IPM 4.1 (LMS 3.1) error in ipmpoller.log

[mladenp]

Hello,

I have problems with generating reports for some collectors (No data available for the selected time period). For those collectors there is two type of errors in log:

[ pet vel 04  09:21:22 CET 2011 ],ERROR,[Poller-1234],com.cisco.nm.ipmng.common.datastructures.HistoricalDataCollector,collectStats,313,Bogus value from device while collecting the stats :nullcom.cisco.nm.ipmng.dal.DALBogusDataException
    at com.cisco.nm.ipmng.dal.ProcessResponseUtil.ProcessObjectIfs(ProcessResponseUtil.java:235)
    at com.cisco.nm.ipmng.dal.ProcessResponseUtil.processGetTableSnmp(ProcessResponseUtil.java:321)
    at com.cisco.nm.ipmng.dal.SnmpDeviceAccess.gettable(SnmpDeviceAccess.java:2256)
    at com.cisco.nm.ipmng.dal.DeviceAccess.gettable(DeviceAccess.java:230)
    at com.cisco.nm.ipmng.common.datastructures.HistoricalDataCollector.collectStats(HistoricalDataCollector.java:255)
    at com.cisco.nm.ipmng.common.datastructures.HistoricalDataCollector.pollDevice(HistoricalDataCollector.java:79)
    at com.cisco.nm.ipmng.poller.server.DataPoller.pollCollector(DataPoller.java:345)
    at com.cisco.nm.ipmng.poller.server.DataPoller.performPolling(DataPoller.java:291)
    at com.cisco.nm.ipmng.poller.server.DataPoller.access$000(DataPoller.java:65)
    at com.cisco.nm.ipmng.poller.server.DataPoller$1.run(DataPoller.java:515)
    at java.lang.Thread.run(Thread.java:595)

[ pet vel 04  09:21:22 CET 2011 ],ERROR,[Poller-1234],com.cisco.nm.ipmng.poller.server.DataPoller,pollCollector,379,Unable to poll the collector CollectorID = 17Name = 1234AdminIndex = 133862Device = 29due to No entries

and second type:

[ pet vel 04  09:54:22 CET 2011 ],ERROR,[Writer-Test2],com.cisco.nm.ipmng.poller.server.DataWriter,writeDataInDB,403,Unable to write the collector due to null

Sometimes is enough to stop/start collector to get reports after some time, sometimes not...

Real time graph for that collectors works OK.

Regards,

Mladen

[Joe Clarke]

The IPMProcess.log should have some more details.  Please post this log.  It would also be helpful to see a packet capture of all SNMP packets between the server and this device when the poller error occurs.

[mladenp]

Hello Joe,

There is a part of IPMProcess.log. I will try to do packet capture.

Regards,

Mladen

[Joe Clarke]

Looking at the error, it appears the device is replying with some bogus (extremely high) data values.  The sniffer trace would be good, but you can also try enabling IPMPOLLER debugging under IPM > Admin > Log Level Settings.  Reproduce the problem, then post the ipmpoller.log.

[mladenp]

Hello Joe,

Here is ipmpoller with debugging. I have pcap with conversations, but I'm worried about posting such things here (snmp communities etc.)...

Regards,

Mladen

[Joe Clarke]

I don't see that you reproduced the polling issue in this log once debugging was enabled.  In order to get an idea of what the bogus value is, I need to see a DB insert that contains the bogus data.

[mladenp]

Hello,

Hm,

"I don't see that you reproduced the polling issue in this log once debugging was enabled." - I'm not sure if we talking about same thing - there is the same error in log for that collector:

[ uto vel 08  09:21:22 CET 2011 ],ERROR,[Poller-1234],com.cisco.nm.ipmng.common.datastructures.HistoricalDataCollector,collectStats,313,Bogus value from device while collecting the stats :nullcom.cisco.nm.ipmng.dal.DALBogusDataException

"In order to get an idea of what the bogus value is, I need to see a DB insert that contains the bogus data." - how can I provide you that ? I have packet capture but I will like to get it to you on some secure way...

Regards,

Mladen

[Joe Clarke · Accepted solution]

Sorry, I missed that.  This is a known limitation of IPM 4.1.  When certain values exceed 999999, a bogus data value exception is thrown.  This problem is fixed in LMS 3.2.  You can upgrade to LMS 3.2 for free by downloading the eval from http://www.cisco.com/go/nmsevals/ .  The eval will become a fully-licensed version when you install it.

[mladenp]

Hello Joe,

Thanks for help.

Do you maybe know bug ID ? I can't find it...

Regards,

Mladen

[Joe Clarke]

Unfortunately, it doesn't look like a bug was filed.  The closest match is      CSCsl76625, but the committed code did not fix this issue.  That was fixed asynchronously.