Kyle Hanson wrote:

>Hi,
>
>I'm quite new to slony...
>
>We have a configuration of one master and one slave.  We had problems for
>the last couple months, thus slony wasn't actively being used(slon process
>executing on server).  Today, we've restarted the server and no errors seem
>to be happening but all I seem to see are messages like:
>
>2006-03-27 13:47:32 AST DEBUG2 syncThread: new sl_action_seq 10306404 - SYNC
>1046062
>2006-03-27 13:47:32 AST DEBUG2 localListenThread: Received event 1,1046062
>SYNC
>2006-03-27 13:47:42 AST DEBUG2 syncThread: new sl_action_seq 10306404 - SYNC
>1046063
>2006-03-27 13:47:42 AST DEBUG2 localListenThread: Received event 1,1046063
>SYNC
>2006-03-27 13:47:52 AST DEBUG2 syncThread: new sl_action_seq 10306404 - SYNC
>1046064
>2006-03-27 13:47:52 AST DEBUG2 localListenThread: Received event 1,1046064
>SYNC
>2006-03-27 13:48:02 AST DEBUG2 syncThread: new sl_action_seq 10306404 - SYNC
>1046065
>2006-03-27 13:48:02 AST DEBUG2 localListenThread: Received event 1,1046065
>SYNC
>2006-03-27 13:48:12 AST DEBUG2 syncThread: new sl_action_seq 10306404 - SYNC
>1046066
>2006-03-27 13:48:12 AST DEBUG2 localListenThread: Received event 1,1046066
>SYNC
>2006-03-27 13:48:22 AST DEBUG2 syncThread: new sl_action_seq 10306404 - SYNC
>1046067
>2006-03-27 13:48:22 AST DEBUG2 localListenThread: Received event 1,1046067
>SYNC
>2006-03-27 13:48:24 AST DEBUG2 syncThread: new sl_action_seq 10306408 - SYNC
>1046068
>2006-03-27 13:48:24 AST DEBUG2 localListenThread: Received event 1,1046068
>SYNC
>2006-03-27 13:48:34 AST DEBUG2 syncThread: new sl_action_seq 10306408 - SYNC
>1046069
>
>2006-03-27 14:27:50 AST DEBUG2 syncThread: new sl_action_seq 1 - SYNC
>1532752
>2006-03-27 14:27:50 AST DEBUG2 localListenThread: Received event 2,1532752
>SYNC
>2006-03-27 14:27:57 AST DEBUG1 remoteListenThread_1: connected to
>'dbname=order_lookup host=192.168.20.5 user=postgres'
>2006-03-27 14:28:00 AST DEBUG2 syncThread: new sl_action_seq 1 - SYNC
>1532753
>2006-03-27 14:28:00 AST DEBUG2 localListenThread: Received event 2,1532753
>SYNC
>
>
>I've been watching the numbers in 'sl_event' and they really just seem to be
>growing on both the master and the slave.  I've also looked at the
>min(ev_timestamp) and it hasn't changed from 20051128 3:44.
>
>Can someone explain what is most likely happening at this point with all
>these messages?
>Is it actually 'catching up'?
>What kind of queries, etc. can I do to determine what slony is doing and
>it's state?
>
>Thanks very much,
>Kyle
>
>  
>
There is a status view, sl_status, which is mostly useful to look at on
the node that is the origin.

It is normal that upon restarting slons after an outage that the slon
for the subscriber will run through a Whole Lot of Events before
deciding to do some work.

What's liable to happen here is that you have some enormous number of
outstanding SYNCs, and it is taking so long to read through them that a
watchdog decides to "whack" the slon because it appears to be doing no work.

What you might do instead is to get the two slons going (one for each
node), and then shut off the watchdog processes (so they don't get
interrupted) and see what you get.

If the nodes haven't actually SYNCed in months, you might find it
quicker and easier to drop replication and reinitialize Slony-I from
scratch.