Notice: No part of this publication may be
reproduced or transmitted in any form or by
any means, electronic or mechanical, including
photocopying and recording, or stored in a
database or retrieval system for any purpose
without the express written permission of
Hitachi Data Systems Corporation (hereinafter
referred to as “Hitachi Data Systems”).
Hitachi Data Systems reserves the right to
make changes to this document at any time
without notice and assumes no responsibility
for its use. Hitachi Data Systems products and
services can only be ordered under the terms
and conditions of Hitachi Data Systems'
applicable agreements. All of the features
described in this document may not be
currently available. Refer to the most recent
product announcement or contact your local
Hitachi Data Systems sales office for
information on feature and product
availability.
This document contains the most current
information available at the time of
publication. When new and/or revised
information becomes available, this entire
document is updated and distributed to all
registered users.
Hitachi Data Systems is a registered
trademark and service mark of Hitachi, Ltd.,
and the Hitachi Data Systems design mark is a
trademark and service mark of Hitachi, Ltd.
Other company and product names mentioned
herein are trademarks of their respective
companies. Mention of third-party products is
for informational purposes only and
constitutes neither an endorsement nor a
recommendation. Hitachi Data Systems
assumes no responsibility with regard to the
performance or use of these products.
This document provides instructions on assessing your snapshot
requirements, designing an implementation to meet those
requirements, and implementing and operating Copy-on-Write
Snapshot software using the Storage Navigator 2 graphical user
interface.
This preface includes the following information:
•Document Revision Level
•Safety and Warnings
•Product Version
•Release Notes
•Referenced Documents
•Document Conventions
•Convention for Storage Capacity Values
•Getting Help
•Comments
Notice: The use of the Hitachi Simple Modular Storage 100and
all Hitachi Data Systems products is governed by the terms of
your agreement(s) with Hitachi Data Systems.
This section provides a history of the revision changes to this document.
•
RevisionDateDescription
MK-97DF8018-00October 2007Initial Release
Safety and Warnings
This document uses the following symbols to draw attention to important
safety and operational information. The Danger, Electric Shock, and ESD
symbols are included only in hardware manuals.
•
SymbolMeaningDescription
TipTips provide helpful information, guidelines, or suggestions for
performing tasks more effectively.
•
NoteNotes emphasize or supplement important points of the main
text.
•
CautionCautions indicate that failure to take a specified action could
result in damage to the software or hardware.
•
Intended Audience
This document is intended for users with the following background:
•Background in data processing and understands RAID storage systems
and their basic functions.
•Familiarity with Hitachi Modular Storage systems.
•Familiarity with operating systems such as the Windows 2000, Windows
Server 2003 operating system, or UNIX.
Product Version
This document applies to Hitachi Simple Modular Storage 100 microcode
versions 1810/A and higher.
Release Notes
Make sure to read the Release Notes before enabling and using this product.
The Release Notes are located on the installation CD. They may contain
requirements and/or restrictions that are not fully described in this
document. The Release Notes may also contain updates and/or corrections
to this document.
The following table provides an overview of the contents and organization
of this document. Click the chapter title in the first column to go to that
chapter. The first page of every chapter or appendix contains a brief list of
the contents of that section of the manual, with links to the pages where
the information is located.
•
Chapter/Appendix
Title
Chapter 1, SnapShot
Overview
Chapter 2, Planning
Provides descriptions of SnapShot components and how
they work together.
Provides detailed planning and design information.
Description
and Design
Chapter 3,
Provides SnapShot requirements.
Requirements
Chapter 4, Enabling
Provides instructions for enabling SnapShot.
and Disabling
SnapShot
Chapter 5,
Provides detailed configuration and testing information.
Configuring
SnapShot
Chapter 6, Using
SnapShot
Chapter 7,
Provides detailed information and procedures for using
SnapShot.
Provides monitoring and maintenance information.
Monitoring and
Maintenance
Chapter 8,
Provides information for correcting system problems.
Troubleshooting
Appendix A,
Operations Using CLI
Appendix B,
Provides detailed Navigator 2 Command Line Interface
instructions for configuring and using SnapShot.
Provides SnapShot specifications.
