Re: Checkpoint Tuning Question

On Mon, Jul 13, 2009 at 3:53 PM, Dan
Armbrust<daniel(dot)armbrust(dot)list(at)gmail(dot)com> wrote:
>> So this thought leads to a couple of other things Dan could test.
>> First, see if turning off full_page_writes makes the hiccup go away.
>> If so, we know the problem is in this area (though still not exactly
>> which reason); if not we need another idea.  That's not a good permanent
>> fix though, since it reduces crash safety.  The other knobs to
>> experiment with are synchronous_commit and wal_sync_method.  If the
>> stalls are due to commits waiting for additional xlog to get written,
>> then async commit should stop them.  I'm not sure if changing
>> wal_sync_method can help, but it'd be worth experimenting with.
>>
>>                        regards, tom lane
>>
>
>
> All of my testing to date has been done with synchronous_commit=off
>
> I just tried setting full_page_writes=off - and like magic, the entire
> hiccup went away.
>
> Thanks,
>
> Dan

I haven't done any other testing for this issue since discovering that
setting "full_page_writes=off" makes my performance hiccup go away.
I'm about to move on to some different tests and reset this setup that
I was using to test out this issue - is there anything else that
anyone would like to have gathered before I axe this system?

I can redo it again later too, it will just take a fair bit longer for
me to reconfigure the test environment.

Thanks,

Dan