From: | Merlin Moncure <mmoncure(at)gmail(dot)com> |
---|---|
To: | Ivan Voras <ivoras(at)freebsd(dot)org> |
Cc: | pgsql-performance(at)postgresql(dot)org |
Subject: | Re: Re: Postgres insert performance and storage requirement compared to Oracle |
Date: | 2010-10-27 11:05:44 |
Message-ID: | AANLkTimdzm-U_10twZ6dVNV9abC-zuZGTs-Nui30rfvk@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers pgsql-performance |
On Wed, Oct 27, 2010 at 6:13 AM, Ivan Voras <ivoras(at)freebsd(dot)org> wrote:
> On 10/26/10 17:41, Merlin Moncure wrote:
>> On Tue, Oct 26, 2010 at 11:08 AM, Leonardo Francalanci <m_lists(at)yahoo(dot)it> wrote:
>>>> temp tables are not wal logged or
>>>> synced. Periodically they can be flushed to a permanent table.
>>>
>>>
>>> What do you mean with "Periodically they can be flushed to
>>> a permanent table"? Just doing
>>>
>>> insert into tabb select * from temptable
>>>
>>
>> yup, that's exactly what I mean -- this will give you more uniform
>
> In effect, when so much data is in temporary storage, a better option
> would be to simply configure "synchronous_commit = off" (better in the
> sense that the application would not need to be changed). The effects
> are almost the same - in both cases transactions might be lost but the
> database will survive.
right -- although that's a system wide setting and perhaps other
tables still require full synchronous fsync. Still -- fair point
(although I bet you are still going to get better performance going by
the temp route if only by a hair).
merlin
From | Date | Subject | |
---|---|---|---|
Next Message | Kevin Grittner | 2010-10-27 14:00:55 | Re: add label to enum syntax |
Previous Message | Ivan Voras | 2010-10-27 10:13:10 | Re: Postgres insert performance and storage requirement compared to Oracle |
From | Date | Subject | |
---|---|---|---|
Next Message | Francisco Reyes | 2010-10-27 12:41:23 | Regression: 8.3 2 seconds -> 8.4 100+ seconds |
Previous Message | Ivan Voras | 2010-10-27 10:13:10 | Re: Postgres insert performance and storage requirement compared to Oracle |