Issues with top

[Knabe, Troy]

Both report load averages at 15+.  But top is still reporting 99.7% idle on the cpus. (there are 2).  The server is responding and everything else points to no load on the box.  Any thoughts on how to diagnose why the high load average?

Thanks


On 6/26/08 9:04 AM, "Andrew Bacchi" <bacchi at rpi.edu> wrote:

Top is only reporting what it sees, it is not the cause.  I sometimes
see behavior as you are reporting.  Often it's due to either a bad
process, poorly written code, or a possible attack to the kernel, ddos.
I've watched CPU0 at 100%, then it will move to CPU1 at 100%, and leave
CPU0 at < 1%, and back again.   I used to see it happen on a test system
after the M programmers wrote a new function that went haywire.

I've never done a search to find the cause on my Linux boxes, since it
usually only lasts for a few minutes.  It does however wreck havoc on my
systems.

Do a 'cat /proc/loadavg' to see the current dynamic load average. or try
'uptime' that too reports load average.  It may be normal for an Oracle
server to have a load this high.  Also look into 'sar' data to get a
history of your machine.

Knabe, Troy wrote:
> I have a development box that has Oracle Application Server on it.  Yesterday we received a complaint that it was "slow" and top showed 0% idle and load averages of 15+ across all three settings.  Then after a minute or two cpu went back to 99% idle but even today the load averages show 15+, but the CPU continues to be 99% idle.  There does not appear to be i/o wait, and after the minute or two yesterday performance of the application is as expected.  Anyone else seen issues with top like this before?
>
> Thanks
> -Troy
>

--
veritatas simplex oratio est
        -Seneca

Andrew Bacchi
Systems Programmer
Rensselaer Polytechnic Institute
phone: 518.276.6415  fax: 518.276.2809

http://www.rpi.edu/~bacchi/

--
redhat-list mailing list
unsubscribe mailto:redhat-list-request at redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list