O rate. Select MB/s if you want to assign an upper limit to the transfer rate.

• To assign an upper limit to the I/O rate, enter the upper limit value in the IOPS column

of the list. To assign an upper limit to the transfer rate, enter the upper limit value in the

MB/s column of the list.

TIP: If you cannot see the IOPS or MB/s column, scroll the list to the left. The column is

located at the right side of the list.

If you selected an PFC group ( ), take the following steps:

• Right-click the selected PFC group and then select Change Upper Limit from the pop-up

menu. The Change Upper Limit dialog box opens.

• To assign an upper limit to the I/O rate, enter the upper limit value and then select IOPS

from the list. Next, select OK. To assign an upper limit to the transfer rate, enter the upper limit value and then select MB/s from the list. Next, select OK.

7. In the Performance Control main window, click Apply. The settings in the window are applied

to the storage system. To confirm an upper limit value specified for each PFC group, select the Storage System folder

in the lower-left tree of the WWN tab. The PFC groups are displayed in the list and you can confirm each upper limit value.

Renaming an PFC group

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control

window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main

window appears.

4. Select the WWN tab.

5. In the lower-left tree, select and right-click an PFC group ( ).

6. Select Rename PFC Group from the pop-up menu. The Rename PFC Group dialog box opens.

7. Enter the new name and select OK.

WWN tab operations 67

8. In the Performance Control main window, click Apply. The settings in the window are applied to the storage system.

Deleting an PFC group

1. Click Reports > Performance Monitor> Performance Control to open the Performance Control window.

2. Click to change to Modify mode.

3. In the Performance Control window, click Performance Control. The Performance Control main window appears.

4. Select the WWN tab.

5. In the lower-left tree, select and right-click an PFC group ( ).

6. Select Delete PFC Group from the pop-up menu.

7. In the Performance Control main window, click Apply. The settings in the window are applied to the storage system.

68 Performance Control operations

11 Estimating cache size

This topic describes how to estimate the cache size required for using Cache Residency.

About cache size

The required cache size for using Cache Residency differs according to operation modes or RAID levels. For example, if the bind mode is set, RAID1 storage systems require twice the size of cache for the user data to use Cache Residency. However, RAID5 or RAID6 storage systems require three times the size of cache. If external volumes are used, twice the cache size for the user data is required to use Cache Residency.

NOTE: If a RAID5 or RAID6 volume area is changed from priority mode to bind mode and no

cache is added, then only 33% of the user data will fit in the area previously assigned for priority mode, and the remaining 67% is used to save read/write data.

If a RAID1 volume area is changed from priority mode to bind mode and no cache is added, then only 50% of the user data will fit in the area previously assigned for priority mode, and the remaining 50% is used to save read/write data.

Changing the mode without cache extension requires reconfiguring Cache Residency.

About cache size 69

If the priority mode or the bind mode is set, the cache size is calculated assuming that one slot has the following values.

• For open-systems volumes:

For OPEN-V, one slot is 264 KB (512 LBAs).◦

◦ For other than OPEN-V, one slot is 66 KB (128 LBAs).

• For mainframe (3390) volumes:

◦ One slot is 66 KB (128 LBAs).

Calculating cache size for open systems

70 Estimating cache size

1. Calculate the converted values of the starting address and the ending address. For all specified LDEVs:

1. For OPEN-V:

Number of LBAs = LDEV size (KB)×2

Convert the LDEV size to the number of LBAs.

Number of slots = ceil (Number of LBA ÷ 512)

Round up the value that is calculated from the formula enclosed by ceil().

Converted value of starting address = 0 Converted value of ending address = (Number of slots × 512) - 1

2. For emulation types other than OPEN-V:

Number of LBAs = LDEV size (KB) × 2

Convert the LDEV size to the number of LBAs

Number of slots = ceil (Number of LBAs ÷ 96)

Round up the value that is calculated from the formula enclosed by ceil().

Converted value of starting address = 0 Converted value of ending address = (Number of slots × 96) - 1

If the volumes are specified:

1. for OPEN-V:

Starting value = floor (Setting value of starting address (LBA) ÷ 512)

Round down the value that is calculated from the formula enclosed by floor(). "Setting value of starting address (LBA)" is the value which is input on the Cache Residency window.

Ending value = floor (Setting value of ending address (LBA) ÷ 512)

Round down the value that is calculated from the formula enclosed by floor(). "Setting value of ending address (LBA)" is the value which is input on the Cache Residency window.

Converted value of starting address = Starting value × 512 Converted value of ending address = ((Ending value + 1) × 512) - 1

2. For emulation types other than OPEN-V:

Starting value = floor (Setting value of starting address (LBA) ÷ 96)

Round down the value that is calculated from the formula enclosed by floor(). "Setting value of starting address (LBA)" is the value which is input on the Cache Residency window.

Ending value = floor (Setting value of ending address (LBA) ÷ 96)

Round down the value that is calculated from the formula enclosed by floor(). "Setting value of ending address (LBA)" is the value which is input on the Cache Residency window.

Converted value of starting address = Starting value × 96 Converted value of ending address = ((Ending value + 1) × 96) - 1

2. Calculate the number of addresses between the starting address and the ending address calculated in step 1.

1. For OPEN-V:

Number of addresses = Converted value of ending address - Converted value of starting address + 1

Calculate the number of LBAs that are used by the user data.

Calculating cache size for open systems 71

2. For emulation types other than OPEN-V:

Number of LBAs = Converted value of ending address - Converted value of starting address + 1

Calculate the number of LBAs that are used by the user data.

Number of slots = Number of LBAs ÷96

Convert the number of LBAs to the number of slots.

Number of addresses = Number of slots × 128

Convert the number of slots with 128 LBA.

3. Calculate the required cache size according to the operation modes, or the RAID levels to use Cache Residency.

1. Where the bind mode is set:

For RAID1 Required cache size = No. of addresses × (512 + 16) × 2 ÷ 1,024

The unit is KB.

For RAID type other than RAID1: Required cache size = No. of addresses × (512 + 16) × 3 ÷ 1,024

The unit is KB.

2. Where the priority mode is set:

Required cache size = Number of addresses × (512 + 16) ÷ 1,024

The unit is KB.

72 Estimating cache size

Calculating cache size for mainframe systems

1. Calculate the converted values of the starting address and the ending address.

1. For all specified LDEVs:

Setting value of ending address (CC) = floor (((LDEV size × 15) - 1) ÷ 15)

Round down the value that is calculated from the formula enclosed by floor(). "Setting value of ending address (CC)" is the value which is input on the Cache Residency window.

Setting value of ending address (HH) = ((LDEV size × 15) - 1) Mod 15

The remainder will be the setting value of ending address (HH). "Setting value of ending address (HH)" is the value which is input on the Cache Residency

window.

Converted value of starting address = 0 Converted value of ending address = Setting value of ending address (CC) × 15 + Setting

value of ending address (HH)

"Setting value of ending address (CC)" and "Setting value of ending address (HH)" is the value which is input on the Cache Residency window.

2. If the volumes are specified:

Converted value of starting address = Setting value of starting address (CC) × 15 + Setting value of starting address (HH)

"Setting value of ending address (CC)" and "Setting value of ending address (HH)" is the value which is input on the Cache Residency window.

Converted value of ending address = Setting value of ending address (CC) × 15 + Setting value of ending address (HH)

"Setting value of ending address (CC)" and "Setting value of ending address (HH)" is the value which is input on the Cache Residency window.

2. Calculate the number of addresses between the starting address and the ending address calculated in step 1:

Number of addresses = Converted value of ending address - Converted value of starting address + 1

Calculate the number of addresses of cache that are used by the user data.

