On Thu, 28 Aug 2008, Scott Marlowe wrote:
On Thu, Aug 28, 2008 at 2:29 PM, Matthew Wakeling <[EMAIL PROTECTED]> wrote:
Another point is that from a business perspective, a database that has
stopped responding is equally bad regardless of whether that is because the
OOM killer has appeared or because the machine is thrashing. In both cases,
there is a maximum throughput that the machine can handle, and if requests
appear quicker than that the system will collapse, especially if the
requests start timing out and being retried.
But there's a HUGE difference between a machine that has bogged down
under load so badly that you have to reset it and a machine that's had
the postmaster slaughtered by the OOM killer. In the first situation,
while the machine is unresponsive, it should come right back up with a
coherent database after the restart.
OTOH, a machine with a dead postmaster is far more likely to have a
corrupted database when it gets restarted.
wait a min here, postgres is supposed to be able to survive a complete box
failure without corrupting the database, if killing a process can corrupt
the database it sounds like a major problem.
David Lang
Likewise, I would be all for Postgres managing its memory better. It would
be very nice to be able to set a maximum amount of work-memory, rather than
a maximum amount per backend. Each backend could then make do with however
much is left of the work-memory pool when it actually executes queries. As
it is, the server admin has no idea how many multiples of work-mem are going
to be actually used, even knowing the maximum number of backends.
Agreed. It would be useful to have a cap on all work_mem, but it
might be an issue that causes all the backends to talk to each other,
which can be really slow if you're running a thousand or so
connections.
--
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
