actionsvr has to be performing, or trying to perform, an action, that you
told it to do in a ruleset. nettl won't be any help, but the nvcorrd logs
with nvcedug -d all should show you a timestamp of when the action was
passed, even though it may not tell you what it was. The nvactiond logs
should indicate at least the actions which complete. So by process of
elimination you should be able to figure out which one is the one that it
still out there, right? And then you could work backward to the ruleset
which caused it. But you have to be willing to watch for the problem and
work at solving it. I don't think it will be easy.
The only other suggestion I can think of is first do ps -ef | grep
actionsvr to find his process id, and then do
dbx -a <process id>
When you get the (dbx) prompt back type in the word where and you'll
get a stack trace of what actionsvr is doing.
When you are done issue detach at the (dbx) prompt. Otherwise you will
kill the process (which you may want to do anyway) if you exit of ctrl-C
out.
Maybe that will help you figure it out.
Otherwise, as always, please call Support.
James Shanks
Tivoli (NetView for UNIX) L3 Support
ADAMCZYK Herbert <herbert.adamczyk@CAIT.CO.AT> on 08/25/98 04:13:08 PM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView et alia <NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: James Shanks)
Subject: actionsvr Performance problem
sometimes our actionsvr process (child process of the actionsvr daemon)
is consuming too much cpu-time for a long period. I tried to check out
the reason, but neither the nvcdebug nor the nettl-trace nor the
nvactiond.*logs could tell me whats going wrong. Does anyone know a
solution to figure out what actionsvr is doing ?
thanx
Herbert Adamczyk
iT-AUSTRIA
Network-Management-Projects
Vienna
|