3. Calculate the required cache size according to the operation modes, or the RAID levels to use Cache Residency.

1. Where the bind mode is set:

For RAID1 Required cache size = (Number of addresses × ((128 × (512 + 16)) × 2) ÷ 1,024

The unit is KB.

2. Where the priority mode is set:

Required cache size = (Number of addresses × (128 × (512 + 16))) ÷ 1,024

The unit is KB.

Cache Residency cache areas

The Cache Residency cache areas have the following parameters:

• The cache areas are dynamic and can be added and deleted at any time.

• The HP XP P9500 supports a maximum of 1,024 addressable cache areas per LDEV and per

storage system.

Calculating cache size for mainframe systems 73

• For OPEN-V volumes, Cache Residency cache areas must be defined in logical blocks using

logical block addresses (LBAs), with a minimum size of 512 LBAs (equivalent to 264 KB). In most cases you will assign an entire open-systems volume for cache residency. If the remaining cache memory is less than 256 MB, Cache Residency is not available.

• For mainframe volumes, Cache Residency cache areas must be defined on contiguous tracks

with a minimum size of one cache slot (or track) (equivalent to 66 KB) and a maximum size of one LVI.

• You can prestage the data to the resident cache area. If prestaging is not used, the data will

be loaded into the Cache Residency area when the first “miss” occurs. If prestaging is used, performance may be affected for a short time while the data is read into Cache Residency cache.

CAUTION: Prestaging of Cache Residency data should not be performed during peak

activity.

• All write I/Os to Cache Residency data are duplex writes, guaranteeing full data integrity.

The Cache Residency data remains fixed in cache until you manually delete it. Deletion of Cache Residency cache areas destages any write data to the affected volumes.

• It is possible to expand the amount of Cache Residency cache without canceling the existing

Cache Residency settings. For details, contact HP Technical Support.

Cache Residency system specifications

types

Unit of cache area allocation

SpecificationItem

OPEN-3, 8, 9, E, L

Virtual LUN volume

For OPEN-V, at least 512 LBAs: Equivalent to 264 KB

For other than OPEN-V, at least 96 LBAs: Equivalent to 66 KB

Per storage system: 16,384Number of cache areas Per LDEV: 4,096

Minimum 512 MBTotal cache capacity

Mainframe systemOpen system

3380-3, 3A, 3B, 3C,OPEN-VSupported device emulation 3390-3, 3A, 3B, 3C, 3R, 9, 9A, 9B, 9C, L,

LA, LB, LC, M, MA, MB, MC, A

Virtual LVI volumeLUN Expansion (LUSE) volumeSupported volume types

At least one cache slot (or track): Equivalent to 66 KB.Up to 1 LDEV.

74 Estimating cache size

12 Managing resident cache

This topic provides instructions for using Cache Residency software to manage resident cache.

Cache Residency rules, restrictions, and guidelines

Rules

• Cache Residency must be enabled on Remote Web Console.

• Administrator or Cache Residency write access to the Remote Web Console software is required

to perform Cache Residency operations. Users without write access can view Cache Residency information for the connected storage system but cannot set or change options.

• Do not attempt to allocate Cache Residency cache beyond the allocated capacity.

• Do not apply Cache Residency settings to volumes reserved for Auto LUN V2.

• Do not attempt to allocate Cache Residency cache redundantly over the cache area that is

already allocated to an LDEV.

• Do not apply or refer to Cache Residency settings to volumes from the host and Remote Web

Console at the same time. You can apply the settings from the host if you use Cache Manager.

• If you specify the Cache Residency setting on the volume during quick formatting, do not use

the prestaging function. If you want to use the prestaging function after the quick formatting processing completes, release the setting and then specify the Cache Residency setting again, with the prestaging setting enabled this time. For information about quick formatting, see the

HP XP P9000 Provisioning for Open Systems User Guide or the HP XP P9000 Provisioning for Mainframe Systems User Guide.

• Do not perform the Business Copy quick restore operation or the Auto LUN V2 operation on

a Cache Residency volume. Also, do not specify the Cache Residency setting on the volume on which the Business Copy quick restore or Auto LUN V2 operation is performed. These operations swap the internal locations of the source and target volumes, which causes a loss of data integrity. For additional information, see the HP XP P9000 Business Copy User Guide, the HP XP P9000 Business Copy for Mainframe User Guide, and/or contact HP Technical Support.

• To set Cache Residency for a LUSE volume, you must set Cache Residency for an LDEV that

is a component of the LUSE volume. To determine the LDEV for which you want to set Cache Residency, you must know the exact number of LBAs in each LDEV that is a component of the LUSE volume.

NOTE: The number of LBAs displayed on the Cache Residency window is different from the

actual number of LDEVs, and does not match the number of LBAs recognized by the host.

To identify the exact number of LBAs in a LDEV, first display the Remote Web Console main window and search for the parity group to which the LDEV belongs according to the LDKC, control unit (CU), and LDEV numbers. For further information about the Basic Information Display window, see the HP XP P9000 Remote Web Console User Guide.

As shown in Figure 18 (page 76), the LUSE volume is composed of three LDEVs, for example, 00:01 (1,000 LBAs), 00:02 (1,002 LBAs), and 00:03 (1,020 LBAs). If you see from the host and want to set Cache Residency to 500 LBAs starting from No. 1,020 LBA, you may set the Cache Residency to 500 LBAs starting from No. 20 LBA of the 2nd LDEV because the first LDEV size is 1,000 LBAs.

Cache Residency rules, restrictions, and guidelines 75

The following operations automatically reset Cache Residency cache:

◦ When LDEVs that is partly or wholly assigned to Cache Residency is deleted.

◦ When the parity group containing LDEVs that are assigned to Cache Residency is deleted.

Figure 18 Example of LBA Value Setting When Using LUSE

Restrictions

• The Cache Residency bind mode is not available to external volumes whose Cache mode is

set to Disable (which is the mode that disables the use of the cache when there is an I/O request from the host).

• You cannot allocate pool-VOLs and V-VOLs for Cache Residency. For further information about

pool-VOLs and V-VOLs, see the HP XP P9000 Snapshot User Guide, the HP XP P9000 Provisioning for Open Systems User Guide, or the HP XP P9000 Provisioning for Mainframe Systems User Guide

• You cannot allocate the journal volumes for Cache Residency. For additional information

about the journal volumes, see the HP XP P9000 Continuous Access Journal User Guide or the HP XP P9000 Continuous Access Journal for Mainframe Systems User Guide.

• You cannot allocate the remote command device for Cache Residency. For further information

about the remote command device, see the HP XP P9000 External Storage for Open and Mainframe Systems User Guide.

• You cannot allocate a quorum disk for Cache Residency.

• You cannot allocate the nondisruptive migration volumes for Cache Residency.

Guidelines

• Performing Cache Residency operations on many LDEVs during host I/O may cause the host

I/O response time to become slow. To avoid degradation of response time, set only one LDEV at a time.

• Deleting data from cache during host I/O may cause the response time of host I/O to become

slow. To avoid degradation of host response time, limit the amount of data you delete in one operation as follows:

If the host timeout period is set to 10 seconds or shorter, limit the total amount of data to:

◦ 1 GB or less for open systems

◦ 1,000 cylinders or less for mainframe systems

76 Managing resident cache

If the host timeout period is set to 11 seconds or longer, limit the total amount of data to:

◦ 3 GB or less for open systems

◦ 3,000 cylinders or less for mainframe systems

Launching Cache Residency

1. Log on to the primary SVP.

2. On the menu bar of the Remote Web Console main window, click Actions > Other function > Cache Residency.

3. In the Cache Residency window, change from View to Modify mode.

Viewing Cache Residency information

The Cache Residency information can be viewed in the following fields in the Cache Residency window:

• CU:LDEV tree

• LDEV information table

• Cache information area

Placing specific data into Cache Residency cache

This procedure places specific data from one or more LDEVs into Cache Residency cache.

