The software described in this document is “commerci al computer software” provided with restricted rights (except
as to included open/free source) as specified in the FAR 52.227-19 and/or the DFAR 227.7202, or successive
sections. Use beyond license provisions is a violation of worldwide intellectual property laws, treaties and
conventions. This document is provided with limited rights as defined in 52.227-14.
The electronic (software) version of this document was developed at private expense; if acquired under an agreement
with the USA government or any contractor thereto, it is acquired as “commercial computer software” subject to the
provisions of its applicable license agreement, as specified in (a) 48 CFR 12.212 of the FAR; or, if acquired fo r
Department of Defense units, (b) 48 CFR 227-7202 of the DoD FAR Supplement; or sections succeeding thereto.
Contractor/manufacturer is SGI, 1140 E. Arques Avenue, Sunnyvale, CA 94085.
TRADEMARKS AND A TTRIBUTIONS
SGI and the SGI logoare registered trademarks of SGI in the United States and/or other countries worldwide.
Windows is a registered trademark of Microsoft Corporation in the United States and/or other countries.
All other trademarks mentioned herein are the property of their respective owners.
This user guide gives you step-by-step instructions on how to install, configure, and connect the
SGI InfiniteStorage 15000 system to your host computer system, as well as to use and maintain the
system.
Who should use this guide
This user guide assumes that you have a working knowledge of the Serial Attached SCSI (SAS) protocol
environments into which you are installing this system.
International St andards
The SGI InfiniteStorage 15000 system complies with the requirements of the following agencies and
standards:
•CE
•UL
•cUL
Potential for Radio Frequency
Interference
USA Federal Communications Commission (FCC)
NoteThis equipment has been tested and found to comply with the limits for a class A digital device, pursuant
to Part 15 of the FCC rules. These limits are designed to provide reasonable protection against harmful
interference when the equipment is operated in a commercial environment. This equipment generates,
uses and can radiate radio frequency energy and, if not installed and used in accordance with the
instruction manual, may cause harmful interference to radio communications. Operation of this
equipment in a residential area is likely to cause harmful interference in which case the user will be
required to correct the interference at his own expense.
Properly shielded and grounded cables and connectors must be used in order to meet FCC emission
limits. The supplier is not responsible for any radio or television interference caused by using other than
recommended cables and connectors or by unauthorized changes or modifications to this equipment.
Unauthorized changes or modifications could void the user’s authority to operate the equipment.
This device complies with Part 15 of the FCC Rules. Operation is subject to the following two
conditions: (1) this device may not cause harmful interference, and (2) this device must accept any
interference received, including interference that may cause undesired operation.
007-5510-002 v
Preface
European Regulations
This equipment complies with European Regulations EN 55022 Class A: Limits and Methods of
Measurement of Radio Disturbance Characteristics of Information Technology Equipment and
EN50082-1: Generic Immunity.
Qualified Personnel
Qualified personnel are defined as follows:
• Service Person: A person having appropriate technical training and experience necessary to be
aware of hazards to which that person may be exposed in performing a task and of measures to
minimize the risks to that person or other persons.
• User/Operator: Any person other than a Service Person.
Safe Handling
• Remove drives to minimize weight.
• Do not try to lift the enclosure by yourself.
Weight Handling Label: Lifting and Tipping
Pinch Hazard Label: Keep Hands Clear
vi 007-5510-002
Chassis Warning Label: Weight Hazard
• Do not lift the drive enclosure by the handles on the power cooling module (PCM); they are not
designed to support the weight of the populated enclosure.
Safety
Preface
ImportantSGI InfiniteStorage 15000 drive enclosures must be always installed in SGI InfiniteStorage 15000
racks. SGI does not authorize or support the use of these drive enclosures in any standalone
benchtop or enclosure-on-enclosure stacking configuration.
If this equipment is used in a manner not specified by the manufacturer, the protection provided
by the equipment may be impaired.
WarningThe SGI InfiniteStorage 15000 MUST be grounded before applying power.
Unplug the unit if you think that it has become damaged in any way and before you move it.
CautionPlug-in modules are part of the fire enclosure and must only be removed when a replacement can be
immediately added. The system must not be run without all units in place. Operate the system with the
enclosure top cover closed.
• In order to comply with applicable safety, emission and thermal requirements no covers should be
removed.
• The drive enclosure unit must only be operated from a power supply input voltage range of 200 V
AC to 240 V AC.
• The plug on the power supply cord is used as the main disconnect device. Ensure that the socket
outlets are located near the equipment and are easily accessible.
WarningTo ensure protection against electric shock caused by HIGH LEAKAGE CURRENT (TOUCH
CURRENT), the SGI InfiniteStorage 15000 must be connected to at least two separate and
independent sources. This is to ensure a reliable earth connection.
• The equipment is intended to operate with two (2) working PCMs. Before removal/replacement of
any module disconnect all supply power for complete isolation.
• A faulty PCM must be replaced with a fully operational module within 24 hours.
007-5510-002 vii
Preface
Power Cooling Module (PCM) Caution Label: Do not operate with modules missing
WarningTo ensure your system has warning of a power failure please discon nect the power from the
power supply, by either the switch (where present) or by physically removing the power source,
prior to removing the PCM from the enclosure/shelf.
• Do not remove a faulty PCM unless you have a replacement unit of the correct type ready for
insertion.
PCM Warning Label: Power Hazards
• The power connection must always be disconnected prior to removal of the PCM from the
enclosure.
• A safe electrical earth connection must be provided to the power cord.
• Provide a suitable power source with electrical overload protection to meet the requirements laid
down in the technical specification.
WarningDo not remove covers from the PCM. Danger of electric shock inside. Return the PCM to your
supplier for repair.
PCM Safety Label: Electric Shock Hazard Inside
viii 007-5510-002
Preface
WarningOperation of the Enclosure with ANY modules missing will disrupt the airflow and the drives will
not receive sufficient cooling. It is ESSENTIAL that all apertures are filled before operating the
unit.
• Drive Carrier Module Caution Label: Drive spin down time 30 seconds
Recycling of Waste Electrical and Electronic
Equipment (WEEE)
At the end of the product’s life, all scrap/ waste electrical and electronic equipment should be recycled
in accordance with National regulations applicable to the handling of hazardous/ toxic electrical and
electronic waste materials.
Please contact your supplier for a copy of the Recycling Procedures applicable to your product.
ImportantObserve all applicable safety precautions, e.g. weight restrictions, handling batteries and lasers
etc., detailed in the preceding paragraphs when dis m an tling and disposing of this equipment
Rack System Precautions
ImportantSGI InfiniteStorage 15000 drive enclosures should only be installed in SGI InfiniteStorage 15000
racks. Mounting and installing these drive enclosures in any other rack is not authorized or
supported by SGI.
The SGI InfiniteStorage 15000 drive enclosures are pre-installed in the rack before shipment. If the drive
enclosures must be re-installed and mounted, the following safety requirements must be considered
when the unit is mounted in a rack.
• The rack stabilizing (anti-tip) plates should be installed and secured to prevent the rack from tipping
or being pushed over during installation or normal use.
• When loading a rack with the units, fill the rack from the bottom up and empty from the top down.
• Always remove all modules and drives, to minimize weight, before loading the chassis into a rack.
WarningIt is recommended that you do not slide more than one enclosure out of the rack at a time, to avoid
danger of the rack tipping over.
• When mounting in a rack, ensure that the enclosure is pushed fully back into the rack.
007-5510-002 ix
Preface
• The electrical distribution system must provide a reliable earth ground for each unit and the rack.
• Each power supply in each unit has an earth leakage current of 1.5mA. The design of the electrical
distribution system must take into consideration the total earth leakage current from all the power
supplies in all the units.
Cover Label
ESD Precautions
CautionIt is recommended that you fit and check a suitable anti-static wrist or ankle strap and observe all
conventional ESD precautions when handling plug-in modules and components. Avoid contact with
backplane components and module connectors, etc.
Data Security
• Power down your host computer and all attached peripheral devices before beginning installation.
• Each enclosure contains up to 60 removable disk drive modules. Disk units are fragile. Handle them
with care, and keep them away from strong magnetic fields.
• All the supplied plug-in modules and blanking plates must be in place for the air to flow correctly
around the enclosure and also to complete the internal circuitry.
x 007-5510-002
Preface
• If the subsystem is used with modules or blanking plates missing for more than a few minutes, the
enclosure can overheat, causing power failure and data loss. Such use may also invalidate the
warranty.
• If you remove any drive module, you may lose data.
– If you remove a drive module, replace it immediately. If it is faulty, replace it with a drive
module of the same type and capacity
• Ensure that all disk drives are removed from the enclosure before attempting to move the rack
installation.
• Do not abandon your backup routines. No system is completely foolproof.
.
007-5510-002 xi
Introduction
Chapter 1
Introduction
The SGI InfiniteStorage 15000 controller is an intelligent storage infrastructure device designed and
optimized for the high bandwidth and capacity requirements of IT departments, rich media, and high
performance workgroup applications.
The controller plugs seamlessly into existing SAN environments, protecting and upgrading investments
made in legacy storage and networking products to substantially improve their performance, availability,
and manageability.
The controller’s design is based on an advanced pipelined, parallel processing architecture, caching,
RAID, and system and file management technologies. These technologies have been integrated into a
single plug-and-play device—the SGI InfiniteStorage 15000—providing simple, centralized, and secure
data and SMNP management.
The SGI InfiniteStorage 15000 is designed specifically to support high bandwidth, rich content, and
shared access to and backup of large banks of data. It enables a multi-vendor environment comprised of
standalone and clustered servers, workstations and PCs to access and back-up data stored in centralized
or distributed storage devices in an easy, cost-effective, and reliable manner.
Each controller orchestrates a coherent flow of data throughout the storage area network (SAN) from
users to storage, managing data at speeds of up to 3000 MB/second (or 3 GB/second). This is
accomplished through virtualized host and storage connections, a DMA-speed shared data access space,
and advanced network-optimized RAID data protection and security—all acting in harmony with
sophisticated SAS storage management intelligence embedded within the controller.
The controller can be “coupled” to form data access redundancy while maintaining fully pipelined,
parallel bandwidth to the same disk storage. This modular architecture ensures high data availability and
uptime along with application performance. With its PowerLUN technology, the system provides full
bandwidth to all host ports simultaneously and without host striping.
1.1Controller Features
The SGI InfiniteStorage 15000 controller incorporates the following features:
• Simplifies Deployment of Complex SANs
The controller provides SAN administration with the management tools required for large number
of clients.
• Infiniband or Fibre Channel (FC-8) Connectivity Throughput
The controller provides up to four (4) individual double data rate four-lane Infiniband or FC-8 host
port connections, including simultaneous access to the same data through multiple ports. Each IB
host port supports point-to-point and switched fab ric operation.
• Highly Scalable Performance and Capacity
The RAID engine provides both fault-tolerance and capacity scalability. Performance remains the
same, even in degraded mode. Internal data striping provides generic load balancing across drives.
The RAID engine can support from 10 drives minimum to 1200 drives maximum. Formatable
capacity is drive capacity.
The controller provides a wide range of management capabilities: Configuration Management,
Performance Management, Logical Unit Number (LUN) Management, Security Administration,
and Firmware Update Management.
• Management Options via RS-232 and Ethernet (Telnet)
A RS-232 port and Ethernet port are included to provide local and remote management capabilities.
SNMP is also supported.
• Data Security with Dual-Level Protection
Non-host based data security is maintained with scalable security features including restricted
management access, dual-level protection, and authentication against authorized listing (up to 256
direct host logins per host port are supported). No security software is required on the host
computers.
• Storage Virtualization and Pooling
Storage pooling enables different types of storage to be aggregated into a single logical storage
resource from which virtual volumes can be served up to multi-vendor host computers. Up to 1024
LUNs are supported.
• SES (SCSI Enclosure Services) Support for Enclosure Monitoring
Status information on the condition of enclosure, disk drives, power supplies, and cooling systems
are obtained via the SES interface.
• Absolute Data Integrity and Availability
Automatic drive failure recovery procedures are transparent to users.
• Hot-Swapable and Redundant Components
The controller utilizes redundant, hot-swappable power supplies and a hot swappable fan module
that contains redundant cooling fans.
1.2The Controller Hardware
The basic controller (See Figure 1–1.) includes:
• A chassis enclosure (with a minimum of 2.56GB cache memory)
• 10 SAS connectors that connect the controller to the drive enclosures
• Connector(s) for host Infiniband or Fibre Channel (FC-8) connection(s)
• Serial connectors for maintenance/diagnostics
• Ethernet RJ-45 connector
2 007-5510-002
Front (behind cover panel)
Rear
Introduction
A
AB
B
AC
FAIL
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
P
PS
S
TEST
CTRL
TEMP
FAN
STATUS
STATUS
STATUS
SYSTEM
DISK
STATUS
STATUSDCSTATUS
HOST 1/2
HOST 1
CLI
STATUS
ACT
1
2
1/2
HOST 2
HOST 3/4
HOST 3
CLI
STATUS
ACT
3
4
HOST 4
TEST
PLACE PIN HERE
ACT
TELNET
LINK
ACT
LINK
LINK
CLI
COM
LINK
MUTEACFAIL
ACT
ALARM
SILENCE
Figure 1–1SGI InfiniteStorage 15000 IB - Front and Rear Views
The controller is a high-performance controller designed to be rack-mounted in standard 19" racks. Each
controller is 3.5" in height, requiring 2U of rack space. The system uses 10 independent SAS drive
channels to manage data distribution and storage for up to 120 disk drives per channel (which can be
limited by drive enclosure type).
007-5510-002 3
1.2.1Power Supply and Fan Modules
Each controller is equipped with two (2) power supply modules and one (1) fan module. The PSU (power
supply unit) voltage operating ranges are nominally 110V to 230V AC, and are autoranging.
The two Power Supply modules provide redundant power . If one module fails, the other will maintain the
power supply and cooling while you replace the faulty module. The faulty module will still be providing
proper air flow for the system so do not remove it until a new module is available for replacement.
The two power supply modules are installed in the lower left and right slots at the front of the unit, behind
the cover panel
The fan module (Figure 1–1) is installed in the front top slot, behind the cover panel, and held in place
by two thumbscrews.
The two LEDs mounted on the front of the power supply module (located on the right and left of the
power supply handle) indicate the status of the PSU:
• Both LEDs will be lit green when the supply is active and the output is within operating limits with
no faults.
(Figure 1–1). Each PSU module is held in place by one thumbscrew.
• The left LED indicates the status of the AC input. The LED is lit green as long as the AC input is
present.
• The right LED indicates the status of the DC output of the power supply. The LED is lit green when
the supply is enabled and the outputs are withing specification. The LED will be off when AC input
is not present, the outputs are disabled (after a SHUTDOWN command), or the outputs are not within
specification. A cooling fan failure will not turn this LED off unless the failure results in a thermal
shutdown of the supply.
The AC switch for each supply is located on the rear of the controller unit.
The fan module contains multiple fans for cooling the controller. It is the primary source of cooling and
must be installed at all times during operation (except when it is being replaced due to a faulty fan).
NOTE :
For more information on fan status, see the description of the Status LEDs on rear panel in
the next section.
Figure 1–2Fan Module (front panel)
4 007-5510-002
1.2.2I/O Connectors and Status LED Indicators
2
e
Figure 1–3 shows the ports at the back of the controller 4 Infiniband (IB) unit.
These ports are for the test of the RAID
engine by the manufacturer or other authorized
personnel only.
Introduction
A
A
AB
B
AC
FAIL
B
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
PGEC
P
PS
S
SYSTEM
STATUS
SHFD
CTRL
STATUS
TEMP
STATUS
DISK
STATUSDCSTATUS
TEST
FAN
STATUS
HOST 1
HOST 2
HOST 1/2
STATUS
1
2
IB LEDs
CTRL
STATUS
TEMP
STATUS
DISK
STATUSDCSTATUS
TEST
FAN
STATUS
HOST 1
HOST 2
HOST 1/2
STATUS
1
2
A
AB
B
AC
FAIL
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
P
PS
S
SYSTEM
STATUS
Figure 1–3I/O Ports on the Rear Panel of the Controller
Host 3Host 1
HOST 3
3
HOST 4
4
PLACE PIN HERE
HOST 3/4
CLI
STATUS
ACT
COM
CLI
ACT
1/2
TEST
TELNET
ACT
LINK
ACT
LINK
LINK
CLI
LINK
ACT
AC
FAIL
MUTE
ALARM
SILENCE
Host 4Host 2
Ethernet
HOST 3
3
HOST 4
4
PLACE PIN HERE
HOST 3/4
CLI
STATUS
ACT
COM
CLI
ACT
1/2
TEST
TELNET
ACT
LINK
ACT
LINK
LINK
CLI
LINK
ACT
AC
FAIL
MUTE
ALARM
SILENCE
RS-23
interfac
AC fail LEDAC fail LEDCOMStatus LEDs
007-5510-002 5
Host port LEDs
CTRL
STATUS
TEMP
STATUS
DISK
STATUSDCSTATUS
TEST
FAN
STATUS
HOST 1
HOST 2
HOST 1/2
STATUS
1
2
HOST 3
3
HOST 4
PLACE PIN HERE
HOST 3/4
CLI
STATUS
ACT
4
COM
CLI
ACT
1/2
TEST
TELNET
ACT
LINK
ACT
LINK
LINK
CLI
AC
LINK
FAIL
ACT
MUTE
ALARM
SILENCE
A
AB
B
AC
FAIL
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
P
PS
S
SYSTEM
STATUS
Host port LEDs
Figure 1–4Host Port LEDs
The four HOST ports are used for IB or FC-8 host connections. You can connect your host servers IB
HCA port(s) or FC HCA port(s)directly to these ports. Additionally, you can connect these ports to your
IB or FC switches and hubs.
The
IB LEDs (the Infiniband LEDs) located between the host ports, when solid green, indicate that
there is physical connectivity with the host; when steady amber, they indicate that the subnet manager
is communicating with the host.
On FC-8 models, the
FC LEDs are located next to each FC host port. There are 3 LEDs for each host
port, which indicate if the connection is running at 8 GB (left LED), 4 GB (middle LED), or 2 GB (right
LED). The respective LED will be a solid green to show that there is a physical connection. If the
respective LED is flashing, this indicates data transfer . If the connector is taken from the host port, all 3
LEDs for that port will flash.
The
DISK CHANNEL ports ( jackscrew style connectors) are for disk connections. The ten ports are
labeled by data channels (ABCDEFGHPS). Flashing LEDs indicate activity.
The
RS-232 connector provides local system monitoring and configuration capabilities and uses a
standard DB-9 null modem female-to-male cable.
The
TELNET port provides remote monitoring and configuration capabilities. The ACT (Activity)
LED flashes green when there is Ethernet activity . It is unlit when there is no Ethernet link. The
LINK
LED turns green when the link speed is 1000MB/s, amber when the link speed is 100MB/s, and is unlit
when the link speed is 10 MB/s.
The
LINK port is used to connect single controller units in order to form a couplet via a cross-over
Ethernet cable. The
unlit when there is no Ethernet link. The
ACT (Activity) LED flashes green when there is Ethernet activity. The LED is
LINK LED turns green when the link speed is 1000MB/s,
amber when the link speed is 100 MB/s, and is unlit when the link speed is 10 MB/s.
The
COM port is an RS-232 Interface that uses an RJ-45 cable and connects controller units. The COM
port has two(2) LEDs associated with it:
The
Controller ID Selection Switch (labeled as 1/2) allows the user to configure the units
HDD ACT (Activity) and LINK ACT.
as Unit 1 or Unit 2. Each unit has an activity LED. It is green for the selected unit. The switch is
comprised of two DIP switches. The first DIP switch (indicated by the 1/2 label ) is used to select the unit
configuration. Flip the switch up for Unit 1---down for Unit 2. When two controller controllers are
paired together to form a couplet, one controller must be configured as unit 1 and its partner must be
configured as unit 2.
6 007-5510-002
Introduction
There are two AC Fail LEDs. Each LED is connected to its power supply independent of the other
supply. The LEDs are green to indicate that the AC input to the supply is present. The LEDs turn red
if the AC input to the supply is not present. If this occurs, check the LEDs on the front side of the unit.
If you lose AC power from one supply cord, the LED for that supply outlet will turn red.
Figure 1–5 shows the following status LEDs: System, Controller, Disk, Temperature, DC, and Fan.
CTRL
STATUS
SYSTEM
STATUS
Figure 1–5LED Status Indicators - Rear Panel of the Controller
The SYSTEM STATUS LED is solid green when the entire storage system is operating normally.
The
CTRL (CONTROLLER) STATUS LED is green when the controller is operating normally
and turns red when the controller unit is failed.
The
DISK ST A TUS LED is green when a disk enclosure is operating normally and turns amber when
there is a problem.
The
TEMP ST ATUS LED is green when the temperature sensors (6 total) indicate that the system is
operating normally, amber when one (1) temp sensor indicates an over-temperature condition, and red
when two (2) or more sensors indicate an over-temperature condition.
The
DC STA TUS LED is green when indicating normal operating status. It turns amber if there is a
non-critical power Supply DC fault (that is, a power supply is not installed or is not indicating “Power Good”). It turns red if an on-board supply fails or if there is a critical supply fault. If this occurs, check
the LEDs on the front side of the unit.
DISK
STATUS
TEMP
STATUS
DC
STATUS
FAN
STATUS
The
FAN STATUS LED is solid green when fans are operating normally. A flashinggreen LED
indicates system monitoring activity such that the monitoring is being updated. The LED flashes amber
if one of the fans in the module fails. If 2 or more fans fail, the LED flashes a solid red and the system
will begin the shut down process at 5 seconds, for a total of 30 seconds to complete shutdown.
007-5510-002 7
Table 1–1LED Indicators
IBSolid Green(Infiniband) Physical Connectivity with host
DISK portsFlashing GreenActivity. There is an LED for each of the ten
Telnet ACTFlashing GreenActivi ty
Telnet LINK (Speed)Solid GreenLink Speed=1000 mb/s
UnlitLink Speed=10 mb/s
Com Port HDD ACTOpen
Com Port LINK ACTOpen
Host 1/2 CLI STATUSOpen
Host 1/2 CLI STATUSOpen
Host 3/4 CLI ACTOpen
Host 3/4 CLI ACTOpen
SystemSolid GreenSystem is operating normally
CtrlSolid GreenSystem is operating normally
Solid AmberSystem is shutting down
8 007-5510-002
Table 1–1LED Indicators
Status IndicatorLed ActivityExplanation
DiskSolid GreenAll related disk enclosures are operating
Temp S tatusSolid GreenAll temp sensors operating normally
DCSolid GreenOperating normally
Fan StatusSolid GreenOperating normally
Introduction
normally
Solid AmberThere is a problem with 1 or more of the disk
enclosures
Solid AmberAt least 1 temp sensor has reported over-
temperature conditions
Solid Red2 or more temp sensors has reported over-
temperature condition
Solid AmberNon-critical power supply fault
Solid RedCritical power supply fault
Flashing GreenSystem monitoring activity
Flashing Amber1 fan has failed and needs to be replaced
Solid Red2 or more fans have failed or are undetected and
the system will shutdown in 30 seconds
AC FailSolid GreenOperating normally
Solid RedPower input to supply not present. AC failure
FC (FC-8 only)SolidPhysical connection has been made.
FlashingData is being transferred.
1.2.3Uninterruptible Power Supply (UPS)
Using an Uninterruptible Power Supply (UPS) with the controller is highly recommended. The UPS can
guarantee power to the system in the event of a power failure for a short time, which will allow the system
to power down properly.
SGI offers two types of UPS: basic and redundant. The basic UPS is rack-mountable. It can maintain
power to a five (5) enclosure system for seven (7) minutes while the system safely shuts down during a
power failure. The redundant UPS contains power cells that provide a redundant UPS solution.
NOTE :
The UPS should be installed by a licensed electrician. Contact SGI to obtain circuit and
power requirements.
007-5510-002 9
Controller Installation
Chapter 2
Controller Inst allation
These steps provide an overview of the controller installation process. The steps are explained in detail
in the following sections of this chapter.
1.Unpack the controller system.
2.If it is necessary to install the controller in the 19-inch cabinet(s), contact your service provider.
NOTE :
3.Set up and connect the drive enclosures to the controller.
4.Connect the controller to your Infiniband (IB) or Fibre Channel (FC) switch and host computer(s).
5.Connect your RS-232 terminal to the controller.
6.Power up the system.
7.Configure the storage array (create and format LUNs - Logical Units) via RS-232 interface, Telnet,
8.Define and provide access rights for the clients in your SAN environment. Shared LUNs need to be
9.Initialize the system LUNs for use with your server/client systems. Partition disk space and create
Most controller configurations arrive at sites pre-mounted in a 42U or 45U rack supplied by
SGI.
or GUI.
managed by SAN management software. Individual dedicated LUNs appear to the client as local
storage and do not require management software.
file systems as needed.
2.1Setting Up the Controller
This section details the installation of the hardware components of the controller system.
The SGI InfiniteStorge 15000 must be removed from the shipping pallet using a
!
Warning
NOTE :
007-5510-002 11
minimum of 4 people. The racked unit may not be tipped more than 10 degrees,
either from a level surface or rolling down an incline (such as a ramp).
Follow the safety guidelines for rack installation given in the “Preface”.
Controller Installation
2.2Unpacking the System
Before you unpack your controller, inspect the shipping container(s) for damage. If you detect damage,
report it to your carrier immediately. Retain all boxes and packing materials in case you need to store or
ship the system in the future.
While removing the components from their boxes/containers, inspect the controller chassis and all
components for signs of damage. If you detect any problems, contact SGI immediately.
Your controller ships with the following:
• controller chassis
• two (2) power cords
• RS-232 and Ethernet cables for monitoring and configuration
• cover panel and rack-mounting hardware
!
Warning
Wear an ESD wrist strap or otherwise ground yourself when ha ndling controller
modules and components. Electrostatic discharge can damage the circuit boards.
2.2.1Rack-Mounting the Controller Chassis
For instructions on mounting the controller in the rack, contact your service provider.
2.2.2Connecting the Controller in Dual Mode
For dual mode configuration only:
1.Connect the LINK ports on the two controller units using the supplied cable.
2.Connect the COM ports on the two units using the supplied cable.
12 007-5510-002
2.2.3Connecting the Controller
To set up the disk enclosures and connect them to the controller, do the following.
1.There are 10 disk channels on the controller. They correspond with disk ports. The disk ports are
labeled as follows (Figure 2–1
DISK A = Channel A DISK B = Ch annel B
DISK C = Channel C DISK D = Channel D
DISK E = Channel E DISK F = Channel F
DISK G = Channel GDISK H = Channel H
DISK P = Channel P (parity)
DISK S = Channel S (spare)
Using the 10 copper SAS cables provided, connect these disk ports to your ten disk channels.
):
Controller Installation
Disk ports
CTRL
STATUS
STATUSDCSTATUS
TEST
FAN
TEMP
STATUS
STATUS
DISK
A
AB
B
AC
FAIL
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
P
PS
S
SYSTEM
STATUS
Figure 2–1I/O Connectors
Each controller supports up to four host connections. You may connect more than four client
2.
systems to the controller using switches and you can restrict user access to the LUNs (as described
in Section 2.3 "Configuring the Controller" of this guide).
The Host ports are numbered 1 through 4 as shown in Figure 2–1. Connect your host system(s) or
switches to these ports. For FC-8 models, ensure that the latches on the transceivers are engaged.
2.2.4Selecting SAS- ID for Your Drives
NOTE :
The controller uses a select ID of 1.
HOST 1
HOST 2
Host port 3Host port 1
HOST 1/2
1
2
HOST 3
3
HOST 4
4
PLACE PIN HERE
HOST 3/4
STATUS
ACT
CLI
COM
CLI
STATUS
ACT
1/2
TEST
TELNET
ACT
LINK
ACT
LINK
LINK
CLI
LINK
AC
ACT
FAIL
MUTE
ALARM
SILENCE
Host port 4Host port 2
2.2.5Laying Out your Storage Drives
Tiers, or RAID groups, are the basic building blocks of the controller. A tier can be catalogued as 8+1
or 8+2. In 8+1 mode, a tier contains 10 drives—eight (8) data drives (Channels A through H), one parity
drive (Channel P), and one optional spare drive (Channel S). In 8+2 mode, a tier contains 10 drives—
eight (8) data drives (Channels A through H) and two parity drives (Channel P and S).
The controller can manage up to 120 tiers.
007-5510-002 13
Controller Installation
CLI (RS-232 Interface)
2.2.6Connecting the RS-232 Terminal
Configuration of disks in the enclosures must be in sets of complete tiers (Channels A through P).
Allocating one spare drive per tier gives you the best data protection but this is not required. The spare
drives on the controller are global hot spares.
For first time set-up, you will need access to an RS-232 terminal or terminal emulator (such as Wi ndows
hyperterminal). Then you may set up the remote management functions and configure/monitor the
controller remotely via Telnet.
1.Connect your terminal to the CLI port at the back of the controller using a standard DB-9 female-
to-male null modem cable (Figure 2–2).
A
AB
B
AC
FAIL
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
Figure 2–2Controller CLI Port
2.Open the terminal window .
3.Use the following settings for your serial port:
SettingValue
Bits per second:115,200
Data bits:8
Parity:None
Stop bits:1
Flow Control:None
2.2.6.1Basic Key Operations
P
PS
S
SYSTEM
STATUS
CTRL
STATUS
TEMP
STATUS
DISK
STATUSDCSTATUS
TEST
FAN
STATUS
HOST 1
HOST 2
HOST 1/2
STATUS
1
2
HOST 3
3
HOST 4
PLACE PIN HERE
HOST 3/4
CLI
STATUS
ACT
4
COM
CLI
ACT
1/2
TEST
TELNET
ACT
LINK
ACT
LINK
LINK
CLI
AC
LINK
FAIL
ACT
MUTE
ALARM
SILENCE
The command line editing and history features support ANSI and VT-100 terminal modes. The
command history buffer can hold up to 64 commands. The full command line editing and history only
work on main CLI and telnet sessions when entering new commands. Basic Key Assignments are listed
in Table 2–1 on page 15.
Simple, not full command, line editing only is supported when the:
• CLI prompts the user for more information.
• alternate CLI prompt is active. (The alternate CLI is used on the RS-232 connection during an active
telnet session.)
NOTE :
Not all telnet programs support all the keys listed in Table 2–1 "Basic Key Assignments".
The Backspace key in the terminal program should be setup to send ‘Ctrl-H’.
14 007-5510-002
Table 2–1Basic Key Assignments
KeyEscape SequenceDescription
BackspaceCtrl-H, 0x08deletes preceding character
DelDel, 0x7F or Esc [3~deletes current character
Up ArrowEsc [Aretrieves previous command in the history buffer
Down ArrowEsc [Bretrieves next command in the history buffer
Right ArrowEsc [Cmoves cursor to the right by one character
Left ArrowEsc [Dmoves cursor to the left by one character
HomeEsc [H or Esc [1~moves cursor to the start of the line.
EndEsc [K or
Esc [4~
InsEsc [2~toggles character insert mode, on and off
NOTE: Insert mode is ON by default and resets to ON for each new command.
PgUpEsc [5~retrieves oldest command in the history buffer
PgDnEsc [6~retrieves latest command in the history buffer
moves cursor to the end of the line
Controller Installation
2.2.7Powering On the Controller
NOTE :
1.Verify that the power switches on the two (2) power supply module at the back of each controller
are off.
2.Connect the two AC connectors, using the power cords provided at the back to the AC power
source for each controller unit. For maximum redundancy, connect the two power connectors to
two different AC power circuits for each unit.
3.Check that all your drive enclosures are powered up.
4.Check that the drives are spun up and ready.
5.Turn on the power supplies on the controller unit(s). The controller will undergo a series of system
diagnostics and the bootup sequence is displayed on your terminal.
6.Wait until the bootup sequence is complete and the controller system prompt is displayed.
NOTE :
Systems that have dual controllers (couplets) should have the controllers powered on
simultaneously ensure correct system configuration.
Do not interrupt the boot sequence without guidance from SGI support personnel.
Yo u may now configure the system as described in Section 2.3 "Configuring the Controller".
007-5510-002 15
Controller Installation
2.3Configuring the Controller
This section provides information on configuring your controller.
NOTE :
The configuration examples provided here represent only a general guideline. These
examples should not be used directly to configure your particular controller.
The CLI (command line interface) commands used in these examples are fully
documented in sections 3.1 through 3.8—though exact commands may change depending
on your firmware version. T o access the most up-to-date commands, use the CLI’s online
HELP feature.
2.3.1Planning Your Setup and Configuration
Before proceeding with your controller configuration, determine the requirements for your SAN
environment, including the types of I/O access (random or sequential), the number of storage arrays
(LUNs) and their sizes, and user access rights.
The controller uses either an 8+2 or an 8+1+1 parity scheme. It is a unique implementation that combines
the virtues of RAID 3, RAID 0, and RAID 6 (Figure 2–3
per 8+1 parity group; two parity drives are dedicated in the case of an 8+2 parity group or RAID 6. A
parity group is also known as a Tier.
This RAID implementation exhibits RAID 3 characteristics such as tremendous large block-transfer—
READ and WRITE—capability with NO performance degradation in crippled mode. This capability
also extends to RAID 6, delivering data protection against a double disk drive failure in the same tier with
no loss of performance.
