Jan 012010
 

This will come in handy in case you ever come across this issue. I am seeing a strange occurrence in whm ‘Main >> System Health >> Show Current CPU Usage’ and in the base OS via the command line. (I know whm pulls cpu info from the OS via script) When I run a ‘ps aux’ from the command line I see the cpu usage is 58950531.5 but the ‘ps auxf’ and ‘top’ commands shows normal accurate cpu usage.

ps auxf
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 2 0.0 0.0 0 0 ? S< 2009 0:00 [kthreadd]


ps aux
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 58950531.5 0.0 2872 544 ? S 01:39 0:00 init [3]

Oh, and all of the %cpu usage for every process says 58950531.5 when ps aux is used, ps auxf is normal

I have tried updating the kernel also with no joy
Linux host.domain.net 2.6.9-89.0.9.ELsmp #1 SMP Mon Aug 24 07:56:18 EDT 2009 i686 athlon i386 GNU/Linux
was at 2.6.9-67.0.15.ELsmp.

In researching this error, I did not see any other reference anywhere. I have repaired this issue by renaming the existing ps binary on the affected server to ps.bak and copying (scp) a pristine ps binary from another server to the affected server. On the server with the broken ps binary;

mv /bin/ps /bin/ps.bak

next,

scp -r [email protected]:/bin/ps /bin/ps

scp the file from the working server to the broken server. This will copy a working ps binary to the broken server and repair the issue.

Here is a working copy of the ps file in case you do not have access to another server; ***FILE***

To get the file from here to your server

mv /bin/ps /bin/ps.bak
cd /tmp/
mkdir ps/
cd ps/
wget http://g33kinfo.com/info/wp-content/uploads/2010/01/ps1.zip
unzip ps1.zip
cp ps /bin/ps

Test the file;

ps aux

Bingo, working ps aux!

 Posted by at 5:02 am

 Leave a Reply