Nexus 7000 single supervisor replacement (5.2.1)

[AlexFer]

Hello experts,

I have 4x 7010 R5.2.1 (earmarked for upgrades to R6.2.x), each with a single Supervisor 1, 3 (plus default) VDCs (with one configured for OTV); with a pair in each OTV site. The CMP in one is faulty (but CP is fine) and RMA Supervisor replacement has arrived.

I have no experience with Supervisor replacement and I have no test/staging chassis to license and pre-provision the replacement.

I can see two previous discussions - Nexus 7000 single supervisor replacement & nexus 7000 supervisor replacement but I'm hoping for more specific advice and/or (most preferably) sharing of a proven procedure.

Cisco TAC has advised that I provision dual Supervisor redundancy, insert the replacement Supervisor and let it sync, remove faulty Supervisor and then un-provision Supervisor redundancy. However, I have no experience with dual Supervisor redundancy and feel it'll add extra complexity.

Since there are two 7010s per site, my feeling is that I should proceed with Direct Replacement method (with cues from Replacing a Supervisor Module in a Single-Supervisor System). Method's documentation specifies "Supervisor 1 Release 5.2(3a) or later" so I'm apprehensive about its applicability to my R5.2.1.

Advice (and/or sharing of a proven procedure) much appreciated.

R's, Alex

[Mark Malone · Accepted solution]

Well i can see that TAC are trying to keep your switch up by the sounds of it , a dual sup would prevent any outage , either way on single sup your going to be down which is unacceptable for some companies maybe thats why they suggested it , makes sure you have backed up everything to USB ,new sup will not come with an old image either , licenses is stored on the chassis so you should be good with that shouldn't have to regenerate it , if the 2 7Ks are running ion VPC design in X style your probably fine wont loose anything unless there are orphan connections , dual sup is easy really , it syncs up just transfer image and config off working sup onto second sup and then failover by pulling active or using the cli but its down to preference

[AlexFer]

Hi Mark,
after walking through the HA documentation, it seems NX-OS is predisposed for dual Supervisor operation. - pretty much everything is automatic. So, as per TAC's recommendation, indeed, temporarily going dual for the sake of RMA seems preferable.
R's, Alex

[Mark Malone]

Great , well hope it goes well for you, if you run into issues just post what you see , we should be able to assist