1. In the Cache Residency window, select the desired CLPR from the CLPR: list.

2. In the CU:LDEV tree, select the LDKC and the CU containing the desired LDEV, and then select the desired LDEV.

The LDEV information table shows the information for the selected LDEV. A dash (-) in the Mode column indicates an area not already allocated to Cache Residency cache.

Launching Cache Residency 77

3. Select an unallocated area in the LDEV information table as the area to place specific data from one or more LDEVs into Cache Residency cache. The starting and ending addresses of the selected area appear in the Start and End fields.

NOTE: For OPEN-V LUs, Cache Residency identifies a logical area in units of 512 blocks.

If you enter 0 or 1 as the starting LBA and a value less than 511 as the ending LBA, Cache Residency automatically changes the ending block address to 511.

4. In the Cache Residency window, select options to apply to all selected LDEVs:

1. In the Cache Residency Mode box, select the desired mode (Bind or Priority).

2. Select the desired Prestaging Mode setting (Yes or No). To set the prestaging function,

the Prestaging check box must already be selected.

3. Verify the starting and ending addresses of the area to be placed in Cache Residency

cache in the Start and End fields. Edit as needed. Make sure that the Select All Area box is NOT checked.

CAUTION: Make sure to select the correct options because the options cannot be

changed after data is added to cache. To change between bind/priority modes, or to enable/disable the prestaging function, release the cache extent that you want to change, and then place the data back into Cache Residency cache with the desired settings.

5. If you do not want to apply the same options to any other LDEV, make sure that the Multi Set / Release box is not checked, click Set, and then click OK on the confirmation dialog box. The requested Cache Residency operation appears in blue in the LDEV information table.

To apply the same options and data range to additional LDEVs:

1. On the Cache Residency window, select the Multi Set / Release box, click Set, and then

click OK. The Multi Set dialog box opens showing the data range and options selected on the Cache Residency window.

2. In the Multi Set dialog box, select the desired LDKC and CU image, and select the desired

LDEVs. The data range and options displayed in the dialog box will be applied to all selected LDEVs.

3. Click Set to return to the Cache Residency window. The requested Cache Residency

operations appear in blue in the LDEV information table.

6. Repeat steps 2 through 5 until all desired operations are listed. The Release is unavailable until you apply (or cancel) your requested operations.

78 Managing resident cache

7. Verify the Prestaging setting:

• To enable prestaging, select Prestaging.

• To disable prestaging, clear Prestaging.

8. To start the operations, click Apply. If Prestaging was selected, respond to the Yes/No confirmation. To continue with prestaging, click Yes. To continue without it, click No.

9. Monitor the Cache Residency window to make sure that the operations complete successfully. The cache information area shows the progress of the requested operations.

Placing LDEVs into Cache Residency cache

This procedure places all data on one or more LDEVs into Cache Residency cache.

1. In the Cache Residency window, select the desired CLPR from the CLPR list.

2. In the CU:LDEV tree, select the LDKC and the CU containing the desired LDEV, and then select the desired LDEV.

The LDEV information table shows the information for the selected LDEV. A dash (-) in the Mode column indicates an area not already allocated to Cache Residency cache.

3. In the Cache Residency window, select desired options:

1. In the Cache Residency Mode box, select the desired mode (Bind or Priority).

2. Select the desired Prestaging Mode setting (Yes or No). To set the prestaging function,

the Prestaging check box must already be selected.

3. Check the Select All Area box. Leave the Start and End fields blank.

CAUTION: Make sure to select the correct options, because the options cannot be

changed after data is added to cache. To change between bind and priority modes, or to enable and disable the prestaging function, release the cache area that you want to change, and then place the data back into Cache Residency cache with the desired settings.

4. If you do not want to apply the same options to any other LDEVs, make sure that the Multi Set / Release box is not checked, click Set, and then click OK on the confirmation dialog box. The requested operation appears in blue in the LDEV information table.

To apply the same options to additional LDEVs:

1. In the Cache Residency window, select the Multi Set / Release box, click Set, and then

click OK. The Multi Set dialog box opens showing the data range and options selected on the Cache Residency window.

2. In the Multi Set dialog box, select the desired CU image, and select the desired LDEVs.

The options displayed on the dialog box will be applied to all selected LDEVs.

3. Click Set to return to the Cache Residency window. The requested Cache Residency

operations appears in blue in the LDEV information table.

5. Repeat steps 2 through 4 until all desired operations are listed. The Release is unavailable until you apply (or cancel) your requested operations.

6. Verify the Prestaging setting:

• To enable prestaging, select Prestaging.

• To disable prestaging, clear Prestaging.

7. To start the operations, click Apply:

• If Prestaging was selected, respond to the Yes/No confirmation. To continue with

prestaging, select Yes. To continue without it, select No.

• To cancel the operation, click Cancel and click OK on the confirmation.

Placing LDEVs into Cache Residency cache 79

8. Monitor the Cache Residency window to make sure that the operations complete successfully.

The cache information area shows the progress of the requested operations.

Releasing specific data from Cache Residency cache

This procedure releases specific data areas on one or more LDEVs from Cache Residency cache.

1. In the Cache Residency window, select the desired CLPR from the CLPR list.

2. In the CU:LDEV tree, select the LDKC and the CU containing the desired LDEV, and then select the desired LDEV.

The LDEV information table shows the information for the selected LDEV. The Mode column indicates Prio or Bind for each data area that is allocated to Cache Residency cache.

3. Select the data areas that you want to release from Cache Residency cache. This enables the Release.

4. Click Release, and click OK on the confirmation message. The requested operation is displayed in blue in the LDEV information table.

5. Repeat steps 2 through 4 for each LDEV for which you want to release specific data from Cache Residency cache. The Set is unavailable until you apply (or cancel) your requested operations.

6. Verify the Prestaging setting:

• To enable prestaging, select Prestaging.

• To disable prestaging, clear Prestaging.

7. To start the operations, click Apply:

• If Prestaging was selected, respond to the Yes/No confirmation. To continue with

prestaging, select Yes. To continue without it, select No.

• To cancel the operation, click Cancel and click OK on the confirmation.

8. When the delete confirmation message appears, click OK to begin the deletion, or click Cancel to cancel your request to delete data.

9. Monitor the Cache Residency window to make sure that the operations complete successfully. The cache information area shows the progress of the requested operations.

When the data has been released, the verification window will appear.

Releasing LDEVs from Cache Residency cache

This procedure releases ALL data on one or more LDEVs from Cache Residency cache.

1. In the Cache Residency window, select the desired CLPR from the CLPR list.

2. In the CU:LDEV tree, select the LDKC and the CU containing the desired LDEV, and then select the desired LDEV.

The LDEV information table shows the information for the selected LDEV. The Release is available if the selected LDEV has data that is stored in Cache Residency cache (indicated by Prio or Bind in the Mode column).

3. If you do not want to release any other LDEVs from Cache Residency cache, make sure that the Multi Set / Release box is not checked, click Release, and then click OK on the confirmation dialog box. The requested operation appears in blue in the LDEV information table.

To release additional LDEVs from Cache Residency cache:

80 Managing resident cache

1. Check the Multi Set / Release box, click Release, and then click OK on the confirmation

message.

2. In the Multi Release dialog box, select the desired LDKC and CU image, and select the

desired LDEVs to release from Cache Residency cache.

3. Click Release to return to the Cache Residency window. The requested Cache Residency

operations appear in blue in the LDEV information table.

4. Repeat steps 2 and 3 until all desired operations are listed.

NOTE: The Set is unavailable until you apply (or cancel) your requested operations.

5. Verify the Prestaging setting:

• To enable prestaging, select Prestaging.

• To disable prestaging, clear Prestaging.

