cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2396
Views
0
Helpful
18
Replies

SNMP-Cisco 3560

I have this of my snmp.exe crashing down.I found the problem root coming from core switch Cisco 3560.I tried to remove the core switch configuration from my SNMP configuration and things started to work, I also tried leaving the running the original configuration with the core switch configuration and power down the core switch and things work.With this scenario, what really is causing my snmp.exe to crash?

Please advise.

3 Accepted Solutions

Accepted Solutions

What is different about this other, working switch?

The configuration from this device is fine. I am not familiar with Ovation, but it if I read this config correctly, it looks like you're just polling ifOperStatus from all of the interfaces, and looking to see if any are down (value == 2). This doesn't raise any red flags per se.

I would highly urge you to pursue this problem with the Ovation vendor to see if they can provide more details about why their daemon is crashing. But if you want to dig into what SNMP packets are causing the problem, you will need a good sniffer application. I recommend Wireshark (freely downloadable from http://www.wireshark.org) as it is very powerful, and includes Windows binaries.

The idea is you will need to filter on udp/161 traffic between your Ovation manager and this switch. When the snmp2ovation.exe crashes, compare the time stamp of the crash to the captured packets. To do this, start a trace using Wireshark with a capture filter of:

host cs1 and udp port 161

Then have snmp2ovation.exe start polling the device. When it crashes, stop the trace, and do the above mentioned analysis.

View solution in original post

Wireshark must be running on the Ovation server as that is where the SNMP traffic flow is occurring. You will need to use the capture filter I provided previously to capture the packets between the manager and the cs1 switch. It would also be useful to see the flow between the working, cs2 switch.

I didn't not notice any relevant configuration differences between these two switches, but configuration is only a small part. What software is the backup switch running? Please also provide an snmpwalk of the ifTable from both switches.

View solution in original post

The problem here is that the manager is requesting ifOperStatus.2, but this instance does not exist on this switch. Since both switches reply with the same noSuchName error, I cannot say why cs1 causes the manager to crash while cs2 does not. That is a question for the Ovation vendor.

What I can tell you is that you will need that snmpwalk of the ifTable to know exactly which ifIndex instances should be polled from these switches. Hopefully, when you poll the correct indexes, the crash will go away.

View solution in original post

18 Replies 18

avmabe
Level 3
Level 3

What software are you using?

It's a proprietary software used in DCS(Distributed Control System)called snmp2ovation.exe that talks to another proprietary software called Ovation.The other switches Cisco 2950 are working and I could see them updating if Cisco 3560 is switched off.

What do you think could be the problem?

Thanks,

June

What SNMP configuration do you have on the 3560? what version of IOS is it running? what is snmp.exe actually polling that causes it to crash? This latter item could be obtained by running a sniffer trace of the SNMP traffic to the switch, and correlating the the packets in the trace to the time the process crashes. Alternatively, a "debug snmp packet" on the switch may be used in the same way.

You may also check to see if there are any logs on the manager side that could reveal what was being polled when the process crashed.

Thanks for your reply...

Attached are the snmp_config file and the switch 3560 configuration for your reference.C3560-ipservices-mz.122-25.SED1.bin is the IOS version.

For your info, there's another Cisco 3560 switch being used as a backup but not causing the snmp2ovation.exe to crash.

I am just new with SNMP and have never used the tools that you had suggested.Is there a literature from Cisco or software that I could download?

Please advise.

What is different about this other, working switch?

The configuration from this device is fine. I am not familiar with Ovation, but it if I read this config correctly, it looks like you're just polling ifOperStatus from all of the interfaces, and looking to see if any are down (value == 2). This doesn't raise any red flags per se.

I would highly urge you to pursue this problem with the Ovation vendor to see if they can provide more details about why their daemon is crashing. But if you want to dig into what SNMP packets are causing the problem, you will need a good sniffer application. I recommend Wireshark (freely downloadable from http://www.wireshark.org) as it is very powerful, and includes Windows binaries.

The idea is you will need to filter on udp/161 traffic between your Ovation manager and this switch. When the snmp2ovation.exe crashes, compare the time stamp of the crash to the captured packets. To do this, start a trace using Wireshark with a capture filter of:

host cs1 and udp port 161

Then have snmp2ovation.exe start polling the device. When it crashes, stop the trace, and do the above mentioned analysis.

The other switches are cisco 2950.

I'll try your advice and let you know of the result later.

Thank you very much for your help.

You mentioned you had another 3560 that was not causing this process to crash. I was more interested in what was different about that switch vs. the one that is causing the process to crash.

Attached is the configuration of the backup Cisco 3560. I tried the wireshark but did not have success.Can you give me some short procedure to configure wireshark on capturing the SNMP packets between ovation manager and cs1?Do I need to have the wireshark running at ovation manager or I could have that running on my laptop?

Please advise.

Thanks again...

Wireshark must be running on the Ovation server as that is where the SNMP traffic flow is occurring. You will need to use the capture filter I provided previously to capture the packets between the manager and the cs1 switch. It would also be useful to see the flow between the working, cs2 switch.

I didn't not notice any relevant configuration differences between these two switches, but configuration is only a small part. What software is the backup switch running? Please also provide an snmpwalk of the ifTable from both switches.

C3560-ipservices-mz.122-25.SED1.bin is the IOS version. I am not familiar with snmpwalk.Can you email me a short procedure how to use it?One thing I noticed is that I can't ping or telnet cs2 when cs1 is down...Is this behavior have any relevance to my SNMP issue?

Please advice.

Thanks...

See http://net-snmp.sourceforge.net for a good SNMP toolkit as well as some useful SNMP tutorials. Without knowing more about what this SNMP manager is doing, I cannot say for certain if this is relevant. The sniffer trace showing the actual SNMP data responsible for the crash is the most useful thing to see at this point. From that, it should be easy to determine if the crash is an SNMP issue on the switch, or a problem with Ovation.

Thanks a lot for your help.I will continue with this troubleshooting tomorrow and try to get the information that you need.I'll let you know of my findings tomorrow as I have to leave the customer site due to some work schedule restriction.

Thanks again...

I was able to get some information before I get done for the day using the wireshark.I am jsut nor really sure if I did it right.Anyway,attached I got from wireshark and will appreicate very much if you could still take a look on it.I tried to run the snmp program and in both case it crash.I also tried to run it with cs1 powered down but I did not get any result.I'll try snmpwalk tomorrow and will send you the information that I'll get.

Thanks again for all your help.Have a good evening.

The problem here is that the manager is requesting ifOperStatus.2, but this instance does not exist on this switch. Since both switches reply with the same noSuchName error, I cannot say why cs1 causes the manager to crash while cs2 does not. That is a question for the Ovation vendor.

What I can tell you is that you will need that snmpwalk of the ifTable to know exactly which ifIndex instances should be polled from these switches. Hopefully, when you poll the correct indexes, the crash will go away.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: