Hardware advice - Mailing list pgsql-performance

From Alexandru Coseru
Subject Hardware advice
Date
Msg-id 00e601c71700$86f23e90$c7646b56@alex
Whole thread Raw
In response to Propagating outer join conditions  ("Aaron Birkland" <birkie@gmail.com>)
Responses Re: Hardware advice  (Josh Berkus <josh@agliodbs.com>)
Re: Hardware advice  (Sven Geisler <sgeisler@aeccom.com>)
List pgsql-performance
Hello..

I'm waiting for my new system , and meanwhile , i have some questions.
First , here are the specs:


The server will have kernel 2.1.19  and it will be use only as a postgresql
server  (nothing else...  no named,dhcp,web,mail , etc).
Postgresql version will be 8.2.
It will be heavily used on inserts , triggers on each insert/update  and
occasionally  selects.

System:  SuperMicro 7045B-3
CPU:   1 Dual Core Woodcrest ,2.66 Ghz , 4Mb cache , FSB 1333Mhz
RAM:   4 Gb   (4 x 1 Gb modules) at 667Mhz
RAID CTRL:  LSI MegaRAID SAS 8408E
DISKS:  8 x  SATA II  7200 rpm , NCQ ,  8 Mb cache        with Supermicro 8
Sas  enclosure

Based on the needs , i'm planning an update of the drives to 15.000 rpms
SAS.  (pretty expensive now)


Question 1:
    The RAID layout should be:
            a)  2 hdd in raid 1 for system and pg_xlog  and 6 hdd in raid10
for data ?
            b)  8 hdd in raid10  for all ?
            c)  2 hdd in raid1 for system  , 2 hdd in raid1 for pg_xlog , 4
hdd in raid10 for data ?
    Obs: I'm going for setup a)  , but i want to hear your thoughts as well.


Question 2:  (Don't want to start a flame here..... but here is goes)
        What filesystem should i run for data ?      ext3 or xfs ?
        The tables have ~ 15.000 rel_pages each.  The biggest table has now
over 30.000 pages.

Question 3:
        The block size in postgresql is 8kb.      The strip size in the raid
ctrl is 64k.
         Should i increase the pgsql block size to 16 or 32 or even 64k ?



As soon as the system will be delivered , i'm planning some benchmarks.

Regards
    Alex


pgsql-performance by date:

Previous
From: "Aaron Birkland"
Date:
Subject: Re: Propagating outer join conditions
Next
From: Tom Lane
Date:
Subject: Re: Propagating outer join conditions