[Pgpool-general] psql hangs via pgpool but not direct

[louis fridkis]

We are running pgpool 2.1 in replication mode on two hosts. Pgpool and
Postgres are running on each host. The hosts are running under:

Linux 2.6.18-53.el5 #1 SMP Mon Nov 12 02:14:55 EST 2007 x86_64 x86_64 x86_64
GNU/Linux



pgpool.conf contains:



port = 9999

max_pool = 4

connection_life_time = 0

replication_timeout = 5000

load_balance_mode = true

replication_stop_on_mismatch = false

replicate_select = false

insert_lock = false



Our automatic test failed because psql -p 9999 stopped working (i.e. hangs)
on

the primary host. It still worked on the secondary host and psql -p 2345

(direct access to postgres without pgpool) worked on both. We determined
that

the application, which uses java was still working fine and the DBs were
still

in sync. The pgpool log had no entries for about an hour prior, which is

unusual since normally there are several messages per minute.



After we issued the shutdown command for pgpool we saw errors like:

2009-02-24 20:06:41 ERROR: pid 13095: pool_flush_it: write failed (Broken
pipe) offset: 0 wlen: 9

2009-02-24 20:06:42 LOG:   pid 6233: ProcessFrontendResponse: failed to read
kind from frontend. frontend abnormally exited



We would appreciate any suggestions as to how we can understand why pgpool

acted this way.



Thanks,



Lou Fridkis

Postgres DBA

Fox Interactive Media -

Publisher Network Portal

2500 Broadway, 2nd Floor

Santa Monica, CA 90404

(310) 633-9925
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://pgfoundry.org/pipermail/pgpool-general/attachments/20090226/0e6bea5a/attachment.html