cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1249
Views
15
Helpful
15
Replies

Discovery problem after upgrade to lms2.6

sem7433
Level 1
Level 1

After upgrade to lms 2.6 i got problem with device discovery.

I got following message in discovery.log - FATAL ERROR ServiceModule: undefined property: ServiceModuleList

15 Replies 15

nhabib
Level 9
Level 9

This is usually due to a corrupt DeviceDiscovery.properties file, which is usually located in CSCOpx/campus/etc/cwsi/

Please attach it.

This is my devicediscovery.properties file:

Discovery.subnets.exclude=

nameserver.resolveBySysName=false

nameserver.resolveByName=true

Discovery.seed=192.168.63.2:192.168.73.2:192.168.96.254:192.168.98.254:192.168.99.254:192.168.209.1:192.168.209.3:192.168.209.4:194.71.67.125:192.168.200.14:192.168.200.17:192.168.43.4

nameserver.useloopbackaddress=false

nameserver.usedns=true

Discovery.router=on

snmp.encryptCommunity=false

snmp.EnableMultiple=true

LogMsg.deviceList=DeviceDiscovery.properties

LogMsg.trace=

LogMsg.logfile=E:\\CSCOpx\\log\\discovery.log

LogMsg.logFileSize=10000

LogMsg.state=enable

LogMsg.debug=

I have already try to change to devicediscovery.properties.org file but it dosnt help. In this case i got following message in discovery.log

2006/11/10 15:26:07 main MESSAGE DiscoveryMain: Properties will be read from E:\CSCOpx\campus\etc\cwsi\DiscoveryStatus.properties

2006/11/10 15:26:07 main MESSAGE DiscoveryMain: Properties will be read from E:\CSCOpx\campus\etc\cwsi\DiscoveryStatus.properties

2006/11/10 15:26:07 main MESSAGE DBConnection: Created new Database connection [hashCode = 2549855]

2006/11/10 15:26:07 main ERROR SMFFactoryLoader: java.lang.NullPointerException

2006/11/10 15:26:07 main ERROR DiscoveryMain: AniStaticConfigException: Error accessing devices--root path , packageName com.cisco.nm.ani.server.devices

As I suspected, your DeviceDiscovery.properties file is corrupt.

Please attach (not copy and paste) your DeviceDiscovery.properties.orig and I will attempt to fix it for you.

Hi

Here is my DeviceDiscovery.proberties.orig file.

Here is a new DeviceDiscovery.properties with a proper DEVICESROOT, and all of your previous settings restored. Place this file back into E:\CSCOpx\campus\etc\cwsi, then you can restart Device Discovery.

Note: you might want to hold on to a copy of this file just in case you need it again.

It work's. Thanks for the help.

In which way did you add your seed devices ?

I import my devices from a file.

sorry for the long delay, but I was very busy the last days..

I did assume your answer- and this is for all the Cisco people reading this thread:

I am sure there is a device limit in importing seed devices from a file on windows ( I didn`t made any test on solaris)

with the discribed effect of making DeviceDiscovery.properties unusable. I think the reason is related to the fact that all seed devices are in a row. The import goes without any problems, but any further changes in the GUI that touch DeviceDiscovery.properties destroy that file (especially adding one more with the ADD option).

I think the max number for devices added this way are 250 to 300.

I made the same experience when I tried to speed up disco in that way !

This should either be documented or a BUG should be rised against this!!

MArtin

There is no limit to the number of seed devices one can have (well, 5000 is the device limit, so you can't go above that). There were bugs in earlier versions of Campus Manager 4.0 that caused DeviceDiscovery.properties corruption. None of them were related to the number of seed devices. You can add all 5000 through the GUI if you want.

thank you joe for your answer.

2 times I made *some* tests and have had always a corrupt properties file. I hope I have time to to recreate this the next days and will give you an update on this.

That was what I did in the past:

I imported the first seeds from a file (around 800 devices) and added afterwards 2 by entering the IP address. I got a corrupted properties file; same sw , same patchlevel and I reduced the number of devices in the file to 200 and everything was OK.

I will do these test again... and I bet...

ok, I won?t bet - I will just do the tests...

and bet afterwards :-)

MArtin

Id have to agree with mermel here. I have had the same problem on solaris. Its not a set number that it corrupts on though. On one of our implementations, the file corrupts after adding around 500+ seeds, where as our other implementation is fine with the exact same seeds.

Ive never bothered to raise a TAC as the problem is easily fixable (look at the working file and compare with the broke one).

I just know that it does happen (on LMS 2.5.1 no less) and it is somehow related to seeds.

Nick

These are the first reports of seed import related corruption that I've heard, and this is very disconcerting. I have since tested Campus 4.0.4 on Windows and Solaris (LMS 2.5.1) as well as Campus 4.0.7 on Windows and Solaris (LMS 2.6), and I cannot reproduce any DeviceDiscovery.conf corruption.

This is what I did. I created a seed file with 508 seeds (basically two class C subnets). I imported that file, via the GUI, into Campus Manager which already had five seed devices. The result was a list of 513 seed devices in the GUI. I then clicked the OK button followed by the Apply button in the previous screen. I did not run discovery since these were bogus addresses.

I then opened the DeviceDiscovery.properties file on the server, and noted that the Discovery.seed line was in tact, and contained all 513 seeds. I saw no other evidence of corruption in the file.

When you say that this file was corrupted, what kind of corruption did you witness? What symptoms of corruption did you encounter?

Currently I can not make any test cause my lab-server has a disc failure...

I observed the behaviour an a customer site, I will ask if they still have one of the corrupted files.

I yet reported this in March this year when I observed the failure:

http://forum.cisco.com/eforum/servlet/NetProf?page=netprof&forum=Expert%20Archive&topic=Network%20Infrastructure&CommCmd=MB%3Fcmd%3Dpass_through%26location%3Doutline%40%5E1%40%40.1dda9020/118#selected_message

MArtin