On 12-03-07 05:57 AM, Laura Passigni wrote:
> Hi,
> I have a big database (about 1350GB) with Slony replica (1 master and 1
> slave).
> Last week I have had a network crash between master and slave server.
> Unfortunately there was a bulk import in progress in the master
> database. As result the slave server was not in sync with the master for
> few days.
> There are now about 80.000 records in the sl_event table and the sync is
> very slow. I have seen in the log file that the sync of each event
> requires about 1500/2000 seconds. It is a very huge time.
> Is there something that I can do? Is there some faster way to sync the
> databases?
> Thank you,
> Laura
>

It is possible that your issue is what we have been referring to as bug 
167.  This has been fixed in slony 2.1.0.  What happens is that when 
sl_log_1 or sl_log_2 grows beyond a certain size the query turns into a 
sequential scan.

Other than upgrading,  you need to push the sync interval up so you 
generate fewer syncs, and increase the sync_groupsize so it processes 
more syncs together.  This might allow slony to catch up.  Once it is 
caught up your sl_log tables will get truncated and you can go back to 
normal settings/operations.


>
>
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at lists.slony.info
> http://lists.slony.info/mailman/listinfo/slony1-general