On Tue, 2005-04-26 at 18:15 +0100, Jamie L. Penman-Smithson wrote:
> On Tue, 2005-04-26 at 11:33 -0400, Christopher Browne wrote:
> > It's not forcibly necessary that the subscriber's slon be running when
> > SUBSCRIBE SET is issued; that generates an event, and puts in the queue
> > to be processed, which can take place later.
<snip>
> > But yes, indeed, it's good to check in the subscriber's slon's logs to
> > find the COPY_SET event.
> 
> There isn't a COPY_SET event, which is probably the problem. Just plenty
> of messages like this:
> 
> 2005-04-26 18:12:15 BST DEBUG2 remoteWorkerThread_1: SYNC 12413
> processing
[..]

I think the problem is that the nodes were not active in the sl_node
table:

 no_id | no_active | no_comment  | no_spool
-------+-----------+-------------+----------
     1 | t         | Master Node | f
     2 | t         | Slave node  | f

Changing no_active to true (on the master and slave) temporarily fixed
the problem - the changes that had been made on the master were
propagated to the slave. However, after restarting the slon daemons and
doing an INSERT on the master, this doesn't get propagated.

I've got this in sl_subscribe, which suggests that a COPY_SET command
was [finally] received, but now it's not doing anything.. again..

 sub_set | sub_provider | sub_receiver | sub_forward | sub_active
---------+--------------+--------------+-------------+------------
       1 |            1 |            2 | f           | t

 st_origin | st_received | st_last_event |     st_last_event_ts      |
st_last_received |    st_last_received_ts     |
st_last_received_event_ts  | st_lag_num_events |   st_lag_time
-----------+-------------+---------------+---------------------------+------------------+----------------------------+----------------------------+-------------------+-----------------
         1 |           2 |         42002 | 2005-04-27 13:53:22.64638 |
41859 | 2005-04-27 13:28:11.548989 | 2005-04-27 13:28:11.538713 |
143 | 00:25:19.149926


 con_origin | con_received | con_seqno |       con_timestamp
------------+--------------+-----------+----------------------------
          1 |            2 |     41859 | 2005-04-27 13:28:11.548989
          2 |            1 |     21873 | 2005-04-27 13:28:12.164697

2005-04-27 13:40:18 BST CONFIG storeNode: no_id=1 no_comment='Master
Node'
2005-04-27 13:40:18 BST DEBUG2 setNodeLastEvent: no_id=1 event_seq=41863
2005-04-27 13:40:18 BST CONFIG storePath: pa_server=1 pa_client=2
pa_conninfo="dbname=mail host=lorien user=slony password=q67p4o5"
pa_connretry=10
2005-04-27 13:40:18 BST CONFIG storeListen: li_origin=1 li_receiver=2
li_provider=1
2005-04-27 13:40:18 BST CONFIG storeSet: set_id=1 set_origin=1
set_comment='All mail tables'
2005-04-27 13:40:18 BST WARN   remoteWorker_wakeup: node 1 - no worker
thread
2005-04-27 13:40:18 BST DEBUG2 sched_wakeup_node(): no_id=1 (0 threads +
worker signaled)
2005-04-27 13:40:18 BST CONFIG storeSubscribe: sub_set=1 sub_provider=1
sub_forward='f'
2005-04-27 13:40:18 BST WARN   remoteWorker_wakeup: node 1 - no worker
thread
2005-04-27 13:40:18 BST DEBUG2 sched_wakeup_node(): no_id=1 (0 threads +
worker signaled)
2005-04-27 13:40:18 BST CONFIG enableSubscription: sub_set=1
2005-04-27 13:40:18 BST WARN   remoteWorker_wakeup: node 1 - no worker
thread
2005-04-27 13:40:18 BST DEBUG2 sched_wakeup_node(): no_id=1 (0 threads +
worker signaled)
2005-04-27 13:40:18 BST DEBUG2 main: last local event sequence = 21941
2005-04-27 13:40:18 BST CONFIG main: configuration complete - starting
threads
2005-04-27 13:40:18 BST DEBUG1 localListenThread: thread starts
2005-04-27 13:40:19 BST DEBUG1 main: running scheduler mainloop
2005-04-27 13:40:19 BST DEBUG1 cleanupThread: thread starts
2005-04-27 13:40:19 BST DEBUG4 cleanupThread: bias = 35383

Also.. on the master, I have the transaction in sl_log_1:

 log_origin | log_xid | log_tableid | log_actionseq | log_cmdtype |
log_cmddata 
------------+---------+-------------+---------------+-------------+-----------------------------------------------------------------------------------------------------
          1 | 1999565 |           4 |             5 | I           |
(alias,dest,username,status) values
('foobar at example.com','test at silverdream.org','example.com','1')
(1 row)

I'm running out of ideas now..

Thanks in advance,

-j
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://gborg.postgresql.org/pipermail/slony1-general/attachments/20050427/98099df3/attachment.bin