I see this most clearly on systems where the seedfile is excluding things
by oid or by name wildcard. These systems usually have a lot of stub
objects in the database (the ones that go away when you do an ovtopofix
and then come back to the database but do not appear on the map.) At
netmon startup, every object is checked against the snmp configuration
settings. That takes a couple of minutes on a system of your size. Then
it goes into a configuration poll. That configuration poll is not only against
the things that are in your maps, but also against those stub objects. They
are re-evaluated to see if they now meet the criteria. (This is a design
feature, and is also done during the 'daily' configuration poll.) In most cases
I am excluding non-snmp objects from discovery, so most of the stubs are for
non-snmp devices. Then the timeout I have configured comes into play.
If during startup you watch the output of netmon -a 4, you will see it crawl
through this list of objects at a pretty slow pace, depending on the timeout.
Three things you can do:
1) Reduce the number of stub objects whenever possible by using address
ranges to do your exclusions. These do not result in stubs.
2) Turn off configuration polling - either all the time, or just when you are in
a
hurry to get past it and onto discovery.
3) Increase the number of concurrent snmp queries, as Mark explained the
other day, with the -Q flag in netmon.lrf. By default it uses 16. I bumped one
up to 32 this week with no ill effects on a 16MB tokenring.
You might also check and see if you are having slow response from your
DNS, if you are using one. If lots of your objects are unnamed, this might
be a factor. To speed up the fail-off on the name lookups, try putting
RES_RETRY=1 and RES_TIMEOUT=1 in /etc/environments and rebooting.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
(248) 552-4968 Voicemail, Fax, Pager
---------------------- Forwarded by Leslie Clark/Southfield/IBM on 09/29/99
07:29 AM ---------------------------
"Engelsen, Bjørn" <bjorn.engelsen@DNB.NO> on 09/29/99 05:20:25 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:
Subject: Netmon "sleeping" for 1,5 to 2 hrs after startup
Lately our netmon has stopped without any obvious reason.
After restarting it (or restarting all daemons) , it seems to do no polling
for almost two hours.
Unmanage/manage of an object turns it into blue, and it does not get back to
green.
I have seen in this forum earlier that netmon tends to use up to 20 mins
before doing any polling, but 2 hours ?
No of objects in database: 6468
ovwdb cache: 8000
Netview version: 5.1.1
Framework: 3.6
Platform: Solaris 2.6
Is there any troubleshooting we coud do here ?
Regards,
Bjorn Engelsen
Den norske Bank ASA
Networked Systems Management
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
This email with attachments is solely for the use of the individual or
entity to whom it is addressed. Please also be aware that DnB cannot
accept any payment orders or other legally binding correspondence with
customers as a part of an email.
This email message has been virus checked by the virus programs used
in the DnB Group.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
|