American Megatrends StorTrends 1100i How to use

StorTrends ITX Software Version 2.6
Copyright (c) 2002-2006 American Megatrends, Inc.
Please visit www.ami.com for the latest notes
regarding the release version you are using.
===============================================================================
Hardware Requirements:
American Megatrends StorTrends ITX 1100i/3100i
*******************************************************************************
Installation Steps:
StorTrends ITX 3100i (Broadcom Controller)
1. Backup your old data, perform a system reboot and enter the Broadcom BIOS
Utility (CTRL+R)
2. Delete all previous RAID arrays and initialize all physical disks
3. Check the version number of the BIOS.
If the version is not RC-200-XXX, then you need to update it to
the supported version (2.0).
Please get the supported version from
http://www.broadcom.com/products/raid_2.0.1_driver_download.php
Download the file firmware_v2.0.1.zip.
Reboot system and go back inside Broadcom BIOS Utility
4. Select Controller Options menu item and select Toggle Board ID to off (default
is on)
5. Automatic reboot of system is followed.
6. Again, go inside Broadcom BIOS Utility and create a RAID 50 of size 3.4GB
using *all* available disks in the system. For maximum safety and performance
choose Distributed Spare and Read-only cache for this array
7. Connect an external USB CD-ROM device to the box and boot the system with
the StorTrends ITX 2.6 CD, and it will prompt you through installation.
NOTES: Make sure that the boot sequence is set to the USB CD-ROM device first.
Not all USB CD-ROM devices may be supported by the BIOS.
For an updated list of supported CD-ROM devices, contact AMI.
StorTrends ITX 1100i (Marvell SATA Controller)
1. Boot the system with the StorTrends ITX 2.6 CD using the internal CD
drive, and it will prompt you through installation.
*******************************************************************************

Release Notes for StorTrends ITX 2.6 - Build 2.6.0006
======================================================================
Build Date: MMM DD, 2007
Build Version No: 2.6.0006
==========================
Features Supported
------------------
1. Advanced Storage stack supporting the following features
a. Thin/Exact Provisioning
b. Provision tracking
c. High speed Read-Only and Writable snapshots
d. Instantaneous rollback of snapshots
e. Advanced caching
f. Supports up to 8 TB Volume size (64 bit LBA)
g. Supports up to 32 Volumes per box
h. Synchronous Replication
j. Volume Expansion
k. Snapshot Assisted Replication
l. High Availability Failover
2. Synchronous replication of SAN volumes.
3. SAN volume cloning within a single box.
4. High-Availability grouping across SAN volumes.
NOTE: There is a deviation from CLI guide for creation of Sync Repl Pair
and HA grouping. This deviation will be only in this build and once Mono support
is added to HA, it will be reverted. Accordingly the CLI command for creation of
has an additional parameter in the build:
Usage:
create -l/--local <local-primary contName_volName>
-r/--remote <remote-secondary contName_volName>
-p/--priority <resync priority percentage in number of outstanding IOs>
-i/--localhost <primary host IP address>
-m/--remotehost <secondary host IP>
-t/--mgmttgt <remote management target name>
-d/--mode [sync | ha] syncreplica
Specifying “sync” will create Sync. Replication. Specifying “ha” will create an
HA pair.
5. Snap assisted replication of Consistency groups, consisting of NAS or SAN
volumes.
6. Extensive Command Line interface (CLI) for advanced configuration &
scripting.
7. Automated Mirrored OS installation for Software RAID installation
(StorTrends ITX 1100i)
8. Hot Swap support
9. Active Directory support.
10. Journaling file system, with fast recovery support.
11. CIFS/NFS/APPLE/HTTP/FTP based access to storage
12. ACL based security control for the users and groups
13. Quota support
14. Scheduled Snapshot support for NAS and SAN
15. SNMP, Email and Event Log based alert system
16. Full ISCSI Target functionality.
17. Supports RFC 3720 of ISCSI IETF specification
18. Supports ISNS protocol and can register with up to 16 servers

