I had a similar problem with NetView 5.1.1 pulling a 340000 entry ARP table
off one of my routers. CPU would spike at 100% with IP SNMP on the router
getting nearly all the CPU during that time frame. It would take NetView
over 45 min, using SNMP v1 to pull the ARP table entries off of that router.
Ended up doing an SNMP View command on the router to tell the router
not to allow anyone to pull his ARP table via SNMP.
First thing I would check on the NetView systems is the entry for that
device in xnmsmpconf and see if they match.
Don't know why two different NetView systems would get different results
when polling the router. You can try using iptrace to get a trace of both
systems demand polling the router and comparing the MIB variables that are
being polled.
__________________________
Thanks,
Sean Davidson
Sr. Network Systems Engineer
Publix Super Markets, Inc.
P.O. Box 32015
Lakeland, Fl. 33802-2015
Email - sean.davidson@publix.com
Voice - (863) 686-8754 x6889
Fax - (863)616-5895
-----Original Message-----
From: Owens, Blaine C [mailto:bowens@EASTMAN.COM]
Sent: Friday, January 07, 2000 1:49 PM
To: NV-L@UCSBVM.ucsb.edu
Subject: Re: polling for default routing entries
Gib, I'm afraid I was in a hurry and didn't read your original post close
enough. I was thinking of the difference between a production ROUTER with
large tables and a test ROUTER with small tables. I see now that you are
looking at the same router (same tables) with two different NetViews.
Anybody with a clear head have any suggestions?
Blaine Owens
Eastman Chemical Company
Phone - (423)-229-3579
Fax - (423)-229-1188
bowens@eastman.com
> -----Original Message-----
> From: Fendrick, Gib (CC-MIS) [SMTP:Gib.Fendrick@CONAGRA.COM]
> Sent: Friday, January 07, 2000 1:30 PM
> To: NV-L@UCSBVM.ucsb.edu
> Subject: Re: polling for default routing entries
>
> Blaine,
> Any idea tho why the demand poll and load it puts on the router could
> be so different between the two NetView systems?
>
> Thanks,
> Gib Fendrick
> ConAgra, Inc.
>
> -----Original Message-----
> From: Owens, Blaine C [mailto:bowens@EASTMAN.COM]
> Sent: Friday, January 07, 2000 12:25 PM
> To: NV-L@UCSBVM.UCSB.EDU
> Subject: Re: polling for default routing entries
>
>
> You are innocent - I have found this behavior to be normal for our Cisco
> routers ( the larger the tables the greater the impact).
>
> Blaine Owens
> Eastman Chemical Company
> Phone - (423)-229-3579
> Fax - (423)-229-1188
> bowens@eastman.com
>
> > -----Original Message-----
> > From: Fendrick, Gib (CC-MIS) [SMTP:Gib.Fendrick@CONAGRA.COM]
> > Sent: Friday, January 07, 2000 1:13 PM
> > To: NV-L@UCSBVM.ucsb.edu
> > Subject: polling for default routing entries
> >
> > Hi all.
> >
> > I have two Netview/AIX 5.1.1 systems running (production and
> development).
> > When we do a Demand Poll on the prod system, it will take about 4
> minutes.
> > And the CPU on the router will shoot up to approx 60% during this time.
> > The Demand poll display spends most of this time on 'get default routing
> > entry'.
> >
> > When we do a Demand Poll on the development system, it takes about 15
> > seconds.
> > NO spike in CPU on the router. No waiting for 'get default routing
> > entry'.
> > (Actually spends most
> > of this time waiting to time out on an interface that's down).
> >
> > It seems to have something to do with the routing table, which on the
> > router
> > in question is about
> > 1500 route entries. Is there something we might have done that is
> > influencing how NetView retrieves
> > this table? I've check the SNMP Configuration with the GUI on both
> > systems, they are the same.
> >
> > Anyone have any suggestions?
> >
> > Thanks,
> > Gib Fendrick
> > ConAgra, Inc.
|