From: | Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Zdenek Kotala <Zdenek(dot)Kotala(at)Sun(dot)COM>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: New FSM patch |
Date: | 2008-09-12 15:27:45 |
Message-ID: | 48CA8A71.3020005@enterprisedb.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Tom Lane wrote:
> Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> writes:
>> Let me describe this test case first:
>> - The test program calls RecordAndGetPageWithFreeSpace in a tight loop,
>> with random values.
>
> What's the distribution of the random values, exactly? In particular,
> how do the request sizes compare to available free space per-page?
The request, and "old avail" sizes are in the range of 0-8100
(random()%8100).
> The design intent for FSM was that we'd not bother to record pages that
> have less free space than the average request size, so as to (usually)
> avoid the problem of uselessly searching a lot of entries. I can't tell
> whether your test case models that behavior at all. If it does then
> there may be something else that needs fixing.
Probably not. The test case starts with a table that's practically
empty, so all pages are put into the FSM.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
From | Date | Subject | |
---|---|---|---|
Next Message | Heikki Linnakangas | 2008-09-12 15:33:46 | Re: Synchronous Log Shipping Replication |
Previous Message | Heikki Linnakangas | 2008-09-12 15:23:26 | Re: New FSM patch |