Re: [squid-users] Squid overloading when RAID drive cache in use?

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Thu, 10 Jul 2003 17:43:47 +0200

On Thursday 10 July 2003 16.17, Peter Smith wrote:

> I am wondering if having cache_dir drives on a RAID controller that
> has Read/Write cache turned on might cause problems?

Depends on the RAID level and the load you plan on putting on the
RAID.

> I'm fairly sure that Squid manages the latency, etc of its
> cache_dir drives.

Nope, but if a lot of requests get queued for the drive then Squid
backs off (applies to aufs and diskd drivers only, not ufs).

> The drives that my Squids use are all on RAID controllers as single
> volumes. However I recently found that if I enable Read/Write
> cache on the cache_dir drives that load on the processor goes off
> the scale.

Do this RAID controller have embedded CPU for the buffer management
etc, or do it use the main CPU?

> Could it be that Squid gets such a quick response from
> the drive that it thinks the drive is super fast and thus slams it,
> causing it to run out of Read/Write cache and then gets overloaded
> as the requests backlog?

Not very likely.

But on the other hand, if your Squid previously was throttled by the
disk I/O due to the lack of buffers then enabling the cache may speed
things up, allowing Squid to catch up on the traffic and thus use
more CPU.

-- 
Donations welcome if you consider my Free Squid support helpful.
https://www.paypal.com/xclick/business=hno%40squid-cache.org
If you need commercial Squid support or cost effective Squid or
firewall appliances please refer to MARA Systems AB, Sweden
http://www.marasystems.com/, info@marasystems.com
Received on Thu Jul 10 2003 - 09:44:01 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:17:55 MST