SNMP Timeouts in Cat6500

Unanswered Question
Apr 29th, 2009

Hello everyone. Have someone seen this behavior before?

I have a cisco WS-C6509-E with IOS IP BASE version 12.2(18)SXF11 and a WS-SUP32-GE-3B. The problem I'm having is that, Internal as well as External Monitoring Servers are giving us SNMP timeouts, so they report the device as down. But when you log into the device, everything looks fine, cpu is normal usage, pings from and to the monitoring servers working fine, traceroutes from and to the monitoring servers work as well. We have all SNMP from all monitoring servers permitted in the firewalls. Routing is working fine, but still, I get lots of timeouts.

Has someone encounter something similar?

Best Regards

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Joe Clarke Wed, 04/29/2009 - 09:47

What objects are being polled? What version of SNMP are you using? What is the current timeout value? What kind of delay do you have in the network when it comes to SNMP traffic?

daniel.pimentel Wed, 04/29/2009 - 10:03

Hi clarke, I'm using CiscoWorks and a monitoring tool called Entuity, not sure which objects are being pulled from CiscoWorks but I suppose interface status, cpu, memory utilization. Also I normally use snmpwalk to test SNMP, We are currently using SNMP v1, I have more devices (Cat6500) and when I do snmpwalks to them, they all answer pretty fast, so it makes me think I'm not having a delay with SNMP traffic. I've tryed different timeout values with snmpwalk, from 10 seconds to 60 seconds. sometimes the switch will respond but soon after, it will issue a timeout response.

Joe Clarke Wed, 04/29/2009 - 10:14

You'll need to get specific examples of the timeouts. This is best done with a sniffer trace on the management server. It would also be helpful to see the output of debug snmp packet on the switch when the timeouts are occurring. This will help prove that the requests are making it to the switch.

daniel.pimentel Tue, 05/19/2009 - 07:37

Finally the SNMP behaviour could be reproduced. I couldn't do the debug and the packet trace the other day because it was working fine by the afternoon.

In the debug packets I see this.

May 19 11:15:16: SNMP: Packet received via UDP from 165.10.88.94 on Vlan101

May 19 11:15:16: SNMP: Get-next request, reqid 1748197598, errstat 0, erridx 0

mib-2 = NULL TYPE/VALUE

May 19 11:15:16: SNMP: Response, reqid 1748197598, errstat 0, erridx 0

system.1.0 = Cisco Internetwork Operating System Software

May 19 11:15:16: SNMP: Packet 311 sent via UDP to 165.10.88.94

May 19 11:15:17: SNMP: Packet received via UDP from 165.10.88.94 on Vlan101

May 19 11:15:17: SNMP: Get-next request, reqid 1748197598, errstat 0, erridx 0

mib-2 = NULL TYPE/VALUE

May 19 11:15:17: SNMP: Response, reqid 1748197598, errstat 0, erridx 0

system.1.0 = Cisco Internetwork Operating System Software

May 19 11:15:17: SNMP: Packet 311 sent via UDP to 165.10.88.94

May 19 11:15:18: SNMP: Packet received via UDP from 165.10.88.94 on Vlan101

May 19 11:15:18: SNMP: Get-next request, reqid 1748197598, errstat 0, erridx 0

mib-2 = NULL TYPE/VALUE

May 19 11:15:18: SNMP: Response, reqid 1748197598, errstat 0, erridx 0

system.1.0 = Cisco Internetwork Operating System Software

May 19 11:15:18: SNMP: Packet 311 sent via UDP to 165.10.88.94

May 19 11:15:19: SNMP: Packet received via UDP from 165.10.88.94 on Vlan101

May 19 11:15:19: SNMP: Get-next request, reqid 1748197598, errstat 0, erridx 0

mib-2 = NULL TYPE/VALUE

May 19 11:15:19: SNMP: Response, reqid 1748197598, errstat 0, erridx 0

system.1.0 = Cisco Internetwork Operating System Software

May 19 11:15:19: SNMP: Packet 311 sent via UDP to 165.10.88.94

May 19 11:15:20: SNMP: Packet received via UDP from 165.10.88.94 on Vlan101

May 19 11:15:20: SNMP: Get-next request, reqid 1748197598, errstat 0, erridx 0

mib-2 = NULL TYPE/VALUE

May 19 11:15:20: SNMP: Response, reqid 1748197598, errstat 0, erridx 0

system.1.0 = Cisco Internetwork Operating System Software

May 19 11:15:20: SNMP: Packet 311 sent via UDP to 165.10.88.94

In the packet tracer I see lots of requests coming from the 165.x.x.x server and occasionally one answer from the switch. I also see other servers requesting SNMP, so maybe the switch is getting overflooded with SNMP requests?

yjdabear Wed, 04/29/2009 - 09:52

I get intermittent SNMP timeouts against various devices (not just 6500s) at remote WAN sites. In my case, the devices are busy dealing with snmpsets from another source when they start timing out against the snmpgets. I simply chained a ping-test routine after that to filter out the "device downs".

Actions

This Discussion