[Pgcluster-general] Replication server falls down after random intervals

[Leif Madsen]

Afternoon everyone,

I have a 3 DB, 1 replication server (miami7) setup. OpenSER is
performing updates to it's local DB (miami6).

It appears I have things configured correctly because for about 30
mins everything will be running along gloriously, when eventually the
DB that OpenSER is updating will give back errors such as:

miami6 /usr/sbin/openser[4808]: PG[517] begin_transaction FATAL ERROR:
 This query is not permitted when all replication servers fell down  ,
'select password,rpid from subscriber where username='my_username''!

I'm wondering if maybe there is a configuration error I'm not seeing?
It seems unlikely because it does seem to work great for a while, it
just seems to lose connection with the replication server. I am
running pgreplicate -n -v so I can see the transactions going on, and
there doesn't appear to be any errors, it just seems to stop receiving
traffic...

On the replication server (miami7) I have ports 7001, 8301, 8101,
8001, and 5432 open (this box is running both the replication server
and a clusterDB).

On miami6, I have the same ports open (as I wasn't positive which ones
I needed at first), but IPtables seems to say I only need 5432 since
it is the only one of the listed ports above that has packets hitting
the rule.

I've changed the hostnames to miami6 and miami7, but in my
configuration they are FQDNs. I've posted the configurations to
pastebin.ca at the following:

miami6 (clusterDB): http://www.pastebin.ca/375925
miami7 (clusterDB / replication server): http://www.pastebin.ca/375927

Any help in solving this would be greatly appreciated!

-- 
Leif Madsen.
http://www.leifmadsen.com
http://www.oreilly.com/catalog/asterisk