How to simulate disk or volume error

[Etzion Bar-Noy]

Hi.
During my work, I need to simulate disk or volume error for my QA team. The only option I see in front of me now is to disconnect the hot-swap disks (although I don't know how the system will react to their return). Can you advise of any better way to do so?

Thanks!
Etzion

[Steven E. Protter · Accepted solution]

Your option will work.

If you pull an unmirrored boot disk, your system will crash. Otherwise the reaction of the system will depend on what is on the disk.

pvdisplay -v can give you some idea what the content of the indivdual disk is.

I used a tape printer to label my disks for just the scenario you are in, so I would know which device was being pulled, so I could check it.

When a hot swap is pulled and replaced an lbolt goes into the dmesg output in /var/adm/syslog/syslog.log

That message will remain in the log until the system is rebooted.

SEP

[Fred Ruffet]

Other options :
. disconnect SCSI cable
. poweroff SCSI array or FC Hub/switch
. dd on a disk (if no mirror, may produce more damage than you really want)
. attack the disk with a screwdriver (vendor's warranty may not apply :)

Regards,

Fred

[Etzion Bar-Noy]

Fred, it doesn't seem like simulation to me :-)
I could, of course, kill the disk with acid, but then again, it would not be simulation we'll talk about.

Thanks you both for your answers.

Etzion