Extremely High Load Average problem

Extremely High Load Average problem

Post by Blas Pedro Uberuag » Sat, 10 Jul 1999 04:00:00



Hi Everyone,

I am running AIX 4.3.2 on an IBM 590 that exports several disks to
roughly 50 machines, most of them dual processor 43P-260s.  I am seeing
problems with the load of the server.  At times, it gets as high as
100.  When this happens, nfsd seems to be the process dominating the cpu
time.  I've tried everything I can think of to determine the problem
here and fix it, but so far, nothing has helped.  If anyone has seen
similar problems and has any suggestions, I would greatly appreciate
them.

Thank you.

 
 
 

1. Odd OpenSolaris extremely high load average bug

So far, I have tested OpenSolaris (Community release based on nv16) and
Schillix under their own separate VMware Workstation 5.0 virtual machines.

Oddity: when system finishes booting and logins becomes available, I do
'uptime' and see load averages between 1500 to 2500 for one minute, a
bit lower for five minutes, and lower for fifteen minutes average.

System does not feel slow. vmstat 1 shows idle almost 100%, and ps agrees.

So clearly this is a bug. I don't know enough about kmdb/mdb yet to
better troubleshoot it, though I am reading usr/src/uts/common/os/clock.c.

I am curious if anyone else has seen this bug on 'real' hardware (not
within VMware), or if they have seen it under their own VMware virtual
machine.

I just checked -- I also see this bug with Solaris 10/x86 FCS in VMware, too.

-Dan

2. via82cxxx_audio not recording....

3. System Load average: How much load average is good for?

4. libxlf90 files

5. Proftpd - High Load averages and strange performance problems.

6. tun1 not working

7. load average on WebServer is high

8. MS-LINUX

9. Surprisingly high load averages on an NFS server

10. Average load very high

11. SVR4 machine, HIGH load average

12. High load average, but why???

13. High load average with multithreaded server