On Sat, Feb 18, 2012 at 8:59 AM, Gregg Jaskiewicz <gryzman@gmail.com> wrote:
> What is a likelihood of a deadlock occurring, caused (or helped by)
> auto vacuum.
> This is on 8.3.
>
> The table with deadlocks was quite busy with updates, etc.
The only locking issues (not deadlocks) I've ever had with autovacuum
were priority inversion issues with slony, where a slony execute
script was stuck behind an autovacuum, and something else was then
stuck behind the slony execute script.
I'm afraid we'll need more info to determine what was happening. It's
a good idea to make a snapshot of what's in pg_stat_activity and
pg_locks when these things happen so you have something to look at
after the fact.