http://www.slony.info/bugzilla/show_bug.cgi?id=81

--- Comment #4 from Christopher Browne <cbbrowne at ca.afilias.info> 2010-03-31 14:40:51 PDT ---
(In reply to comment #0)
> First reported 08/2008 by Shahaf Abileah, then found by Bill Moran at CFI.
> 
> http://www.mail-archive.com/slony1-general@lists.slony.info/msg02754.html
> 
> http://www.nabble.com/duplicate-key-sl_nodelock-pkey-td22246434.html

The report by Shahaf Abileah doesn't seem to me to indicate an error; if you
try to run a slon on 2 servers, then there's no way to find, via /usr/bin/ps,
that there were two instances.  That case doesn't seem like a bug to me - it
looks instead like the slon is working correctly.  The only possible change
would be to catch the exception and give a more explicit warning to the effect
that "it looks like you have a slon running already!"

Bill Moran's report isn't about the same thing.

In his case, the attempt to start another slon leads to the same symptom (e.g.
- slon complains that it can't get the exclusive lock on the tuple).

The *problem*, however, seems to be that the "old" slon has gotten somehow
'wedged,' and I'd think it likely that the problem is that it's waiting for a
lock to be granted.

What's likely necessary to diagnose the issue is a query against pg_locks on
the node being managed showing what the "stalled" connection is waiting on.

-- 
Configure bugmail: http://www.slony.info/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
You are the assignee for the bug.