Mirror faultted and performance issue

[Rosana Ramos]

Hi!
I have an hp server running HPUX 11.0, which has OS file system in mirror. A few days ago, one disk failed and the server has a big performance issue. We must split the mirror in order to get the server working. I supposed that if a disk failed the server would work fine.. doesn't it? Has anyone any idea why this happened?
Thanks!
Ross

[Torsten. · Accepted solution]

Hi,

it depends "how" the disk dies...

If the disk is just dead, the server works fine, but if the disk is "almost dead" and produces some SCSI bus resets every second, this can slow down your performance of course.

[James George_1]

Hi

You should call your h/w vendor and get the disk replaced.

Rgds/ James

[Rosana Ramos]

The disk was almost dead, we received some messages like these:

vmunix: DIAGNOSTIC SYSTEM WARNING:
vmunix: If the diaglogd daemon is not active, use the Daemon Startup command
vmunix: in stm to start it.
vmunix: If the diaglogd daemon is active, use the logtool utility in stm
vmunix: to determine which I/O subsystem is logging excessive errors.
vmunix: SCSI: Async write error -- dev: b 31 0x012000, errno: 5, resid: 8192,
vmunix: blkno: 9851336, sectno: 19702672, offset: 1497833472, bcount: 8192.
vmunix: DIAGNOSTIC SYSTEM WARNING:
vmunix: The diagnostic logging facility is no longer receiving excessive
vmunix: errors from the I/O subsystem. 65 I/O error entries were lost.

These messages were the cause of the performance issue?. Should we monitor this kind of messages in order to avoid any issue int he future?

The disk was replaced and the server is working fine, I just want to know if I was wrong thinking that the server should works fine even if the disk fail

I really appreciate your help.

[sathish kannan]

Hi Ross,

The errors are coming from c0t1d2 disk and they are write errors. "mstm" logtool will help you but i would recommed you to log a call with HP and check the disk.

Regards
Sathish