Hi all,
here's a progress report on our slow Alpha 2100, for any of you that are
interested.
With the help of Chris Jankowski of DEC (who is on this list, THANK YOU!)
it appears that our alpha is being trashed by a rapidly forking process
that forks and dies, forks and dies. Possibly spawned by init (but I
cannot find any evidence of what process yet).
What we are seeing, in summary:
* High kernel time on the system, ie: 80%+ of CPU processing during peak times
is kernel based.
* High turnover of process ID's, here is an example:
root 26287 0.0 0.0 1.33M 168K ttyr9 S <+ 22:25:39 0:00.03 grep gre
Mon Aug 21 22:25:53 EST 1995
root 25953 0.0 0.0 1.33M 168K ttyr9 S <+ 22:25:53 0:00.02 grep gre
That's 31,666 processes in 14 seconds !
Interestingly an upgrade of the firmware from version 3.9.something to
4.0 has made the load a little better. OS is still V3.2a.
I am going to pass the matter on to Digital support tomorrow.
(When the system boots, the problem is NOT there, however it surfaced
within the first few hours of uptime. This would lead one to believe
it was a user process, but I am not so sure as after shutting down the
system to single user mode, the same problem kept occurring with the
only visable processes running being the shell, the kernel and init).
If I find an answer with DEC I'll let you all know.
-richard
--
Richard A. Muirden, Sys. Admin |Fan of Shostakovich, "Star Trek" and BOEING!
Mailto: richard_at_rmit.EDU.AU |Fly the Friendly Skies of United Airlines !
Phone: (+61 3) 9660 3814 |I created alt.fan.shostakovich! Fly: UA,AN,WN
http://www.rmit.edu.au/richard |Can *YOU* beat my 114 Shost CD's? :-)
Jolt Can Tower Counter: 26 cans
* 1995: Remembering 20 years since the death of Shostakovich (1906-75) *
Received on Mon Aug 21 1995 - 14:45:08 NZST