SnapShot
Specifications
GlossaryProvides definitions for terms and acronyms found in this
document.
IndexProvides locations to specific information in this document.
Referenced Documents
These documents contain information that is related to the topics in this
document and can provide additional information about them.
•Hitachi Storage Navigator 2 Command Line Interface (CLI) User’s Guide
(MK-97DF8038), used to develop scripts.
•Hitachi Command Control Interface (CCI) software for Modular Storage
User's Guide (MK-97DF8016).
•Hitachi Storage Navigator Modular 2 program Help, which explains
SnapShot operations using the Navigator 2 GUI application.
The following table describes the typographic conventions used in this
document.
•
ConventionDescription
BoldIndicates text on a window, other than the window title, including
menus, menu options, buttons, fields, and labels. Example: Click
OK.
ItalicIndicates a variable, which is a placeholder for actual text provided
by the user or system. Example: copy source-file target-file. Note:
Angled brackets (< >) are also used to indicate variables.
screen/code
< > angled
brackets
[ ] square
brackets
{ } braces
| vertical bar Indicates that you have a choice between two or more options or
underlineIndicates the default value. Example: [ a | b ]
Indicates text that is displayed on screen or entered by the user.
Example: # pairdisplay -g oradb
Indicates a variable, which is a placeholder for actual text provided
by the user or system. Example: # pairdisplay -g <group>
Note: Italic font is also used to indicate variables.
Indicates optional values. Example: [ a | b ] indicates that you can
choose a, b, or nothing.
Indicates required or expected values. Example: { a | b } indicates
that you must choose either a or b.
arguments. Examples:
[ a | b ] indicates that you can choose a, b, or nothing.
{ a | b } indicates that you must choose either a or b.
Convention for Storage Capacity Values
Storage capacity values for hard disk drives (HDDs) in Hitachi Data
Systems’ storage products are calculated based on the following values:
•1 KB = 1,000 bytes
2
•1 MB = 1,000
•1 GB = 1,000
•1 TB = 1,000
Storage capacity values for hard disk drives (HDDs) in the Hitachi Simple
Modular Storage 100 are calculated based on the following values:
•1 KB (kilobyte) = 1,024 bytes
•1 MB (megabyte) = 1,024
•1 GB (gigabyte) = 1,024
•1 TB (terabyte) = 1,024
For further information on the Hitachi storage system, refer to the user's
guide for the system (for example, Hitachi Simple Modular Storage Hardware Guide, MK-96DF8061).
If you have questions after reading this guide, contact an Hitachi Data
Systems authorized service provider[H1] or visit the Hitachi Data Systems
support website, http://support.hds.com.
Support Contact Information
If you purchased this product from an authorized Hitachi Data Systems
reseller, contact that reseller for support. For the name of your nearest
Hitachi Data Systems authorized reseller, refer to the Hitachi Data Systems
support web site for locations and contact information.
To contact the Hitachi Data Systems Support Center, please visit the Hitachi
Data Systems website for current telephone numbers and other contact
information. http://support.hds.com
Please provide at least the following information about the problem:
•Product name, model number, part number (if applicable) and serial
number
•System configuration, including names of optional features installed,
host connections, and storage configuration such as RAID groups and
LUNs
•Operating system name and revision or service pack number
•The exact content of any error message(s) displayed on the host
•The circumstances surrounding the error or failure
•A detailed description of the problem and what has been done to try to
•Confirmation that the Hitachi Data Systems Hi-Track remote monitoring
Comments
Your comments and suggestions to improve this document are greatly
appreciated. When contacting HDS, please include the document title,
number, and revision. Please refer to specific section(s) and paragraph(s)
whenever possible.
• E-mail: doc.comments@hds.com
• Fax: 858-695-1186
• Mail: Technical Writing, M/S 35-10
system(s)
solve it
feature has been installed and tested.
Hitachi Data Systems
10277 Scripps Ranch Blvd.
San Diego, CA 92131
Thank you! (All comments become the property of Hitachi Data Systems
Corporation.)
Snapshot creates virtual copies of data volumes within Hitachi
Simple Modular Storage systems. These copies can be used for
recovery from logical errors. They are identical to the original
volume at the point in time they were taken.
