CSM redundancy Issue.Plz help

fariha zain · ‎03-16-2011

Hi All,

I have two Cat Cisco Swithces 6500 in my network.

Switch 1 with CSM1 ,Sup1 &2--------------------------------------------------------------Switch 2 with CSM2 & Sup 1&2

Issue:One of the supervisor module in switch 1 failed and never came back online which made the CSM module in inactive state.

And as per redundancy Switch 2 became active and CSM2 started processin the requests;

The problem is that when the supervisor module in slot 2(Sup2) become active the CSM module changed state from inactive to active.(Which is wrong ,It should be in Standby state.)

So both the modules were in Active/Active state which caused some rservers to go up and down.

Then I manullay reseated the CSM1 module after which it stated as standby which is correct.

Questions:

1) What caused the CSM1 to come up as active state when sup2 on switch 1 was active?

2) Is this a bug?

3) Any config mistake and needs correction?

Any help would be appriciated.

REgards

Fari

fariha zain · ‎03-17-2011

Can someone please help me in regards to my issue?

Thanks

Fari

Christopher Miles · ‎03-21-2011

Hi Fari,

The CSM state is not defined by which SUP is the active, but by the heartbeats that are being sent and received between the two CSM modules and which IP addresses are being polled. At the time of the issue this syslog would have looked something like this on both CSMs.

Mar  x 00:00:24.128 JST: %CSM_SLB-4-REDUNDANCY_WARN: Module X FT warning: Standby is
Active now (no heartbeat from active unit)
Mar  x 00:00:25.128 JST: %CSM_SLB-6-REDUNDANCY_INFO: Module X FT info: State Transition
Active -> Standby
Mar  x 00:00:26.128 JST: %CSM_SLB-6-REDUNDANCY_INFO: Module X FT info: State Transition
Standby -> Active

If heart beats have not been seen for a period of time then both sides will assume the other side has gone down and will become active. Normally the CSM logs should provide some information as to why the connection between the two chassis failed.

When you reseated the module whatever issues where blocking the heart beat packets must have cleared and the modules were happily communicating again.

Generally heartbeat connections are good to be on their own dedicated cable between the boxes, things like high CPU, flooded links etc can effect the processing or delivery of these packets.