From: Dave Cramer <pg@fastcrypt.com>
To: William Yu <wyu@talisys.com>
Cc: pgsql-performance@postgresql.org
Subject: Re: Low Performance for big hospital server ..
Date: Mon, 03 Jan 2005 19:58:44 -0500
Message-ID: <41D9EA44.3020504@fastcrypt.com> (raw)
In-Reply-To: <crc87h$2ehh$1@news.hub.org>
References: <1104634472.41d762681b548@webmail.moph.go.th>
<20050102140828.GA21826@pobox.com>
<1104683293.41d8211d62221@webmail.moph.go.th>
<41D827C1.3050707@fastcrypt.com>
<1104717243.41d8a5bb9cb9b@webmail.moph.go.th>
<crb01v$hp4$1@news.hub.org>
<41D95042.70705@fastcrypt.com>
<crc87h$2ehh$1@news.hub.org>
William Yu wrote:
> Dave Cramer wrote:
>
>>
>>
>> William Yu wrote:
>>
>>> amrit@health2.moph.go.th wrote:
>>>
>>>> I will try to reduce shared buffer to 1536 [1.87 Mb].
>>>
>>>
>>>
>>>
>>> 1536 is probaby too low. I've tested a bunch of different settings
>>> on my 8GB Opteron server and 10K seems to be the best setting.
>>
>>
>>
>> Be careful here, he is not using opterons which can access physical
>> memory above 4G efficiently. Also he only has 4G the 6-10% rule still
>> applies
>
>
> 10% of 4GB is 400MB. 10K buffers is 80MB. Easily less than the 6-10%
> rule.
>
Correct, I didn't actually do the math, I refrain from giving actual
numbers as every system is different.
>
>>> To figure out your effective cache size, run top and add free+cached.
>>
>>
>>
>> My understanding is that effective cache is the sum of shared
>> buffers, plus kernel buffers, not sure what free + cached gives you?
>
>
> Not true. Effective cache size is the free memory available that the
> OS can use for caching for Postgres. In a system that runs nothing but
> Postgres, it's free + cached.
You still need to add in the shared buffers as they are part of the
"effective cache"
Dave
>
> ---------------------------(end of broadcast)---------------------------
> TIP 8: explain analyze is your friend
>
>
--
Dave Cramer
http://www.postgresintl.com
519 939 0336
ICQ#14675561
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-performance@postgresql.org
Cc: pg@fastcrypt.com, wyu@talisys.com
Subject: Re: Low Performance for big hospital server ..
In-Reply-To: <41D9EA44.3020504@fastcrypt.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox