Re: We've been affected by a pg_upgrade bug. What do we do next? - Mailing list pgsql-general

From Alvaro Herrera
Subject Re: We've been affected by a pg_upgrade bug. What do we do next?
Date
Msg-id 20140716152916.GH11811@eldon.alvh.no-ip.org
Whole thread Raw
In response to Re: We've been affected by a pg_upgrade bug. What do we do next?  (Alvaro Herrera <alvherre@2ndquadrant.com>)
Responses Re: We've been affected by a pg_upgrade bug. What do we do next?  (Shaun Thomas <sthomas@optionshouse.com>)
List pgsql-general
Alvaro Herrera wrote:
> Shaun Thomas wrote:
> > This:
> >
> > 23334|4293964274||||53c68b69.5b26|2014-07-16 09:25:45 CDT|ERROR:  could not access status of transaction 7150346
> > 23334|4293964274||||53c68b69.5b26|2014-07-16 09:25:45 CDT|DETAIL:  Could not open file "pg_multixact/offsets/006D":
Nosuch file or directory. 
> >
> > Been doing it every ten seconds since yesterday.
> >
> > My log line prefix is this:
> >
> > log_line_prefix = '%p|%x|%u|%d|%r|%c|%t|'
> >
> > So the user, database, and connection source are all blank. The PID is different every time, too. Not sure what to
makeof that. 
>
> It's an autovacuum worker, which is expected.  Just get rid of the 0000
> file and all should be well.

BTW if you do a hexdump of the 0000 file, it should be all zeroes.

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


pgsql-general by date:

Previous
From: Shaun Thomas
Date:
Subject: Re: We've been affected by a pg_upgrade bug. What do we do next?
Next
From: Shaun Thomas
Date:
Subject: Re: We've been affected by a pg_upgrade bug. What do we do next?