). Like RAID 3, a dedicated parity drive is used
Tier Configuration
Striping
across tiers
when a LUN
is created
across
multiple tiers
However, Like RAID 5, this RAID implementation does not lock drive spindles and does allow the disks
to re-order commands to minimize seek latency, and the RAID 0-like functionality allows multiple tiers
to be striped, providing “PowerLUNs” that can span hundreds of disk drives. These PowerLUNs support
very high throughput and have a greatly enhanced ability to handle small I/O (particularly as disk
spindles are added) and many streams of real-time content.
LUNs can be created on just a part of a tier, a full tier, across a fraction of multiple tiers, or across multiple
full tiers. A minimum configuration for tiers of drives require either 9 drives in an 8+1 configuration or
10 drives in an 8+2 configuration. When configured in 8+1+1 mode, the tenth data segment is reserved
for global hot spare drives. When configured in 8+2 mode, spares may reside on each data segment and
are global only to that data segment.
16 007-5510-002
Controller Installation
The controller supports various disk drive enclosures that can be used to populate the 10
<ABCDEFGHPS> disk channels in both SAS 1x and SAS 2x modes. Each chassis has a limit to the tiers
that can be created and supported. Refer to the specific disk enclosure user guides for further information.
You can create up to 1024 LUNs in a controller . LUNs can be shared or dedicated to individual users,
according to your security level setup, with Read or Read/Write privileges granted per user. Users only
have access to their own and “allowed-to-share” LUNs. Shared LUNs need to be managed by SAN
management software. Individual dedicated LUNs appear to users as local storage and do not require
external management software.
NOTE :
In dual mode, LUNs are “owned” by the controller unit on which they are created. Hosts only
see the LUNs on the controller to which they are connected, unless cache coherency is enabled.
For random I/O applications, use as many tiers as possible and create one or more LUNs. For
applications that employ sequential I/O, use individual or small grouping of tiers. If you need guidance
in determining your requirements, contact SGI support.
2.3.2Configuration Interface
You can use the Command Line Interface (CLI) to configure the controller system. This user guide
provides information for setup using the CLI.
2.3.3Login as Administrator
The default Administrator account name is admin and its default password is password. (See
Section 3.1.3 "Administrator and User Logins" for information on how to change the user and
administrator passwords.) Only users with administrator rights are allowed to change the configuration.
To login:
1.At the login prompt, type:
login admin
<Enter>
2. At the password prompt, type:
password
<Enter>
2.3.4Setting System Time & Date
The system time and date for the controller are factory-configured for the U.S. Pacific Standard Time
(PST) zone. If you are located in a different time zone, you need to change the system date and time so
that the time stamps for all events are correct. In dual mode, changes should always be made on Unit 1.
New settings are automatically applied to both units.
T o set the system date, at the prompt, type:
date mm dd yyyy
where mm represents the two digit value for month, dd represents the two digit value for day , and yyyy
represents the four digit value for year.
007-5510-002 17
<Enter>
Controller Installation
For example, to change the system date to March 1, 2009, enter:
date 3 1 2009
<Enter>
To set the system time, at the prompt, type:
time hh:mm:ss
<Enter>
where hh represents the two digit value for hour (00 to 24), mm is the two digit value for minutes, and
ss represents the two digit value for seconds
For example, to change the system time to 2:15:32 p.m., enter:
time 14:15:32
<Enter>
NOTE :
The system records time using the military method, which records hours from 00 to 24, not
in a.m. and p.m. increments of 1 to 12.
2.3.5Setting Tier Mapping Mode
When the controller system is first configured, it is necessary to select a tier mapping mode for the
attached enclosures.
The controller currently supports SAS drive enclosures.
To display the current mapping mode, type:
Figure 2–4Tier Changemap Screen
T o change the mapping mode (Figure 2–4):
tier map
<Enter>
.
1.Enter:
tier changemap
<Enter>
2.Enter the appropriate mapping mode.
•For 10 and 20 box solutions, choose SAS_2X.
•For 5 box solutions, choose SAS_1X.
3.For the changes to take effect, enter:
18 007-5510-002
restart
<Enter>
.
2.3.6Checking Tier St atus and Configuration
Use the tier command to display your current tier status. Figure 2–5 illustrates the status of a system
containing 80 drives on 8 tiers with both parity modes of 8+1 and 8+2 tiers. The plus sign (+) adjacent
to the tier number indicates that the tier is in 8+2 mode.
Each letter under the “Disk Status” column represents a healthy drive at that channel (as s hown in Figure
2–5). Verify that all drives can be seen by the controller.
“Unhealthy” drives appear as follows:
• A blank space indicates that the drive is not present (or detected) at that location.
• A period (.) denotes that the disk was failed by the system.
• A question mark (?) indicates that the disk has failed the diagnostics tests or is not configured
correctly.
• The character “r” indicates that the disk at that location is being replaced by a spare drive.
After entering the tier command, perform the following steps if necessary:
1.If a drive is not displayed at all (that is, it is “missing”), check to ensure that the drive is properly
seated and in good condition. To search for the drive, enter:
disk scan
<Enter>
2.If the same channel is missing on all tiers, check the cable connections for that channel.
3.If “automatic disk rebuilding” is not enabled, enable it by entering:
tier autorebuild=on
<Enter>
4.To display the detailed disk configuration information for all of the tiers (Figure 2–6) enter:
tier config
<Enter>
007-5510-002 19
Controller Installation
2.3.6.1Heading Definitions
Figure 2–6Current Tier Configuration
• Total LUNs. LUNs that currently reside on the tier.
• Healthy Disk. The “health” of the spare disk currently being used (if any is being used) to replace
a disk on the listed tier. The health indication for the spare channel that is physically on the listed
tier is found under SP H.
• F indicates the failed disk (if any) on the tier.
• R indicates the replaced disk (if any) on the tier.
• Sp H indicates if the spare disk that is physically on the tier is healthy.
• Sp A indicates if the spare disk that is physically on the tier is available for use as a replacement.
• Spare Owner indicates the current owner of the physical spare, where ownership is assigned when
the spare is used as a replacement. “RES-#” will appear under the Spare Owner heading while a
replacement operation is underway to indicate that unit “#” currently has the spare reserved.
• Spare Used on indicates the tier (if any) where this physical spare is being used as a replacement.
• Repl Spare from indicates the tier (if any) whose spare disk is being used as a replacement for this
tier.
NOTE :
Tiers are 8+1 mode by default.
2.3.7Cache Coherency and Labeling in Dual Mode
Use the DUAL command to check the status of the units that are healthy and verify that the “Dual”
(COM2) and “Ethernet” (LINK) communication paths between the two controller units are established
(Figure 2–7).
20 007-5510-002
Controller Installation
Figure 2–7Dual Controller Configuration
If you require multi-pathing to the LUNs, enable cache coherency . If you do not require multi-pathing,
disable cache coherency.
To enable/disable the cache coherency function, enter the following (ON enables, OFF disables):
dual coherency=on|off
<Enter>
You may change the label assigned to each unit. This allows you to uniquely identify each unit in the
system. Each unit can have a label of up to 31 characters long.
1.To change the label, enter:
dual label
<Enter>
2.Select which unit you want to re-label (see Figure 2–8).
3.When prompted, type in the new label for the selected unit. The new name is displayed.
15000 [1]: dual label
Enter the number of the Unit you wish to rename.
LABEL=1 for Unit 1, Test System[1]
LABEL=2 for Unit 2, Test System[2]
Unit: 1
Enter a new label for Unit 1, or DEFAULT to return to the default label.
Up to 31 characters are permitted.
Current Unit name: Test System[1]
New Unit name: System[1]
15000 [1]:
Figure 2–8Labeling a Controller Unit
2.3.8Configuring the Storage Arrays
When you have determined your array configuration, you need to create and format the LUNs. You have
the option of creating a 32-bit or a 64-bit address LUN.
In the example below, 2 LUNs (32-bit addressing) are created:
• LUN 0 on Tiers 1 to 8 with capacity of 8192MB each.
• LUN 1 on Tiers 1 and 2 with capacity of 8192MB each.
NOTE :
007-5510-002 21
You may press e at any time to exit and cancel the command completely.
Controller Installation
NOTE :
In dual mode, LUNs are “owned” by the controller unit where they are created. Hosts only
see the LUNs on the unit to which that they are connected, unless cache coherency is enabled.
1.To display the current cache settings, type:
cache
<Enter>
2.Select a cache segment size for your array. For example, to set the segment size to 128KBytes,
type:
cache size=128
<Enter>
This setting can also be adjusted on-the-fly for specific application tuning: see section 3.2.12
"Couplet Controller Configuration (Cache/Non-Cache Coherent)". The default setting is 1024.
3.Type:
lun
<Enter>
.
The Logical Unit Status chart should be empty, as no LUN is present on the array.
4.To create a new LUN, type:
where
X is the LUN number. Valid LUN numbers are 0..1023.
If only
lun add is entered, you are prompted to enter a LUN number.
lun add=x
<Enter>
5.Yo u will be prompted to enter the parameter values for the LUN. In this example:
- Enter a label for the LUN (you can include up to 12 characters). The label may be changed later
using the LUN LABEL command.
- Enter the capacity (in Mbytes) for a single LUN in the LUN group:
- Enter the number of tiers to use:
8
<Enter>
8192
<Enter>
- Select the tier(s) by entering the Tier number. Enter each one on a new line. Tiers are numbered
from 1 through 125.
1
2
3
4
5
6
7
8
- Enter the block size in Bytes: 512
NOTE :
512 is the recommended block size. A larger block size may give better performance.
<Enter>
<Enter>
<Enter>
<Enter>
<Enter>
<Enter>
<Enter>
<Enter>
<Enter>
However, verify that your OS and file system can support a larger block size before changing the block
size from its default value.
This message will display: Operation successful: LUN 0 added to the system
6.When you are asked to format the LUN, type:
22 007-5510-002
y
<Enter>
Controller Installation
After you have initiated LUN format, the message Starting Format of LUN is displayed.
You can monitor the format progress by entering the command
LUN (see Figure 2–9).
Upon completion, this message: Finished Format of LUN 0 displays.
System Capacity 2240096 Mbytes, 2207328 Mbytes available.
Logical Unit Status
Status
(Mbytes)
8192Ready0
Block
Size
512
Tiers
8
Figure 2–10 LUN Status - Ready
8.To create the LUN 1, type:
lun add=1
<Enter>
9.Enter these parameters:
- Enter a label for the LUN 1
- For capacity, enter the value in MBytes:
- Enter the number of tiers to use:
2
<Enter>
8192
<Enter>
- Select the tier(s) by entering the Tier number. Enter each one on a new line and press the
key. The tiers are numbered from 1 through 125.
1
2
- Enter the block size in Bytes: 512
- When asked to form at the LUN, type:
NOTE :
LUN format is a background process and you can start adding the next LUN as soon as the
<Enter>
<Enter>
<Enter>
y
<Enter>
format for the previous LUN has started.
<Enter>
007-5510-002 23
Controller Installation
2.3.9Setting Security Levels
After you have formatted all the LUNs, you can define users’ access rights. Configurations come in two
types:
• authorized user
• host port zoning
The Authorized User configuration is highly recommended for use in a SAN environment-- your data
is completely secured and no accidental plug-in is allowed to do damage such as data change or deletion.
Authorized users have access only to their own and “allowed to share” data. Administrators can also
restrict users’ access to the host ports and their read/write privileges to the LUNs. Another advantage of
this configuration is that the users see the same LUN identification scheme regardless of the host port
connection.
The Host Port Zoning configuration provides the minimum level of secu rity. The LUN mappings
change according to the host port connection. The read-only and read/write privileges can be specified
for each LUN.
The place holder LUN feature allows the controller administrator to map a zero-capacity LUN to a host
or group of hosts (via zoning or user authentication). The administrator can then create a real LUN and
map it to the host(s) to replace the place holder LUN in the future. In most cases, the host does not have
to reboot since it already mapped to the place holder LUN.
NOTE :
Support of place holder LUNs is dependent upon the OS (operating system), the driver, and
the Host Card Adapter (HCA-IB), or Host Bus Adapter.
2.3.9.1User Authentication (Recommended for SAN Environment)
Each user connected to the controller is identified by a World Wide Name (WWN) or GUID, and is
given a unique user ID number. The controller can store configurations for up to 512 users and the
security settings apply to all host ports.
Below is an example for adding two users to a system containing two LUNs (numbered 0 and 1). Each
user has an internal LUN 1 is shared and “read-only.” Both users see the shared LUN as LUN 0 and they
see their own LUN as LUN 1. User 1 has access to host ports 1 and 4 while User 2 only has access to
host port 2.
Prior to adding any users, verify that no “anonymous” access is allowed to the system:
1.Enter:
zoning
<Enter>
24 007-5510-002
2.Check to ensure that the LUN Zoning chart is empty (Figure 2–11).
3.Specify a new Host User’s world wide name, enter
user audit=on
user add
<Enter>
The controller reports which users are connected.
<Enter>
.
s.
4.Specify a 64-bit world wide name or GUID, taken from the list of available anonymous users.
5.Enter an alias name for the user. The name may contain up to 12 characters. Type in a name and
press
6.Host users can have their port access zoned. Enter
<Enter>
.
y to specify host port zoning.
7.For Unit 1, enter each active port on a new line and then exit. For this example, type:
1
4
e
<Enter>
<Enter>
<Enter>
8.For Unit 2, enter each active port on a new line and then exit. For this example, type:
1
4
e
<Enter>
<Enter>
<Enter>
Host users are limited to accessing specific LUNs, as follows:
• a host user may have its own unique LUN mapping, or
• a host user may use the anonymous LUN mapping.
The anonymous user LUN mapping is handled by the port ZONING command. In either case, the LUN
mapping applies on all the ports for which the user has been zoned.
9.Enter
y to specify the unique LUN mapping (Table 2–2).
007-5510-002 25
Controller Installation
10. Enter a new unique LUN mapping for this user. Options are shown in Table 2–2 on page 26.
Table 2–2LUN Mapping Options.
OptionDescription
G.1GROUP.LUN number
Pplace-holder
RRead-Only. Place before the GROUP.LUN
NClear current assignment
<cr>No Change
EEXIT
?Display detailed hel p te xt.
11. Connect user 2 and repeat steps 2--10 to specify the host port zoning and LUN mappings with the
following changes:
- For active host port (step 6), enter port 2 only.
- For LUN mapping:
External LUN 0 is mapping to internal LUN: R1
External LUN 1 is mapping to internal LUN: 1
External LUN 2 is mapping to internal LUN: q
NOTE :
In this scheme, users 1 and 2 have their own custom LUN identification scheme. The internal
LUN 1 that is shared by the users needs to be manag e d by SAN management software.
The individual dedicated LUN appears to the user as local storage and does not require
external management software.
Host Port Zoning (Anonymous Access) should only be used for non-SAN environment. Users are given
“general admission” to the data.
Anonymous Access (host port zoning) provides only the minimum level of
security.
!
Warning
One zoning configuration is supported for each of the host ports. Any unauthorized user accessing the
storage is considered “anonymous” and granted zoning access for the host port to which they are
connected. Given below is an example for adding LUN zoning to host port 1. External LUN 1 is mapped
to internal LUN 0 and it is read-only for the users.
Controller Installation
1.To edit the default zoning on a host port, type:
displayed
2.Select a host port (1..4):
.
1
<Enter>
zoning edit
. The current settings are
<Enter>
3.Specify the internal LUN (0..1023) to be mapped to the external LUN. The new settings will display
.
4.Repeat steps 1–3 to configure other host ports.
007-5510-002 27
Chapter 3
CLI (RS-232 Interface)
Controller Management
3.1Managing the Controller
The controller provides a set of tools that enable administrators to centrally manage the network storage
and resources that handle business-critical data. These include Configuration Management, Performance
Management, Remote Login Management, Security Administration, and Firmware Up date
Management. Bundled together, this is called the controller’s Administrator Utility.
3.1.1Management Interface
SAN management information for the controller can be accessed locally through a serial interface, or
remotely through Telnet.
Controller Management
NOTE :
A controller may have only one active login (serial or Telnet) at any given time.
Locally - Serial Interface
Any RS-232 terminal or terminal emulator (such as Microsoft Windows HyperTerminal) can be used to
configure and monitor the controller.
1.Connect your terminal to the CLI port at the back of the controller using a standard DB-9 female-
to-male null modem cable (Figure 3–1).
