Crashed hard disk

[Fabian Briseño]

Hello Guys.

We had a strange problem yesturday (The problem is already solved), It turns out that the primary OS disk on one of our production servers crashed, the alternate took over, but here is were it get weird, once the alternate took over the NFS moutn points in this server started haging and the system was slow to innaccesible.

And once the new disk (with no information completely blank) was put in the server the system started working again fne.

My question is why did the NFS mount points hang, I see no relation between the crashed disk and the mount points, Once the alternate disk takes over everything should be the same right ?

[Steven E. Protter · Accepted solution]

Shalom Fabian,

The answer depends on a few factors. I don't think there would have been a nfs issue had the disk crashed say during a boot and the system booted off the alternate drive.

This appears to have been a crash while running and there seems to have been some delay in the alternate disk taking over, enough to make NFS, which is touchy a problem.

You should have been able to clear the issue by starting and stopping the three nfs services on the system, core, client and server (if in use).

You are correct that when the alternat disk takes over this should not occur. I'd check the configuration and mirroring of the system, there appears to have been a problem uncovered. lvlnboot -v to start.

Some people don't mirror every logical volume on the boot disk and this could have also contributed to the issue. Its quite common to not mirror swap and if swap were inacessible during the outage this would explain a lot.

If you have time to check, the system logs when the event occurred might shed some light on the situation.

SEP

[Fabian Briseño]

Thanks for your response steven.

Your answers were very helpfull.