| From: | Fujii Masao <masao(dot)fujii(at)gmail(dot)com> |
|---|---|
| To: | Andres Freund <andres(at)2ndquadrant(dot)com> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: too low NAPTIME_PER_CYCLE /too many wakeups in walreceiver.c |
| Date: | 2012-05-30 16:43:06 |
| Message-ID: | CAHGQGwH9QJmffoTYc7QBRjaQe3c_5aRML2MRU7aK+SM9o9iQDw@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, May 30, 2012 at 9:57 PM, Andres Freund <andres(at)2ndquadrant(dot)com> wrote:
> Hi,
>
> Currently the walreceiver wakeups NAPTIME_PER_CYCLE=100 miliseconds in idle
> state. This is rather frequent. I don't really see any reason to do so.
> A nice fix would be to latchify that with WaitLatchOrSocket + a SetLatch in
> the signal handler for shutdown but that seems to be too invasive at that
> point. So I suggest simply increasing NAPTIME_PER_CYCLE to 500 or 1000ms?
Or calculate the sleep time from wal_receiver_status_interval in each
cycle and use it?
Regards,
--
Fujii Masao
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Ants Aasma | 2012-05-30 16:43:22 | Re: slow dropping of tables, DropRelFileNodeBuffers, tas |
| Previous Message | Merlin Moncure | 2012-05-30 16:26:06 | Re: 9.2beta1, parallel queries, ReleasePredicateLocks, CheckForSerializableConflictIn in the oprofile |