100% CPU Usage running linux commands

[KarloChacon]

hi 

the application guy told me that his server was executing linux commands with some delay...let's say 1-2-3 seconds after the command is execute the output is shown 

checking the system we found that when the delay took place CPU usage for that command in top was 100%.... this happens for bash, chkconfig, w, df, du among others in fact when you start working with command load average is normal and after executing commands load average is the double amount in CPU Usage

has someone seen this before? a know issue?
something to check?


important to note is this the Red Hat Release is 5.4 Beta which I think is not good but the system was working fine until last week
by the way is it recommended to upgrade from Red Hat 5.4 Beta to 5.4 final release? or even 5.5?

I mean going from beta to a final release is OK?

thanks a lot guys

[Matti_Kurkela]

Check the health of the server's disks ASAP. One possible reason for that kind of behavior is that the system disk is failing but still manages to provide the requested information after a number of retries.

(If the system disk is mirrored but the health of the mirror is not monitored, one half of the mirror might already have completely failed, and the other half might now be failing too...)

Things to check:

- dmesg (any disk errors or other serious-looking things visible?)

- any hardware RAID status, if you have them (usually hardware-type-specific utility, e.g. for HP SmartArray RAID controllers, "hpacucli ctrl all show config detail"  would list all information on every SmartArray controller installed on the system. You would check the "Status:" fields on all listed logical & physical drives.)

- if you have NFS-mounted filesystems, are the NFS servers overloaded? (If your PATH environment variable includes NFS-mounted directories, and the NFS server is overloaded or down, you might see slowdown with many commands.)

[KarloChacon]

I don't see anything related with disks on dmesg 

space on disks look ok

Filesystem                                      Size  Used Avail Use% Mounted on

/dev/mapper/VolGroup00-LogVol00  28G   3.4G  23G 13%   /
/dev/sda1                                       99M  13M   82M 14%   /boot
tmpfs                                             3.9G  0       3.9G 0%    /dev/shm

/dev/mapper/VolGroup00-LogVol02  40G   34G   3.7G 91%  /mxapps

see  dmesg attached

and no everything is local disks no NFS I will have tro send someone to check any fail status on disk, this is a IBM server I and I think dont hace the RAID utilities

thanks a lot 

[Alzhy]

post output of:

sar 5 5

sar -d 5 5

vmstat 5 5

free