This guide provides instructions for planning and designing,
configuring and testing, and using and monitoring SnapShot. In
this chapter, see:
Hitachi’s Copy-on-Write Snapshot software creates virtual backup copies of
any data volume within the Simple Modular Storage systems with minimal
impact to host service or performance levels. These snapshots are suitable
for immediate use in decision support, software testing and development,
data backup, or rapid recovery operations.
SnapShot minimizes disruption of planned or unplanned outages for any
application that cannot tolerate downtime for any reason or that requires
non-disruptive sharing of data. Since each snapshot captures only the
changes to the original data volume, the amount of storage space required
for each Copy-on-Write Snapshot is significantly smaller than the original
data volume.
The most probable types of target applications for Copy-on-Write Snapshot
are:
•Database copies for decision support/database inquiries
•Non-disruptive backups from a Copy-on-Write Snapshot V-VOL
•Periodic point-in-time disk copies for rapid restores in the event of a
corrupted data volume
Hardware and Software Configuration
A typical SnapShot hardware configuration includes a Hitachi Simple
Modular Storage system, a host connected to the storage system and
software to configure and manage SnapShot (management software). The
host is connected to the storage system via iSCSI connections. The
management software is connected to the storage system via a
management LAN.
SnapShot employs primary volumes (P-VOLs), virtual volumes (V-VOLs),
data pool, and Hitachi Storage Navigator Modular 2 Graphical User
Interface. Advanced user functionality is available with Storage Navigator 2
Command-Line interface (CLI) and Hitachi Command Control Interface
(CCI). Figure 1-1 shows a typical SnapShot configuration.
The following sections describe how these components work together.
How SnapShot Works
SnapShot creates a virtual duplicate volume of another volume. This
volume “pair” is created when you:
• Select a volume that you want to replicate
• Identify another volume that will contain the copy
• Associate the primary and secondary volumes
• Create a snapshot of the primary volume data in the virtual (secondary)
volume.
Until the new or re-synchronized pair is split, all data written to the primary
volume is also copied to the virtual volume. When the pair is split, the
primary volume continues being updated, but the snapshot in the virtual
volume remains as it was at the time of the split. The pair can be made
identical again by re-synchronizing changes from primary-to-secondary or
secondary-to-primary.
A volume pair is a relationship established by SnapShot between two
volumes. A pair consists of a production volume, which contains the original
data and is called the primary volume (P-VOL), and from 1 to 32 virtual
volumes (V-VOLs), which contain copies of the P-VOL. The P-VOL and its VVOL(s) are located in the same Simple Modular Storage system.
A V-VOL is a mirror image of the P-VOL at the time of the snapshot. Unlike
the P-VOL, which contains the actual data, the V-VOL is made up of pointers
to the data.
To maintain the snapshot image of the P-VOL as new data is being written
to the P-VOL, SnapShot copies data being replaced to the data pool. V-VOL
pointers are updated to reference the original data's new location in the
pool. Figure 1-2 illustrates SnapShot volumes and data pool interaction.
After a snapshot, the V-VOL maintains a static copy of the P-VOL even as
new writes are updating the P-VOL. To do this, before an updated block is
written to the P-VOL, SnapShot sends the data that is being replaced to the
data pool. Pointers in the V-VOL are updated to locate the replaced data that
is now located in the data pool, thus maintaining the V-VOL point-in-time
image of the P-VOL.
The data pool’s function in the SnapShot process is illustrated in Figure 1-2.
Up to 64 data pools can be set for each controller. Each data pool can be
assigned to two or more P-VOLs plus the differential data of two or more VVOLs.
Differential Management LUs (DM-LU)
The DM-LU is an exclusive volume used for storing SnapShot information
when the array system is powered down. The DM-LU is treated the same as
other volumes in the storage system, but is hidden from a host. The DM-LU
is set at the factory at 10 GB. User configuration is not required
Consistency Group (CTG)
Application data often spans more than one volume. With SnapShot, it is
possible to manage operations spanning multiple volumes as a single group.
In a “consistency group” (CTG), all primary logical volumes are treated as
a single entity.
Managing SnapShot primary volumes as a consistency group allows
multiple operations to be performed on grouped volumes concurrently.
