killing a hung postgres process brings down the Postgres database server on MAC OS X!

[Qing Zhao]

Hi, there,

We are using PostgrSQL 7.3.4 on MAC OS X G5 with Dual prpcessors and
8 GB RAM.

The postgreSQL server is our backend that stores information for a
JBoss server.
Every once in a while, we see a process (thread) that processing a
request from
the JBoss server gets timed out. Supposedly JBoss should clean up
after itself when it times out.
But it did not. If we do "kill -9 pid" to get rid of the deserted
process, it bring down the PG database
server entirely.

Is there a way to kill the dead processes/thread without really affect
the PG server?

Thanks!

Qing Zhao


---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings

[Rosser Schwarz]

while you weren't looking, Qing Zhao wrote:

[...]

Post by Qing Zhao
Is there a way to kill the dead processes/thread without really affect
the PG server?

See: http://www.postgresql.org/docs/7.3/static/postmaster-shutdown.html

I have occasionally had to kill -9 a backend, but it's never been
pleasant. Avoid if at all possible.

/rls

[Tom Lane]

If we do "kill -9 pid" to get rid of the deserted process, it bring
down the PG database server entirely.

It's supposed to.

Is there a way to kill the dead processes/thread without really affect
the PG server?

It will time out by itself eventually (order of a couple hours probably
--- it depends on your TCP stack). You can try a SIGINT or SIGTERM.
SIGINT is safe but may or may not cause the backend to quit. SIGTERM
will make it quit, but I'm not prepared to guarantee that SIGTERM won't
have bad side effects. I'd suggest waiting...

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ***@postgresql.org so that your
message can get through to the mailing list cleanly