FI-6296UP - Mgmt Issue

[Ron Baduach]

Hi Guys ,

We got pair of FI-6296UP , connected to Nexus 5020 [also mgmt ports connected to N5K] .

I got ping to FI-B , but not to FI-A . [although in the nexus the ports are up]

Don't know what exactly to check on the both FIs in order to see what is the problem .

I'm afraid to restart the FI-A , because this is production environment .

How can i check what is the issue ?

[BTW-

1. ping between FI-A to FI-B failed

2. ping between FI-A to FI floating address working

3. ping between FI-B to FI floating address not working]

Big thanks in advance ,

[Kirk J]

Greetings.

From FI-B local-mgmt context, can you ping anything else on the subnet (i.e default gateway)?

From FI-B local-mgmt context, run #show pmon state.

If there appear to be no failed services, or evidence of service restarts/cores, depending on your code level you may want to look at the symptoms of https://bst.cloudapps.cisco.com/bugsearch/bug/CSCuv31912/?reffering_site=dumpcr 

This is all of course assuming your upstream switch port is still set as an access port, with the correct vlan, etc.

thanks,

Kirk...

[Ron Baduach]

Hi Kirk , 

1. from FI-B local-mgmt , i can ping anywhere .

2. show pmon state on FI-B [take a look on the screenshot]

3. i run "show mgmt-ip-debug ip-tables | wc -l" on both fabric, and here is the results :

on FI-A [the un-pingable] - 100093

on FI-B - 206

what you suggest me to do ?

[Kirk J]

The screenshot shows the sam_dme process has crashed a couple of times.

Please check in your UCSM GUI admin tab, and see if there are any recent files listed in the 'cores' section.  If there are, you will want to generate a UCSM tech support and open a case with TAC to have the cores analyzed.

If you have a maintenance window you may want to attempt a reboot of FI-A to see if FI-A mgmt node IP comes back online.

Thanks,

Kirk

[Ron Baduach]

Hi Kirk , thanks for your reply .

I can't access the UCSM GUI , because the active FI [FI-A] is down ....

I'm afraid of reboot FI-A , afraid that the hosts will lost connectivity .

[Kirk J]

Greetings.

If FI-A is your actual Primary, then you may have to force the cluster lead shift to B.

I would highly recommend calling in to TAC to review your scenario prior to doing this though.

Thanks,

Kirk...