Write order is guaranteed across application logical volumes, since
snapshots can be taken at the same time.
SnapShot Interface—Storage Navigator Modular 2
Use Storage Navigator Modular 2 (Navigator 2) on the workstation to
perform SnapShot tasks. Configure, operate, and monitor SnapShot
functions with either of two interfaces: a graphical user interface (GUI) or a
Command Line Interface (CLI). Instructions in this guide refer to both
interfaces.
•
CAUTION! Storage Navigator 2 CLI is provided for users with significant
storage management expertise. Improper use of this CLI could void your
Hitachi Simple Modular Storage 100 system warranty. Please consult with
your reseller before using CLI.
A backup insures that a volume with bad or missing data can be
restored. With SnapShot you create copies of your production
data that can be used for backup and other uses.
Creating a backup system that fully supports business continuity
is best done when SnapShot is configured to match your business
needs.
This chapter guides you in planning a configuration that meets
organization needs and the workload requirements of your host
application.
The SnapShot planning effort consists of finding the number of V-VOLs your
organization requires for the P-VOL, the V-VOL(s)’ lifespan — how long they
must be held before being updated again — the frequency that snapshots
are taken, and the size of the data pool. The answers are determined by
analyzing the organization’s business needs and measuring the write
workload that is generated by the host application.
The plan and design workflow consists of the following:
•Assess business needs.
•Determine how often a snapshot should be taken.
•Determine how long the snapshot should be held.
•Determine the number of snapshot copies required per P-VOL.
•Measure production system write workload.
•Size the data pool. (For a description of the data pool, see Data Pools
on page 1-6.
These objectives are addressed in detail in this chapter. Two other tasks are
required before your design can be implemented. These are also addressed
in this chapter.
•When you have established your SnapShot system design, the system’s
maximum allowed capacity must be calculated. This has to do with how
the Storage system manages segments.
•Equally important in the planning process are the ways that various
operating systems interact with SnapShot. Make sure to review the
information at the end of the chapter.
Assessing Business Needs
Business needs have to do with how long back-up data needs to be retained
and what the business or organization can tolerate when disaster strikes.
These organizational priorities help determine the following:
•How often a snapshot should be made (frequency)
•How long a snapshot (the V-VOL) should be held (lifespan)
•The number of snapshots (V-VOLs) that will be required for the P-VOL.
Establishing How Often a Copy Is Made (Copy Frequency)
The frequency that copies need to be made is determined by how much data
can be lost in a disaster before business is significantly impacted.
To determine how often a snapshot should be taken
•Using knowledge of your business, decide how much data could be lost
in a disaster without significant impact to the business.
Ideally, a business desires no data loss. In the real world, disasters occur
and data is lost. You or your organization’s decision makers must decide
the number of business transactions, the number of hours required to
key in lost data, and so on.
•If losing 4 hours of business transaction is acceptable, but not more,
backups should be planned every 4 hours. If 24 hours of business
transaction can be lost, backups may be planned every 24 hours.
Determining how often copies should be made is one of the factors used to
determine data pool size. The more time that elapses between snapshots,
the more data accumulates in the data pool. Copy frequency may need to
be modified to reduce the data pool size
Selecting a Reasonable Time Between Snapshots
The length of time between snapshots, if too short or too long, can cause
problems.
•When short periods are indicated by your company’s business needs,
consider also that snapshots taken too frequently could make it
impossible to recognize logical errors in the storage system. This would
result in snapshots of bad data. How long does it take to notice and
correct such logical errors? The time span for snapshots should provide
ample time to locate and correct logical errors in the storage system.
•When longer periods between snapshots are indicated by business
needs, consider that the longer the period, the more data accumulates
in the data pool. Longer periods between backups require more space
in the data pool.
This effect is multiplied if more than one V-VOL is used. If you have two
snapshots of the P-VOL, then two V-VOLs are tracking changes to the P-VOL
at the same time.
Establishing How Long a Copy Is Held (Copy Lifespan)
Copy lifespan is the length of time a copy (V-VOL) is held, before a new
backup is made to the volume. Lifespan is determined by two factors:
•Your organization’s data retention policy for holding onto backup
copies.