6. To start the operations, click Apply:

• If Prestaging was selected, respond to the Yes/No confirmation. To continue with

prestaging, click Yes. To continue without it, click No.

• To cancel the operation, click Cancel and click OK on the confirmation.

7. Monitor the Cache Residency window to make sure that the operations complete successfully.

The cache information area shows the progress of the requested operations.

Changing mode after Cache Residency is registered in cache

If Cache Residency is registered in the cache, the following mode options appear gray and are unavailable for change:

• Cache Residency Mode (Bind, Priority)

• Prestaging Mode (Yes, No)

To change the mode options:

1. Release the specific data from the Cache Residency cache. For details, see “Releasing specific

data from Cache Residency cache” (page 80).

2. Restore the data with the new settings. For details, see “Placing specific data into Cache

Residency cache” (page 77).

Changing mode after Cache Residency is registered in cache 81

13 Troubleshooting

This topic provides references to sources of troubleshooting and contact information for HP Technical Support.

Troubleshooting resources

For troubleshooting information on the XP P9500, see the HP XP P9000 Owner Guide. For troubleshooting information on the Remote Web Console software, see the HP XP P9000

Remote Web Console User Guide. For information on Remote Web Console error codes, see the HP XP P9000 Remote Web Console

Messages.

82 Troubleshooting

14 Support and other resources

Contacting HP

For worldwide technical support information, see the HP Support Center:

http://www.hp.com/go/hpsc

Before contacting HP, collect the following information:

• Product model names and numbers

• Technical support registration number (if applicable)

• Product serial numbers

• Error messages

• Operating system type and revision level

• Detailed questions

Subscription service

Receive, by email, support alerts announcing product support communications, driver updates, software releases, firmware updates, and customer-replaceable component information by signing up at http://www.hp.com/go/myadvisory.

To change options for support alerts you already receive, click the Sign in link on the right.

Documentation feedback

HP welcomes your feedback. To make comments and suggestions about product documentation, please send a message to

storagedocsfeedback@hp.com. Include the document title and manufacturing part number. All

submissions become the property of HP.

Related information

The following documents provide related information:

• HP XP P9000 Remote Web Console Messages

• HP XP P9000 Remote Web Console User Guide

You can find these documents on the Manuals page of the HP Business Support Center website:

http://www.hp.com/support/manuals

In the Storage section, click Disk Storage Systems for hardware or Storage Software for software, and then select your product.

HP websites

For additional information, see the following HP websites:

• http://www.hp.com

• http://www.hp.com/go/storage

• http://www.hp.com/service_locator

• http://www.hp.com/support/manuals

• http://www.hp.com/support/downloads

• http://www.hp.com/storage/whitepapers

Contacting HP 83

Conventions for storage capacity values

XP P9000 disk arrays use the following values to calculate physical storage capacity values (hard disk drives):

• 1 KB (kilobyte) = 1,000 bytes

• 1 MB (megabyte) = 1,0002bytes

• 1 GB (gigabyte) = 1,0003bytes

• 1 TB (terabyte) = 1,0004bytes

• 1 PB (petabyte) = 1,0005bytes

• 1 EB (exabyte) = 1,0006bytes

XP P9000 disk arrays use the following values to calculate logical storage capacity values (logical devices):

• 1 KB (kilobyte) = 1,024 bytes

• 1 MB (megabyte) = 1,0242bytes

• 1 GB (gigabyte) = 1,0243bytes

• 1 TB (terabyte) = 1,0244bytes

• 1 PB (petabyte) = 1,0245bytes

• 1 EB (exabyte) = 1,0246bytes

Typographic conventions

Table 7 Document conventions

Blue text: Table 7 (page 84)

Bold text

Monospace text

ElementConvention

• Cross-reference links and e-mail addresses

• A cross reference to the glossary definition of the term

in blue text

email addressesBlue, bold, underlined text

Website addressesBlue, underlined text: http://www.hp.com

• Keys that are pressed

• Text typed into a GUI element, such as a box

• GUI elements that are clicked or selected, such as menu

and list items, buttons, tabs, and check boxes

Text emphasisItalic text

• File and directory names

• System output

• Code

• Commands, their arguments, and argument values

Monospace, italic text

WARNING! Indicates that failure to follow directions could result in bodily harm or death.

84 Support and other resources

• Code variables

• Command variables

Emphasized monospace textMonospace, bold text

CAUTION: Indicates that failure to follow directions could result in damage to equipment or data.

IMPORTANT: Provides clarifying information or specific instructions.

NOTE: Provides additional information.

TIP: Provides helpful hints and shortcuts.

Typographic conventions 85

A Export Tool

This topic explains how to export the monitoring data collected on your storage system into files.

About the Export Tool

Use the Export Tool to export the monitoring data (statistics) shown in the Monitor Performance window to text files. You can also use the Export Tool to export monitoring data on remote copy operations performed by Continuous Access Synchronous, Continuous Access Synchronous Z, Continuous Access Journal, and Continuous Access Journal Z. After exporting monitoring data to text files, you can import that data into desktop publishing applications, such as Microsoft Word, or into spreadsheet or database applications for analysis.

Example of a text file

The following example is of a text file imported into spreadsheet software.

NOTE: In this LU_IOPS.csv file, the last four digits of a table column heading (such as 0001

and 0002) indicate a LUN. For example, the heading CL1-A.00(1A-G00).0001 indicates the port CL1-A, the host group ID 00, the host group name 1A-G00, and the LUN 0001.

If you export monitoring data about concatenated parity groups, the resulting CSV file does not contain column headings for the concatenated parity groups. For example, if you export monitoring data about a concatenated parity group named 1-3[1-4], you will be unable to find 1- 3[1-4] in column headings. To locate monitoring data about 1-3[1-4], find the 1-3 column or the 1-4 column. Either of these columns contains monitoring data about 1-3[1-4].

Installing the Export Tool

• “System requirements” Page 86

• “Installing the Export Tool on a Windows system” (page 87)

• “Installing the Export Tool on a UNIX system” (page 87)

System requirements

The following components are required to use the Export Tool (for more information, see the HP P9000 Remote Web Console User Guide):

• A Windows system or a UNIX system

The Export Tool runs on Windows systems and UNIX systems that can run the Remote Web Console software. If your Windows or UNIX system is unable to run Remote Web Console, your system is unable to run the Export Tool.

86 Export Tool

NOTE: If a firewall exists between the Remote Web Console computer and the SVP, see

Chapter 2 of the HP P9000 Remote Web Console User Guide. In the section “Setting up TCP/IP for a firewall”, the RMI port numbers listed are the only direct communication settings required for the Export Tool.

• The Java Runtime Environment (JRE)

To be able to use the Export Tool, you must install Java Runtime Environment on your Windows or UNIX system. If your system runs Remote Web Console, JRE is already installed on your system and you can install the Export Tool. If your system does not run Remote Web Console but contains an appropriate version of JRE, you can install the Export Tool on your system.

The JRE version required for running the Export Tool is the same as the JRE version required for running Remote Web Console.

• A user ID for exclusive use of the Export Tool

Before you can use the Export Tool, you must create a user ID for exclusive use of the Export Tool. Assign only the Storage Administrator (Performance Management) role to the user ID for the Export Tool. It is recommended that you do not assign any roles other than the Storage Administrator (Performance Management) role to this user ID. The user who is assigned to the Storage Administrator (Performance Management) role can do the following:

◦ Save the monitoring data into files

◦ Change the gathering interval

◦ Start or stop monitoring by the set subcommand

For details on creating the user ID, see the HP P9000 Remote Web Console User Guide.

• The Export Tool program

CD-ROM Disc 2, which is named Host PP, contains the Export Tool software. For instructions on installing the Export Tool, see:

◦ “Installing the Export Tool on a Windows system” (page 87)

◦ “Installing the Export Tool on a UNIX system” (page 87)

Installing the Export Tool on a Windows system

The Export Tool program is a Java class file and is located in the export\lib folder.

