I have a weekly task set up to VACUUM FULL a fairly large (~300M row ~50GB) database. The intent is to free up disk space from especially-large inserts that aren't covered by the regular reclamation from a daily VACUUM. Recently, I've been getting the following error:
(OperationalError) out of memory
DETAIL: Failed on request of size 330.
The PG version is 8.3.5; can't upgrade to 8.4 at this time. The machine is running a reasonably patched RHEL5, and should have at least 1-2GB RAM free at the time VACUUM is run, plus some additional GB of swap.
Some googling reveals others with the same problem, including one earlier mailing list discussion that recommended using CLUSTER instead. As I understand it, though, CLUSTER requires at least as much free disk space as the size of the largest table, which occupies most of the total size of the database.
Is there any documentation on the memory & disk usage requirements of the more intensive commands, like VACUUM, VACUUM FULL, CLUSTER, etc.? I would find this very useful, since, regarding my VACUUM v.s. CLUSTER decision, for example, I seem to be groping around without really understanding the situation. I've always been under the impression that VACUUM FULL required few resources besides time.
I realize that the long-term solution is probably to partition the table, but in the short-term, is there anything I can do about this? Thanks very much,
David
