On 9/22/2014 12:33 PM, Luke Coldiron wrote:
It is possible and that is part of what I am trying to discover however I am very familiar with the system / code base and in this case there is a single process updating the timestamp and a single process reading the timestamp. There are no other user processes programmed to interact with this table outside of potentially what Postgres is doing.
ANY other connection to the same postgres server, even to a different database, that has an open long running transaction (most frequently, "Idle In Transaction") will block autovacuum from marking the old tuples as reusable.
-- john r pierce 37N 122W somewhere on the middle of the left coast -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general