Re: P200 + 128M + 1.1.15 + Network Appliance = SLOOOOOOOOOOWWWWWWWW

From: Kyle R VanderBeek <>
Date: Mon, 25 Aug 1997 10:29:41 -0400 (EDT)

The only thing I'd be concerned about in this capacity is that, last I
checked, Linux's NFS implementation runs in user space and is consequently
slower than many others. I'd be willing to bet that this is why your dual
PPro system is running so much better than your single Pentium systems.
Check load averages/top and see just who is sucking the most power: is it
squid or nfsd?

        Kyle VanderBeek
        Systems Engineer/Webmaster/Postmaster Calvin College
             I'd love to, but the monsters haven't turned blue yet,
                          and I have to eat more dots.

On Sun, 24 Aug 1997, Jason Lixfeld wrote:

> Any ideas?! It's just the basic install.. I have 2 other server running
> on this same netapp. One is a Dual PPro 200, and the other two are
> P200MMX's all systems have 128MB EDO RAM. The Dual PPro has 256M.
> I'm not doing anything out of the ordinary.. I have told Squid to use 7
> of 29Gigs on the Network Appliance via NFS, and all 3 servers are reading
> and writing that 7Gig spool via NFS. The Dual PPro FLIES.. but the two
> regular Pentium 200's crawl, and I have no ideas why. IDENTICAL
> squid.conf files EXCEPT the Dual PPro used 128M of mem and the Pentium
> 200's use 64 (I tried cutting that number back but it didn't help).
> BTW: All of the machines are running RedHat 2.0.30.
> I have 5 other proxy machines that are like P100's with 64M each writing
> to their own local 2Gig spool, but I want to try to utilize this NetApp..
> HELP!! =-)
Received on Mon Aug 25 1997 - 07:31:25 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:36:49 MST