xfs
[Top] [All Lists]

Re: productionserver

Subject: Re: productionserver
From: "Eric Sandeen" <sandeen@xxxxxxx>
Date: Thu, 23 Aug 2001 20:49:36 -0500
Cc: "Linux XFS Mailing List" <linux-xfs@xxxxxxxxxxx>
References: <Pine.BSI.4.10.10108232336290.21190-100000@xxxxxxxxxxxxx> <002f01c12c25$adceae20$50824e40@xxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
Steve managed to hit a spam filter somehow, so forwarding his message
back out...

--- begin Steve's message ---

> Hmm yes, however our disk lights are lit a lot. A saw my co worker test
> run a testconversion from guilders to euro in our logistics package
> kept the disk lit for over 40 minutes on the testbox.
>
> Now this onvolves actually ploughing trough the complete 6GB+ of the
> databases and changing fields and values. We are talking milions of
> records that needed to be changed. Afterall this was just a desktop
> machine but still this is going faster then the now in production
machine.

  I forgot to mention - the slowdown when updating the entire database
may
not be indicative of real-world slowdowns, since you certainly can't
cache
the entire 6 GB.  But in the real world, are you really doing things
that
affect that much data?  If you're not, then the disk cache much more
effectively makes up for the overhead of RAID 5.  If you are doing
things
like that in production, never mind. : )

steve

--- end Steve's message ---


<Prev in Thread] Current Thread [Next in Thread>