| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Craig Ringer <craig(at)postnewspapers(dot)com(dot)au> |
| Cc: | tanjunhua <tanjh(at)riso(dot)co(dot)jp>, PostgreSQL general <pgsql-general(at)postgresql(dot)org> |
| Subject: | Re: ***UNCHECKED*** Re: memory leak occur when disconnect database |
| Date: | 2009-07-21 15:02:42 |
| Message-ID: | 13164.1248188562@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Craig Ringer <craig(at)postnewspapers(dot)com(dot)au> writes:
> On Tue, 2009-07-21 at 10:13 -0400, Tom Lane wrote:
>> It's not unusual for "top" to show the postmaster's child processes as
>> "postmaster" as well. Depends on the platform and the options given
>> to top.
> Ah. Thanks for clearing that one up. That'd make more sense, since it'd
> mean the postmaster, bgwriter, and whichever of the stats collector/wal
> writer exist in 8.1 were running.
Hmm ... actually an idle 8.1 cluster should have four processes. Mine
looks like this:
UID PID PPID C STIME TTY TIME COMMAND
postgres 1643 1 0 Jul 17 ? 0:00 postmaster
postgres 1646 1643 0 Jul 17 ? 0:43 postgres: writer process
postgres 1647 1643 0 Jul 17 ? 0:00 postgres: stats buffer process
postgres 1648 1647 0 Jul 17 ? 0:00 postgres: stats collector process
If the OP had turned off stats collection then there would be only two,
so it still doesn't add up. Maybe stats collection off and one live
backend? Or maybe he missed the stats collector process --- notice in
the above that that's not a direct child of the postmaster, but a
grandchild. Possibly top would show it with a different name.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Dennis Gearon | 2009-07-21 15:08:03 | Re: commercial adaptation of postgres |
| Previous Message | Sam Mason | 2009-07-21 14:47:30 | Re: element from an array by its index |