[Pgcluster-general] Replication: PgCluster

[K, Niranjan (NSN - IN/Bangalore)]

Hi Mitani,

I have the following environment:
Server1 - Replicator_1 (Active) & Cluster_1
Server2 - Cluster_2 & Replicator_2 (Coldstandby Standby)

While executing test_app on the Server_1, I bring down  the interface in
the Server_2. With this somehow the 'Postmaster' listening socket is
getting connection request for database 'template1'. Do you have a clue
which component is sending this request?. This is causing the subprocess
to wait for SQL statements infinitely.
Futher to the above, the subprocess which was handling the DB session
which was working fine before the interface was brought down, the
subprocess hangs infinitely at the recv() system call in 'secure_read'
(waiting for the SQL). But at the client side, the PQexec() would have
also hung. So there is a deadlock situation. Coud you please let me know
if there are known issues regarding this. 

regards,
Niranjan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://pgfoundry.org/pipermail/pgcluster-general/attachments/20080703/64fa5d97/attachment.html