pg.ddx.io  pgsql-admin@postgresql.org mailing list archive  
help / color / mirror / Atom feed
Relations summary at end of vacuum
2+ messages / 2 participants
[nested] [flat]

* Relations summary at end of vacuum
@ 2004-10-18 17:39 Doug Y <dylists@ptd.net>
  2004-10-18 18:06 ` Re: Relations summary at end of vacuum Tom Lane <tgl@sss.pgh.pa.us>
  0 siblings, 1 reply; 2+ messages in thread

From: Doug Y @ 2004-10-18 17:39 UTC (permalink / raw)
  To: pgsql-admin

Hi,
  What exactly are the "Relations"  that are referred to in the summary 
at the end of a vacuum? I have two DBs that are supposed to be mirrored 
copies of each other, but my relation & page counts are different for 
each after a vacuum:

DB_a:
INFO: free space map: 109 relations, 70413 pages stored; 71696 total 
pages needed
DETAIL: Allocated FSM size: 1000 relations + 100000 pages = 647 kB 
shared memory.

DB_b:
INFO: free space map: 281 relations, 57954 pages stored; 61904 total 
pages needed
DETAIL: Allocated FSM size: 1000 relations + 100000 pages = 647 kB 
shared memory.

The schemas and DBs are set up exactly the same, so I'm rather confused 
as to the discrepancy. Data checks (sums on key columns, etc) show that 
the data does appear to be in sync. I checked pg_tables, pg_indexes, and 
both have the same info.

If this is normal behavior, I'd just like some additional info to set my 
mind at ease.

Thanks



^ permalink  raw  reply  [nested|flat] 2+ messages in thread

* Re: Relations summary at end of vacuum
  2004-10-18 17:39 Relations summary at end of vacuum Doug Y <dylists@ptd.net>
@ 2004-10-18 18:06 ` Tom Lane <tgl@sss.pgh.pa.us>
  0 siblings, 0 replies; 2+ messages in thread

From: Tom Lane @ 2004-10-18 18:06 UTC (permalink / raw)
  To: Doug Y <dylists@ptd.net>; +Cc: pgsql-admin

Doug Y <dylists@ptd.net> writes:
>   What exactly are the "Relations"  that are referred to in the summary 
> at the end of a vacuum?

Tables and indexes.

> I have two DBs that are supposed to be mirrored copies of each other,
> but my relation & page counts are different for each after a vacuum:

IIRC, a relation won't get entered into the free space map at all if
no useful free space is found by VACUUM.  So if the only thing that ever
happens on the mirror is insertion of new rows, I could believe that
there would be many fewer active FSM entries than for the master.

However the fact that the database with fewer FSM relations knows about
more free pages within those tables seems a bit at variance with that
idea.  Maybe your mirror is insert-only for some tables and not others,
and you've been lax about vacuuming the mirror?

			regards, tom lane



^ permalink  raw  reply  [nested|flat] 2+ messages in thread


end of thread, other threads:[~2004-10-18 18:06 UTC | newest]

Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2004-10-18 17:39 Relations summary at end of vacuum Doug Y <dylists@ptd.net>
2004-10-18 18:06 ` Tom Lane <tgl@sss.pgh.pa.us>

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox