RE: [squid-users] Sizing a pretty Large Squid Install

> > so the actual peak is higher). Hardware specs:
> > dual PIII/450
> > 1 Gb RAM
> > 6 9.1Gb (not sure about RPM, might even be 7.2k)
> Ultra2-wide spindles,
> > 1 for system, 5 for diskd-based (10 cache_dirs) cache on ext2
>
> you do 2 cache_dirs per drive????

Yes.
The idea is to exploit (if possible) elevator optimization in moving
the drive heads, which requires having multiple outstanding
requests on the same drive. Diskd only has one active disk operation
at any given moment AFAIK.
Probably mostly irrelevant under normal conditions, but it can help
decrease consistency checks and crash recovering.

As you could have noticed from the text you didn't quote,
the squid version I'm using is 2.3DEVEL4-NTLM, which is broken in
a few ways (thus the need to recover from time to time), but
I NEED the NTLM and 2.5 is not yet ready - currently
my test setup crashes after (on average) .5M hits - it wouldn't
last 40 minutes under my load. Robert and I have been chasing
this bug for way too long now, and still haven't got a solution.

-- 
	/kinkie