Re: pg_multixact not getting truncated

From: Andres Freund <andres(at)2ndquadrant(dot)com>
To: Josh Berkus <josh(at)agliodbs(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org, Jim Nasby <Jim(dot)Nasby(at)BlueTreble(dot)com>, Alvaro Herrera <alvherre(at)commandprompt(dot)com>
Subject: Re: pg_multixact not getting truncated
Date: 2014-11-08 11:52:46
Message-ID: 20141108115246.GD4510@alap3.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2014-11-07 16:09:44 -0800, Josh Berkus wrote:
> On 11/05/2014 11:15 AM, Josh Berkus wrote:
> > On 11/05/2014 10:40 AM, Jim Nasby wrote:
> >> Can you post the contents of pg_multixact/members/?
> >
> > Well, not as of last week, obviously.
> >
> > https://gist.github.com/jberkus/d05db3629e8c898664c4
> >
> > I haven't pasted all the filenames, because, well, look at the count. I
> > also added the contents of the /offsets directory, for full information.
> >
> > Note that we've added 400 multixact files since my first email.
>
> So, just did a quick survey of our managed services customers. 4 out of
> 6 have pg_multixact bloat. Basically it looks like this:
>
> User1 (the one above): 1.1GB
> User2: 64K
> User3: 929MB
> User4: 1.3GB
> User5: 301MB
> User6: 48K
>
> Nobody (out of 6 more) running a version older than 9.3 had a
> pg_multixact larger than 128K.
>
> None of these users used pg_upgrade, so that's not the source of this
> problem. Instead, as demonstrated earlier, we are simply not truncating
> pg_multixact sufficiently.
>
> So this looks like a pretty serious bug.

I've not seen actual evidence of a bug here. What's their multixact
related settings? How large is pg_clog?

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2014-11-08 11:54:58 Re: pg_multixact not getting truncated
Previous Message Petr Jelinek 2014-11-08 11:17:25 Re: Sequence Access Method WIP