1. Log on with administrator privileges.

2. Create a new folder for the Export Tool application (for example, C:\Program Files\ monitor). If this folder already exists, skip this step.

3. Insert the Export Tool CD-ROM into the CD drive.

4. Locate the \program\monitor\win_nt folder on the CD-ROM, and copy the self-extracting file export.exe from the CD-ROM into the new folder you just created.

5. Double-click export.exe to start the installation. The Export Tool is installed, and a new folder named “export” is created.

Installing the Export Tool on a UNIX system

The Export Tool program is a Java class file and is located in the lib directory.

1. Log on as a superuser. You do not need to remove a previous installation of Export Tool. The new installation overwrites

the older program.

2. Create a new directory for the Export Tool program (for example, /monitor).

3. Mount the Export Tool CD-ROM.

Installing the Export Tool 87

4. Go to the /program/monitor/UNIX directory on the CD-ROM, and copy the export.tar file to the new directory you just created.

5. Decompress the export.tar file on your system. The Export Tool is installed into the installation directory.

Using the Export Tool

To be able to export monitoring data, you must first prepare a command file and a batch file, and then you can run the Export Tool to export monitoring data.

• “Preparing a command file” (page 88)

• “Preparing a batch file” (page 91)

• “Running the Export Tool” (page 92)

Preparing a command file

Before you run the Export Tool, you must write scripts for exporting monitoring data. When writing scripts, you need to write several subcommands in a command file. When you run the Export Tool, the subcommands in the command file are executed sequentially, and then the monitoring data is saved in files.

Example of a command file

svpip 158.214.135.57 ; Specifies IP address of SVP login expusr passwd ; Logs user into SVP show ; Outputs storing period to standard ; output group PhyPG Long ; Specifies type of data to be ; exported and type of ; storing period group RemoteCopy ; Specifies type of data to be ; exported short-range 201210010850:201210010910 ; Specifies term of data to be ; exported for data stored ; in short range long-range 201209301430:201210011430 ; Specifies term of data to be ; exported for data stored ; in long range outpath out ; Specifies directory in which files ; will be saved option compress ; Specifies whether to compress files apply ; Executes processing for saving ; monitoring data in files

A semicolon (;) indicates the beginning of a comment. Characters from a semicolon to the end of the line are comments.

The scripts in this command file are explained as follows:

• svpip 158.214.135.57

• login expusr passwd

88 Export Tool

This script specifies that you are logging into SVP whose IP address is 158.214.135.57. You must log into SVP when using the Export Tool.

The svpip subcommand specifies the IP address of SVP. You must include the svpip subcommand in your command file. For detailed information about the svpip subcommand, see “svpip” (page 96).

This script specifies that you use the user ID expusr and the password passwd to log into SVP.

The login subcommand logs the specified user into SVP. You must include the login subcommand in your command file. For detailed information about the login subcommand, see “login” (page 98).

CAUTION: When you write the login subcommand in your command file, you must specify

a user ID that should be used exclusively for running the Export Tool. See “System requirements”

(page 86) for reference.

• show

The show subcommand checks SVP to find the period of monitoring data stored in SVP and the data collection interval (called gathering interval in Performance Monitor), and then outputs them to the standard output (for example, the command prompt) and the log file.

Performance Monitor collects statistics by the two types of storing periods: in short range and in long range. The show subcommand displays the storing periods and the gathering intervals for these two types of monitoring data.

The following is an example of information that the show subcommand outputs:

Short Range From: 2012/10/01 01:00 - To: 2012/10/01 15:00 Interval: 1min. Long Range From: 2012/09/01 00:00 - To: 2012/10/01 15:00 Interval: 15min.

Short Range indicates the storing period and gathering interval of the monitoring data

stored in short range. Long Range indicates those of the monitoring data stored in long range. In the above example, the monitoring data in short range is stored every 1 minute in the term of 1:00-15:00 on Oct. 1, 2012. Also, the monitoring data in long range is stored every 15 minutes in the term of Sep. 1, 2012, 0:00 through Oct. 1, 2012, 15:00. When you run the Export Tool, you can export monitoring data within these periods into files.

All the monitoring items are stored in short range, but a part of monitoring items is stored in both the short range and long range. For details on monitoring items that can be stored in long range, see “long-range” (page 111).

The use of the show subcommand is not mandatory, but it is recommended that you include the show subcommand in your command file. If an error occurs when you run the Export Tool, you might be able to find the error cause by checking the log file for information issued by the show subcommand. For detailed information about the show subcommand, see “show”

(page 98).

• group PhyPG Long and group RemoteCopy

The group subcommand specifies the type of data that you want to export. Specify a operand following group to define the type of data to be exported. Basically, monitoring data stored in short range is exported. But you can direct to export monitoring data stored in long range when you specify some of the operands.

The example script group PhyPG Long in “Preparing a command file” (page 88) specifies to export usage statistics about parity groups in long range. Also, the script group RemoteCopy specifies to export statistics about remote copy operations by Continuous Access Synchronous and Continuous Access Synchronous Z in short range. You can describe multiple lines of the group subcommand to export multiple monitoring items at the same time.

For detailed information about the group subcommand, see “group” (page 99).

• short-range 201210010850:201210010910 and long-range

201209301430:201210011430

The short-range and long-range subcommands specify the term of monitoring data to be exported. Use these subcommands when you want to narrow the export-target term within

Using the Export Tool 89

the stored data. You can specify both the short-range and long-range subcommands at the same time. The difference between these subcommands is as follows:

◦ The short-range subcommand is valid for monitoring data in short range. You can

use this subcommand to narrow the export-target term for all the monitoring items you can specify by the group subcommand.

Specify a term within "Short Range From XXX To XXX" which is output by the show subcommand.

◦ The long-range subcommand is valid for monitoring data in long range. You can use

this subcommand only when you specify the PhyPG, PhyLDEV, PhyProc, or PhyESW operand with the Long option in the group subcommand. (The items that can be saved by these operands are the monitoring data displayed in the Physical tab of the Performance Management window with selecting longrange.)

Specify a term within "Long Range From XXX To XXX" which is output by the show subcommand.

In the sample file in “Preparing a command file” (page 88), the script short-range 201210010850:201210010910 specifies the term 8:50-9:10 on Oct. 1, 2012. This script is applied to the group RemoteCopy subcommand in this example. When you run the Export Tool, it will export the statistics about remote copy operations by Continuous Access Synchronous and Continuous Access Synchronous Z in the term specified by the short-range subcommand.

Also, in “Preparing a command file” (page 88), the script long-range

201209301430:201210011430 specifies the term from Sep. 30, 2012, 14:30 to Oct. 1, 2012, 14:30. This script is applied to the group PhyPG Long subcommand in this example.

When you run the Export Tool, it will export the usage statistics about parity groups in the term specified by the long-range subcommand.

If you run the Export Tool without specifying the short-range or long-range subcommand, the monitoring data in the whole storing period (data in the period displayed by the show subcommand) will be exported.

◦ For detailed information about the short-range subcommand, see “Short-range”

(page 109).

◦ For detailed information about the long-range subcommand, see “long-range”

(page 111).

• outpath out

This script specifies that files should be saved in the directory named out in the current directory.

The outpath subcommand specifies the directory in which files should be saved. For detailed information about the outpath subcommand, see “outpath” (page 113).

• option compress

This script specifies that the Export Tool should compress monitoring data in ZIP files. The option subcommand specifies whether to save files in ZIP format or in CSV format. For

detailed information about the option subcommand, see “option” (page 113).

• apply

The apply subcommand saves monitoring data in files. For detailed information about the apply command, see “apply” (page 114).