19. Supports a maximum of 128 targets per box (including 2 system-generated
targets: mtarget and smatarget)
20. Supports 32 LUNs per Target.
21. Supports ISCSI Error recovery level 0 and 1
22. Supports fine tuning of ISCSI parameters from UI
23. ISCSI traffic can be enabled or disabled on a per NIC basis
24. Supports CHAP based iSCSI authentication
25. ISCSI Authentication based on iSCSI Initiator WWN/FQDN
26. ISCSI Authentication based on local StorTrends iTX username and password
27. Supports mutual CHAP a.k.a Bi-directional authentication
28. Support for StorTrends iTX 3100i 3U SATA box (Broadcom Controller)
29. The Broadcom Hardware RAID controller supports RAID Levels 0, 1,
1N, 10, 10N, 5, 50 and online RAID transformation & expansion,
Distributed, Dedicated and Global spares management and Cache settings.
30. Hardware RAID events from Broadcom (disk removal, RAID degradation etc)
31. Supports UPS Master/Slave mode,
32. Network teaming supporting Round Robin, Balance TLB and 802.3ad modes.
33. Automatic Write back to Write through cache settings on UPS event
generation
34. Auto shutdown of box when CPU/System temperature reaches high critical
threshold.
35. NAS & SAN read only and read write snapshots can be activated and mounted
36. Broadcom Hardware Auto RAID Rebuild support
37. Web based Configuration
38. iSCSI Tape Support. Any Tape device connected through SCSI can be attached to
an iSCSI target.
39. Support for Multiple UPS models. The supported models are
a. APC Smart-UPS 1000
b. EATON Powerware 5125
Known Limitations/Issues in 2.6.0006:
-------------------------------------
Hardware
1. Pulling a USB CD-ROM may cause a kernel oops.
2. When creating a Network Team, the MAC address of the bond is always that of the NIC-
1, irrespective of the NIC specified as Primary for the team.
3. System auto shutdown based on FAN sensors is not supported.
4. 3U: Always install ITX stack with Broadcom read-only cache.
5. While creating a Network Team, network connection will go down for few seconds
6. XFS soft panic at xfs_alloc_read_agf(). It is a known issue in XFS. It is due to XFS
metadata corruption. When this happen, the IO will be still going on. But next time
when ITX is restarted, XFS repair will take long time (depends on file system size)
and some of the file will be lost.
7. Read performance is reduced in Balance RR mode. This is a known issue with
Balance_RR teaming mode due to TCP congestion control problems in network. A managed
switch with proper configuration might fix this issue.
8. Supermicro 3U systems: the Mute button and the BIOS option for 4-second override are
not supported.
9. Broadcom 3U: even if 2 out of the 3 power supplies fail, you will only get a
single event entry regarding the failures.
11. Daylight Savings time: the new DST guidelines effective in 2007, whereby the time
changes occur in March and November, instead of the traditional April and October
dates, are not supported in this release.
12. The iTX software supports both DHCP and Static IP addresses, but it is strongly

recommended that you use static addresses for all servers to minimize the risk of
connectivity problems.
13. May lose a network interface if NICs are connected at different speeds and one
NIC gets updated as static/dhcp. The lost NIC can be reset to its previous
setting.
Initiator
1. The TimeOutValue in Windows Registry under Disk subsystem needs to be increased to
150 seconds or more, to avoid Delayed Write failures when using iSCSI.
2. Support for greater than 2TB iSCSI Disk is possible only with Windows 2003 SP1 and
other OS that support 64 Bit SCSI LBA.
3. Windows might prompt for installing drivers when logged on to targets, which have a
special AMI SCSI processor device. There is no need to install drivers. This device
is for internal communication only. Simply click on Cancel.
4. Windows 2000 NTFS file system on top of an ISCSI Thin provisioned volume might run
out of space in container even though the file system used space is less than the
container size. Alternative is to use Exact provisioned volumes.
5. Sometimes if an iSNS server’s status is displayed as inconsistent, remove and add
the iSNS server back to get the correct status message.
6. ISNS: when you register your iTX device with a new iSNS server, the list of nodes
registered with that server will not become automatically available. You have to use
the “update –t noderegistrations iscsi” CLI command to update that list.
7. The Adaptec 7211C iSCSI HBA initiator is not supported.
High Avalabilty Failover
1. A HA (high-availability) pair can only be accessed using an initiator with the iTX
DSM. Microsoft’s Default MPIO will not work with iTX HA and will result in data
corruption. You will need to uninstall it prior to installing the DSM.
2. The DSM-MPIO requires the latest HAK to be installed as well.
3. HA is supported on the 3100i and handles power failures (bug 7084). The 1100i does
not support HA failover for power loss or local Ethernet switch failure.
4. After failover, the primary state may show up as ‘Remote Unreachable’, although, the
pair has become in-sync and all other aspects of the system are fine.
ManageTrends
1. If ManageTrends displays a HTTP 500 or 404: IO File not found exception, restart the
itxweb service from the ITX shell prompt, by using the following command "service
itxweb restart".
2. If system date time is changed, then page may return timeout error. User needs to
refresh the browser manually.
3. While changing the user password in Mozilla browser, Ignore browser password manager
dialog
Protocols
1. Sometimes after a non-graceful shutdown, Apple Shares (using AFP) may not work
properly. In this case, it is necessary to delete folders .AppleDB and .AppleDouble
from the iTX box manually. These folders are hidden and can be located
in/volumes/container_volumename/apple folder.
2. Only CIFS protocol supports ADS user
3. If you enable https for added security, then the unsecured http access will be
disabled. However, since https is dependent on http, the 2 protocols will be shown
as Enabled.

