Re: does wal archiving block the current client connection?

On Fri, 2006-05-19 at 08:23 -0700, Jeff Frost wrote:
> On Fri, 19 May 2006, Simon Riggs wrote:
>
> > On Thu, 2006-05-18 at 10:08 -0700, Jeff Frost wrote:
> >
> >> May 18 08:00:18 discord postgres[20228]: [129-1] LOG: archived transaction log file "00000001000000000000007F"
> >> May 18 08:00:41 discord postgres[20573]: [254-1] LOG: archived transaction log file "00000001000000000000007F"
> >> May 18 08:00:41 discord postgres[20573]: [255-1] WARNING: could not rename file "pg_xlog/archive_status/00000001000000000000007F.ready" to
> >> May 18 08:00:41 discord postgres[20573]: [255-2] "pg_xlog/archive_status/00000001000000000000007F.done": No such file or directory
> >
> > I'm not clear how you can have two processes both issuing this message,
> > since it is only the archiver process that ever says this.
> >
> > Do you have two archivers running on this system?
> > pid = 20229 and pid = 20573
> >
> > Presumably pg_xlog/archive_status/00000001000000000000007F.done already
> > exists?
> >
> > Are there two postmasters running (at all)?
> > Is there somehow an archiver process running from a previously shutdown
> > postmaster (somehow)?
>
> Ah!! Good catch Simon! There are in fact two postmasters running on this
> system. The normal postmaster running in /var/lib/pgsql on port 5432 and this
> test postmaster running on port 55432. The 8.0.7 postmaster on port 5432
> shows the following for it's archive_command though:
>
> template1=# show archive_command;
> archive_command
> -----------------
> unset
> (1 row)
>
> and the port 55432 postmaster shows:
>
> template1=# show archive_command;
> archive_command
> ------------------------------------------------------
> /usr/local/pgsql-8.1.3/bin/archive_test.sh "%p" "%f"
> (1 row)
>
> Do you think the postmaster on 5432 is trying to archive the other
> postmaster's WAL files somehow?

Some wierd interaction is occurring. The log showed two attempts to
archive the file, the second one each time failing, as you'd expect. But
two archivers should not be accessing the same data directory.

What were those two pids? Which postmasters are they children of?

Give me some more info and I can probably track down a bug...

--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com