When you install the Export Tool, the command.txt file is stored in the installation directory. This file contains sample scripts for your command file. It is recommended that you customize scripts in command.txt according to your needs. For details about subcommand syntax, see “Export

Tool command reference” (page 95).

90 Export Tool

Preparing a batch file

A batch file is used to run the Export Tool. The Export Tool starts and saves monitoring data in files when you execute the batch file.

The installation directory for the Export Tool contains two default batch files: runWin.bat for Windows systems, and runUnix.bat for UNIX systems.

The following examples illustrate scripts in the runWin.bat and runUnix.bat batch files. These batch files include a command line that executes a Java command. When you execute the batch file, the Java command executes the subcommands specified in the command file and then saves monitoring data in files.

Example batch file for Windows systems (runWin.bat):

java -classpath "./lib/JSanExport.jar;./lib/JSanRmiApiEx.jar;./lib/JSanRmiServerUx.jar"

-Xmx536870912 -Dmd.command=command.txt -Dmd.logpath=log

sanproject.getmondat.RJMdMain<CR+LF> pause<CR+LF> Example batch file for UNIX systems (runUnix.bat): #! /bin/sh<LF> java -classpath

"./lib/JSanExport.jar:./lib/JSanRmiApiEx.jar:./lib/JSanRmiServerUx.jar"

-Xmx536870912 -Dmd.command=command.txt -Dmd.logpath=log

sanproject.getmondat.RJMdMain<LF>

In the previous scripts, <CR+LF> and <LF> indicate the end of a command line. If the system running the Export Tool communicates directly with SVP, you usually do not need to

change scripts in runWin.bat and runUnix.bat. However, you might need to edit the Java command script in a text editor in some cases, for example:

• if the name of your command file is not command.txt

• if you moved your command file to a different directory

• if you do not want to save in the log directory

• if you want to name log files as you like

If the system that runs the Export Tool communicates with SVP via a proxy host, edit the Java command script in a text editor. to specify the host name (or the IP address) and the port number of the proxy host. For example, if the host name is Jupiter and the port number is 8080, the resulting command script would be as shown in the following examples:

Example of specifying a proxy host on Windows (runWin.bat):

java -classpath "./lib/JSanExport.jar;./lib/JSanRmiApiEx.jar;./lib/JSanRmiServerUx.jar"

-Dhttp.proxyHost=Jupiter -Dhttp.proxyPort=8080 -Xmx536870912

-Dmd.command=command.txt -Dmd.logpath=log sanproject.getmondat.RJMdMain

<CR+LF> pause <CR+LF> Example of specifying a proxy host on UNIX (runUnix.bat): #! /bin/sh <LF> java -classpath

"./lib/JSanExport.jar:./lib/JSanRmiApiEx.jar:./lib/JSanRmiServerUx.jar"

-Dhttp.proxyHost=Jupiter -Dhttp.proxyPort=8080 -Xmx536870912

-Dmd.command=command.txt -Dmd.logpath=log sanproject.getmondat.RJMdMain

<LF>

In the preceding scripts, <CR+LF> and <LF> indicates the end of a command line. If the IP address of the proxy host is 158.211.122.124 and the port number is 8080, the resulting

command script is as follows: Example batch file for Windows systems (runWin.bat):

java -classpath "./lib/JSanExport.jar;./lib/JSanRmiApiEx.jar;./lib/JSanRmiServerUx.jar"

-Dhttp.proxyHost=158.211.122.124 -Dhttp.proxyPort=8080 -Xmx536870912

Using the Export Tool 91

-Dmd.command=command.txt -Dmd.logpath=log sanproject.getmondat.RJMdMain <CR+LF> pause <CR+LF>

Example batch file for UNIX systems (runUnix.bat):

#! /bin/sh <LF> java -classpath "./lib/JSanExport.jar:./lib/JSanRmiApiEx.jar:./lib/JSanRmiServerUx.jar"

-Dhttp.proxyHost=158.211.122.124 -Dhttp.proxyPort=8080 -Xmx536870912

-Dmd.command=command.txt -Dmd.logpath=log sanproject.getmondat.RJMdMain <LF>

In the above scripts, <CR+LF> and <LF> indicates the end of a command line. For detailed information about syntax of the Java command, see “Java” (page 116).

Running the Export Tool

CAUTION: Running multiple instances of the Export Tool simultaneously is not supported. If you

run multiple instances, the SVP might become overloaded and a timeout error might occur.

To save monitoring data in files, launch the Export Tool by running the batch file: From a system running UNIX, enter the name of the batch file at the command prompt, and then

press the <Enter> key. From a system running Windows, double-click the batch file to run it.

c:\WINDOWS> cd c:\export c:\export> runWin.bat

Dots (...) appear on the screen until the system finishes exporting data. If an internal error occurs, an exclamation mark (!) appears, and then the Export Tool restarts automatically.

Example of command prompt outputs from the Export Tool:

[ 2] svpip 158.214.135.57 [ 3] login User = expusr, Passwd = [****************] : : [ 6] group Port : : [20] apply Start gathering port data Target = 16, Total = 16 +----+----+----+----+----+----+----+----+----+----+

...........................!

.................................

End gathering port data

By default, the system compresses monitoring data files into a ZIP-format archive file. When you want to view the monitoring data, you can decompress and extract the CSV files from the ZIP archive. If your system is not able to extract files from a ZIP archive, you need to obtain software to view the data.

NOTE: You can change the default method of exporting files to an uncompressed format. However,

the resulting files could be significantly larger and take longer to compile. For more information, see option Subcommand “option” (page 113).

For a complete list of files to be saved by the Export Tool, see “Using the Export Tool” (page 88).

File formats

If you specify the nocompress operand for the option subcommand, the Export Tool saves files in CSV format instead of ZIP format (For detailed information, see option Subcommand “option”

(page 113)). When files are saved in CSV format instead of ZIP format, the file saving process could

take longer and the resulting files could be larger.

92 Export Tool

Processing time

Files saved by the Export Tool are often very large. The total file size for all the files can be as large as approximately 2 GB. For this reason, the exporting process might take a lot of time. If you want to export statistics spanning a long period of time, it is recommended that you run the Export Tool multiple times for different periods, rather than run it one time to export the entire time span as a single large file. For example, if you want to export statistics spanning 24 hours, run the tool eight times to export statistics in three 3- hour increments.

The next table provides lists time estimates for exporting monitoring data files using different operands in the group subcommand:

Table 8 Estimate of time required for exporting files

subcommand

RemarksEstimated timeOperand for the group

5 minutesPort

5 minutesPortWWN

This estimate assumes that the Export Tool should save statistics about 128 ports within a 24-hour period.

This estimate assumes that:60 minutesLDEV

• The Export Tool should save statistics about 8,192 volumes within a

24-hour period.

• The Export Tool is used eight times. Each time the Export Tool is used,

the tool obtains statistics within a 3-hour period.

This estimate assumes that:60 minutesLU

• The Export Tool should save statistics about 12,288 LUs within a

24-hour period.

• The Export Tool is used eight times. Each time the Export Tool is used,

the tool obtains statistics within a 3-hour period.

NOTE:

• The estimated time in the table is for the 1-minute interval of data collection. If the interval is

2 minutes, "a 24-hour period" in the table will be "a 48-hour period", because the interval is proportional to the stored time.

• The estimated time that includes the transfer time of the network might take a lot of time

depending on the transmission speed of the network.

• To shorten the acquisition time, specify the option of the group command to narrow acquisition

objects. For details about the group command, see “group” (page 99).

Termination code

If you want to use a reference to a termination code in your batch file, do the following:

• To use such a reference in a Windows batch file, write %errorlevel% in the batch file.

• To use such a reference in a UNIX Bourne shell script, write $? in the shell script.

