System resource useage abnormalities

Tracy R Reed (treed@dt6h1n44.san.rr.com)
Thu, 1 May 1997 00:27:21 -0700 (PDT)

Qmail was running just fine on my P133, but that machine had to go away
and I installed qmail and SmartList on my 266Mhz UDB Alpha. The kernel is
2.0.27 and distrib RedHat 4.1. This machine isn't one of those great
powerhouse Alpha's, but it should be a little better than a P133. Instead,
I am finding that the load average is unusually high and the system
somewhat slow in delivering mail for no good reason. Here's a typical view
from top.

12:01pm up 2 days, 20:34, 8 users, load average: 4.44, 3.27, 2.50
128 processes: 124 sleeping, 1 running, 2 zombie, 1 stopped
CPU states: 9.7% user, 27.8% system, 31.7% nice, 65.0% idle
Mem: 86080K av, 84848K used, 1232K free, 76352K shrd, 10032K buff
Swap: 0K av, 0K used, 0K free 16296K cached

PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
26607 treed 12 0 1160 1160 832 R 1064 9.5 0.6 0:06 top
26913 list 11 0 1176 1176 920 S 784 7.6 0.6 0:01 procbounce
9124 qmailr 2 0 544 544 408 S 504 2.2 0.3 4:21 qmail-rspawn
9122 qmails 1 0 768 768 496 S 680 1.7 0.4 4:11 qmail-send
27034 list 8 0 1136 1136 1024 S 760 1.5 0.6 0:00 subscribe
17946 treed 2 0 2360 2360 1928 S 2288 1.3 1.3 1:37 xsysinfo
374 treed 1 0 928 928 528 S 864 1.2 0.5 42:48 xsysinfo
1 root 0 0 112 112 56 S 24 0.6 0.0 58:11 init
166 root 1 0 10504 10504 1296 S 9224 0.6 6.1 42:28 X
26945 qmailr 0 0 664 664 528 S 592 0.5 0.3 0:00 qmail-remote
26934 qmailr 0 0 664 664 528 S 592 0.5 0.3 0:00 qmail-remote
26944 qmailr 1 0 664 664 528 S 592 0.5 0.3 0:00 qmail-remote
26985 qmailr 1 0 664 664 528 S 592 0.5 0.3 0:00 qmail-remote
27036 qmailr 7 0 672 672 528 S 600 0.5 0.3 0:00 qmail-remote
27000 qmailr 4 0 664 664 528 S 592 0.4 0.3 0:00 qmail-remote
26956 qmailr 0 0 664 664 528 S 592 0.4 0.3 0:00 qmail-remote
27001 qmailr 2 0 672 672 528 S 600 0.4 0.3 0:00 qmail-remote

And here's vmstat output:

[treed@freeside pilot-digest]$ vmstat 10
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
1 0 0 0 2296 10512 16248 1 0 2 7 5 11 15 0 0
1 1 0 0 2192 10512 16248 0 0 0 19 65 132 39 48 13
1 0 0 0 1848 10576 16264 0 0 0 14 57 116 39 50 11
1 0 0 0 2088 10576 16224 0 0 0 14 63 118 44 51 5
3 0 0 0 2600 10608 16224 0 0 0 16 58 125 37 53 11

A very dead system. Nothing unusual happening. Minimal interrupts/second
and context switching. But there's a high load average with 60% cpu idle
time. The top just didn't happen to catch the cpu while waiting on disk,
it's averaging 60%. I'm sending lots of mail through the mailing list.
It's on cablemodem now as opposed to T1 when I had the list on the P133.
Not sure how this would cause such a high load with low cpu useage.

Anyone have any ideas?

----------
Tracy Reed
http://www.ultraviolet.org
"There are no significant bugs in our released software that any
significant number of users want fixed." -Bill Gates