Re: speeding up pg_dump?

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: me(at)alternize(dot)com
Cc: pgsql-novice(at)postgresql(dot)org
Subject: Re: speeding up pg_dump?
Date: 2005-12-27 04:35:18
Message-ID: 19455.1135658118@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-novice

<me(at)alternize(dot)com> writes:
> the pg_dump of our 2.45gb db takes an awfully long 45min during which db
> services come more or less to a hold.
> this does not seem right to me.

Me either. Do you have operations that are taking exclusive locks on
tables? pg_dump takes AccessShareLock (ie, a reader's lock) on every
table in sight, to make sure the tables don't disappear or change schema
underneath it. This doesn't cause any problem for concurrent SELECT,
INSERT/UPDATE/DELETE, nor plain VACUUM ... but it could result in
blocking schema changes, VACUUM FULL, etc. And if so, those would
in turn block everything else.

Looking at pg_locks would confirm or deny this idea.

regards, tom lane

In response to

Responses

Browse pgsql-novice by date

  From Date Subject
Next Message me 2005-12-27 05:39:44 Re: speeding up pg_dump?
Previous Message me 2005-12-27 02:47:53 speeding up pg_dump?