Performance and 72.devel

From: John Summerfield <pgtest(at)os2(dot)ami(dot)com(dot)au>
To: PostgreSQL Bugs <pgsql-bugs(at)postgresql(dot)org>
Subject: Performance and 72.devel
Date: 2001-09-15 12:47:28
Message-ID: Pine.LNX.4.33.0109152039340.14458-100000@dugite.os2.ami.com.au
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

I checked out the latest updates about 14 hours ago.

I've also put together a new box, featuring an Athlon running at 1.3 Mhz. I cloned the OS (Red Hat Linux 7.1) - I'd copied it from one disk to another fairly recently, and so the software setup is pretty well precisely what I've been using all along.

A job which took well over three hours on my Pentium II(at)133 (mostly CPU time) runs in a little over two (mostly I/O), a result that doesn't surprise me a lot.

It crossed my mind that PG is probably using new log files all the time:

2001-09-15 20:26:56 [30787] DEBUG: recycled transaction log file 0000000000000032
2001-09-15 20:29:36 [30788] DEBUG: recycled transaction log file 0000000000000033
2001-09-15 20:29:36 [30788] DEBUG: recycled transaction log file 0000000000000034
2001-09-15 20:35:11 [30791] DEBUG: recycled transaction log file 0000000000000035
2001-09-15 20:35:11 [30791] DEBUG: recycled transaction log file 0000000000000036

and so on.

I think that if it actually reused them instead of deleting old files and creating new ones then on my system (256 Mbytes of RAM) a few log files would fit entirely in cache and it would really fly.

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message pgsql-bugs 2001-09-16 04:52:56 Bug #452: unique indices and multi-row UPDATE problem
Previous Message Tom Lane 2001-09-14 14:41:23 Re: Re: timestamps cannot be created without time zones