I just had a thought. Could this be the NT version of the problem we always
see on Unix, the error reading background file, caused by a zero-byte
file in /usr/OV/bitmaps/C? That is usually fixed by replacing the bad file
with some other file. Take a look at the Cisco files in that directory
after
the integration utility is run.
Also, running out of memory when loading the V2 mibs seems to be the
normal case. I've never seen it NOT do that.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
"GENet network management" <todd.ross@genet.gc.ca>@tkg.com on 10/29/2000
11:50:22 AM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: "NV Mailing list" <NV-L@tkg.com>
cc:
Subject: [NV-L] Fatal netview.exe problem - 6.01 on NT 4(6a) + Ciscoworks
2000
I've been having problems integrating NV 6.01 and Ciscoworks 2000 on NT 4
with SP 6a. Both are installed on the same box - a dual-processor system
wth 2 * 200 Mhz Pentium IIs + 256 mb memory.
The major issue is: after the integration of CW2K, the netview.exe
program
will suddenly eat all the memory on the system. The process virtual memory
jumps to more than 1,000,000 k ( I have a lot of swapfile space), after
which all the physical memory on the system is taken over by netview.exe.
CPU use stays relatively low (10%?) . The task manager (application tab)
reports that the netview process is "Not responding". Attemps to kill the
process via task manager don't work.
The cycle only ends when the process eats almost all the physical memory on
the system. The netview process will suddenly die (or be killed by NT...
don't know which.)
I can trigger the problem at will by starting netview.exe and selecting
Object => New , leting the object wizard start, and selecting Object type
of Cisco Router or Cisco Switch. Other object types work fine. The problem
starts immediately after clicking on one of the Cisco object types - so
soon
that the list of individual objects is never displayed.
The problem will also show up at random when using the submap explorer.
Running up against a submap with a cisco router in it seems to trigger it.
Reinstalling NV 6.01 fixes the memory issue, but loses all the integration
stuff (of course).
Any idea what netview.exe is looking at when it lists object types? Any
parameter file (etc.) that can be checked ?
Also, when running the integration adapter, one iteration of loadmibv2.exe
also blows up with an "out of memory" error. I didn't feel this was
related, but I'll mention it just in case...
Thanks,
Todd
Todd Ross
GENet Network Engineering
Network Services Directorate,
GTIS/PWGSC
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
|