Raid
1. Full or near-full disk capacity is used for Dedicated & Global Spares.
Therefore, Dedicated & Global spares cannot be used as part of other RAID
Logical drives.
2. Hardware RAID Cache setting cannot be switched from Write-Back
to Write-through if the array is in initializing, rebuilding,
expanding or transforming mode.
3. The status of SATA drives is updated only when re-scan operation is initiated.
4. Hardware RAID 0 needs a minimum of 2 disks for a raid 0.
5. Removing two or more disks may cause the system to freeze in StorTrends iTX 3100I as
the OS Array may go offline.
6. When specifying size of RAID LD (only in hardware), the container size cannot be too
small (less than 50GB). This situation is more significant if NAS Thin Provisioned
volumes are created much bigger than 50GB. An example is, having a container of size
50GB and trying to create a NAS volume (Thin Provision) of size 500GB, causes XFS
file system to create kernel panics. This is because when the XFS format happens on
the volume, due to the nature of thin provisioning, disk runs out of space before
format is complete. XFS cannot handle out of space conditions properly. The
general formula is: Minimum size of LD (container) = number of NAS volumes(Thin
Provision) * 50GB.
7. With the Broadcom Controller, if a disk is marked as global spare or dedicated
spare, it cannot be used as part of another array (Logical Drive)
8. An Expanded RAID Logical Drive may not immediately show the new expanded size. A
reboot is required to show the new size. Also, the new expanded LD size will not be
used by the container. The solution is to create another RAID array with the free
space and add it to the container.
9. On StorTrends 1100i platform, OS RAID Auto rebuild is supported only in Hot Swap
mode.
10. A kernel panic may occur while creating and deleting LD recursively using bcadm
command.
11. False reporting of failed array and drive occasionally occurs, and is caused by
Broadcom driver.
12. Drive listing issue in BCRAID can occasionally be seen if there are two drive
failures.
13. If an iTX device is upgraded from iTx 2.0 to the 2.6 software, a raid 0 LD that
was created prior to the upgrade cannot be deleted if it should fail.
Recommend: create redundant arrays only or else transform any raid 0 arrays prior
to the upgrade.
14. A reboot is required after changing a physical disk’s caching mode.
15. On the 1100i, hard drives should be swapped to different slots prior to fresh
install of iTX to insure that old data is completely over written. This would
not be necessary if the user enters the RAID BIOS setup and initializes all
drives prior to install (this takes a very long time).
Replication
1. Local Replication does not support Failover and Failback. Failover and Failback is
only meant for disaster recovery and therefore, it is applicable only for Remote
replication pairs.
2. SAN SAR (snap-assisted replication) can only be accessed via 1 initiator with the
AMI HAK installed.
3. While creating containers, volumes, CG and etc., it is better to use lower case
names if the volume has to participate in a NAS SAR (Bug 7129).
4. Replication link status status display is not always correct (Bug 7258).

