Hi guys,
We have slony 1.1.5 running on top of postgresql 8.1 on Solaris 10 box. We
have master-slave replication, both machine sitting in the same LAN.
Since last night, the number of entries on sl_log_1 didn't get cleared
(although i can see the cleanup thread doing its job).

MONSOON=# select ev_origin, ev_seqno, ev_minxid from
"_MONSOONCLUSTER".sl_event where (ev_origin, ev_seqno) in   (select
ev_origin, min(ev_seqno)  from "_MONSOONCLUSTER".sl_event    where ev_type =
'SYNC'    group by ev_origin);
 ev_origin | ev_seqno | ev_minxid
-----------+----------+-----------
         2 |     6823 | 82606432
         1 |    13836 | 746264948
(2 rows)

The value of ev_minxid has been the same since last night, which explains
why the sl_log_1 keeps growing.
We haven't made any schema changes to the database since mid september. No
node has been added/removed from replication either. In short, nothing
changes since mid september.

This is not the first time we encounter this issue. we had one similar to
this last week. sl_log_1 grows so big that the "FETCH 100 FROM LOG" took
more than 20 seconds to complete and we saw system slow down.
In the past, we just drop and recreate the cluster to solve this issue.

I am wondering if anyone has any sight as to why this is happening, how to
solve this issue and how to prevent this issue from happening again in the
future.
Thanks in advance

J
PS: currently, there are around 500k records in sl_log_1, and we started
seeing the slowness in our system.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gborg.postgresql.org/pipermail/slony1-general/attachments/20061016/6eff330d/attachment.html