Relocating one of the fabric interconnect 6200 without downtime

[SR13]

I am trying to relocate one of the Chassis(2) and subordinate FI to different rack in same location without downtime.

ENVIRONMENT:

FI model: 6200 (EOS)

Chassis Model: UCSB-5108-AC2

2 FI's connected to 2 chassis (blade servers) with high availability in place

As these FI's are EOS i am not able to contact TAC support either.

Steps completed till now: 

-VMs on servers are migrated, servers are powered OFF in chassis 2, decommissioned the chassis-2, and unconfigured the ports 5&6 on FI B(connected to chassis 2).

-Failover is in place

-Backup of ucsm config

I need help with next steps on FI B..

-Do we need to Evacuation traffic now(Configure Evacuation)? or disable Uplink Ports (LAN & SAN) on FI-B and then remove physical cables? 

[Kirk J · Accepted solution]

Greetings.

One thing to keep in mind, is that if you have an FI offline, while you can get discovery of moved hardware to usually complete, the service profile association will usually fail because the vnics related to the down side/path will be unavailable, and you'll get a insufficient hardware/resources error.

I have seen customers successfully maintain uptime with a 'similar' scenario,  with UCSM managed rack servers, and they were able to run long fiber to reconnect the servers and split location FIs.

As long as the remaining chassis/blades (not being moved/decomissioned) can remain up and connected to remaining FI, you should be ok.

Once you get offline chassis moved, along with FI, and the move FI is reconnected/cabled, you should be able to recomission-discover the moved chassis/blades, and then bring those back online.

At this point, if you are planning on relocating/moving the remaining FI, you should be able to repeat the process.

Traffic evac, will down the ports and vnics related to that FI, but so too will disabling the uplink ports.

Kirk...

[SR13]

Thank you Kirk.