Re: [squid-users] Moving to squid from inktomi

On Thursday 27 June 2002 07:13 pm, Marshall wrote:
> > If disk becomes your bottleneck, you're toast. Disk cache is king at
> > these speeds. Consider going with 1 or 2 IDE disks and sink the
> > savings into more RAM.
>
> I guess you are saying go heavy on RAM no matter what, and buy more
> disks if possible? Is there any perf hit to having a cache partition on
> the same disk as OS and swap?

The system won't actively swap and the needed portions of the OS will be
in disk cache, so they won't affect the cache_dir.

> Another scenario might be to add a third disk and keep the RAM at 1GB.
> But I guess it seems ridiculous to have 54GB of storage for 2 to 3 GB of
> cacheable objects. Unless seek time is really the end of the road.
> What do you think?

I think at 1gb you will serve almost entirely out of RAM and at 2gb, you
most certainly will. Past the ramp up stage, nearly all of your physical
disk IO will be writing logs, since there will be few new objects to write
out.

(Piles of disks are common around here because forward proxies can't
possibly hold their whole working set in RAM. They need to fall back to
disk. You don't.)

As Joe said, once you're serving entirely from RAM, CPU and bandwidth are
the only bottlenecks left. For byte shoveling, I highly recommend the
3com NICs over Intel's -- they use about half the CPU to do the same job.

        -- Brian
Received on Thu Jun 27 2002 - 18:32:25 MDT