On Sat, Oct 30, 2004 at 03:14:46PM -0400, Jan Wieck wrote:
> On 10/30/2004 10:20 AM, Ed L. wrote:
> 
> >On Friday October 29 2004 11:55, Ed L. wrote:
> >>I'm seeing these messages as I attempt to start replication with 1.0.5
> >>...
> >>
> >>DEBUG1 remoteWorkerThread_1: connected to provider DB
> >>WARN   remoteWorkerThread_1: transactions earlier than XID 41058487 are
> >>still in progress
> >>WARN   remoteWorkerThread_1: data copy for set 1 failed - sleep 60
> >>seconds
> >
> >After more than 8 hours, nothing has been replicated, and the above
> >messages continue.  This system has over 100 connections; what is
> >the easiest way to identify whose holding up the replication?
> >Would it be possible for slony to actually help in identifying the
> >backend that has the outstanding transactions?
> 
> A combination of pg_locks together with pg_stat_activity (if that's
> enabled) or a ps(1) would probably give you some idea what's sitting
> there in an endless, idle transaction.

Here's my chance to pimp for a thing that lets you JOIN
pg_stat_activity (or whatever) with the process table.  Some wackiness
may ensue if you're not using 8.0, but you can avoid it by using
commandprompt's PL/PerlNG (thanks, Josh!)
<http://www.commandprompt.com/entry.lxp?lxpe=285>.

The rest of the code (warning: untested) is at

<http://fetter.org/pg_ps/pg_ps.sql>

HTH :)

Cheers,
D
-- 
David Fetter david at fetter.org http://fetter.org/
phone: +1 510 893 6100   mobile: +1 415 235 3778

Remember to vote!