HP StorageWorks X9720, Руководство администратора

Страница: 109 / 187

LUN status is failed
A LUN status of failed indicates that the logical drive has failed. This is usually the result of failure
of three or more disk drives. This can also happen if you remove the wrong disk drive when
replacing a failed disk drive.
If this situation occurs, take the following steps:
1. Carefully record any recent disk removal or reinsertion actions. Make sure you track the array,
box, and bay numbers and know which disk drive was removed or inserted.
2. Immediately run the following command:
# exds_escalate
This gathers log information that is useful in diagnosing whether the data can be recovered.
Generally, if the failure is due to real disk failures, the data cannot be recovered. However,
if the failure is due to an inadvertent removal of a working disk drive, it may be possible to
restore the LUN to operation.
3. Contact HP Support as soon as possible.
Apparent failure of HP P700m
Sometimes when a server is booted, the HP P700m cannot access the SAS fabric. This is more
common when a new blade has just been inserted into the blade chassis, but can occur on other
occasions. Symptoms include:
•
The HP P700m reports a POST error—this is visible using the TFT monitor/keyboard.
•
The server crashes when the cciss driver loads— this is visible using the TFT monitor/keyboard.
Sometimes this happens to all servers in the system.
•
No controllers are seen when you run the
exds_stdiag
command.
The underlying causes of these problems differ. However, the recovery process is similar in all
cases. Do not replace the HP P700m until you have worked through the process described here.
In general terms, the solution is to reset the SAS switches and if that fails, reboot each X9700c
controller until you locate a controller that is interfering with the SAS fabric.
If your system is in production, follow the steps below to minimize downtime on the system:
1. Log in to the Onboard Administrator and run the
show bay info all
command. Compare
entries for the affected blade and working blades.
If the entries look different, reboot each Onboard Administrator, one at a time.
Re-seat or replace the P700m in the affected server blade.
2. Run
exds_stdiag
. If
exds_stdiag
detects the same capacity blocks and X9720c controllers
as the other server blades, then the procedure is completed; otherwise, continue to the next
step.
3. If all servers are affected, shut down all servers; if a subset of servers is affected, shut down
the subset.
4. Using OA, log into the SAS switch 1 and reset it.
5. Wait for it to reboot.
6. Reset SAS switch 2.
7. Wait for it to reboot.
8. Boot one affected server.
9. Run the following command:
# exds_stdiag
10. If X9700c controllers can be seen, boot other affected servers and run
exds_stdiag
on
each. If they also see the X9700c controllers, the procedure is completed; otherwise continue
to the next step.
Troubleshooting specific issues 109