What is going on with my system?

From: David Benfell <benfell_at_parts-unknown.org>
Date: Sun, 16 Dec 2007 20:43:02 -0800
Hello all,

My system, running 7.0BETA4, is periodically responding extremely
slowly.  Rebooting seems to resolve the problem, but I also had
to deinstall console-kit-daemon because it was, more often than
not, provoking a kernel page fault during start-up.

You can find all the information I can think of to describe the
system and its troubles at:

http://www.parts-unknown.org/systems/lupin.cybernude.org/

Unfortunately, this web server is the afflicted system, so you
might have trouble getting the information if it happens to be
going through this.  But the last time, I did manage to record
the top output:

last pid: 86007;  load averages: 65.16, 59.96, 52.63    up 0+09:17:16  19:53:46
150 processes: 15 running, 134 sleeping, 1 zombie
CPU states:  0.3% user,  0.3% nice, 97.5% system,  1.9% interrupt,  0.0% idle
Mem: 340M Active, 602M Inact, 227M Wired, 32M Cache, 112M Buf, 298M Free
Swap: 17G Total, 16K Used, 17G Free

  PID USERNAME       THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
 2206 benfell          3   4   19 39844K 24728K fifoor 119:25 3173.58% trackerd
85943 benfell          1   8   20 19112K 16440K wait     0:01  6.93% perl5.8.8
 1205 qmails           1 128    0  3072K  1236K RUN      7:32  2.78% qmail-send
77799 cvsupin          1 130   10  6376K  5780K RUN      0:43  1.03% cvsup
85991 benfell          1  -8   20  5904K  3504K piperd   0:00  0.39% perl5.8.8
 1273 Gdnscache        1  96    0  4096K  2212K RUN      0:43  0.10% dnscache
 2381 root             1 137   10  3500K  1892K RUN      1:54  0.00% zsh
 1208 qmaill           1  -8    0  3072K   656K piperd   0:32  0.00% multilog
 1272 Gdnslog          1  -8    0  3072K   656K piperd   0:20  0.00% multilog
 1215 root             1 131    0  3072K   656K select   0:14  0.00% qmail-lspa
 1277 haldaemon        1  96    0 19484K  4576K select   0:10  0.00% hald
 2382 root             1  -8   10   144M   142M pipewr   0:10  0.00% find
 1577 root             1  96    0  3208K   848K select   0:10  0.00% moused
 1217 qmailq           1  -8    0  3072K   632K piperd   0:07  0.00% qmail-clea
 1216 qmailr           1  96    0  3072K   764K select   0:06  0.00% qmail-rspa
 1207 qmaild           1  20    0  2944K  1208K pause    0:06  0.00% tcpserver
 1224 Gtinydns         1   4    0  2480K   716K sbwait   0:05  0.00% tinydns
 1206 qmaill           1  -8    0  3072K   656K piperd   0:03  0.00% multilog

As you can see, there are a number of troubling indications.  First, the load
averages are off the scale (and I've seen them higher).  Second, the CPU state
is 97.5% system.  Third, check out the "WCPU" figure for trackerd.

The WCPU figure is similar to that I previously saw with javavm.  I managed to
make that problem go away by not allowing firefox to run java.

If I'm going to do any debugging here, you're going to have to treat me like an
idiot.  I'm in way over my head here.  The only reason I went to this version
was to try to correct problems interacting with my Treo (it didn't help) and I'm
really looking forward to 7.0-RELEASE.

Thanks!

-- 
David Benfell, LCP
benfell_at_parts-unknown.org
---
Resume available at http://www.parts-unknown.org/
NOTE: I sign all messages with GnuPG (0DD1D1E3).

Received on Mon Dec 17 2007 - 12:35:02 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:24 UTC