high request queue

[Deniz Cendere]

Hi all,

We are trying to restore some number of datafiles to our system. Datafiles are approximately 100Gb. We started to restore, after a while it started to wait something, after 8 hours it didnt restore the datafile and passed to the other one. When I examined that 8 hours, I saw that no I/O s but very high request queues like 50 to 300 per disk. What do you think of the problem?

Also I have 4 luns and my luns sizes are 800 Gb. Their queue_depth is 8. What must be the queue_depth for these luns?

Thanks

Deniz

[Steven E. Protter]

Shalom,

I've seen this when there is high activity on a system and a log of open filehandles, without the files actually needing to be there.

What is the LUN construction? Raid 5? This could be the result of a bad or near failure disk in the raid queue.

SEP

[skt_skt]

does the syslog reports any type of I/O errors during that time frame..I would recommend an analysis from the SAN side for these four luns.

[Deniz Cendere]

No I/O errors in the system logs, no disk failures. We were able to restore that file later.