[Pgcluster-general] pgreplicate hangs on sem_unlock

[Sky Stebnicki]

I all, thanks in advance for your work and quality support.

We are working on a multi-master solution for PostgreSQL and were extremely happy to find the pgcluster project. However, we are having some trouble on our test system.

pgreplicate hangs after a few queries and the clients return "Query not permitted when all replication servers fell down."

I started pgrepliacte with -n -v and got the following:

.....
2007-08-03 15:45:20 [33583] DEBUG:from_host=10.100.16.45
2007-08-03 15:45:20 [33583] DEBUG:dbName=guaranty
2007-08-03 15:45:20 [33583] DEBUG:userName=guaranty
2007-08-03 15:45:20 [33583] DEBUG:recieve sec=1186181120
2007-08-03 15:45:20 [33583] DEBUG:recieve usec=713792
2007-08-03 15:45:20 [33583] DEBUG:query_size=21
2007-08-03 15:45:20 [33583] DEBUG:request_id=1
2007-08-03 15:45:20 [33583] DEBUG:replicate_id=0
2007-08-03 15:45:20 [33583] DEBUG:recovery_status=0
2007-08-03 15:45:20 [33583] DEBUG:query=PGR_CLOSE_CONNECTION
2007-08-03 15:45:20 [33583] DEBUG:sem_lock [1] req
2007-08-03 15:45:20 [33583] DEBUG:sem_lock [1] got it
2007-08-03 15:45:20 [33583] DEBUG:source host
2007-08-03 15:45:20 [33583] DEBUG:replicate_packet_send_internal():connect db:guaranty port:5432 user:guaranty host:devserv2.guaranty.com query:PGR_CLOSE_CONNECTION
2007-08-03 15:45:20 [33583] DEBUG:start thread_send_cluster()
2007-08-03 15:45:20 [33583] DEBUG:sem_unlock[1]

Then it just hangs.

Here is our test configuration:

Server 1:
Postgresql Node 1 & pgreplicate server

Server 2:
Postgresql Node 2

Both are running FreeBSD 6.0 and pgcluster 1.7.0rc7

Any suggestions?

Sincerely,

Sky


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://pgfoundry.org/pipermail/pgcluster-general/attachments/20070803/e2c746db/attachment.html