http://www.slony.info/bugzilla/show_bug.cgi?id=218

--- Comment #16 from Steve Singer <ssinger at ca.afilias.info> 2011-07-04 07:50:44 PDT ---
(In reply to comment #15)
> (In reply to comment #14)
> > Also remember to remove the 'LOCK @NAMESPACE at .sl_config_lock' lines from
> > slony_funcs.sql - since a) the locking is now being done elsewhere, no point in
> > paying a penalty twice.  b) Leaving it in there is a bit confusing since
> > someone else might come along and think that it is okay to obtain
> > sl_config_lock in the stored functions
> 
> Good point, though the price isn't too terribly high.
> 
> Have done so, and had a successful run of all the regression tests against
> this.
> 
> https://github.com/cbbrowne/slony1-engine/commit/35fbb13b44e2136b808a5cd7e8457f616fcc5ccb

in cleanupThread.c  there is no explicit transaction (that I can see)

This means that the LOCK TABLE  a and the 'select ..cleanupEvent(...) are done
in separate transactions, so the lock is already released before the call to
cleanupEvent.

in remote_worker.c you seem to lock config_lock before event_lock while in
slonik.c you lock event_lock before config_lock.  We should try to always get
these locks in the same order (I am not sure which order is better or if it
matters)

-- 
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.