On Mon, Jan 07, 2002, Duane Wessels wrote:
> > The C code is written such that Q1(64) < Q2(72) < SHM limit(96), and is
> > also how it is documented in squid.conf.default as far as I can tell.
>
>
> No, its was stupid on my part I guess, but it should be Q2 < Q1 < SHM limit.
>
> I've been doing some benchmarks and banging my head on this for a couple of
> days. With Q1 < Q2 (as in the comments and defaults) really bad things
> happen. Squid hits the Q1 limit and hit ratio goes down, response time
> goes WAY up and doesn't come back down for a long time.
Remember that I was highly, highly confused when I moved the diskd
code into the New World Order(tm). I'm not surprised that things got
confused.
Once I've done all of the commloops code, I'll look at playing around
with some auto IO load balancing ideas I've had. I think this is going to
be the best way to do things - set a maximum number of slots/threads,
and then use some latency/queuelength feedback to find out when its
time to "skip" open/create events.
Adrian
Received on Sat Jan 12 2002 - 03:54:00 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:14:45 MST