Re: Replication: slave is in permanent startup 'recovery'

On Wed, April 13, 2011 20:15, Henry C. wrote:
> If I try and execute a long-lived SQL query on the slave, it eventually fails
> with "canceling statement due to conflict with recovery". Replication is
> definitely working (DML actions are propagated to the slave), but something
> is amiss.

Let me ask the question another way.

Following the hints at http://wiki.postgresql.org/wiki/Streaming_Replication:

Between the primary and standby hosts, pg_current_xlog_location() and
pg_last_xlog_receive_location()/select pg_last_xlog_replay_location() show
what appears to be normal activity (ie, replication is happening successfully,
and the numbers indicate health).

primary% ps -ef | grep sender
standby% ps -ef | grep receiver

...show things talking nicely and the numbers match.

ps also shows a startup process recovering logs on the standby (which I assume
is normal behaviour).

However, a SELECT eventually fails with "canceling statement due to conflict
with recovery".

Where else can I check, or what else can I do to determine what the problem is?

Thanks