Miscellaneous
1. The total number of NAS shares is limited to 10000.
2. UID range for local user is from 1001 to 100000.
3. UID range for local group is from 1001 to 100000.
4. NDMP backup is not supported in this release.
5. NT Backup is not supported for iSCSI tape.
6. Seen only one time: Unidentified network conditions caused a hard kernel panic
requiring the iTX device to be rebooted manually. This has been identified as coming
from the IP stack of the Linux kernel itself, and is not otherwise caused by the iTX
hardware or software. (Bug 7082)
7. For NAS SAR pair creation, on remote side ( Passive Box ) volumes should NOT be
present. Based on the remote side volume name input, SAR pair creation will create
the remote side volumes and NAS SAR pair also. If volumes present, return “Volume
already exists” error code.
Snapshot
1. To disable the snapshot schedule for a volume, set its time interval to 0.
2. Apple Share access for a NAS Read Only Snapshot is not supported (using AFP) and
will not work. Alternative is to use CIFS access from MAC Clients.
3. If a writeable snapshot is activated as Read-only, then it will not accept any write
IO. In order to perform write on writeable snapshot, it has to be activated as Read-
Write, i.e. with the –w switch.
4. As a best practice measure, it is recommended that 20% of container capacity be
reserved (left free) for snapshots.
5. Folder permissions (ACL) cannot be set for NAS Read only snapshots.
6. With replication pairs, the secondary/remote volume still requires manual
configuration for SAN scheduled snapshots. If not configured, the snapshots will get
taken on the primary volume’s schedule, but the maximum set for the number of
snapshots for the volume will be the default of 254.
Volume
1. Containers can be created only with a 64KB chunk size. The RAID Stripe size should
also be of 64KB minimum.
2. If the container runs out of space and any volumes which have not been pre-allocated
space (Thin-Provisioned volumes) require additional space, you might get an out of
space error when new writes happen to the volume. Note that the writes will fail
ONLY if new territory needs to be allocated. If space (territory) has already been
allocated, the writes will not fail.
Solution: Configure alerts (email or snmp) and when thus notified, add new storage
capacity to the container by expanding it before it runs out of space.
3. Event entries are not supported for SAN volumes since iTX does not host the file
system. However, capacity events for containers associated with the SAN volumes do
get logged.
iTX Open Bugs:
5524 – Sev3 - Shares of volumes/containers that have been deleted do not get
automatically deleted sometimes.
5865 – Sev3 - LTP test suite returns some failures.
6521 – Sev3 - cli commands performance degradation with iometer running.
6524 – Sev3 - If you repeatedly add/remove non-existent isns servers, the isns service
gets stopped.
6600 – Sev3 - Need warning message when user wants to rollback to snapshot of a volume
that is in a rep pair.
6601 – Sev3 - SAR: if you expand the Active volume, the Passive volume continues to

show the pre-expanded size.
6750 – Sev3 - QLogic 4010c on Suse Linux: get a connection error when trying to access
iSCSI space.
6753 – Sev3 – Incorrect output when adding ADS user to LOCAL group.
6754 – Sev3 - iSWAT: iSCSI Conformance test keeps failing.
7089 – Sev3 - Unable to delete a failed raid 0 LD using the CLI command.
7091 – Sev3 - Replication status does not return back to “Active/Data is being
replicated” after rollbackjoin.
7097 – Sev3 - Creating snapshots by VSS fails because of presence of only one type
(RT/WT) of oldest snapshot.
7122 – Sev3 - iSCSI boot option not working with the QLogic 4050C HBA.
7146 – Sev3 – HA failover occurs after NIC-2 failure
7171 – Sev3 – CLI: should show a more appropriate message if a user attempts to
rollback a volume that is in a SAR pair to a snapshot.
7208 – Sev3 – iTX displays schedule information for deleted volume.
7259 – Sev3 – SAR status shows “Link Failed” when link has not failed.
7260 – Sev3 - A snapshot created in primary before failover is mounted after issuing
failover and failbackjoin in secondary.
7267 – Sev3 - In NAS one to many replication Rollback for one SAR gives Success and for
other Resource is Busy.
7316 – Sev3 – Delete with –a option fails to delete primary volume in a sync rep pair.
6484 – Sev4 - (Recommendation): Need CLI command to get volume and snapshot counts.
6582 – Sev4 - (Recommendation): Requested change to ‘cli list -t pd raid’ display.
6765 – Sev4 – Plextor USB DVD ROM drives not compatible with iTX hardware.
6772 – Sev4 – Unable to manually create writeable snapshots.
7246 – Sev4 - The schedule set for volumes in a CG do not take into account max
snapshots supported, snapshots are not taken according to the schedule.
ManageTrends Open Bugs:
7038 – Sev2 – Not able to create SAR pair with Zlib compression.
7198 – Sev2 – Creating 32 SAR pairs requires 32 logins.
7325 – Sev2 – No support for iSCSI tape configuration in GUI.
6871 – Sev3 – ManageTrends displays wrong compression values after updating.
7031 – Sev3 – No keyboard support for most of the pages in ManageTrends 2.6.
7288 – Sev3 - Displaying snap information as "undefined" in CDT page.
7289 – Sev3 - Irrelevant message is getting displayed in CDT page.
7296 – Sev3 - There is no snapshot information available for selected CG pair in CDT.
7299 – Sev3 - Sometimes snapshots are not replicating after updating replication
schedule.
7306 – Sev3 - Not able to delete CG pair after updating.
7311 – Sev3 – Sometimes failback does not work properly using the GUI.
6916 – Sev4 – Buttons should be disabled to change disk cache while LD is being
initialized.
7285 – Sev4 - After completing the HA pair creation MT is not displaying any meassage
or status.