Re: PostgreSQL as a local in-memory cache

On Tue, Jun 29, 2010 at 9:39 PM, Bruce Momjian <bruce(at)momjian(dot)us> wrote:
> Jignesh Shah wrote:
>> On Tue, Jun 29, 2010 at 2:45 PM, Bruce Momjian <bruce(at)momjian(dot)us> wrote:
>> > Tom Lane wrote:
>> >> Bruce Momjian <bruce(at)momjian(dot)us> writes:
>> >> >>> I asked on IRC and was told it is true, and looking at the C code it
>> >> >>> looks true. ?What synchronous_commit = false does is to delay writing
>> >> >>> the wal buffers to disk and fsyncing them, not just fsync, which is
>> >> >>> where the commit loss due to db process crash comes from.
>> >>
>> >> >> Ah, I see. ?Thanks.
>> >>
>> >> > I am personally surprised it was designed that way; ?I thought we would
>> >> > just delay fsync.
>> >>
>> >> That would require writing and syncing to be separable actions. ?If
>> >> you're using O_SYNC or similar, they aren't.
>> >
>> > Ah, very good point. ?I have added a C comment to clarify why this is
>> > the current behavior; ?attached and applied.
>> >
>> > --
>> > ?Bruce Momjian ?<bruce(at)momjian(dot)us> ? ? ? ?http://momjian.us
>> > ?EnterpriseDB ? ? ? ? ? ? ? ? ? ? ? ? ? ? http://enterprisedb.com
>>
>>
>> Though has anybody seen a behaviour where synchronous_commit=off is
>> slower than synchronous_commit=on  ? Again there are two cases here
>> one with O_* flag and other with f*sync flags. But I had seen that
>> behavior with PostgreSQL 9.0 beta(2 I think) though havent really
>> investigated it much yet .. (though now I dont remember which
>> wal_sync_method flag) . Just curious if anybody has seen that
>> behavior..
>
> I have trouble believing how synchronous_commit=off could be slower than
> 'on'.
>
> --
>  Bruce Momjian  <bruce(at)momjian(dot)us>        http://momjian.us
>  EnterpriseDB                             http://enterprisedb.com
>
>  + None of us is going to be here forever. +
>

Hi Bruce,

Let me clarify the problem a bit.. If the underlying WAL disk is SSD
then it seems I can get synchronous_commit=on to work faster than
synchronous_commit=off.. Yes sounds unintuitive to me. But the results
seems to point in that direction. It could be that it hit some other
bottleneck with synchronous_commit=off reaches that
synchronous_commit=on does not hit (or has not hit yet).

Brads point of wal buffers could be valid. Though typically I havent
seen the need to increase it beyond 1024kB yet.

Hopefully I will retry it with the latest PostgreSQL 9.0 bits and see
it happens again.
More on that later.

Regards,
Jignesh