Hi,
My database stopped responding last night (Postgres 8.1.4).
at 2 am, a vacuum began running:
/usr/local/pgsql/bin/vacuumdb -afz
When I came in to work this morning, I could query some tables
but not others. There were many(100) processes like:
postgres 11791 6901 0 02:07 ? 00:00:00 postgres: www-data yadb2 127.0.0.1(40883) SELECT waiting
I began running an update yesterday that may or may not have completed.
I was moving data out of one table into a new table, and setting
an ID field to point to the new table. That may or may not have
completed (it was done by this morning). It is not a regular occurence
and today was the first time I saw this hanging behavior, so it's
probably related.
The postmaster was using a lot of CPU.
I stopped the postmaster with a SIGINT, it stopped quickly
and came back up automatically (I'm using daemontools) and
everything is fine.
I'm guessing it's something related to table locks.
Any pointers on what I should be looking for to prevent this from
happening again? What information I should be tracking to figure
out what is exactly happening?
Thanks very much,
j