CTRL
STATUS
STATUSDCSTATUS
TEST
FAN
TEMP
STATUS
STATUS
DISK
HOST 1
HOST 2
HOST 1/2
1
2
HOST 3
3
HOST 4
4
PLACE PIN HERE
HOST 3/4
STATUS
ACT
CLI
COM
CLI
STATUS
ACT
1/2
TEST
TELNET
ACT
LINK
ACT
LINK
LINK
CLI
LINK
AC
ACT
FAIL
MUTE
ALARM
SILENCE
A
AB
B
AC
FAIL
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
P
PS
S
SYSTEM
STATUS
Figure 3–1Controller CLI Port
2.Open your terminal window and use these settings for your serial port:
SettingValue
Bits per second:115,200
Data bits:8
Parity:None
Stop bits:1
Flow Control:None
007-5510-002 29
3.With the controller ready, press <Enter> to get the controller prompt.
NOTE :
T o change the baud rate on controller , see section 3.8.3 "Changing Baud Rate for the CLI
Interface" in this guide.
Remotely - Telnet
T o configure and monitor the controller remotely, connect the controller to your Ethernet network. Refer
to Section 4.1, "Remote Management of the Controller" for information on how to set up the controller’s
network interface.
3.1.2Available Commands
Use the Help command to display the available commands within the utility. To get help information on
a command, type the command followed by a question mark.
For example, type: cache?
to display help on cache options on the system.
<Enter>
3.1.3Administrator and User Logins
The login command allows the user to log into a (new) terminal or Telnet session at a specific security
user level—administrative or general purpose. You will need Administrator access on the controller in
order to change the system configurations.
For RS-232 terminal session, the general purpose user does not require login. For a Telnet session, you
are required to login as either an administrator or a general purpose user. If you login as an administrator ,
you will have access to all the management and administrative functions. You can obtain status
information and make changes to the system configuration.
At the general purpose user access level, you are only allowed to view status and configuration
information. If the controller determines that the individual does not have the proper privileges, it will
return a message (where the “user entered command” represents a command keyed in by the user):
<user entered command>: Permission denied
3.1.3.1Login
To login to the system, do the following:
1.To login (Figure 3–2), enter: login
The prompt will display Enter a login name:
2.Enter a login name.
This prompt will display Enter the password:
3.Enter a password (see Figure 3–2).
<Enter>
30 007-5510-002
Controller Management
NOTE :
15000 [1]: login
Enter a login name: admin
Enter the password: ********
Successful CLI session login.
New owner : admin.
New security level: Administrative.
Figure 3–2Login Screen
3.1.3.2Logout
To logout of the system, enter:
For a terminal session, you are returned to the general purpose user level.
For Telnet, the current session is disconnected.
3.1.3.3Password
The default administrator account name is “admin” and its password is “password.”
Similarly, the default user account name is “user” and its password is “password.”
logout
<Enter>
Entering the PA SSWORD command allows the administrator to change the login names and passwords
for administrative and general purpose users
regardless of the name or password changes.
15000 [1]: password
Enter new name to replace <admin>:
Enter old password: ********
Enter new name to replace <user>:
Enter old password: ********
Figure 3–3Password Configuration Screen
Login names and passwords can be changed using the P ASSWORD command, via RS-232 or Telnet
(see Section 3.1 .3 in this guide). By default, the administrator name is “admin” and its password is
“password”. Similarly, the default user name is “user” and its password is “password.” If a user forgets
the password, entering PASSWORD DEF AUL TS while logged in as “admin”, will restore all passwords
and user names to the default values.
3.1.3.4Who Am I
T o display the owner and the security level of the current terminal or T elnet session (Figure 3–4), enter:
whoami
<Enter>
(Figure 3–3). The associated privileges remain the same
007-5510-002 31
15000 [1]: whoami
CLI session:
Current owner : admin.
Current security level: Administrative.
Figure 3–4WHOAMI Screen
3.2Configuration Management
The controller provides uniform configuration management across heterogeneous SANs. Status of host
ports and storage assets are continuously being monitored.
Table 3–1Controller Limits
Item
Limit
Number of LUNs
Total Number of Users
Number of LUNs Per User
Number of LUNs per port (zoning)
FC logins per port
Number of IB Logins per HCA (2 ports)
Max RDMA size (IB only)
Max Msg size (IB only )
Max Msg Depth (IB only )
Max number of tiers per LUN
Max number of tiers
Max size of 32-bit LUN
Max size of 64-bit LUN
Granularity of LUN size
Support LUN block sizes
Active host commands
Max queued commands per host port (does not
include active host commands)
1024
512
255
255
512
256
256 KB
4 KB
32
8
120
0xFFFF0000 (blocks)
168 TB
2 MB x number of tiers
512 bytes, 1K, 2K, 4K
32 per port
512
Max commands per disk
32 (can be lower for SATA
3.2.1Configure and Monitor Status of Host Ports
The status information of the host ports can be obtained at any time.
The HOST command displays the current settings and status for each host port
Figure 3–6
). It also displays a list of the users currently logged into the system. An unauthorized user is
given the user name Anonymous.
32 007-5510-002
(Figure 3–5 and
Controller Management
The PORT=X|ALL parameter specifies the specific host port(s) (1 to 4) to be affected when used in
combination with any of the other parameters: ID, TIMEOUT, SPEED (for FC only), or WWN. The
default is to apply changes to ALL host ports.
Figure 3–5FC Host Ports Configuration Screen
Figure 3–6IB Host Ports Configuration Screen
3.2.1.1Host ID
HOST ID=<new ID> changes the hard loop ID of a host port. The system selects a soft ID if the hard
loop ID is already taken by another device. This parameter is entered as an 8-bit hex value. The default
value is EF.
Host WWN
HOST WWN=X|DEFAUL T overrides the system ID and specifies a different World Wide Name (WWN)
for a host port. This parameter is entered as an 64-bit hex value. Default WWN is based on the serial
number of the unit.
3.2.1.2Host Status
HOST ST A TUS displays the loop status of each host port (Figure 3–7).
007-5510-002 33
HOST ST A TUSCLEAR resets the error counts.
Figure 3–7IB Host Ports Status Screen
3.2.1.3Host IB Users
HOST ibusers displays additional information on the Infiniband (IB) users logged into the controller.
(Figure 3–8).
Figure 3–8Host IB Users Screen
3.2.1.4Host Port Speed
HOST SPEED lets you display and change the port speed on the host port(s). Y ou are prompted for the
desired speed as well as for the choice of host port(s).
3.2.2Configure and Monitor Status of Storage Assets
Disk and Channel Information
The DISK command displays the current disk configuration and the status of the ten disk channels
(ABCDEFGHPS) on the controller (Figure 3–9).
34 007-5510-002
Controller Management
Figure 3–9Disk Channel Screen
If the channel status is “acquiring loop synchronization,” this may indicate a channel problem. Refer to
4.2.2, "Recovering from Drive Failures" for recovery information.
Entering DISK INFO=<tier><channel> retrieves information about a specific disk (tier, channel).
DISK LIST displays a list of the disks installed in the system and indicates how many were found.
DISK SCAN checks each disk channel in the system for any new disks and verifies that the existing disks
are in the correct location. DISK SCAN also starts a rebuild operation on any failed disks which pass the
disk diagnostics.
DISK ST A TUS displays the loop status of each disk channel and a count of the SAS errors encountered
on each channel (Figure 3–10).
DISK ST A TUSCLEAR=<tier><channel> resets the channel error counts.
DISK DEFECTLIST=<tier><channel> displays the defect list information for a specified disk
. The tier
is in the range of <1..125>. The channel is one of the following: A, B, C, D, E, F, G, P, or S. The list is
classified into two types: G (Grown) and P (Permanent) (Figure 3–11).
The G list consists of the sectors that have become bad after the disk has left the factory and which can
be added to at any time. The P list consists of the bad sectors that are found by the disk manufacturer .
007-5510-002 35
Figure 3–11 Disk Defect List Screen
DISK FAIL=<tier><channel> instructs the system to fail the specified disk at the physical tier in the
range of <1 ...125> and channel in the range of <ABCDEFGHPS>. When a non-SP ARE disk is specified
and it is failing, the disk will not cause a multi-channel failure. The disk is marked as failed. An attempt
is made to replace it with a spare disk. When a SPARE disk is specified and it is currently in use as a
replacement for a failed disk, the disk that the spare is replacing is put back to a failed status and the spare
is released, but marked as unhealthy and unavailable.
DISK PLS=<tier><channel> requests/displays the PHY Link Error Status Block information for the
specified drive.
The disk is specified by its physical tier and channel locations, 'tc', where:
• 't' indicates the tier in the range <1..120>, and
• 'c' indicates the channel in the range <ABCDEFGHPS>.
If neither the tier nor the channel are specified, the PLS information is requested from all drives. If only
the tier is specified, the PLS information is requested from all the drives on the specified tier.
T able 3–2 on page 37 shows the types of PHY errors. Note that SA T A and SAS drives report PHY errors
differently.
36 007-5510-002
Controller Management
Table 3–2 PHY Link Error Status Block Information
SATA AAMUX PHY ERRORS
H-RX The number of SATA FIS CRC errors received on the host port of the AAMUX
H-TX The number of SATA R_ERR primitives received on the host port indicating a problem
with the transmitter of the AAMUX
H-LinkThe number of times the PHY has lost link on the host port.
H-Disp The number of frame errors for the host port of the AAMUX. These include: code error,
disparity error, or realignment
O-RX The number of SATA FIS CRC errors received on the other host port of the AAMUX
O-TXThe number of SATA R_ERR primitives received on the other host port indicating a
problem with the transmitter of the AAMUX
O-Link The number of times the PHY has lost link on the other host port.
O-DispThe number of frame errors for the other host port of the AAMUX. These include: code
error, disparity error, or realignment.
D-RXThe number of SATA FIS CRC errors received on the device port of the AAMUX
D-TX The number of SA TA R_ERR primitives received on the device port indicating a problem
with the transmitter of the AAMUX.
D-Link The number of times the PHY has lost link on the device port.
D-DispThe number of frame errors for the device port of the AAMUX. These include: code error,
disparity error, or realignment.
SAS PHY ERRORS
InvDW Invalid DWORD Count - The number of invalid dwords received outside of the PHY
reset sequence.
RunDisRunning disparity Count - The number of dwords containing running disparity errors
received outside of the PHY reset sequence
LDWSYNLoss of DWORD synchronization count - The number of times the PHY has lost
synchronization and restarted the link reset sequence
PHYRESPHY Reset Problem count - The number of times the PHY reset sequence has failed.
NOTE :
SATA drives have an Active/Active MUX (AAMUX) installed. Error counts are read
directly from the AAMUX.
For other DISK parameters, see section 3.2.10 "Automatic Drive Rebuild" in this guide.
3.2.2.1Tier View
Tiers (also known as RAID groups) are the basic building blocks of the controller. In an 8+1 mode, a
tier contains 10 drives: eight (8) data drives (Channels A through H), one(1) parity drive (Channel P),
and an optional spare drive (Channel S). In an 8+2 mode, a tier contains 10 drives, but the setup is
different: eight (8) data drives (Channels A through H), and two (2) parity drives (Channel P and S).
Drives that have the same SAS ID across all ten channels are put on the same tier. Tiers are automatically
added to the system when the disks are detected. A tier will automatically be removed if it is not in use
by any of the LUNs and all of the disks in the tier are removed or moved to another location.
Entering TIER lets you display the current status and configuration of the tiers in the system.
The tiers’ total and available capacities are shown under the “Capacity” and “Space A vailable” columns
respectively.
The TIER command shows the status of each disk on the tier as follows:
• A letter <ABCDEFGHPS> represents a healthy disk at that location.
• A space indicates that the disk is not present or detected.
• A period (.) denotes that the disk was failed by the system.
• The symbol “?” indicates that the disk has failed the diagnostic tests or is not configured correctly .
• The character “r” indicates that the disk was failed by the system and replaced by a spare disk.
• The symbol “!” indicates that the disk is in the wrong location.
NOTE :
The rate of rebuild and format operations can be adjusted with the commands, Tier
Delay=x and Tier Extent=x.
38 007-5510-002
3.2.2.2Tier Configuration
TIER CONFIG displays the detailed tier configuration information for all of the tiers (Figure 3–13).
Figure 3–13 Tier Configuration Screen
The headings for the Tier Configuration screen indicate the following values or conditions for the tiers.
Total LUNs lists the number of LUNs that currently reside on the tier.
Controller Management
NOTE :
The health indication for the spare channel under the 'Healthy Disks' heading is an
indication of the health of the spare disk (if any) that is currently being used to replace
a disk on the listed tier. The health indication for the spare channel that is physically
on the listed tier is found under the 'Sp H' heading.
These headings indicate the respective conditions on the tier.
Ffailed disk (if any).
Rreplaced disk (if any).
Sp Hwhether the spare disk that is physically on the tier is healthy.
Sp A whether the spare disk that is physically on the tier is available for use as a
replacement.
Spare Ownercurrent owner of the physical spare tier , where ownership is assigned when the
spare is used as a replacement.
Spare Used on tier (if any) on which this physical spare is being used as a replacement.
Repl Spare from tier (if any) whose spare disk is being used as a replacement. Tiers are in 8+1
mode by default.
NOTE :
RES-# will display under the Spare Owner heading while a replacement operation is
underway to indicate that unit '#' currently has the spare reserved.
TIER CONFIG=ALL displays tier configuration and replacements for both 8+1 and 8+2 modes. (Figure
3–14).
007-5510-002 39
Figure 3–14 Tier Configuration ALL Screen
3.2.2.3LUN View
Entering the LUN command displays the current status of the LUNs (
). “Ready” indicates that the LUN is in good condition. The percentage of completion is displayed if the
LUN is being formatted or rebuilt. A status of “Unavailable” may result from multiple drive failures.
“Ready [GHS]” indicates that a spare drive has been successfully swapped for one of the drives the tier.
Figure 3–15 LUN Status Screen
LUN LIST displays a list of all valid LUNs in the system. The list shows the capacity, owner, status, and
LUN RESERV A TIONS displays a list of all valid LUNs in the system and shows which LUNs currently
have a SCSI reservation and which initiator holds the reservation (Figure 3–18).
LUN RELEASE releases any SCSI reservations and registrations on a LUN.
No SCSI Reservations0
No SCSI Reservations1Ready
No SCSI Reservations2Ready
3.2.2.6Adding/Removing Storage Assets
The controller supports up to 120 tiers. New tiers can be added without affecting system operations.
DISK SCAN checks each disk channel in the system for any new disks. New tiers are automatically
added to the system when the disks are detected. A tier is automatically deleted if it is not in use by any
of the LUNs and all of the disks in the tier are removed or moved to another location.
PortLUNStatus
007-5510-002 41
3.2.2.7Status of Drive Enclosures
The SES command displays the failures reported by the enclosure (Figure 3–19), through the SCSI
Enclosure Services (SES). It also provides a means to access SES specific functions such as disk,
channel, and LUN. Drive failures are not displayed using the SES command; you must use the TIER
command to view drive status.
15000 [1]: ses
EncID:50050CC0000033C8: Power Supply 1:DC Power Failure
Figure 3–19 Displaying the Current Disk Enclosure Failures
If your enclosures provide redundant SES communication paths, the error is reported twice. In Figur e
3–19, EncID is the Enclosure Logical Indentifier of the enclosure that reported the failure. The last
four digits of the WWN are the last four digits of the enclosure’s serial number.
SES ON saves the SES state to the parameter blocks, and starts up the SES monitors.
SES OFF saves the SES state to the parameter blocks, and shuts down the SES monitors.
3.2.2.8Display SES Devices Information
SES SHOWDEVICES displays all the SES devices on all channels.
SES SHOWALL displays all configuration information for all the SES devices on all channels.
SES SHOW=<tier><channel> displays the configuration information and the status information
returned from an SES Enclosure Status page for the SES device for the specified drive in the range of
<1..120> and <ABCDEFGHPS>.
3.2.2.9Visual Indication of Drive
SES IDDISK=<tier><channel> provides a visual indication of the specified drive
(<1..120><ABCDEFGHPS>). The status LED of the drive blinks until the command SES ID=OFF is
issued. The SES ID=OFF command restores the system to its original visual state.
3.2.2.10Visual Indication of Tier
SES IDTIER=<tier> provides a visual indication of the specified tier <1..125>. The status LED of the
drives blinks until the command SES ID=OFF is issued, which restores the system to its original visual
state.
3.2.2.11Visual Indication of Channel
SES IDCHANNEL=<channel> provides a visual indication of the specified channel
<ABCDEFGHPS>. The status LED of the drives blinks until the command SES ID=OFF is issued,
which restores the system to its original visual state.
3.2.3Tier Mapping for Enclosures
The controller supports various drive enclosures. When the system is first configured, it is necessary to
select a tier mapping mode so that the position of the tiers in the system are changed to conform with the
layout of your drive enclosures. The tier mapping information also allows the controller to properly light
the enclosure fault LEDs.