• To use such a reference in a UNIX C shell script, write $status in the shell script.

A reference to a termination code is used in the following example of a Windows batch file. If this batch file executes and the Export Tool returns the termination code 1 or 3, the command prompt displays a message that indicates the set subcommand fails.

java -classpath "./lib/JSanExport.jar;./lib/JSanRmiApiEx.jar;./lib/JSanRmiServerUx.jar"

-Xmx536870912 -Dmd.command=command.txt -Dmd.logpath=log

sanproject.getmondat.RJMdMain<CR+LF> if %errorlevel%==1 echo THE SET SUBCOMMAND FAILED<CR+LF> if %errorlevel%==3 echo THE SET SUBCOMMAND FAILED<CR+LF> pause<CR+LF>

Using the Export Tool 93

Log files

In the previous scripts, <CR+LF> indicates the end of a command line.

When the Export Tool runs, the Export Tool creates a new log file on your system. Therefore, if you run the Export Tool repeatedly, the size of free space on your system will be reduced. To secure free space on your system, you are strongly recommended to delete the Export Tool log files regularly. For details about the location of the log files, see “Java” (page 116).

The Export Tool returns a termination code when the Export Tool finishes.

Table 9 Termination codes returned by the export tool

MeaningTermination

code

The Export Tool finished successfully.0

Error handling

When an internal error occurs during export processing, an exclamation mark (!) appears to signal an error. By default, the Export Tool makes up to three more attempts at processing. You can change the maximum number of retries by using the retry subcommand. For detailed information about the retry subcommand, see “retry” (page 97).

If export processing does not finish within three retries or if an internal error occurs other than those listed in the following table, the Export Tool stops. If the Export Tool stops, quit the command prompt, and then run the tool again.

For more information, see “Troubleshooting the Export Tool” (page 133).

Errors for which export tool retries processing

An error occurred when the set subcommand (see “set” (page 114)) executed, because an attempt to switch to Modify mode failed. Some other user might have been logged on in Modify mode.

One of the following two errors occurred;2

• A command file has been corrupted or could not be read.

• An error occurred when a command was parsed.

An error occurred due to more than one reason. One of the reasons is that an attempt to switch to Modify mode failed when the set subcommand (see “set” (page 114)) executed. Some other user might have been logged on in Modify mode.

The Storage Administrator (Performance Management) role is not assigned to the user ID.4

94 Export Tool

Cause of errorError message ID

An error occurred during SVP processing.0001 4001

Because SVP is busy, the monitoring data cannot be obtained.0001 5400

An administrator is changing a system environment file.0001 5508

Array is refreshing, or the settings by the user are registered.0002 2016

The storage system is in internal process, or some other user is changing configuration.0002 5510

Now processing.0002 6502

Another user has lock.0002 9000

A service engineer is accessing the storage system in Modify mode.0003 2016

SVP is not ready yet, or an internal processing is being executed.0003 2033

An error occurred during SVP processing.0003 3006

The storage system status is invalid.0405 8003

Cause of errorError message ID

An internal process is being executed, or maintenance is in progress.5205 2003

SVP is now updating the statistics data.5205 2033

SVP is now updating the statistics data.5305 2033

The storage system status is invalid.5305 8002

Export Tool command reference

This topic provides the syntax of the Export Tool subcommands that you can write in your command file and the command that should be used in your batch file. “Subcommand list” (page 96) lists the subcommands explained in this topic. The Java command is explained in “Java” (page 116).

Export Tool command syntax

This topic explains the syntax of Export tool subcommands that you can write in your command file. This topic also explains the syntax of the Java command that should be used in your batch file.

Conventions

The following conventions are used to explain syntax:

DescriptionConvention

italics

{ }

Syntax descriptions

Indicates characters that you must type exactly as they are shown.bold

Indicates a type of an operand. You do not need to type characters in italics exactly as they are shown.

Indicates one or more operands that can be omitted.[ ] If two or more operands are enclosed by these square brackets and are delimited by vertical bars

(|), you can select one of the operands.

Indicates that you must select one operand from the operands enclosed by the braces. Two or more operands are enclosed by the braces and are delimited by vertical bars (|).

Indicates that a previously used operand can be repeated....

Vertical bar delimiter, indicating you can select one of the operands enclosed in square brackets.|

Indicates you can write this script...This syntax...

connect 123.01.22.33connect ip-address

destinationdestination [directory] destination c:\temp

compresscompress [yes|no] compress yes compress no

answer yesanswer {yes|no} answer no

portsports [name][...] ports port-1 ports port-1 port-2

Export Tool command reference 95

Writing a script in the command file

When you write a script in your command file, be aware of the following:

• Ensure that only one subcommand is used in one line.

• Empty lines in any command file will be ignored.

• Use a semicolon (;) if you want to insert a comment in your command file. If you enter a

semicolon in one line, the remaining characters in that line will be regarded as a comment. Following are examples of comments in a command file:

;;;;;;; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ;;; COMMAND FILE: command.txt ;;;; ;;;;;;;;;; ;;;;;;;;;;;;;;;;;;;;;;;;;;;;; svpip 158.214.135.57 ; IP address of SVP login expusr "passwd" ; Log onto SVP

Viewing the online Help for subcommands

You can display the online Help to view the syntax of subcommands when you are working at the command prompt. To be able to view the online Help, you must use the help subcommand of the Export Tool. For more information about how to use the help subcommand, see “help” (page 115).

Subcommand list

FunctionSubcommand

“show” (page 98)

“Short-range” (page 109)

“long-range” (page 111)

“set” (page 114)

Specifies the IP address of SVP to be logged in.“svpip” (page 96)

Makes settings on retries of export processing.“retry” (page 97)

Logs the specified user into SVP.“login” (page 98)

Checks SVP to find the period of monitoring data stored in SVP and the data collection interval (that is called “gathering interval”), and then outputs them to the standard output and the log file.

Specifies the type of data that you want to export.“group” (page 99)

Specifies the term of monitoring data to be exported for short-range monitoring data.

Specifies the term of monitoring data to be exported for long-range monitoring data.

Specifies the directory in which files should be saved.“outpath” (page 113)

Specifies whether to save files in ZIP format or in CSV format.“option” (page 113)

Saves monitoring data in files.“apply” (page 114)

Starts or ends monitoring of the storage system, and specifies the gathering interval in short-range monitoring.

Displays the online help for subcommands.“help” (page 115)

Starts the Export tool and writes monitoring data into files.“Java” (page 116)

svpip

Description

This subcommand specifies the IP address or the host name of SVP.

Syntax

svpip {ip-address|host-name}

96 Export Tool

retry

Operands

DescriptionOperand

Specifies the IP address of SVP.ip-address If SVP is managed with IPv6 (Internet Protocol Version 6), you must specify the ip-address operand

to match the format of IPv6. If the Export Tool runs on Windows XP, the interface identifier (for example, "%5") must be added to the end of the specified IP address.

Specifies the host name of SVP.host-name Alphanumeric characters, hyphen, and period can be specified. Underscore (_) cannot be

specified. The host name can include a hyphen but must be enclosed by double quotation marks (").

Example

The following example specifies the IP address of SVP as 158.214.127.170.

svpip 158.214.127.170

Description

This subcommand makes settings on retries of export processing. When an internal error occurs during export processing, the Export Tool stops processing and then

retries export processing. By default, the Export Tool can retry processing up to three times, but you can change the maximum number of retries by using the retry subcommand.

By default, the interval between one retry and the next retry is two minutes. You can change the interval by using the retry subcommand.

The retry subcommand must execute before the login subcommand executes.

