STM output for disk read/write error

[HP_novice_1]

Currently I run STM on some of my harddisk and the return result show some count for the read/write error. So the question is that is there any guildline on how many read/write error will it be consider that the disk need to be replaced. As my system uptime is very critical, thus need to know any threshold for the error. If possible, I try not to replace unless of no choice.

[Steven E. Protter]

Shalom,

I don't want any read-write errors on a disk. Zero.

I would run some further tests, get a backup done and preemptively replace the disk. If the disk is merely busy, I might decide otherwise.

http://docs.hp.com/en/7161/LVM_OLR_whitepaper.pdf

That document is the holy grail on this subject.

SEP

[Vivek Bhatia]

Hi Goh,

Have a look at the doc below .It will tell you all the ways to diagnose and replace the disks.

http://docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad_WP.pdf

Regards
Vivek Bhatia

[Ganesan R]

Hi,

If you suspect any disk i/o error, run dd to check the consistency. If errors are there, replace it without second thought.

Example:
# dd if=/dev/rdsk/c1t5d0 of=/dev/null bs=1024k

[HP_novice_1]

Thanks for all the reply. But most of the reply is on the action when a disk fail. What I am looking for is that base on STM output on read/write error, can we conclude that disk is failing and need to be replace ? If not what will be the guildline for the count of read/write error before decide on changing the disk.