Re: O_DSYNC broken on MacOS X?

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: Dave Page <dpage(at)pgadmin(dot)org>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: O_DSYNC broken on MacOS X?
Date: 2010-09-30 20:27:34
Message-ID: AANLkTi=AVOoz-_gB1fNANmyKPLHd1o4tA+TBS136MCE1@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thu, Sep 30, 2010 at 4:09 PM, Dave Page <dpage(at)pgadmin(dot)org> wrote:
> On Thu, Sep 30, 2010 at 8:26 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>> Jayant Kumar did some benchmarking of InnoDB vs. PostgreSQL and PG
>> came out 5 times faster.  The benchmark isn't very thoroughly
>> described, but it turns out not to matter.
>>
>> http://jayant7k.blogspot.com/2010/09/database-speed-tests-mysql-and.html
>>
>> Apparently, the reason we're faster is that wal_sync_method =
>> open_datasync, which is the default on MacOS X, doesn't actually work.
>
> That might be true, but if you check the comments, Jayant replied to say:
>
> @Andrew : I am running linux - ubuntu 10.04 - kernel 2.6.32-24. Linux
> does not support fsync_writethrough
> http://www.westnet.com/~gsmith/content/postgresql/TuningPGWAL.htm
>
> So I don't think that invalidates his benchmark. Something else might
> of course...

Oh, I missed that. Actually, I wasn't really so concerned with
whether his benchmark is correct. I *am* concerned about being broken
out of the box on MacOS X.

(I also suspect problems with the benchmark. It's hard to believe
we're 5x faster than InnoDB on an apples-to-apples comparison on
trivial queries. I'd believe 20% either way, but 5x is a lot. But
that's a question for another day.)

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise Postgres Company

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2010-09-30 20:33:09 Re: Patch to reindex primary keys
Previous Message Gurjeet Singh 2010-09-30 20:23:00 Re: git diff --patience