In my original query I was wondering if there was a simple way to
identify a process which would be causes excessively high IOWAIT
percentages.
It was pointed out to me that IOWAIT is really just another way
of saying DISK IO BOTTLENECK (although this can be tty or
network traffic, this appears to be a less common problem). So, the
process is to simply use the standard tools like top, proctool, or
iostat
to look for heavy disk usage, either in the form of disk device
utilization or
high paging rates. A couple of people pointed out that it's difficult
to figure out what process is causing all the disk activity. Several
strongly suggested using proctool because of all of the statistics
it provides. Also "iostat -x 30" will help isolate the disk device.
Because the iowaits are kernel oriented they don't show up as
CPU intensive processes. It still comes down to a lot of guesswork...
what process might be using this disk? If it's paging or swapping
causing the IOWAITs then memory hogging apps could be to blame.
Thanks to...
Kevin.Sheehan@uniq.com.au (Kevin Sheehan {Consulting Poster Child})
russell@mds.lmco.com (Russell David)
Michael Hill <Hill.Michael@tci.com>
Matthew Stier <Matthew.Stier@tddny.fujitsu.com>
rros44@tsg.cbot.com (Rusty Rose)
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | Dan Penrod - Network Administrator | | Paradyne / Networking Services | | 8545 126th Ave. N., Mailstop LG130, Largo, FL 8597 | | vox:813/530-8597 fax:530-2480 dan@paradyne.com | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:05 CDT