* Michal Taborsky - Internet Mall <michal.taborsky at mall.cz> [070419 13:00]:
> Hello everyone.
> 
> We have a pretty simple setup. Two servers, one database replicated from master to slave, everything in one replication set, about 200 tables. Slony version is 1.2.9, PostgreSQL version is 8.1.5.
> 
> We loaded "a bit" of data into one of the tables on master, about 6 million records. Which generated 6 million sl_log_1 rows. The problem now is, that the slon on slave issues a "fetch 100 from LOG;" command, 
> which takes about 2 minutes to complete! The servers are loaded, but not overloaded (they are both dual-dual-core Xeons with 4G RAM).
> 
> Thre replication is virtually stopped at the moment, the sl_log_1 is growing bigger and bigger, because the primary server does some OLTP work.
> 
> Can anything be done to remedy this? The problem obviously is with the FETCH, but I thought there were supposed to be some indexes on that table to speed-up this.

drop the replication, restart it from scratch.

Basically, slony is prone to getting overwhelmed by to many changes in
a given time interval. E.g. one "update table set count=count+1;" will
generate one update per row.

> The primary server may not be taken offline. The secondary could be, but only as a last resort, if there are no other ways.
Restart the replication from scratch.

Andreas