Re: [SQU] multiple spindles

From: Robert Collins <[email protected]>
Date: Sat, 30 Dec 2000 12:24:06 +1100

RAID 5 is bad for write intensive tasks. It requires 4 disk io's per write (and that's with the best implementation). While squid
doesn't automatically detect dead cache_dirs at the moment, I believe that the time to restart squid after commenting out a
cache_dir entry is more than paid for by the performance gain of using non RAID 5 disks.

striping without parity may work as fast as separate disks, but you lose the resilience of multiple cache dirs in you do that... so
I'd always run with multiple cache_dirs.

Rob

----- Original Message -----
From: "Florin Andrei" <florin125@hotmail.com>
To: <squid-users@ircache.net>
Sent: Saturday, December 30, 2000 8:54 AM
Subject: [SQU] multiple spindles

> I think i saw somewhere in the docs (but i'm not 100% sure) that it is
> better to have several cache_dir's on multiple independent HDDs, than just
> one cache_dir on multiple stripped (RAID ...) HDDs. Because in the past, i
> really saw corrupted cache partitions, i tend to agree to that.
> But there are several people that say that if you're running Squid on Irix
> and you use XFS (or some other commercial Unix with a good filesystem),
> there's no need to put multiple cache_dir's, so just let's RAID all those
> HDDs together under a single cache_dir. They say that XFS is too reliable to
> crash.
> Now, i really don't think that XFS is so crash-proof, and i believe that
> there are some other advantages of using multiple cache_dirs. Can someone
> enlighten me on this?
>
> Happy New Year everyone!
>
> --
> Florin Andrei
>
> _________________________________________________________________
> Get your FREE download of MSN Explorer at http://explorer.msn.com
>
> --
> To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
>
>

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
Received on Fri Dec 29 2000 - 18:16:11 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:57:09 MST