On 3/4/2005 4:06 PM, Josh Berkus wrote:

> Chris,
> 
>> Is there some old transaction lingering around perhaps <IDLE in
>> transaction>?
> 
> Yeah, that was the first thing I thought of.  No.    What else would cause 
> this?   I'm at kind of a loss on why Postgres can't free up the rows.

PostgreSQL can't free up rows that are deleted but (thanks to MVCC) 
still visible to someone. And a long running, old transaction is the 
only thing that can cause this.

> 
>> That table will presumably need a VACUUM FULL when you get a chance.
>> (That will slow replication while it's running, but shouldn't have much
>> impact on users; it doesn't block inserts into replicated tables...)
> 
> Will try, but I'm afraid that something is still blocking on the table.
> 
> BTW, what process is it that requires an ExclusiveLock on pg_listener?

Look with ps or via pg_stat_activity what process 18883 is trying to do, 
or what 19205 actually is chewing on?


Jan

-- 
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#================================================== JanWieck at Yahoo.com #