Package: procps
Version: 1:3.2.7-3
Severity: normal
We're seeing some strange output from top, with both the Debian packaged
version 3.2.7 and the latest 3.8beta1.
Here are the first few lines of the output of top -b:
last pid: 25676; load avg: 0.06, 0.02, 0.00; up 55+13:53:25 07:15:12
75 processes: 1 running, 74 sleeping
CPU states: 0.5% user, 0.0% nice, 0.0% system, 99.5% idle, 0.0% iowait
Kernel: 1251 ctxsw, 578 intr
Memory: 6300M used, 1672M free, 446M buffers, 3561M cached
Swap: 644K used, 1960M free, 8K cached
PID USERNAME THR PRI NICE SIZE RES SHR STATE TIME CPU COMMAND
14357 zope 5 20 0 2186M 2036M 4696K sleep 7:59 20.00% python2.4
188 root 1 15 -5 0K 0K 0K sleep 3:10 0.00% kswapd0
20203 root 1 16 -4 10M 696K 324K sleep 2:13 0.00% udevd
10 root 1 15 -5 0K 0K 0K sleep 0:48 0.00% events/1
9 root 1 15 -5 0K 0K 0K sleep 0:46 0.00% events/0
[this is top 3.8b1 but we see the same numbers in 3.2.7]
The system this dump was taken from is a dual-core Athlon system that was
under heavy load on a single thread. There are three things that are weird
here:
1. the overall CPU states show the system being 99.5% idle, and yet the top
process consumes 20% of CPU. That surely is just wrong ;)
2. even under very heavy load, the process rarely consumes more than 20%
CPU, when I would expect to see it use 100% of one core, if it was CPU
bound. But if it's IO bound I would expect to see more iowait, hence...
3. the iowait is shown at 0.0%, which is very odd.
We have over 20 servers with this identical spec and OS build and all show
the same behaviour here.
We've spent the last couple of days exhaustively testing this particular
application because it was showing very poor performance. Our conclusion
is that the process is suffering from high iowait. Even under heavy
sustained load the process rarely shows greater than 20% CPU consumed.
During these periods when it shows this behaviour, strace shows the process
is spending all of it's time in repeated calls to recvfrom(2) - from which
I would expect to see a high iowait. Yet iowait is never reported at
higher than 1-2%.
So our conclusion is that top and vmstat are both lying to us about
iowait!
-- System Information:
Debian Release: 4.0
APT prefers proposed-updates
APT policy: (500, 'proposed-updates'), (500, 'stable')
Architecture: amd64 (x86_64)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.24-etchnhalf.1-amd64
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
Versions of packages procps depends on:
ii libc6 2.3.6.ds1-13etch8 GNU C Library: Shared libraries
ii libncurses5 5.5-5 Shared libraries for terminal hand
ii lsb-base 3.1-23.2etch1 Linux Standard Base 3.1 init scrip
Versions of packages procps recommends:
pn psmisc <none> (no description available)
-- no debconf information
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