•Secondary business uses of the backup data.
When you determine the snapshot’s lifespan, you can then calculate the
number of V-VOLs your system requires.
Lifespan Based on Backup Requirements
•If the snapshot is to be used for tape backups, the minimum lifespan
must be => the time required to copy the data to tape. For example:
Hours to copy a V-VOL to tape = 3 hours
V-VOL lifespan => 3 hours
•If the snapshot is to be used as a disk-based backup available for online
recovery, you can determine the lifespan by multiplying the number of
generations of backup you want to keep online by the snapshot
frequency. For example:
Generations held = 4
Snapshot frequency = 4 hours
4 x 4 = 16 hours
V-VOL lifespan = 16 hours
Lifespan Based on Business Uses
•If you use snapshot data (the V-VOL) for testing an application, the
testing requirements determine the amount of time a snapshot is held.
•If snapshot data is used for development purposes, development
requirements may determine the time the snapshot is held.
•If snapshot data is used for business reports, the reporting
requirements can determine the backup’s lifespan.
Establishing the Number of V-VOLs that Are Required
V-VOL frequency and lifespan determine the number of V-VOLs your system
needs per P-VOL.
For example: Suppose your data must be backed up every 12 hours, and
business-use of the data in the V-VOL requires holding it for 48 hours. In
this case, your SnapShot system would require 4 V-VOLs, since there are
four 12-hour intervals during the 48-hour period. This is illustrated in
The data pool holds data from the P-VOL that is being replaced. By holding
this original data in the data pool, the mirror image of the V-VOL is
maintained at the time of the snapshot.
You determine the size of the data pool that your system requires by:
•Measuring the amount of write workload that passes from the host
application to the P-VOL. Write workload is the megabytes per second
that are written to the primary volume over a specific time.
•Calculating the amount of data that would accumulate during the
lifespan of your V-VOL
•Multiplying times the number of V-VOLs
•
Figure 2-2: Write workload and data pool size
Measuring Workload Data
To set up SnapShot, you must measure the amount of data that changes in
your production system. The amount of data written to the primary volume
indicates how large the data pool must be.
Workload data is collected using performance monitoring software on your
operating system—preferably during the busiest time of month, quarter,
and year. The goal is to collect data that shows your system’s actual
workloads during high peaks and spikes, when more is changing and the
demands on the system are greatest.
To collect workload data
1. Using your operating system’s performance monitoring software, collect
the following:
• Disk-write bytes/second for every physical volume that will be
replicated.
• Collect this data at 10 minute intervals.
• Collect this data over a 4-6 week period that includes high peaks
•
and spikes, and when the demands on the system are greatest.
2. At the end of the period, convert the data to MB/second, if it is not
already so, and import into a spreadsheet tool. Figure 2-3 shows
collected raw data, in megabytes per second in 10 minute segments.
•
Figure 2-3: Raw Data Example in MB/sec
3. Using the copy frequency established earlier, calculate averages over the
collection period. Most spreadsheet tools have an average function. For
example:
If copy frequency is 1 hour, then calculate 60 minute rolling averages
using the values in 6 10-minute intervals.
If copy frequency is 4 hours, then calculate 240 minute rolling averages
using the values in 24 10-minute intervals.
Figure 2-4 illustrates 60-minute rolling averages.
•
Figure 2-4: Rolling Averages Calculated Using V-VOL Frequency
This instructs the tool to calculate the average value in cells B2
through B7 and populates C4 with that data.
c. Copy the value in C4.
d. Highlight cells C5 to the last C cell in the last row of workload data in
the spreadsheet.
e. Right-click the highlighted cells and select the paste option.
Excel maintains the logic and increments the formula values initially
entered in C4. It then calculates all the point in time averages and
populates the C cells.
Figure 2-5 illustrates rolling averages graphed over raw
measurement data averages.
•
Figure 2-5: Rolling Averages Graphed Over Raw Averages
4. Locate the maximum rolling average (RA) value in the C column. Using
this peak value and the following formula, calculate the cumulative peak data change over the lifespan of a copy (V-VOL):
(RA peak MB/sec) x (V-VOL lifespan seconds) = (Cumulative data over V-VOL lifespan)
For example, if the RA peak is 25 MB/sec, and the V-VOL lifespan is 3600
seconds (1 hour), then:
25MB/sec x 3600 seconds = 90,000 MB
The cumulative data over a V-VOL’s 1-hour lifespan is 90,000 MB.