42 007-5510-002
Controller Management
TIER MAP displays the current mapping mode fo r the disk s in the array.
TIER CHANGEMAP changes the current tier mapping for the disks in the array. To change the current
tier mapping, do the following:
1.Enter: tier changemap
<Enter>
2.Select the appropriate mapping mode for your drive enclosures and press
3.For the changes to take effect, enter: restart
NOTE :
The CHANGEMAP command should only be used when the system is first configured.
Changing the mapping mode will alter all the tier information, making LUN information
inaccessible.
3.2.4System Network Configuration
These commands do the following:
NETWORK displays the current network interface settings.
NETWORK USAGE displays the address resolution protocol map, ICMP (ping), general network, and
IP, TCP, and UDP lay e r statistics.
NETWORK IP=<new address> changes the IP address. (The system must be restarted before the
changes will take effect).
NETWORK NETMASK=<aaa.bbb.ccc.ddd> changes the netmask.
NETWORK GA TEW A Y=<aaa.bbb.ccc.ddd> sets the current gateway in the network routing table to
the supplied Internet address. The gateway is where IP datagrams are routed when there is no specific
routing table entry available for the destination IP network or host.
<Enter>
.
<Enter>
.
NOTE :
GA TEW AY=<no Internet address> clears out the current gateway.
NETWORK PRIV A TE displays the MAC address for the private network device.
3.2.4.1Telnet
NETWORK TELNET=ON|OFF enables/disables the Telnet capability on the controller. The system
must be restarted before the changes will take effect.
NOTE :
NOTE :
007-5510-002 43
T o only temporarily affect Telnet session availability during a concurrent power-cycle,
refer to the TELNET command in Section 3.6 "Remote Login Management" in this
guide.
Telnet connections are “clear text.” If Telnet connections are used, you may expose
controller passwords to third parties. For greater security, turn off Telnet access if it is
not required.
NETWORK TELNETPORT=<port number> changes the T elnet port number for the current controller .
The system must be restarted before the changes will take effect. Valid port numbers are 0 to 32768;
however, the results may be unpredictable if the port number chosen is already in use (on this unit) by
either the GUI or SYSLOG facilities. The default port number is 23.
3.2.4.2SNMP & Syslog
NETWORK SNMP=ON|OFF enables and disables the SNMP functionality. The system must be
restarted before the changes will take effect.
NETWORK LIMIT_SNMP=ON|OFF specifies whether the SNMP functionality will only report
component-level information, or all levels of information, The default setting is OFF. NETWORK TRAPIP=<aaa.bbb.ccc.ddd> changes the destination IP address for SNMP trap packets.
The system must be restarted before the changes will take effect.
NETWORK SYSLOG=ON|OFF enables and disables the Syslog capability.
NOTE :
NETWORK SYSLOGIP=<aaa.bbb.ccc.ddd> changes the destination IP address for syslog packets,
Both controllers in the couplet pair will share the same syslog destination IP address but each controller
can specify a different destination port.
NETWORK SYSLOGPORT=<port number> changes the destination port number for syslog packets
for the current controller. Both controllers in the couplet pair will share the same syslog destination IP
address but each controller can specify a different destination port. Valid ports are 0 to 32768. However
the results may be unpredictable if the port number chosen is already in use (on this unit) by the TELNET
facilities. Default port number is 514.
NOTE :
Refer to Chapter 4,"Controller Remote Management and Troubleshooting" in this guide for information
on how to set up Telnet and SNMP functionality on your host computer.
NETWORK SYSLOG should be enabled , since it is the best way to find out what
occurred in the event of a problem. However, since some problems can produce a large
amount of output, it is a good idea to have your syslog program configured to rotate
based on log size rather than date.
The controller sends syslog messages via the local 7 (23) facility.
3.2.4.3API Server Connections
NETWORK API_SERVER=ON|OFF enables/disables the API server capability. The system must be
restarted before the changes will take effect.
NOTE :
NETWORK API_PORT=<port number> specifies the API Server port number for the current
controller. The sys tem must be restarted before the changes will take ef fect. Valid ports are 0 to 32768.
The results may be unpredictable if the port number chosen is already in use (on this unit) by either the
TELNET or SYSLOG facilities. The default port number is 8008.
44 007-5510-002
To affect the API Server connection availability only temporarily during the current
power-cycle, see Section 3.8.2, "API Server Connections".
3.2.4.4Displaying and Editing the Routing Table
The ROUTE command displays the current routing table of the system (Figure 3–20) and allows the
administrator to change it. The routing table describes how the controller can communicate with the
hosts on other networks.
ROUTE ADD=<aaa.bbb.ccc.ddd> GATEW AY=<aaa.bbb.ccc.ddd> adds gateways to the routing
table. Up to 6 permanent routes can be added to the tables. For example, to indicate that the machine with
Internet address 91.0.0.3 is the gateway to the destination network 90.0.0.0, enter: ROUTE
ADD=90.0.0.0 GA TEW A Y=91.0.0.3
ROUTE DEL=<aaa.bbb.ccc.ddd> GA TEW AY=<aaa.bbb.ccc.ddd> deletes gateways from the routing
table.
ROUTE GATEW AY=<aaa.bbb.ccc.ddd> sets the current gateway in the network routing table to the
specified Internet address. The gateway is where IP datagrams are routed when there is no specific
routing table entry available for the destination IP network or host. If an empty gateway value is
provided, then the current gateway is cleared.
3.2.5Restarting the Controller
3.2.5.1System Restart
REST ART performs a restart on the controller on which the command is issued. This command prepares
the system to be restarted. The system halts all I/O requests and saves the data to the disks before
restarting. The restart process may take several minutes to complete.
NOTE :
If cache coherency is enabled, restarting a controller unit will cause the partner
controller to fail the restarting unit. Once the reboot is complete, you will have to heal
the controller unit.
REST ART DELAY=X (where “X” is minutes) delays a restart of a unit between 0 and 255 minutes.
007-5510-002 45
REST ART DUAL restarts both units.
REST ART KILL stops a timed restart that is in progress.
3.2.5.2System Shutdown
SHUTDOWN shuts down the controller unit.
If you need to power down the controller, use SHUTDOWN prior to shutting off power. This will cause
the controller to flush its cache, abort all format and rebuild operations, and proceed with an orderly
shutdown.
All hosts actively using the controller should be safely shutdown and all users logged out before using
this command. The controller will halt all I/O requests and save the data to the disks.
NOTE :
T o perfor m a hard res tart of t he unit by cycling the power, use: SHUTDOWN REST ART=X, where X is
a value between 1 and 1023 seconds before the unit powers up again. If the number is not specified, the
default is 15 seconds.
NOTE :
SHUTDOWN DELA Y=X delays a shutdown of a unit between 0 and 255 minutes (where x is minutes
delayed).
SHUTDOWN DUAL shutdowns both units.
SHUTDOWN KILL stops a timed shutdown that is in progress.
Use SHUTDOWN whenever you power down the controller for maintenance.
SHUTDOWN flushes any data left in the cache and prepares the controller for an orderly
shutdown. For couplet controller configuration, issue SHUTDOWN to both controllers.
If SHUTDOWN RESTART is used in conjunction with the DUAL parameter, the
restart will only affect unit where it was issued on and not both.
3.2.6Setting the System’s Date and T ime
Valid date setting s are bet ween years 200 0 and 2104. In dual mode, settings should always be done on
Unit 1. Changes will automatically be applied to both units. Settings are automatically adjusted for leap
years.
3.2.6.1System Date
DA TE displays the current system date.
Yo u can also change the system date. At the prompt, type:
date mm dd yyyy
where mm represents the two digit value for month, dd represents the two digit value for day , and yyyy
represents the four digit value for year.
For example, to change the date to March 14, 2009, type:
date 03 14 2009
46 007-5510-002
<Enter>
<Enter>.
3.2.6.2System Time
TIME displays the current system time.
You can also change the system time. At the prompt, type:
Controller Management
time hh:mm:ss
<Enter>
where hh represents the two digit value for hour (00 to 24), mm represents the two digit value for
minutes, and ss represents the two digit value for seconds.
For example, to change the system time to 2:15:32 p.m., type:
time 14:15:32
NOTE :
The system records time using the military method, which records hours from 00 to 24,
<Enter>
not in a.m. and p.m. increments of 1 to 12.
3.2.7Saving the Controller’s Configuration
The SA VE command can be used to save the system configuration to non-volatile memory
(Figure 3–21).
15000 [1]: save
Saving system parameters. Done.
Figure 3–21 Saving System Parameters Screen
Backup copies of the system configuration are also saved on the disks. The system will automatically
save and update the backup copies when changes are made to the system configuration or status.
The SA VE ST ATUS command, in addition to saving the parameter blocks to non-volatile memory and
on the disks, displays the current status of the system parameters
Figure 3–22 Current System Parameters Status Screen
(Figure 3–22).
Normally, the system must determine which copy of the parameter blocks is more recent, the one on the
disks or the internal copy. When the system reboots, it will load the more recent copy.
3.2.8Restoring the System’s Default Configuration
The DEFAULTS command may be used to restore the system to its default configuration.
007-5510-002 47
The DEFAULT command will delete all LUN configuration and data
!
Warning
The system will halt all I/O requests, delete all the LUNs and restore all the parameters back to their
default values. This is a destructive operation which will delete all the data stored in the system.
The system will ask if you want to erase all the configuration information stored on the disks. This will
prevent the system from retrieving the backup copies of the configuration settings from the disks after
the system is restarted. After the default settings have been loaded, the system will ask if you want to
begin reconfiguration by scanning for the disks. New LUNs can be created after the disks have been
added back to the system.
unconditionally. Do not issue this command without guidance from SGI.
3.2.9LUN Management
The controller creates centrally-managed and vendor-independent storage pooling. It enables different
types of storage to be aggregated into a single logical storage resource from which virtual volumes
(LUNs) can be served up to multi-vendor host computers. The networked storage pools will provide the
framework to manage the growth in storage demand from web-based applications, database growth,
network data-intensive applications, and disaster tolerance capabilities.
3.2.9.1Configuring the Storage Array
The storage array may consist of up to 120 tiers, depending on individual disk enclosure’s numbering
scheme. The tiers can be combined, used individually, or split into multiple LUNs. A LUN can be as
small as part of a tier or as big as the whole system. LUNs can be shared or dedicated to individual users.
Up to 1024 LUNs are supported in total. LUNs are “owned” by the controller via which they are created.
You can add and remove LUNs without affecting system operations. Use the LUN command to display
the current Logical Unit Status (Figure 3–23).
NOTE :
Figure 3–23 Logical Unit St atus Screen
In dual mode, LUNs will be “owned” by the controller unit on which they are created.
Hosts will only see the LUNs on the controller to which they are connected, unless cache
coherency is used.
3.2.9.2Creating a LUN
LUNs can be added to the system based on two commands.
48 007-5510-002
To add a 32-bit LUN that will not exceed 2 TB, type:
Controller Management
lun add=X
To add a 64-bit LUN, that exceeds 2 TB:
lun add64=X
For both cases, “X” is the Logical Unit with a range of <0..1023>.
For either case, the system prompts you for all the necessary information to create the LUN and indicates
if the LUN was successfully added to the system.
The required LUN information includes:
• Capacity (in MBytes) - default is to use all available capacity
• Number of tiers - default is to use all tiers
• Block size (in Bytes) - default is 512Bytes
• Label - may contain up to 12 characters
3.2.9.3Formatting a LUN
A LUN must be formatted before it can be used.
To format a LUN, use LUN FORMAT. Specify the LUN <0..1023> when prompted. This performs a
destructive initialization on the specified LUN by over-writing all the data on the LUN with zeroes.
The rate of format can be adjusted using the DELA Y and EXTENT parameters of the LUN command.
<Enter>
<Enter>
3.2.9.4Interrupting a LUN Format Operation
If you need to interrupt a format operation, for any reason, use these commands:
• LUN P AUSE pauses the current format operations.
• LUN RESUME releases the paused format operations.
• LUN STOP aborts all the current format operations.
3.2.9.5Changing a LUN Label
To change the label of a LUN:
1.Type:
lun label
2.Select the LUN to change <0..1023> and press Enter.
3.T ype in the new label and press Enter. A LUN label may contain up to 12 characters (Figure 3–24).
T o change the owners hip of a LUN from one contr oller to its part ner (when the units are i n dual mode),
enter:
where x is the Logical Unit number <0..1023> (Figure 3–25). If a LUN is on a tier that is shared by other
LUNs, the controller will prompt and then move the other dependent LUNs as well.
15000 [1]: lun move=0
LUN 0 is owned by this 15000.
Do you want to move ownership to the OTHER 15000? (y/n):
Figure 3–25 Moving a LUN
3.2.9.7Deleting a LUN
LUN DEL=x (where “x” is the LUN <0.1023>) deletes a LUN from the system. You can only delete a
LUN that is owned by the controller unit onto which you are logged.
SCSI Reservations
LUN RELEASE=x allows you to release all SCSI reservations on a LUN. The command LUN
RESERV ATIONS can be used to view the current SCSI reservations on all of the LUNs in the system.
The LUN to be released can be specified by “x” where “x” is in the range <0..1023>.
LUN ST ART lets you start all the LUNs that have been stopped by a SCSI START/STOP request. This
parameter is not related to the LUN STOP command.
lun move=x
<Enter>
3.2.10Automatic Drive Rebuild
The controller’s automatic drive failure recovery procedures ensure that absolute data integrity is
maintained while operating in degraded mode. In the event of a drive failure, the controller will
automatically initiate a drive rebuild using a spare drive if the “autorebuild” function has been enabled.
50 007-5510-002
Controller Management
Use the TIER command to display the current setting (Figure 3–26). The rebuild operation can take up
to several hours to complete, depending on the size of the disk and rate of rebuild.
TIER AUTOREBUILD=ON|OFF enables/disables the automatic disk rebuild function. A disk will only
be replaced by a spare disk if it fails and Autorebuild is ON (ON being the default setting). This function
should always be enabled so that data can be reconstructed on the spare drive when a drive failure occurs.
After the failed drive is replaced, data will be automatically copied from the spare drive to the
replacement drive.
3.2.10.1Manual Drive Rebuild
DISK REBUILD=<tier><channel> initiates a rebuild on a specific drive. This operation will reconstruct
data on the replacement drive and restore a degraded LUN to healthy status.
3.2.10.2Drive Rebuild Verify
DISK REBUILDVERIFY=ON|OFF determines if the system will send SCSI Write with Verify
commands to the disks when rebuilding failed disks. This feature is used to guarantee that the data on the
disks is rebuilt correctly. Default is OFF. This feature will increase the time it takes fo r rebuilds to
complete.
3.2.10.3Manual Drive Replace
To replace the specified failed drive with a spare drive, enter: DISK REPLACE=<tier><channel>
A Replace operation is used to temporarily replace a failed disk with a healthy spare disk.
3.2.10.4Interrupting a Rebuild Operation
To interrupt a Rebuild operation, use these commands:
• TIER P AUSE pauses the current rebuild operations.
• TIER RESUME releases the paused rebuild operations.
• TIER STOP aborts all the current rebuild operations.
3.2.11SMART Command
Use the SMART command to identify failing drives before they fail.
007-5510-002 51
SMART ENABLE enables SMART on all the disk drives installed in the system and updates the
parameter blocks on the disk. This enables the Information Exception and the Temperature warnings.
However, the user can skip the update part and enter SMART DISKUPDATE later to write the
parameter blocks to the disks.
SMART DISABLE disables SMART on all the disk drives installed in the system and updates the
parameter blocks on the disk. However, the user can skip the update part and call and enter SMART
DISKUPDATE later to write the parameter blocks to the disks.
SMART ST A TUS displays the SMAR T ENABLE status of a specified disk. The disk is specified by its
physical tier and channel locations, 'tc', where:
• 't' indicates the tier in the range <1..125>
• 'c' indicates the channel in the range <ABCDEFGHPS>.
Displays the Information Exception and the Temperature warnings if enabled for a specified disk; (i.e.
SMART ENABLE).
SMART UPDA TE updates the parameter blocks that correspond to the SMART configuration.
SMART DA TA=tc displays the SELF TEST log. Also reads and displays the SMAR T information of the
specified disk. The disk is specified by its physical tier and channel locations, 'tc', where:
• 't' indicates the tier in the range <1..125>
• 'c' indicates the channel in the range <ABCDEFGHPS>.
SMART SELFTEST=tc|testtype starts a specified self test on a specified hard disk. The disk is specified
by its physical tier and channel locations, 'tc', where:
• 't' indicates the tier in the range <1..125>
• 'c' indicates the channel in the range <ABCDEFGHPS>.
There are 3 tests to choose from: default test, Background short test, and Background long test. All the
tests are supported on the SAS drives, while only the Background short test is supported on the SATA
drives.
