| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | "Nicholson, Brad (Toronto, ON, CA)" <bnicholson(at)hp(dot)com> |
| Cc: | "pgsql-general(at)postgresql(dot)org" <pgsql-general(at)postgresql(dot)org> |
| Subject: | Re: Stat estiamtes off - why? |
| Date: | 2013-08-29 17:14:59 |
| Message-ID: | 9309.1377796499@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
"Nicholson, Brad (Toronto, ON, CA)" <bnicholson(at)hp(dot)com> writes:
> I'm seeing something odd in my DB stats. This is PG 9.2.4 on CentOS. I know the stats collector can be off at times, but I am curious as to why it would be off in such a case. I know that under heavy load the stats collector can be off sometimes, but this system was barely doing anything (CPU ~ 90% idle). PG didn't shutdown at all either.
Most likely there was some DML happening concurrently with a VACUUM or
ANALYZE at some point in the past. VACUUM/ANALYZE update the live and
dead tuple counts based on what they saw, but a transaction that added
such tuples (that were counted by the maintenance op) could commit after
that happens, and then you end up with double-counted tuples.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andreas Kretschmer | 2013-08-29 19:11:07 | question about age() |
| Previous Message | Nicholson, Brad (Toronto, ON, CA) | 2013-08-29 16:24:10 | Stat estiamtes off - why? |