Syntax

retry [time=m] [count=n]

Operands

DescriptionOperand

time=m

Specifies the interval between retries in minutes, where m is a value within the range of 1 to

59. If this operand is omitted, the interval between retries is two minutes.

Specifies the maximum number of retries.count=n If n is 0, the number of retries is unlimited. If this operand is omitted, the maximum number of retries is 3.

Example

If the following command file is used, the interval between retries is 5 minutes and the maximum number of retries is 10.

svpip 158.214.135.57 retry time=5 count=10login expusr passwd show group Portshort-range 201204010850:201204010910outpath outoption compressapply

Export Tool command reference 97

login

Description

This e subcommand uses a user ID and a password to log the specified user in SVP. The svpip subcommand must execute before the login subcommand executes. The login subcommand fails if monitoring data does not exist in SVP.

Syntax

Operands

DescriptionOperand

Specifies the user ID for SVP.userid If the user ID includes any non-alphanumeric character, the user ID must be enclosed by double

quotation marks ("). Be sure to specify a user ID that should be used exclusively with the Export Tool. For detailed

information, see “System requirements” (page 86).

Specifies the password of the user.password If the password includes any non-alphanumeric character, the password ID must be enclosed

by double quotation marks (").

show

Example

This example logs the user expusr into SVP whose IP address is 158.214.127.170. The password is pswd.

svpip 158.214.127.170 login expuser pswd

Description

This subcommand outputs the following information to the standard output (for example, to the command prompt):

• the period during which monitoring data was collected onto SVP (storing period)

• the interval at which the monitoring data was collected (gathering interval)

Performance Monitor collects statistics by the two types of storing periods: in short range and in long range. In short-range monitoring, the monitoring data between 8 hours and 15 days is stored in SVP, and in long-range monitoring, the monitoring data up to 3 months is stored in SVP. For details on the two storing periods, see “Short-range” (page 109) and “long-range” (page 111).

Storing periods output by the show subcommand are the same as the information displayed in the

Monitoring Term area of the Monitor Performance window.

Figure 19 The monitoring term area

The login command must execute before the show subcommand executes.

98 Export Tool

group

Syntax

show

Outputs

The show subcommand displays the storing period and the gathering interval for these two types of monitoring data: in short range and in long range. For example, the show subcommand outputs the following information:

Short Range From: 2012/10/01 01:00 - To: 2012/10/01 15:00 Interval: 1min. Long Range From: 2012/09/01 00:00 - To: 2012/10/01 15:00 Interval: 15min.

Short Range indicates the storing period and gathering interval of the monitoring data stored

in short range. Long Range indicates those of the monitoring data stored in long range. When you run the Export Tool, you can export the monitoring data within these periods into files. If you use the short-range or long-range subcommand additionally, you can narrow the term of data to be exported (see “Short-range” (page 109) or “long-range” (page 111)).

From indicates the starting time for collecting monitoring data. To indicates the ending time for collecting monitoring data.

Interval indicates the interval at which the monitoring data was collected (gathering interval). For example, Interval 15min. indicates that monitoring data was collected at 15-minute intervals.

Description

The group subcommand specifies the type of monitoring data that you want to export. This command uses an operand (for example, PhyPG and PhyLDEV above) to specify a type of monitoring data.

Table 10 (page 100) shows the monitoring data that can be saved into files by each operand, and

the saved ZIP files. For details on the monitoring data saved in these files, see the tables listed in the See column.

Export Tool command reference 99

Table 10 Operands of the group subcommand and saved monitoring data

PhyPG

PhyLDEV

PhyExG

PhyExLDEV

PhyProc

PhyESW

GUI operationOperand

Select Parity Groups from Object list in

Performance Objects field in Monitor Performance window.

Select Logical Device from Object list in

Performance Objects field in Monitor Performance window.

Select Controller from Object list in Performance Objects field in Monitor Performance window.

Select Access Path from Object list in Performance Objects field in Monitor Performance window.

the file

Usage statistics about parity groups

Usage statistics about volumes

external volume groups

Usage conditions about external volumes

Usage statistics about MPs and data recovery and reconstruction processors

Usage statistics about access paths, write pending rate, and cache

PhyPG_dat.ZIP

PhyLDEV_dat.ZIP

PhyExG_dat.ZIPUsage conditions about

PhyExLDEV_dat/PHY_ExLDEV_XXXXX.ZIP

PhyProc_dat.ZIP

PhyESW_dat.ZIP

SeeSaved ZIP fileMonitoring data saved in

Table 12 (page

118)

LDEV

Port

PortWWN

PPCGWWN

Select Parity Group from Object list in

Performance Objects field in Monitor Performance window.

Select Logical Device from Object list in

Performance Objects field in Monitor Performance window.

list in Performance Objects field in Monitor Performance window.

Select WWN from Object list in Performance Objects field in Monitor Performance window.

list in Performance Objects field in Monitor Performance window.

Select WWN from Object list in Performance Objects field in Monitor Performance window.

groups, external volume groups, or V-VOL groups

Statistics about volumes in parity groups, in external volume groups, or in V-VOL groups

adapters connected to ports

are connected to ports

PG_dat.ZIPStatistics about parity

LDEV_dat/LDEV_XXXXX.ZIP

Port_dat.ZIPStatistics about portsSelect Port from Object

PortWWN_dat.ZIPStatistics about host bus

LU_dat.ZIPStatistics about LUsSelect LUN from Object

PPCGWWN_dat.ZIPAll host bus adapters that

Table 13 (page

121)

Table 14 (page

122)

Table 16 (page

125)

Table 17 (page

125)

Table 18 (page

125)

Table 19 (page

126)

RemoteCopy

100 Export Tool

Usage Monitor tab in the Continuous Access

copy operations by Continuous AccessSynchronous and Synchronous and

RemoteCopy_dat.ZIPStatistics about remote

Table 21 (page

127)

HP XP RAID Manager Software User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual

HP XP P9000 Performance for Open and Mainframe Systems User Guide

Contents

1 Performance overview

Performance Monitor overview

Performance Control overview

Performance of high-priority hosts

Upper-limit control

Threshold control

Cache Residency overview

Prestaging data in cache

Priority mode (read data only)

Bind mode (read and write data)

2 About performance monitoring

Monitoring storage system resources

Placing data in cache

Prestaging specific data

Priority mode

Bind mode

3 Interoperability of Performance Monitor and other products

Cautions and restrictions for monitoring

Cautions and restrictions for usage statistics

Using Performance Control

WWN monitoring data

Microprogram replacement

Changing the time setting of SVP

Using Performance Control

4 Monitoring WWNs

Viewing the WWNs that are being monitored

Adding new WWNs to monitor

Removing WWNs to monitor

Adding WWNs to ports

Editing the WWN nickname

Connecting WWNs to ports

Deleting unused WWNs from monitoring targets

5 Monitoring CUs

Displaying CUs to monitor

Adding or removing CUs to monitor

Confirming the status of CUs to monitor

6 Monitoring operation

Performing monitoring operations

Starting monitoring

Stopping monitoring

7 Setting statistical storage ranges

About statistical storage ranges

Viewing statistics

Setting the storing period of statistics

8 Working with graphs

Basic operation

Objects that can be displayed in graphs

Usage rates of MPs

Usage rate of a data recovery and reconstruction processor

Usage rate of cache memory

Write pending statistics

Access paths usage statistics

Throughput of storage system

Size of data transferred

Response times

Cache hit rates

Back-end performance

Hard disk drive usage statistics

Hard disk drive access rates

Business Copy usage statistics

Detailed information of resources on top 20 usage rates

9 Changing display of graphs

Graph operation

Changing displayed items

Changing a display period

Adding a new graph

Deleting graph panel

10 Performance Control operations

Overview of Performance Control operations

If one-to-one connections link HBAs and ports

If many-to-many connections link HBAs and ports

Port tab operations

Analyzing traffic statistics

Setting priority for ports on the storage system