SMART ABORTSELFTEST=tc aborts a self test that has been launched using the SMAR T SELFTEST
command. The abort event will logged to the self test log.
The disk is specified by its physical tier and channel locations, 'tc', where:
• 't' indicates the tier in the range <1..125>
• 'c' indicates the channel in the range <ABCDEFGHPS>.
SMART ABORTSELFTEST=ALL. This command only works for SAS disks. Aborts background self
tests on all disks.
SMART CLEAR=tc|all Clears SMART trips on specified drives or all drives.
• 't' indicates the tier in the range <1..125>
• 'c' indicates the channel in the range <ABCDEFGHPS >
• ‘ALL’ indicates all drives.
SMART LOG=tc|ALL reads the self test log from the specified disk and displays it.
• 't' indicates the tier in the range <1..125>
• 'c' indicates the channel in the range <ABCDEFGHPS>.
52 007-5510-002
Controller Management
SMART TEST=ON enables the test bit in the Information exception mode page for all the disks installed.
Setting the test bit simulates a F ALSE SMART trip condition which raises a FALSE check condition to
the controller. Currently, th is parameter is valid only with Fibre Channel disks.
SMART TEST=OFF disables the test bit in the Information exception mode page for all the disks
installed.
SMART INTERV ALTIME displays the interval (in hours) in which the SMART Information will be
polled.
SMART INTERV AL TIME=h sets the interval (in hours) that SMART Information polling will occur. 'h'
indicates the interval in hours in the range <1..24>.
There are two primary couplet controller configurations: cache coherent and non-cache coherent.
The DUAL command displays information about couplet system configuration
Figure 3–27 Couplet Controller Configuration
3.2.12.1Cache Coherent
In this configuration, each controller can access all LUNs. The couplet controller communication occurs
over the internal UART and private Ethernet. If the controllers detect an Ethernet failure, controller 2
will be failed. (This means that an external event can cause a controller to fail even though the controller
may be perfectly fine.) Therefore, it is mandatory that the controller Ethernet resides on a private
Ethernet segment.
NOTE :
Data cache is not copied from one controller to another. If a controller fails, all “dirty”
data in cache will be lost. Thus if power failures are a concern, writeback cache should
be disabled.
(Figure 3–27).
Non-Cache Coherent
In this configuration, the couplet controller communication occurs over the internal UART. Each
controller owns LUNs and tiers. Spare drives are “owned” by individual controller units, according to
tier ownership.
In healthy situations, the controller cannot access LUNs or tiers owned by the other controller. However ,
if the other controller is failed, the healthy controller will have access to all LUNs and tiers.
Users, via mapping, can be assigned any combination of LUNs. In a healthy environment, users will only
see LUNs owned by the controller to which they are connected.
007-5510-002 53
For example, a user is given access to internal LUNs 5, 6, and 7, which are mapped to external LUNs 0,
1, and 2, respectively. controller 1 owns LUNs 0 and 1 while controller 2 owns LUN 2 . The user is
physically connected to controller 1, thus, they will only see LUNs 0 and 1. The user will not be able to
access LUN 2. If the user was physically connected to controller Unit 2, the reverse would be true: only
LUN 2 would be accessible. When a controller fails, the user will be given access to all mapped LUNs
regardless of the physical connection.
Data cache is not copied from one controller to another. If a controller fails, all “dirty” data in cache will
be lost. Thus if power failures are a concern, writeback cache should be disabled.
DUAL COHERENCY=ON|OFF enables/disables the cache coherency function.
Default is dual
coherency disabled which is the non-cache coherent configuration.
DUAL TIMEOUT=X allows you to set the cache coherency timeout for cache node requests in seconds.
Valid range is <0...255>. Defau lt is zero (0) seconds. The timeout value should be less than the host
timeout value (HOST TIMEOUT=X). A timeout value of 0 allows for only one retry.
NOTE :
In dual mode, LUNs will be “owned” by the controller unit on which they are created.
Hosts will only see the LUNs on the controller to which they are connected, unless cache
coherency is enabled.
3.2.12.2Fail / Restore the Other Controller Unit in the Couplet Pair
To fail the other controller unit in the system (for example, in order to perform maintenance), enter:
dual fail
The healthy controller unit wil l take ownership of all the LUNs/tiers from the failed controller unit.
To restore the other controller unit in the system to healthy status after failure recovery, enter:
dual heal
<Enter>
<Enter>
Ownership of LUNs/tiers are transferred back to the formerly failed controller unit.
3.2.12.3Labeling the Controller Unit(s)
Y ou may change the label assigned to each controller unit. This allows you to uniquely identify each unit
in the controller system. The CLI prompt for each controller is built by adding a colon (:) and a space at
the end of the label. Each controller can have a label up to 31 characters long.
To change the label:
1.Type: dual label=1|2
2.Select which unit you want to rename (Figure 3–28).
3.When prompted, type in the new label for the selected unit. The new name is displayed.
NOTE :
If you type DEFAUL T for the new label, the label for the unit is restored to its default
setting.
54 007-5510-002
<Enter>.
Figure 3–28 Labeling a Controller Unit
3.2.12.4Singlet
The DUAL SINGLET command sets the system in the singlet mode. System recognizes only Unit 1. This
command:
• disables cache coherency
• heals unit 1 if it is failed
• fails unit 2 before attempting to remove it.
15000 [1]: dual label
Enter the number of the unit you wish to rename.
LABEL=1 for unit 1, Test System[1]
LABEL=2 for unit 2, Test System[2]
unit: 1
Enter a new label for unit 1, or DEFAULT to return to the default label.
Up to 31 characters are permitted.
Current unit name: Test System[1]
New unit name: System[1]
Controller Management
To set the system in singlet mode not couplet mode, type: dual singlet
NOTE :
The system may automatically add unit 2 if it is connected to the system. Therefore, we
advise you to power off and remove unit 2 from the system after the Dual Singlet
command is completed.
3.3Performance Management
The controller optimizes performance operations due to its extensive monitoring and reporting
capability.
3.3.1Optimizing I/O Request Patterns
The controller manages pre-fetch and cache efficiency through the LUN.
Display Current Cache Settings
The CACHE command displays the current cache settings for each LUN in the system( Figure 3–29).
Yo u can use the LUN=x option to specify which LUN to chan ge. If no LUN is specified, changes will
be applied to all the LUNs. V alid LUN values are 0 to 1023. The default value will apply changes to all
LUNs.
3.3.1.1Cache Segment Size
A large cache segment size may give better performance for large I/O requests and a small cache segment
size may give better performance for small I/O requests. For optimal performance, the cache segment
size should be larger than the average host I/O request size. You may use the STA TS LENG TH command
to determine the average host I/O request size. The cache segment size should not be changed during
heavy I/O conditions because the system will temporarily halt all I/O requests while the changes are
taking effect.
Use the CACHE SIZE=x command to set the cache segment size for the specified LUN in kilobytes
(kbs). Valid segment sizes are 128, 256, 512, 1024, and 2048 kilobytes (kbs). The default value is 1024.
This command should not be issued under heavy I/O conditions because the system will momentarily
halt all I/O requests while the changes are taking effect.
3.3.1.2Writeback Cache Settings
Writeback caching allows the system to increase the performance of handling write I/O requests by
storing the data in cache and saving the data to the disks at a later time.
CACHE WRITEBACK=ON|OFF enables or disables writeback caching for the specified LUN. Default
setting is ON.
CACHE WRITELIMIT=x specifies the maximum percentage of the cache that can be used for writeback
caching. The system will force all writeback requests to be flushed to the disks immediately if the
percentage of writeback data in the cache exceeds this value. Valid range is <0...100>. Default value is
75.
3.3.1.3Prefetch Settings
When the system receives a request, it can read more data than has been requested. PREFETCH tells the
system how much data to look ahead. This will improve performance if your system needs to perform
sequential reads. For random I/O applications, however, use the smallest prefetch value.
CACHE PREFETCH=x sets the prefetch that will occur on read commands for the specified LUN. Valid
range is 0 to 65535. Default setting is 1.
56 007-5510-002
If the MF (Multiplication Factor) parameter is OFF, the system will only prefetch the number of blocks
specified by PREFETCH after every read command. If the MF parameter is ON, then the system will
multiply the transfer length of the command by the prefetch value to determine how much data will be
prefetched. A prefetch value of less than 8 is recommended when the MF parameter is ON.
CACHE MF=ON|OFF enables/disables the MF bit on the specified LUN. Default is ON.
The Maximum Prefetch Ceiling paramete r sets the maximum pref etch ceiling in blocks for prefetches
on read commands. It sets an upper limit on prefetching when the MF parameter is ON. The system will
automatically limit the amount of prefetching if the system is running low on resources.
CACHE MAX=x (where X is a range from 0...65535) sets the maximum prefetch ceiling in blocks for
prefetches on Read commands for the specified LUN. V alid range is 0 to 65535. Default setting is 65535.
3.3.1.4Cache Settings Reset
CACHE DEFAULTS loads the default settings for all of the cache parameters for the specified LUNs.
Disk Configuration Settings
The DISK command displays the current disk configuration settings (Figure 3–30).
Controller Management
Figure 3–30 Disk Configuration Setting Screen
The writeback cache and disk timeout settings can be configured manually.
DISK TIMEOUT=x sets the disk timeout for an I/O request in seconds. Valid range is 1 to 512 seconds.
Default setting is 68 seconds.
DISK CMD_TIMEOUT=x sets the Retry Disk timeout (in seconds) for an I/O request. The retry timeout
value indicates the maximum amount of time that is allotted to receive a reply for each retry of an I/O
request. If the I/O request does not complete within this time, it is aborted and potentially retried: if there
is still time remaining in the overall disk timeout to allow for another retry, it is retried; if not, it
completes with an error status.
NOTE :
The DISK CMD_TIMEOUT value must be smaller than or equal to DISK TIMEOUT.
Valid range is 1 to 512 seconds.
007-5510-002 57
3.3.2Audio/Visual Settings of the System
The audio and visual (AV) settings of the system and the disks can be tuned to provide better
performance and a lower latency. The writeback and prefetch settings for each LUN are changed with
the CACHE command.
The AV command displays information about the audio/visual settings of the system (Figure 3–31).
Figure 3–31 Current Audio/Visual Settings
AV FASTAV=ON|OFF enables/disables the disk fast audio/video read options for streaming data. When
enabled, the system will start the data transfer for read operations before all of the disk commands have
finished. This feature reduces the latency for read operations but the system will be unable to check the
integrity of the data. This parameter is saved on a LUN by LUN basis. Use LUN=X command to change
the settings for a single LUN. Default setting is OFF.
NOTE :
When FASTAV mode is enabled, the controller no longer checks data in real-time.
Changing the disk parameters can adversely affect the I/O operation of the system. This
parameter should only be adjusted when the system is idle. Default setting is OFF.
FASTAVTIMEOUT=x sets the timeout before the FASTAV option activates on a host read command.
The FASTAV mechanism is not used until the host command takes longer than the timeout value. A
value of zero indicates that the system starts the data transfer as soon as a minimum number of drives are
ready. This value is in 100 millisecond increments. The range for “x” is 0 to 255. The default is 50.
ORDEREDQUEUE=x enables the use of ordered tags when communicating with the drives. The value
“x” indicates the number of disk commands that can be sent before an ordered tag must be sent to the
disks. Valid range is 0 to 255. Default is 0.
UA=ON enables the initial Unit Attention condition when an initiator logs into the system; the system
reports a Unit Attention condition on the first SCSI command after the initiator logs in. Default is ON.
UA=OFF disables the initial Unit Attention condition when an initiator logs into the system; the system
automatically clears the unit attention condition when an initiator logs in.
58 007-5510-002
Controller Management
RC=ON|OFF enables the Read Continuous (RC) option for Audio/Video streaming data; the system
starts the data transfer for read operations after RCTIMEOUT is reached, even if the disks commands
have not finished. Use this to reduce the latency for read operations in Audio/Visual environments where
latency is more important than data integrity. This parameter is saved on a per-LUN basis. Use in
combination with the LUN=x parameter to change the settings for a single LUN. Enabling this feature
automatically enables FASTAV.
WARNING! This feature allows the system to return invalid data to the initiator.
RCTIMEOUT=x Default setting: disables the Read Continuous option for Audio/Video streaming data.
Note: This parameter is saved on a per-LUN basis. Use in combination with the LUN=x parameter to
change the settings for a single LUN.
LUN=x sets the host command timeout for the Read Continuous option for Audio/Video streaming data.
Set to 0 to disable the Read Continuous feature in the system. This value is in 100 millisecond
increments. The range for 'x' is 0 to 255. The default is 8.
FAILCC=ON instructs the host ports to report a check condition for all SCSI commands when the unit
is in a failed state. This command should only be used in AV environments when a check condition is
required instead of taking the unit off the loop.
FAILCC=OFF This is the default setting. Host ports will NOT report a check condition for all SCSI
commands when the unit is in a failed state.
3.3.3Locking LUN in Cache
Locking a LUN in data cache will keep all of the data for the LUN in the cache for faster access. Once
a LUN is locked, the data that is gathered to service read and write commands will stay permanently in
the cache. The controller will continue to fill up the cache until 50% of the total cache is filled with data
from locked LUNs, while the other 50% of the cache is reserved to service I/O for unlocked LUNs.
Initial Cache
50% of Data Cache used
to service Unlocked LUNs
For example, when a host issues a read command for data from LUN 1 that has been locked in cache, the
following will occur:
• The controller reads data from disks, locks data in cache, and sends data to host
• Any reads of the same data will be serviced from cache, which provides faster access than reading
from disks.
Cache allocation after I/O completes
Unlocked LUN data
* Unallocated cache can be used for unlocked LUNs’ or locked LUNs’ data.
Once cache has been allocated to a locked LUN, however, it cannot be used
by an unlocked LUN.
Once the size of the locked LUNs exceeds 50% of the total cache, the controller must create cache space
to process a new I/O by removing older data from the locked portion of cache. The Least Recently Used (LRU) algorithm is used to determine which locked data to remove from cache.
50% of Data Cache used
to service Locked LUNs
Unallocated
cache*
Data for
LUN 1
007-5510-002 59
For example, LUNs 0 to 3 are locked in cache and all 50% of the total cache has been filled by data from
LUN 0, 1, and 2.
Initial Cache
Unlocked LUN data
Data for
LUN 0
Data for
LUN 1
Data for
LUN 2
When a host issues a read command for data from LUN 3, the following will occur:
• The controller unit determines which data to remove from the locked portion of cache, using the
LRU algorithm. The LRU algorithm is thus: If LUN 0 has not been accessed for 1 hour, LUN 1
has not been accessed for 30 minutes, and LUN 2 has not been accessed for 2 minutes, then
LUN 0’s data will be removed from cache because it is the least recently used data.
• The controller reads data from disks, locks data in cache, sends data to host.
• Any reads of same data will be serviced from cache (until data is removed from cache due to its
being the least recently used data).
Cache allocation after I/O completes
Unlocked LUN data
Data for
LUN 3
Data for
LUN 1
Data for
LUN 2
3.3.3.1Locking / Unlocking a LUN
To lock a LUN in the data cache, enter:
LUN LOCK=X
<Enter>
where “X” is the Logical Unit number <0..1023> (Figure 3–32
The controller monitors pre-fetch and cache efficiency, request distribution, transaction, and transfer
rates by port.
The ST ATS command displays the Performance S tatistics for the host ports, disk channels, and cache
memory (Figure 3–33
60 007-5510-002
). It will show the read and write performance of each of the host ports.
Controller Management
Figure 3–33 System Performance Statistics Screen
Read Hits shows the percentage of Read I/O requests where the data was already in the cache. Prefetch
Hits shows the percentage of Read I/O requests where the data was already in the cache due to
prefetching. Prefetches shows the percentage of host Read I/O requests to the disks due to prefetching.
The bottom of the screen displays the Read and Write performance of the disks. Disk Pieces shows the
total number of disk I/O requests from the host ports. The system will combine several host I/O requests
into a single disk I/O request. The histogram at the lower right shows how often this is occurring for reads
and writes. BDB Pieces is the number of host I/O blocking and deblocking requests.
Cache Writeback Data shows the percentage of the cache which contains writeback data that must be
written to the disks. Cache Rebuild Data shows the percentage of the cache in use for rebuild operations.
Cache Data Lock shows the percentage of the cache which is locked by the locked LUNs.
ST A TS CLEAR resets all the statistics back to zero.
007-5510-002 61
ST A TS DELAY displays a histogram of the time it takes for the host and disk I/O requests to complete
in 100 msec intervals (Figure 3– 34).
ST A TS HOSTDELAY displays a histogram of the time delay between when the last data transfer is set
ready and the host command completes (Figure 3–35). The host ready delay information is shown in
100msec intervals.
