No, Gord, they should stop. You are right that is how it is designed. But when
you restart them all, who had the problem?
It is possible that actionsvr hung first and that nvcorrd hung in response
trying to send him a message and wait for a response. Do you get the drift?
It does not follow that because you restarted nvcorrd, that he is the one who is
the source of the problem. For that matter, nvcorrd could have hung because
trapd quit talking to him, and the connection gets re-established upon start up.
There is no simple way to know unless you investigate with the help of someone
who understands how it all works together.
James Shanks
Tivoli (NetView for UNIX) L3 Support
Gord Michaels <gord_michaels@HOTMAIL.COM> on 08/06/99 10:42:55 AM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on NetView
<NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: James Shanks/Tivoli Systems)
Subject: Re: Events stop coming into Events Display...
Thanks James. Your response is helpfull.
But just one thing to claify, are you saying that when I ovstop nvcorrd, the
actionsvr daemon and the nvserverd daemon should NOT also stop?? I thought
that they were dependent on nvcorrd and this behaviour would be natural??
Gord.
>From: James Shanks <James_Shanks@TIVOLI.COM>
>Reply-To: Discussion of IBM NetView and POLYCENTER Manager on NetView
> <NV-L@UCSBVM.ucsb.edu>
>To: NV-L@UCSBVM.ucsb.edu
>Subject: Re: Events stop coming into Events Display...
>Date: Fri, 6 Aug 1999 09:59:37 -0400
>
>There is no damage if that is what you mean. But stopping nvcorrd means
>that
>his in-memory cache of recent events is destroyed. If you left the event
>windows up while you did this, then you perhaps won't notice anything. But
>if
>you start a new window after that it won't even show the last event you had
>before the restart. Just something to be aware of. Each restart of
>nvcorrd
>means that the event cache gets rebuilt and old ones are gone as far as the
>display is concerned. Also, if you were running any pass-on-match or
>reset-on-match rulesets then they no longer have any old events to match
>with.
>It all starts over.
>
>I would do a netstat -a, and a netstat -an, before you restart nvcorrd next
>time
>since Support will want to see if there is a socket holdup. And I would
>turn on
>the nvcorrd trace, nvcdebug -d all, while you wait, so there is some record
>of
>what nvcorrd was doing when he hung. And finally, I think Level 2 will
>probably
>recommend a dbxprof (a utility they provide) to get a call stack of nvcorrd
>when
>it is hung. That's assuming that nvcorrd is the real problem. Notice that
>when
>you stopped him, both nvserverd and actionsvr stopped as well. That's why
>someone has to look at this.
>
>And it may still turn out to be an AIX socket problem.
>
>
>
>James Shanks
>Tivoli (NetView for UNIX) L3 Support
>
>
>
>Gord Michaels <gord_michaels@HOTMAIL.COM> on 08/06/99 09:37:17 AM
>
>Please respond to Discussion of IBM NetView and POLYCENTER Manager on
>NetView
> <NV-L@UCSBVM.UCSB.EDU>
>
>To: NV-L@UCSBVM.UCSB.EDU
>cc: (bcc: James Shanks/Tivoli Systems)
>Subject: Re: Events stop coming into Events Display...
>
>
>
>
>
>Hello James.
>
>I am using AIX 4.2.1 and Netview 5.1.1.
>I will contact support.
>Right now, when the problem occurs, I do the following:
>
>1. ovstop nvcorrd - this will also stop actionsvr and nvserverd.
>2. ovstart nvcorrd - will start nvcorrd but may not restart actionsvr
> and nvserverd.
>3. ovstart - this will restart actionsvr and nvserverd.
>
>This seems to temporarily solve the problem without shutting down the
>Netview Server completely.
>
>Do you see any potential problems with this temporary procedure?
>
>Sincerely,
>
>Gord Michaels.
>
> >From: James Shanks <James_Shanks@TIVOLI.COM>
> >Reply-To: Discussion of IBM NetView and POLYCENTER Manager on NetView
> > <NV-L@UCSBVM.ucsb.edu>
> >To: NV-L@UCSBVM.ucsb.edu
> >Subject: Re: Events stop coming into Events Display...
> >Date: Fri, 6 Aug 1999 09:20:55 -0400
> >
> >Folks -
> >
> >What operating system are we talking about here? AIX? Solaris? NT?
>Let's
> >make
> >sure we are comparing apples to apples.
> >
> >The only bug I know of in this area was on Solaris 2.5.1 and we fixed
>that.
> > It
> >was easier than waiting for SUN to fix it, because it really is there
> >problem,
> >and they fixed it in 2.6. If you are on Solaris 2.5.1 and cannot go to
> >Solaris
> >2.6, where SUN has fixed a host of their other problems, then I urge you
> >to
> >call Support and get a new copy of trapd for Solaris if you cannot wait
>for
> >5.1.2.
> >
> >But otherwise, having the events display stop updating is not normal
> >behavior
> >and you need to call Support and get some help in diagnosing the problem.
> >Otherwise, whatever is wrong in your environment will likely stay wrong
>and
> >this
> >problem will surface over and over again until it is fixed. We cannot
>even
> >tell
> >if it is a code bug until someone looks at it in detail.
> >
> >Restarting the daemons may work, until you hit the condition again. But
> >the
> >size of the trap log should have nothing to do with it.
> >
> >James Shanks
> >Tivoli (NetView for UNIX) L3 Support
> >
> >
> >
> >"Desai, Aditya" <DesaiA@USA.REDCROSS.ORG> on 08/05/99 03:15:42 PM
> >
> >Please respond to Discussion of IBM NetView and POLYCENTER Manager on
> >NetView
> > <NV-L@UCSBVM.UCSB.EDU>
> >
> >To: NV-L@UCSBVM.UCSB.EDU
> >cc: (bcc: James Shanks/Tivoli Systems)
> >Subject: Re: Events stop coming into Evenst Display...
> >
> >
> >
> >
> >
> >We are experiencing the same problem, It was not bad before the patch,
>but
> >after the patch the control desk events stop updating, some time twice in
> >one day.
> >
> >-----Original Message-----
> >From: Gord Michaels [mailto:NV-L@UCSBVM.ucsb.edu]
> >Sent: Thursday, August 05, 1999 12:35 PM
> >To: gord_michaels@HOTMAIL.COM; NV-L@UCSBVM.ucsb.edu
> >Subject: Events stop coming into Evenst Display...
> >
> >
> >Hello All.
> >
> >We have recently migrated to Netview 5.1.1.
> >
> >After the migration, I noticed that the size of my trapd.log file was set
> >to
> >26MB. I reduced this to 10MB because I felt that 26 MB was too much to
>surf
> >through with vi.
> >
> >To be honest, I cannot recall what it was set to prior to migration.
> >
> >Shortly after I did this, I noticed a strange problem. My events would
>stop
> >coming into my Events Display. Even though my trapd.log was still
>receiving
> >them! The only way to fix this was to perform a "ovstop and ovstart".
>After
> >that, all would be fine for a week or more, and now it has happened
>again!
> >
> >Has anyone ever experienced this problem before? Could it be related to
> >changing the trapd.log size to 10Mb from 26Mb??
> >
> >Also, I only have one rulset in the foreground and none in the
>background.
> >My maps still work fine and all else seems to be working fine, I am just
> >suddenly stop reeiving events in my Events Display.
> >
> >Any help appreciated.
> >
> >Sincerely,
> >
> >Gord Michaels.
> >
> >
> >
> >______________________________________________________
> >Get Your Private, Free Email at http://www.hotmail.com
>
>
>______________________________________________________
>Get Your Private, Free Email at http://www.hotmail.com
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
|