HP StorageWorks 9100 Extreme Data Storage System administration guide V1.0.2 (AN540-96018, February 2010)

repair it using psfsck. However, if the underlying logical disk drive is failed, then this repair is
impossible because the data is lost.
Normally, the utility filesystem is based on an "adm" LUN on the first storage block. During normal
operation, the HP ExDS9100 system software makes a backup copy of the utility filesystem and copies
it to a filesystem on one of the "admspare" LUNs. To recover after the original utility filesystem is lost,
mount the “admspare” LUN instead of the original corrupt or lost “adm” LUN as follows:
1. Shut down all servers except the first server. You cannot use exdsmgr to do this; however, (pdsh
is available).
2. Identify the psv devices in question as shown in the following example:
In this example, psv1 is the original utility filesystem and psv3 is the backup.
3. Make sure the Matrix software does not try to mount the original utility filesystem as shown in
the following example (where psv1 is the original):
# mx fs unmount --persistent psv1 ALL_SERVERS
4. Mount the "admspare" LUN on the /var/mxso mountpoint (that is, replacing the original
with the backup; where psv4 is the backup):
5. Reboot the first server. The server should boot and the utility filesystem should now be located
on a different storage block.
6. Reboot the other servers.
POST error messages
For an explanation of server error messages, see the "POST error messages and beep codes" section
in the HP ProLiant Servers Troubleshooting Guide at http://www.hp.com/support/manuals.
Identifying failed I/O modules on an X9700cx chassis
When an X9700cx I/O module (or the SAS cable connected to it) fails, the X9700c controller attached
to the I/O module reboots and if the I/O module does not immediately recover, the X9700c controller
stays halted. Because there are two X9700cx I/O modules, it is not immediately obvious which I/O
module has failed. In addition, the X9700c controller may halt or appear to fail for other reasons.
This document describes how to identify whether the failure condition is on the X9700cx I/O module
or elsewhere.
Failure indications
A failed or halted X9700c controller is indicated in a number of ways as follows:
• An email alert or message in /var/log/messages such as the following:
2009-06-12 12:37:33 glory1 218 crit storage[10162]: array SGA832000D: array: 2
controller P89A40A9SUQ077: no longer seen from glory1 (controller may be powered off, halted,
failed or disconnected)
• The exds_stdiag report could indicate a failed or halted X9700c controller.
9100 Extreme Data Storage System administration guide 133