After monitoring our newly installed DS25 Version 5.1b (with patches),
once every couple of days, the load averages climb above 12 which cause
a sendmail rejecting connections message (default value in the sendmail
config file). After collecting several days of data (collect utility),
and graphing it with collgui, there is no apparent miss-behaving
process, excessive disk use, lack of memory, or high network activity
directly related with the spike in load averages. These spikes last for
up to 5 hours and then magically return to a normal levels (1.5 - 3.5
load averages) within an hour or so. CPU idle time is normal (75 - 90%)
during these load average spikes. There is nothing extra running in our
cron tabs nor are there any significant events happening with our
application. I am running out of things to monitor which might help
explain what is causing the high load averages. Is there actually a
performance problem somewhere or is the load average values being
miss-interrupted?
Any help would be appreciated.
John
Received on Wed Feb 23 2005 - 23:32:07 NZDT