"Arthur Ward" <award@dominionsciences.com> writes:
>> It looks to me like the guy doing VACUUM is simply waiting for the other
>> guy to release a page-level lock. The other guy is running a deferred
>> trigger and so I'd expect him to be holding one or two page-level locks,
>> on the page or pages containing the tuple or tuples passed to the
>> trigger. Nothing evidently wrong there.
> If I remember what I was working on the other day when this whole thing
> started, I think it was a single backend and a checkpoint that collided.
> I'll trace that combination, assuming it happens again.
A checkpoint would also have reason to wait for a page-level lock, if
the stuck backend was holding one. I am wondering though if the stuck
condition consistently happens while trying to fire a trigger? That
would be very interesting ... not sure what it'd mean though ...
>> The real question is why does vsyslog() have anything to block on, when
>> it's running in an unthreaded process? Seeing that you are using
>> plpython, I wonder if Python is confusing matters somehow.
> Oof. I'm using plpython all over the place; I don't think this problem has
> happened in any location that can work without it easily. :-/
It looks to me like your plpython code is all dead in the water, seeing
that your Python installation is refusing creation of rexec. (AFAIK the
only workaround is to downgrade Python to a version that allows rexec.)
If you're using it all over the place, how come you haven't noticed
that??
regards, tom lane