Re: [squid-users] Squid seem to be slower than our Novell BorderManger

From: Mathew Thomas <[email protected]>
Date: Fri, 12 Sep 2003 14:14:20 +1000

Hi Adam,

I am not raiding the cache disk - mounting as six cache disk. I can't see any bottleneck

[root@sproxy1 ~]$ vmstat 10
   procs memory swap io system cpu
 r b w swpd free buff cache si so bi bo in cs us sy id
 0 0 0 0 7805568 90572 19360 0 0 4 5 27 2 0 0 100
 0 0 0 0 7805540 90596 19360 0 0 0 4 101 7 0 0 100
 0 0 0 0 7805496 90596 19368 0 0 0 0 125 17 0 0 100
 0 0 0 0 7805264 90612 19380 0 0 0 12 175 50 0 0 100
 0 0 0 0 7804996 90748 19388 0 0 0 35 187 49 0 0 100
 0 0 0 0 7804584 90804 19412 0 0 0 18 243 89 0 1 98
 0 0 0 0 7804268 90976 19436 0 0 0 35 181 58 0 0 100
 0 0 0 0 7803964 91084 19460 0 0 0 36 156 46 0 0 100

Thank you for your help
Mathew

>>> Adam Aube <aaube@firstindependent.net> 12/09/03 11:28:40 >>>
>I used top and the system was not bottlenecking.

Top only checks CPU usage - and Squid (like most servers on modern
hardware) will usually bottleneck on CPU last.

I suggested procinfo, vmstat, and sar because they give you info
on memory, disk, and CPU - and Squid will tend to bottleneck on disk
before anything else.

Also, you mentioned that the Squid server had 6 36-GB SCSI disks
- are you using any sort of RAID?

Adam
Received on Thu Sep 11 2003 - 22:14:46 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:19:40 MST