15000 [1]: stats hostdelay
Host Command Ready Delay Statistics
Time
seconds Reads Writes
0.100000000
0.200010000
0.300012100
0.400120200
0.500000000
0.600010200
0.700010200
0.800000000
0.900000000
1.000000200
1.100000000
1.200000100
1.300002100
1.400002100
1.500000000
1.600000000
Figure 3–35 Host Delay Statistics Screen
Port 1
Port 2
Reads Writes
Port 3
Reads Writes
Port 4
Reads Writes
ST ATS TIERDELA Y=<tier> displays a histogram of the time it takes for the disk I/O request to complete
for all the disks in the specified tier (Figure 3–36). If no tier is specified, all valid tiers will be displayed.
ST A TS DISK displays a histogram of the disks in the system that have taken an unusually long time to
complete an I/O request (Figure 3–37). The count is incremented for a disk if that disk takes longer than
the other disks to finish an I/O request. This command is used to determine if a disk in the array is slowing
down system performance.
Normally all the disks in a tier should have similar counts. A disk with a significantly higher count
indicates that the disk may be slower than the other disks or have problems.
Most commands should be in column 0 or 4 for the best performance.
x0x1x2x3x4x5x6x7
Figure 3–40 Host Command Offsets Screen
ST A TS REPEAT=OFF|MBS|IOS allows you to enable/disable the repeating statistics display where
MBS displays MB/s, IOS displays IO/s, and OFF turns off (both) the repeating displays.
3.3.4Resources Allocation
Background Format/Rebuild Operations
Format and rebuild operations are background processes; their rates can be adjusted to minimize their
impact on system performance.
Figure 3–41 Displaying the Current Rebuild Parameters
TIER displays the current rebuild parameter settings for the system (Figure 3–41).
The TIER DELA Y parameter controls the amount of system wait time before rebuilding the next chunk
of data. This parameter slows down the rebuild and format operations so they will not affect the
performance of the system. TIER DELA Y=0 will remove many delays so the rebuild and format
operations will go as fast as possible, but this could significantly affect the performance of the system.
NOTE :
66 007-5510-002
A delay value less than 1 (<1) is not recommended.
TIER DELA Y=x is used to set the system rebuild/format delay. This value is in 100 millisecond
increments. The range is 0 to 1000. The default setting is 30 milliseconds (3 seconds).
The REBUILD EXTENT parameter determines how much data to rebuild or format at one time. A small
EXTENT value will slow down the rebuild and format operations so they will not affect the performance
of the system. Increasing the EXTENT value will allow more data to be rebuilt in a single pass. The
recommended setting is to use the default value of 32 MBytes (MBs) and only adjust DELAY to match
your user load.
TIER EXTENT=X (where X is a value from 1 to 128) sets the system rebuild/format extent in MBs. The
range is 1 to 128 MBs. Default is 32MBs.
3.3.4.1Background LUN Verify Operations
LUN VERIFY displays the current setting for background verify on all LUNs.
LUN VERIFY=X turns on background verify for LUN X, where X is a Logical Unit <0..1023>.
LUN VERIFY=ON|OFF prompts you for a list of LUNs where the background verify will be turned either
ON or OFF .
LUN VERIFY=ON will both turn on the background verify for the specified LUN(s), as well as start up
the verify operation(s).
Controller Management
LUN VERIFY=OFF only turns off the Background Verify setting for the specified LUN(s). Therefore,
any Verifys that are already active on the LUN(s) will not terminate until after the completion of that
Verify's current iteration. To stop all verify operations immediately, use the LUN STOP command.
NOTE :
LUN DELA Y= X sets the system Verify Delay value to X, where x is a value from 0 to 1000. The Verify
Delay value determines how long a verify operation will pause after it reaches the verify extent. This
parameter slows down the verify operation so that it will not affect the performance of the system (except
in the case where X is set to 0, as described below).
DELA Y=X will remove all delays so that the verify operation will go as fast as possible; however, this
will slow down the performance of the system. This value is in 100 millisecond increments. The range
for X is 0 to 1000. Default is 40.
LUN EXTENT=X sets the system verify extent value X in Mbytes. The verify extent determines how
much data can be verified before the verify operation must pause. This parameter slows down the verify
operation so that it will not affect the performance of the system. Increasing the extent value will allow
more data to be verified in a single pass. The range for X is 1 to 128 MBs. Default is 32 MBs.
It is recommended that you run LUN VERIFY in continuous mode, since it can help
increase disk reliability.
3.3.4.2Background TIER Verify Operations
TIER VERIFY verifies LUNs on a tier by tier basis. TIER VERIFY differs from LUN VERIFY in that the
number of simultaneous Tier Verifys is limited to a value that is set by TIER MAXVERIFIES (default =
2) parameter. The valid range is 1 to 16. If a tier is marked for continuous verification, once the
verification completes, the next sequential tier marked for verification, not presently being verified, will
start.
007-5510-002 67
TIER VERIFY Displays a summary of verifications.
To enable Tier Verify (Figure 3–42):
1.At the prompt, type TIER VERIFY=ON
<Enter>
.
2.The system will ask which tier you wish to verify. Enter the tier number or type a for “All.”
3.The system will ask if you want run the Tier Verify operation continuously or not. Type y to run
continuously or N to run just once. The default is N.
Figure 3–42 Tier Verify ON Screen
To disable Tier Verify (Figure 3–43):
1.At the prompt, type TIER VERIFY=OFF
<Enter>
.
2.The system will ask which tier you wish to verify. Enter the tier number or type a for “All.”
Tier Verify will be disabled off after the next iteration has completed.
68 007-5510-002
Controller Management
Figure 3–43 Tier Verify OFF Screen
TIER VERIFY=X A specified tier will be verified if possible.
These API and CLI commands will affect the TIER verification process:
• The CLI commands TIER PAUSE, TIER RESUME, TIER STOP.
• The API command TIER STATUSCHANGE.
LUN operations (add, delete, move) can affect TIER VERIFY operations. As tiers are verified, only
LUNs that are valid and formatted are verified.
If a tier is owned by the other unit, and it is healthy, the user is notified that the verification cannot occur
due to the ownership of the tier. The user can then retry the verification on the other unit.
The verification of LUNs on a tier is performed in the order of addressing on the tier. Only valid and
formatted LUNs can be verified.
3.3.4.3Rebuild Journaling
The rebuild journaling feature is intended to speed the recovery from disk-side loss of communication
problems. A loss of communication includes, but is not limited to, SAS expanders, hardware/software
failures, and SAS cable failures, and SFP failures.
Currently, when the controller encounters a loss of communication with a drive or a group of drives on
a SAS link, the software fails the drive(s) and continues operation. When cache coherency is enabled,
if either controller encounters a loss of communication with a drive, the firmware will fail the drive. This
allows a controller unit to maintain operation during disk-side events. Once the loss of communication
is resolved, however, the controller must rebuild all the af fected drives. For large installations, a loss of
007-5510-002 69
communication, such as a cable failure, can cause the controller to fail numerous disk drives. Once the
loss of communication is resolved, the time to rebuild all the failed drives can take many weeks.
The rebuild journals contain bitmaps that indicate which portions of the disks in a tier have been updated
with new data while a disk was failed or replaced. The system uses the information in the journals to
reduce the rebuild time of drives that have not been swapped out. This can dramatically lower rebuild
time, since only portions of the tier may have been updated while the drive was failed or replaced.
The granularity of the journal will be 4MB of data on a single disk or 32MB of host data. Thus a single
host write will force the system to rebuild a minimum of 4MB of data on the disk. A new host write into
a 4MB section that has already been journaled will not cause a new journal entry. The system will
automatically update journals when disks are failed or replaced regardless of whether journaling is
enabled.
To ensure that the journals are correct, the system carefully monitors the state of the journals and will
automatically invalidate or disable the journals if it detects a condition where the journal cannot be used
or journal information could potentially be lost.
The following summarizes the limitations that apply to journaling:
• Rebuild journaling will automatically be disabled if the failed disk is swapped with a new disk. The
system will track the serial number of the disks when they are failed and will force a rebuild of the
entire disk if the serial number changes.
• Rebuild journaling will not be used when a failed disk is replaced by a spare. The rebuild journal
can be used when rebuilding a replaced disk that has not been swapped.
• The system will invalidate the journal on tiers that have failed or replaced disks on boot up. This is
required because the system does not save the journal information.
• Rebuild journaling will be managed by the controller that owns the tier. If a controller is failed, then
the journals on the tiers owned by that controller will be invalidated.
• The system tracks the original owner of a tier when a drive is failed so changing the ownership of
the tier will disable use of the journal for rebuilds on that tier.
• Rebuild journaling will be disabled when rebuilding disks that are failed due to a change in the
parity mode of the tier.
• Use of the rebuild journal will be temporarily disabled if the system is rebuilding a LUN that is a
backup LUN in a mirror group.
70 007-5510-002
Controller Management
T o display the information about the rebuild journal, use the TIER JOURNAL command (Figure 3–44).
To display the information for a specified tier, use the TIER JOURNAL=t command, where t is the
specified tier. This screen will give detailed information about the status of the journal and a display of
all the journal entries for the tier. This screen will also give more detailed information on the status of
the journal and will indicate why it was disabled or invalidated.
Figure 3–44 Sample Tier Journal Command Screen
The status field indicates the current status of the journal:
• Ready - The journal is waiting for updates.
• Active - A disk is failed and the journal has updates.
• All other statuses indicate why the journal cannot be used.
The Rebuild OK field indicates if a rebuild can use the journal:
• Off - Journaling not enabled. Use JOURNAL=ON to enable.
• Yes - Journaling can be used when rebuilding.
• No - Journaling cannot be used.
The rebuilds will only use the journals if the “Rebuild OK” field indicates “Yes”. In order to use
journaling on rebuilds, the operation must be manually started using DISK REBUILD=tc where 't'
indicates the tier, and “c” indicates the channel or REBUILD=ALL which will start a rebuild on all disks.
The TIER JOURNAL=ON|OFF command enables/disables use of the journals during rebuild operations.
The system will automatically update the journals when disks are failed or replaced regardless of this
setting. This parameter only indicates if the journal can be used during the rebuild. The default is OFF.
3.3.4.4SES Device Monitoring Rate
The SES device monitoring rate can be adjusted to minimize its impact on system performance. SES
M_WAIT displays the current setting in seconds
15000 [1]: ses m_wait
SES timer m_wait = 6 seconds
Figure 3–45 SES Device Monitoring Rate
SES M_WAIT=x sets the SES device monitoring rate for the system in seconds. Valid range is 4 to 90.
The default monitoring rate is 6 seconds.
(Figure 3–45).
007-5510-002 71
NOTE :
Improper use of the SES M_WAIT command can prevent the SES monitors from
detecting an enclosure fault before the enclosure automatically shuts down.
3.3.4.5Host Command Timeout
The Host Command Timeout parameter allows the system to free up resources and make them
available to other users if the request from a particular user cannot be completed. This helps to improve
performance in a SAN environment where there are a lot of users accessing the storage.
HOST TIMEOUT=X (where X is value range 1..512) lets the host command timeout for an I/O request
in seconds. Valid ran ge is 1 to 512 seconds. Defau lt setting is 75 seconds.
3.4Security Administration
The controller’s dual-level, non-host based data security is maintained with scalable features including
restricted management access and authentication against authorized listing. No security software is
required on the host computers. (Refer to Section 3.1.3,"Administrator and User Logins" for information
regarding Telnet and serial port security.)
Each authorized user will have its customized LUN identification scheme which applies to all host ports
(Figure 3–46).
Internal
LUN Map
WWN 1
WWN 1
External
LUN Map
Figure 3–46 Mapping Internal LUNs to External LUNs
01234567
01234
LUN 0
LUN 3
LUN 2
LUN 5
LUN 4
LUN 1
Read-only and read/write privileges can be specified for each LUN and for each user.
The “place holder” LUN feature allows the controller administrator to map a zero capacity LUN to a host
or group of hosts (via zoning or user authentication). The administrator can then create a real LUN and
map it to the host(s) to replace the “place holder” LUN in the future. In most cases, the host will not have
to reboot since it already mapped to the “place holder” LUN.
NOTE :
Support of place holder LUNs is dependent upon the operating system, the driver, Host
Card Adapter (HCA), and host bus adapter.
72 007-5510-002
3.4.1Monitoring User Logins
The AUDIT function continuously monitors logins to the controller and provides alerts in the event of
unauthorized login attempts (Figure 3–47
Host Int 15:04:07 User Logout Client1, port:4 S_ID:000004
Host Int 15:04:47 Authenticated Login Client10, port:3 S_ID:000002
Figure 3–47 User Login Messages
USER AUDIT=ON|OFF enables/disables the user auditing function. When enabled, the system will
display a message when a user logs in or out. Default is OFF.
USER CONNECTIONS displays a list of all the currently connected users and the host port to which the
user is connected (Figure 3–48)
.
Controller Management
).
Figure 3–48 User Connections Screen
3.4.2Zoning (Anonymous Access)
This type of configuration provides the first-level protection. The LUN identification scheme can be
customized for each host port. Any unauthorized user accessing the controller will be considered
“anonymous” and granted the zoning rights for the host port to which they are connected.
The ZONING command will display the current settings for the host ports
indicates which internal LUNs the users will be able to access (with Read-only and Read/Write
privileges) and where the internal LUN will appear to the users. In Figure 3–49, only internal LUN 1
can be accessed and it is read-only. It will appear as LUN 0 to the users.
Figure 3–49 Current Zoning Configuration Screen
ZONING EDIT lets you change the settings for the host ports. You will be asked to select a host port to
change and enter the mapping for each LUN
to all the LUNs.
(Figure 3–50). The default configuration is to deny access
. The LUN Zoning chart
ZONING DEFAULT restores the zoning of a host port back to its default settings.
007-5510-002 73
Figure 3–50 Edit Zoning Configuration Screen
3.4.3User Authentication
The controller creates correspondence between users (W orld Wide Name or GUIDs), storage LUNs, and
permissions. The system can store configurations for up to 512 users in total, and the settings apply to
all host ports.
Each authorized user will only have access to their own and “allowed-to-share” data determined by their
customized LUN identification scheme. Administrators can also restrict users’ access to the host ports
and their Read/Write privileges to the LUNs. Unauthorized users will be given the “host port zoning”
rights as defined in Section 3.4.2,"Zoning (Anonymous Access)".
USER displays the current settings for all authorized users
(Figure 3–51). Each user is identified by their
64-bit W orld W i de Name (or GUID) and is given a unique user ID number. The Ports column indicates
which host ports, on each controller, the user is allowed. The LUN Zoning chart indicates which internal
LUNs the user will have access to (with read-only and read/write privileges), and where the internal LUN
is displayed to the user.
Figure 3–51 User Settings Scre en
To configure/change the settings, use these commands:
• USER ADD adds a new user and defines the user’s access rights
• USER EDIT edits the access rights of an existing user
• USER DELETE deletes an existing user from the system.
See Section 2.3.9, "Setting Security Levels", subsection entitled User Authentication (Recommended
for SAN Environment) for further information on how to add a new user.
74 007-5510-002
3.5Firmware Update Management
SGI periodically releases firmware updates to enhance features of their products. Contact SGI technical
support to obtain the latest firmware files.
3.5.1Displaying Current Firmware V ersion
The VERSION command displays version information of the controller’s hardware and firmware
(Figure 3–52).
Figure 3–52 Version Information Screen
Controller Management
3.5.2Firmware Update Procedure
TFTP enables the administrator to download the new controller firmware from a TFTP server to the
controller. When us ing the TFTP command, a TFTP server must be running and a copy of the firmware
file image must be on the TFTP server. This command will “fail” the current controller and should not
be used during active I/O.
Follow these steps to update the firmware files.
1.Collect and save the output of the following commands before you update the firmware:
VERSION AV CACHE
DISK DISK LIST DUAL
HOSTHOST STATUS LOG
LUN LUN CONFIG NETWORK
ST A TSST A TS DELAY STA TS TI ER DELA Y
TIER TIER CONFIG
2.Copy the new firmware file to your TFTP server.
3.Connect to the controller via Telnet or serial (CLI port).
4.Enter TFTP
5.You will be asked to confirm action (Figure 3–53). Enter y to continue.
007-5510-002 75
Figure 3–53 Downloading Controller Firmware
6.Enter the TFTP server’s IP address: TFTP <IP_address >
7.Enter the firmware path and filename: TFTP <filename>
8.For the couplet controller configuration, connect and log into the other controller. Repeat Steps 4-
7 above to update the firmware.
9.Enter REST ART to restart the unit(s).
NOTE :
10. (For dual mode only): After both controllers are back on-line, use the DUAL command to verify
that both controller units are healthy. If either controller shows failed, login to the healthy
controller and issue the DUAL HEAL command.
