| From: | "Tomas Vondra" <tv(at)fuzzy(dot)cz> |
|---|---|
| To: | "Gerhard Wohlgenannt" <wohlg(at)ai(dot)wu(dot)ac(dot)at> |
| Cc: | "Tomas Vondra" <tv(at)fuzzy(dot)cz>, "Andy Colson" <andy(at)squeakycode(dot)net>, pgsql-performance(at)postgresql(dot)org, "Heinz-Peter Lang" <heinz(at)langatium(dot)net>, "Gerhard Wohlgenannt" <wohlg(at)ai(dot)wu-wien(dot)ac(dot)at>, "Weichselbraun, Albert" <albert(dot)weichselbraun(at)wu(dot)ac(dot)at> |
| Subject: | Re: Sudden drop in DBb performance |
| Date: | 2011-09-06 09:04:38 |
| Message-ID: | f84f86dd397a1c8872c8c5645d1c9cfc.squirrel@sq.gransy.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
On 6 Září 2011, 10:26, Gerhard Wohlgenannt wrote:
> Thanks a lot to everybody for their helpful hints!!!
>
> I am running all these benchmarks while the VMs are up .. with the
> system under something like "typical" loads ..
>
> The RAID is hardware based. On of my colleagues will check if there is
> any hardware problem on the RAID (the disks) today, but nothing no
> errors have been reported.
>
> please find below the results of
> iostat -x 2
> vmstat 2
>
> hmm, looks like we definitely do have a problem with I/O load?!
> btw: dm-19 is the logical volume where the /var (postgresql) is on ..
Well, it definitely looks like that. Something is doing a lot of writes on
that drive - the drive is 100% utilized, i.e. it's a bottleneck. You need
to find out what is writing the data - try iotop or something like that.
And it's probably the reason why the bonnie results were so poor.
Tomas
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Stefan Keller | 2011-09-06 13:45:57 | Fwd: Summaries on SSD usage? |
| Previous Message | Gerhard Wohlgenannt | 2011-09-06 08:55:57 | Re: Sudden drop in DBb performance |