Re: Mount database on RAM disk? - Mailing list pgsql-performance

From Christopher Browne
Subject Re: Mount database on RAM disk?
Date
Msg-id m33bqnl0qa.fsf@mobile.int.cbbrowne.com
Whole thread Raw
In response to Re: Mount database on RAM disk?  ("Merlin Moncure" <merlin.moncure@rcsonline.com>)
List pgsql-performance
> On 8 Jul 2005, at 20:21, Merlin Moncure wrote:
>> ditto windows.
>>
>> Files cached in memory are slower than reading straight from memory
>> but not nearly enough to justify reserving memory for your use.  In
>> other words, your O/S is a machine with years and years of
>> engineering designed best how to dole memory out to caching and
>> various processes.  Why second guess it?
>
> Because sometimes it gets it wrong. The most brutal method is
> occasionally the most desirable. Even if it not the "right" way to do
> it.

The fact that cache allows reads to come from memory means that for
read-oriented activity, you're generally going to be better off
leaving RAM as "plain ordinary system memory" so that it can
automatically be drawn into service as cache.

Thus, the main reason to consider using a RAM-disk is the fact that
update times are negligible as there is not the latency of a
round-trip to the disk.

That would encourage its use for write-heavy tables, with the STRONG
caveat that a power outage could readily destroy the database :-(.
--
let name="cbbrowne" and tld="acm.org" in String.concat "@" [name;tld];;
http://cbbrowne.com/info/rdbms.html
Rules of  the Evil  Overlord #153.  "My Legions of  Terror will  be an
equal-opportunity employer. Conversely, when  it is prophesied that no
man  can defeat  me, I  will  keep in  mind the  increasing number  of
non-traditional gender roles." <http://www.eviloverlord.com/>

pgsql-performance by date:

Previous
From: Alex Stapleton
Date:
Subject: Re: Mount database on RAM disk?
Next
From: Simon Riggs
Date:
Subject: Re: cost-based vacuum