Hi admins,
I'm running version 8.1.9 of postgres, and using autovacuum, and I'm
worried that I'm heading towards a transaction ID wraparound failure.
autovacuum is running because it is recovering the disk space of updated
rows (of which there are many each second).
But age(datfrozenid) is over 1.2 billion and steadily increasing.
From what I can understand of the manual, once the age nears 2 billion
there will be problems.
At what age (default for version 8.1) should autovacuum kick in and
reassign the frozen XIDs so that the age is reduced?
Some more info:
postgres=# show autovacuum_freeze_max_age;
ERROR: unrecognized configuration parameter "autovacuum_freeze_max_age"
postgres=# SELECT datname, datfrozenxid FROM pg_database;
datname | datfrozenxid
----------------+--------------
postgres | 1075383673
template1 | 1075383787
template0 | 499
nga_cane_entry | 1082334308
(4 rows)
postgres=# SELECT datname, age(datfrozenxid) FROM pg_database;
datname | age
----------------+-------------
postgres | 1213693293
template1 | 1213693179
template0 | -2005890829
nga_cane_entry | 1206742658
(4 rows)
So, is autovacuum doing it's job?
Many thanks,
Rob Newton