| From: | Noah Misch <noah(at)leadboat(dot)com> |
|---|---|
| To: | hubert depesz lubaczewski <depesz(at)depesz(dot)com>, pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Huge spikes in number of connections doing "PARSE" |
| Date: | 2011-03-15 04:14:11 |
| Message-ID: | 20110315041411.GB24933@tornado.gateway.2wire.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Mon, Mar 14, 2011 at 10:21:27PM +0100, Martijn van Oosterhout wrote:
> On Mon, Mar 14, 2011 at 07:49:46PM +0100, hubert depesz lubaczewski wrote:
> > These 60 were summarized, and output is available here:
> > http://www.depesz.com/various/locks.summary.txt
> >
> > as you can seem, in 48 cases backend process was in semop(), which relates
> > directly to my previous findings with ps/wchan.
>
> It's unfortunate you don't have debug symbols enabled, which makes
> these traces somewhat unreliable. So you get odd things like index_open
> calling index_close.
That trace is legitimate. If index_open()'s call into the relcache needs to
consult the catalogs, it will in turn scan pg_class_oid_index.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tory M Blue | 2011-03-15 05:36:06 | pgstat wait timeout |
| Previous Message | Noah Misch | 2011-03-15 04:13:21 | Re: Huge spikes in number of connections doing "PARSE" |