faulty disk

[lawrenzo_1]

I have been alerted to a faulty disk through monitoring software,

ioscan shows no hardware and I have located the disk to be in vg01.

what else can I do to gather more information and check on the faulty disk?

Thanks in advance.

[Sridhar Bhaskarla · Accepted solution]

Hi Lawrenzo,

Check for "POWERFAILED" messages in syslog.log. 'vgdisplay -v' should show you the status of the failed disks as "unavailable".

If this is not a mirror disk, then you may have to restore the data from the backups. If the disk is not hot-pluggable, then you will have to shutdown the system, replace the disk and bring the system. vg01 will not be activated due to failed disk. Once the system is backup, Do

#vgcfgrestore -n /dev/vg01 /dev/rdsk/cxtydz
#vgchange -a y vg01

Data will be lost on all the filesystems that share (lvdisplay output) this disk. You may have to do a 'newfs' on those logical volumes, mount them and then restore the data.

Sometimes, reseating the disks may help. You can try it before you attempt the above procedure.

-Sri

-Sri

[lawrenzo_1]

Super - thanks a millions Sri, I have a few idea's where to look and this helps me much more.

Lawrenzo.

[Steven E. Protter]

Another common syslog message involves the term lbolt.

That error will occur when a powerfail has not happened. It will be pretty obvious in syslog.log that a disk has failed.

SEP

[Isaac Loven_4]

diskinfo /dev/dsk/cxtxdx

[Indira Aramandla]

Hi Lawrenzo,

We had a similar alert about disk on our system.

When I checked with ioscan it did not report any hardware failure and showed as claimed.

As other suggested I checked in the syslog for any error messagaes (like power failure...). But syslog did not report any message about the disk.

The first test is to see if there is any amber/red light on the disk array or the disk.

If there is not such light then the disk as a whole is fine but it could be that the disk has few stale blocks, which is of no good.

The next test is to check with dd.
dd if=/dev/dsk/c6t9d0(faulty disk) of=/dev/null bs=1024k
and watch the lcd lights on the disk and see which ones are out of sync.

You can also try
pvdisplay -v /dev/dsk/c6t9d0 | more
Then use cstm tool to check the error log and you might find someting like this
Error Logs
Total Retries: 1333 Buffer Overruns: N/A
Read Reverse Errors: N/A Buffer Underruns: N/A
Write Errors: 2 Non-Medium Errors: 5
Verify Errors: 0

In this case you must log a call with HP to replace the faulty disk. In which case as Sri mentioned you will have to do the following after replacing the disk.

vgcfgrestore -n /dev/vg01 /dev/rdsk/c6t9d0
vgchange -a y /dev/vg01
vgsync /dev/vg01

Indira A

[lawrenzo_1]

thanks