5. Calculate the base data pool size for your primary/virtual volumes by
multiplying the MB size of one V-VOL in Step 4 by the number of V-VOLs,
which was established earlier. For example:
90,000 MB
x 4 V-VOLS = 360,000 MB
This is the base data pool size for a SnapShot system in which the copy
frequency is 1 hour, the copy lifespan is 4 hours, and the number of
copies (V-VOLs) is 4.
6. It is highly recommended that a safety factor of 20%, be calculated. Do
so using the following formula:
(Base data pool size) x 1.2. For example:
360,000 MB x 1.2 = 432,000 MB
7. It is also advisable to factor in annual increases in data transactions. Do
this by multiplying the base pool size by the percentage of expected
annual growth. For example:
432,000 MB x 1.2 (20 percent growth rate for per year)
= 518,400 MB
This is the size of the data pool with growth factored for the first year.
8. Repeat this step for each year the solution will be in place. For example:
518, 400 MB x 1.2 (20 percent growth rate for second year)
= 622,080 MB
This is the size of the data pool with growth factored for the second year.
Rule of Thumb Calculation
When measurements of host workload has not been performed, Hitachi
suggests the change rates shown in Table 2-1.
•
Table 2-1: Workload Rates when No Measurement
Snapshot lifespan
1-4 hours10%
4-8 hours15%
8-12 hours20%
12-24 hours25%
Data pool calculation using the suggested change rates in Table 2-1 is:
Data Pool size = (P-VOL x % of changed data x 2.5 safety rate)
x a number of V-VOLs
For example:
Suggested write workload
change rate
1. P-VOL = 1 TB. 1 snapshot per 24 hours. 25% of 1 TB = 250 GB.
2. Multiply the initial calculation by the Hitachi safety factor of 2.5. In the
example above: 2.5 x 250 GB = 625 GB. This is t h e base d a ta pool s ize.
3. Multiply the base data pool size by the number of V-VOLs. Thus:
4 V-VOLs x 625 GB = 2500 GB (2.5 TB).
Data Pool Key Points
•The data pool must be on same controller as the P-VOL and V-VOL(s).
•Data pool capacity should be at least 20 GB.
•Up to 64 volumes can be assigned to a data pool.
•When a volume is assigned to a data pool, it is no longer recognized by
a host.
Calculating Maximum Capacity
Simple Modular Storage manages capacity for in-system replication
systems in segments of 15.75 GB for P-VOLs and 3.2 GB for data pools. As
a result, your local replication system’s managed capacity must be
calculated using the formulas provided below and compared to the
maximum supported capacity.
To calculate SnapShot capacity
1. List the size of each P-VOL in the Storage system. For example:
P-VOL 1 = 100 GB
P-VOL 2 = 50 GB
2. Calculate managed P-VOL capacity, using the formula:
ROUNDUP (P-VOL capacity / 15.75) * 15.75
For example:
P-VOL1: ROUNDUP (100 / 15.75) = 7
7 * 15.75 = 110.25 GB, the managed P-VOL Capacity
P-VOL2: ROUNDUP (50 / 15.75) = 4
4 * 15.75 = 63 GB, the managed P-VOL Capacity
3. For each P-VOL, list the data pools and their sizes. For example:
P-VOL1 has 1 data pool whose capacity = 100 GB
P-VOL2 has 1 data pool whose capacity = 60 GB
4. Calculate managed data pool capacity, using the formula:
ROUNDUP (data pool capacity / 3.2) * 3.2
For example:
P-VOL 1 data pool: ROUNDUP (100 / 3.2 = 32)
32 * 3.2 = 102.4 GB, the managed data pool capacity
P-VOL 2 data pool: ROUNDUP (60 / 3.2 = 19)
19 * 3.2 = 60.8 GB, managed data pool capacity
5. Calculate maximum capacity using the following equation:
(Total P-VOL capacity) / 5 + (Total Data Pool capacity) < = 800 GB