RESTART can be done at a later time.
3.6Remote Login Management
TELNET ENABLE allows the administrator to temporarily enable the establishment of a remote T elnet session. Use the TELNET command to display the current setting.
TELNET DISABLE allows the administrator to temporarily disable the establishment of a remote T elnet
session.
NOTE :
TELNET STATS allows the administrator to view vari ous st atistics maintained on remote Telnet
sessions (Figure 3–54). These statistics are kept from the time that the system is powered on.
Telnet capability is reset to ON after a controller restart. To turn off Telnet access
permanently, use the NETWORK command.
76 007-5510-002
Controller Management
Figure 3–54 Telnet Statistics
The administrator is strongly advised to perform any commands affecting the system’s configuration
from the CLI UART only (and not from a Telnet session), and to only perform such commands after
issuing the TELNET DISABLE command, so that remote users cannot log into the system in the middle
of an administrative command.
3.6.1When a Telnet Session is Active
Whenever a remote Telnet session is active, the current RS-232 console switches to a CLI sub-shell
which allows the administrator to enter a very limited sub-set of the CLI commands. The following
message is displayed on the console when a T elnet ses sion is initiated from a remote site (Figure 3–55).
15000 [1]:
New TELNET Session initiated from IP address: 010:123:139:005
[Remote TELNET session ON] Local SUBshell 15000 [1]:
Figure 3–55 Telnet Session Initiated
007-5510-002 77
Within the CLI subshel l, the TELNET command allows the administrator to view information regarding
the currently active Telnet session (Figure 3–56).
Figure 3–56 Telnet Session Information
TELNET KILL=m lets the system administrator terminate the remote Telnet session (Figure 3–57). The
KILL parameter may also be specified with TELNET KILL=m, where m indicates the number of minutes
that will be allowed to elapse before the remote Telnet session is terminated. The valid range is
<0..15> minutes. Default is 1 minute. An administrative login is required before the command is
processed.
[Remote TELNET session ON] Local SUBshell 15000 [1]: telnet kill=1
-- WARNING --
Any CLI command that may be in progress on the remote Telnet site
will need to be completed locally after the remote session has been
terminated.
Enter the administrative (or higher) login name: admin
Enter the appropriate password: ********
-- Please wait for the remote TELNET session to be terminated. --
.........................
Telnet Session termination.
Figure 3–57 Terminating a Telnet Session
The remote user is given a warning that the administrator has killed his session, and indicates to him the
amount of time (if any) that he has remaining
(Figure 3–58). An m value of 0 (zero) is an immediate
KILL. The remote user will be notified, but most likely will be unable to read the entire warning message
before the session ends.
15000 [1]:
-- The System Administrator will terminate this TELNET Session in 1 minute --
-- The System Administrator will terminate this TELNET Session in 30 seconds -Connection closed by foreign host.
Figure 3–58 Telnet Session Being Terminated
NOTE :
If a user is in the middle of running a CLI command at a remote Telnet site when the
administrative KILL is issued, the command will continue on the CLI console.
78 007-5510-002
3.7System Logs
3.7.1Message Log
All controller events are logged and saved in non-volatile memory. The log will automatically roll over
when it is full.
LOG displays the log of previous system messages.
LOG CLEAR clears the log of all previous messages.
LOG CHECKCONDITION displays the Check Condition log.
LOG CHECKCONDITION=MORE will display additional information concerning the check condition.
LOG CHECKCLEAR clears the Check Condition log, enter LOG CHECKCLEAR.
LOG QUIET= ON|OFF . This Administrator command enables a “quiet mode” on the CLI where
Message Log statements will still be logged, but not displayed.
LOG QUIET will display the current state of the Log Quiet mode. There should be the word “Quiet” at
the CLI prompt when the Log Quiet mode has been enabled.
3.7.2System and Drive Enclosure Faults
Controller Management
Use the FAULTS command to display a list of all current disk , system, and drive enclosure faults or
failures (Figure 3–59).
Figure 3–59 Current System Faults
To display the current SDRAM memory faults (ECC- error controller counters), use F AULTS
MEMORY command. To clear the values in the memory faults (ECC) statistics, use F AUL TS
MEMCLEAR command.
To display the current status of the host and disk SFPs, use the FAUL TS SFP command.
NOTE :
007-5510-002 79
A transmitter fault and a loss of signal on a disk channel or host port may indicate that
there is no connection at the corresponding connector.
To display the number of LUN array parity errors detected by the system, use the FAULTS ARRA YPARITY command. The system saves the counts for each tier of all the LUNs. To clear the count
of LUN array parity errors in the system, use the FAUL TS ARRAYPARITYCLEAR command.
FAULTS BUSPARITY displays the number of bus parity and data path errors detected by the system.
FAULTS BUSPARITYCLEAR clears the count of errors.
You may set a parameter (ECCSHUTDOWN) that allows the system to automatically shutdown if it
encounters an unrecoverable error. Use the FAULTS ECCSHUTDOWN=on command to enable
automatic shutdown for unrecoverable ECC errors. This is the default setting. To disable and allow the
system to continue to run in spite of unrecoverable ECC errors, use the F AULTS ECCSHUTDOWN=off.
The EXCEPTIONSHUTDOWN command paramet er allows the system t o automat ically shutdo wn if it
encounters a task exception.
FAULTS
default setting.
The FAULTS EXCEPTIONSHUTDOWN=OFF disables automatic shutdown and allows the system to
continue to run in spite of task exceptions.
EXCEPTIONSHUTDOWN=ON enables automatic shutdown for task exceptions. This is the
3.7.3Displaying System Uptime
UPTIME displays the total time the system has been operational--or “uptime”(also known as Power on
Hours), as well as the total time since the last system restart. The uptime is displayed as
YY:DDD:HH:MM where YY is the number of years, DDD is the number of days, HH is the number of hours, and MM is the number of minutes (Figure 3–60) the system has been continually in operation.
Figure 3–60 Display System Uptime
3.7.4Saving a Comment to the Log
COMMENT <text of message> allows you to echo a message to the screen. The message is saved in the
LOG and is also sent to syslog if it is enabled. Any printable text can be entered.
80 007-5510-002
3.8Other Utilities
3.8.1APC UPS SNMP T rap Monitor
APC_UPS displays the status of the APC UPS SNMP trap monitor (Figure 3–61).
15000 [1]: apc_ups
APC UPS SNMP trap monitor is off.
No APC UPS faults detected via SNMP trap.
Figure 3–61 APC UPS SNMP Trap Monitor Status
APC_UPS CLEAR_FAULTS will delete all pending APC UPS faults from the fault list. All APC UPS
events that disabled writeback caching will be cleared.
3.8.2API Server Connections
The API command displays the current status of the API connections (Figure 3–62).
Controller Management
15000 [1]: api
API Server connections are currently -- ENABLED--
Figure 3–62 Displaying Status of API Connections
API DISABLE temporarily enables/disables the establishment of connections to the API server. When
disabled, users at remote locations will be unable to establish a new API connection until an
API ENABLE command is issued. This command only provides control over API connections during the
current power cycle.
API ST A TS displays the collected statistics on API connections
(Figure 3–63).
API CLEARST A TS resets the collected statistics.
Figure 3–63 API Server Connection Statistics
3.8.3Changing Baud Rate for the CLI Interface
The CONSOLE command displays the current serial console setting (Figure 3–64) of the controller.
007-5510-002 81
Figure 3–64 Displaying the Serial Console Setting
CONSOLE BAUD changes the baud rate of the CONFIG port of the controller (Figure 3–65).
15000 [1]: console baud
Select the new serial console baud rate from choices below:
1 - 9600
2 - 19200
3 - 38400
4 - 57600
5 - 115200 <- Current setting
e - escape out of this command
Enter selection:
Figure 3–65 Changi ng the Baud Rate
3.8.4CLI/Telnet Session Control Settings
Yo u may change the CLI’s and Telnet’s various session control settings. The SETTINGS command
displays the current setting (Figure 3–66).
Figure 3–66 Current Session Control Setti ngs
SETTINGS DEFAULTS resets all the CLI and Telnet session control settings to their default values.
SETTINGS LINES=<number of lines> sets the number of lines displayed at a time in a page of screen
information. Pages provide a way to control the amount of information displayed to the user at one time.
You will be prompted to either press a specified key in order to scroll from one page to the next, or (in
certain circumstances), to terminate the display. Valid range is 0 to 512 lines, where 0 indicates that no
paging is to be performed on the output information. Default setting is 0.
SETTINGS PROMPTINFO=ON enables extra status information in the CLI prompt. OFF disables extra
status information in the CLI prompt. When enabled, the CLI prompt indicates whether the system is
booting (BOOTING), failed (FAILED) or waiting for a restart (RESTART NEEDED). ON is the default.
3.8.5Disk Diagnostics
The DISK DIAG=tc command performs a series of diagnostics tests on the specified disk. The disk is
specified by its physical: tier (t) in the range <1..125>, and channel (c) in the range <ABCDEFGHPS>.
82 007-5510-002
Controller Management
3.8.6Disk Reassignment and Miscellaneous Disk Commands
The DISK REASSIGN=tc 0xh command allows for the reassigning of defective logical blocks on a disk
to an area of the disk reserved for this purpose. The disk is specified by its: tier (t) in the range <1..125>,
and channel (c) in the range <ABCDEFGHPS> 0xh is the hexadecimal value of the LBA (Logical Block
Address) to be reassigned.
The DISK LLFORMA T=tc command allows the user to perform a low level format of a disk drive. The
disk is specified by its: tier (t) in the range <1..125>, and channel (c) in the range <ABCDEFGHPS>.
The DISK AUTOREASSIGN=ON command is the default setting. When enabled bad blocks are
reassigned when a medium error occurs on a healthy tier, the DISK AUTOREASSIGN=OFF command
disables this feature and bad blocks are NOT reassigned when a medium error occurs on a healthy tier.
The DISK MAXCMDS=x command sets the maximum command queue depth to a tier of disks in the
range of 1 to 32 commands per tier. The default is 32 commands.
3.8.7SPARE Commands
Use the SPARE commands to display information about the spare disks in the system or to change the
configuration settings for background diagnostics in the system. The information displayed pertains to
the current spare configuration settings as well as task status.
The SP ARE CLI commands are for background diagnostics. The intent of these commands is to test
otherwise idle spare disks at least one (1) time per month to validate that they are continuing to function
properly, and are truly available to be swapped in as a replacement disk. It is testing of the “hot” spares.
They are intended to run in the background and SPARE operations are a lways at lower priority than an y
other kind of I/O in the system.
The SP ARE INFO=tc command displays the information and status about a specific spare disk in the
system. The disk is specified by its physical tier and channel locations, “tc.” The “t” indicates the tier
in the range <1..125>, and the 'c' indicates the channel in the range <ABCDEFGHPS>.
The SP ARE CLEAN=tc command erases any previous test data stored on the disk indicated. The disk
is specified by its physical tier and channel locations, “tc.” The “t” indicates the tier in the range
<1..125>, and the 'c' indicates the channel in the range <ABCDEFGHPS>.
The SP ARE COVERAGE=x command sets the spare diagnostic coverage of the blocks being tested as
a percent of the total number of blocks available for test. Note that increasing the coverage to higher
numbers means that more blocks on the disk will be tested for better coverage, but it also will take a
longer time for the test to complete. This parameter can be tuned to provide an optimal test time for a
single disk in the system such that all spares are tested in a reasonable amount of time. The parameter
is limited to a discrete set of values. The valid parameters for “x” are [1, 5, 10, 20, 40, 80, 100] Percent.
Default is 1 Percent.
The SP ARE EXTENT=x command sets the spare diagnostic extent in Mbytes. The diagnostic extent
determines how much data can be tested before the test must sleep. This parameter sl ows down the test
operations so they will not affect the performance of the system. Increasing the extent will allow more
data to be tested in a single pass. Any changes applied to extent will affect tests in progress as well as
future testing. The valid range for 'x' is 1..32 Mbytes. Default is 8 Mbytes.
The SP ARE DELAY=x command sets the system spare diagnostics delay. The test delay determines
how long a test operation will pause after it reaches the test extent. This parameter slows down the spare
007-5510-002 83
test so it will not affect the performance of the system. Any changes applied to delay will affect tests
in progress as well as future testing. This system spare diagnostic delay value is given in 100 millisecond
increments. The valid range for 'x' is 0..100. The default is 0.
The SP ARE P ATTERN=x command sets the syst em spare diagnostics pattern. The test pattern
determines the pattern written to the disks during the test. The system supports the following patterns:
• UNIQUE Includes unique information including timestamp
• AA 0xAA is written to each byte
• 55 0x55 is written to each byte
• FF 0xFF is written to each byte
• 00 0x00 is written to each byte
• COUNTUPA pattern of counting up is written to each byte
• COUNTDOWN A pattern of counting down is written to each byte
The default is UNIQUE. Note that the tests in progress are not affected by this parameter setting.
Changing the pattern only applies to tests started after the parameter was modified.
The SPARE ST ART command starts the spare diagnostics task if it is not running. Note that this will start
diagnostics on both units in a dual system, as this is a system parameter.
The SP ARE STOP command aborts any ongoing diagnostic operations. Note that this command will
stop them and then the task will be idle until the SPARE RESTART command is executed. Note that
this will stop diagnostics on both units in a dual system as this is a system parameter.
The SP ARE P AUSE command pauses but does not stop any ongoing diagnostic operation, only on the
unit from which the command is run. If a test is being run from the other unit in a dual, the pause
command will NOT affect that test.
The SP ARE RESUME command releases any paused diagnostic operations and allows them to continue
only on the unit from which the command is run. If a test has been paused on the other unit in a couplet,
the SP ARE RESUME command will NOT affect that test.
84 007-5510-002
Chapter 4
Telnet port
Controller Remote Management and
T roubleshooting
4.1Remote Management of the Controller
The controller can be managed locally through the RS-232 interface, or remotely via Telnet. The
Administrative Utility is the same regardless of the management interface (RS-232 or Telnet).
The controller supports SNMP and allows the system to be remotely monitored.
4.1.1Network Connection
Connect the Telnet port on the back of the controller to your Ethernet network (Figure 4–1). Then set the
IP addresses, login names, and passwords as described below.
Troubleshooting
A
AB
B
AC
FAIL
C
CD
D
E
EF
F
DISK CHANNELS
G
GH
H
P
PS
S
SYSTEM
STATUS
CTRL
STATUS
Figure 4–1Telnet Port on the Controller
NOTE :
Currently, the controller does not support network configuration protocols such as DHCP or
BOOTP.
4.1.2Network Interface Set Up
For first time set up, you will need to connect to the CLI (RS-232) port in order to change the IP address
and/or network settings.
To set up the network interface:
1.Use the NETWORK command to display the current settings
2.Change the controller’s IP address for your network environment: network IP=<new IP address>.
TEMP
STATUS
DISK
STATUSDCSTATUS
TEST
FAN
STATUS
HOST 1
HOST 2
HOST 1/2
1
2
HOST 3
3
HOST 4
4
PLACE PIN HERE
HOST 3/4
STATUS
ACT
CLI
COM
CLI
STATUS
ACT
1/2
TEST
TELNET
ACT
LINK
ACT
LINK
LINK
CLI
LINK
AC
ACT
FAIL
MUTE
ALARM
SILENCE
(Figure 4–2).
3.Change the netmask of the controller (if needed): network netmask=<new netmask>.
4.Enable the Telnet capability (if needed): network telnet=ON.
007-5510-002 85
NOTE :
T elnet connections are clear text. If Telnet connections are used, you may expose controller
passwords to third parties. For higher security, we recommend that you disengage Telnet access if it is
not required.
5.Decide whether the SNMP functionality should be enabled.
To enable or disable SNMP, use the appropriate version of the command network SNMP=on|off.
Figure 4–2Current Network Configuration Screen
NOTE :
If you are using an external system console option like Telnet, the SNMP function should
be enabled.
6.If the SNMP function is enabled, enter the IP address of the computer to be used to monitor the
SNMP traps:
network trapip=<computer’s IP address>
7.Decide whether the Syslog capability should be enabled. To enable (ON) or disable (OFF) the
Syslog, enter:
network syslog=on|off
NOTE :
The SGI default is the Syslog capability enabled.
8.If the SYSLOG function is enabled, enter the destination IP address for the Syslog packets:
network SYSLOGIP=<destination IP address>
Ensure your destination computer supports the SYSLOG feature. For example, on UNIX systems,
the SYSLOG application must be properly installed and running.
9.The default destination port number for Syslog packets is 514. To change it, enter:
network SYSLOGPORT=<port number>
10. Set up the routing table. This table describes how the controller communicates with the hosts on
other networks. Use the ROUTE command to display the current settings
(Figure 4–3).
86 007-5510-002
Loading...
+ hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.