On our CiscoWorks 2000 installation, running on Solaris with LMS3.2 and RME4.3.1, we noticed a few weeks ago, that the job "Inventory Collection" needs more and more time to finish. At the moment the job execution time is as high as 8 hours, compared to 3 hours before. When we look at the network traffic during job execution, we see after the collection phase (~3h) several hours with no traffic together with a very high CPU load. I mean to remember that this job used to complete around 15 minutes after data collection phase. Restarting the application and rebooting the system didn't help.
When we look in to the IC_Server.log, we find devices with a completion time of over 20'000 seconds. Increasing from day to day.
Example from log: [ Wed Jun 08 13:34:15 MEST 2011 ],INFO , [Thread-43],com.cisco.nm.rmeng.inventory.ics.core.CollectionController,498,Total time is 26832 seconds for .....".
Affected devices are Cisco Nexus 5548P. We tried to delete/add these devices. That helps for a few days, then the runtime starts to increase again.
Are we hitting a bug? And is it possible, as a workaround, to exclude devices from only the Inventory Collection?
Thanks for advice
You can try upgarding the LMS from 3.2 to LMS 3.2.1 and I hope will help you !!!
Below is the download link for LMS 3.2.1 :
The system is now running for 2 weeks with the patch. And it looks like this patch solved our problem. The runtime for "inventory collection" is not increasing any more.
thanks a lot and regards