Hello guys,
I have a pair of Nexus 9508 in vPC topology that need a downgrade, the software on their is pretty buggy and has a couple of deal breaking features.
The problem is that the 9508 is a good platform switch with line cards and supervisors :) , you hear me coming.
The cluster they ordered here only has 1 supervisor on both chassis, this eliminates the ISSU upgrade/downgrade.
So i have to think of another plan to upgrade/downgrade.
Since it's a vPC cluster i always have a separate path
What do you other nexus specialist recommend?
Scenario 1: I just downgrade/upgrade one member in the vPC (secondary first) and when the switch reboots, the traffic will fail-over to the primary unit. The primary unit will lose all connectivity and there should be a clean fail over and recovery.
Most likely i will have to check the hold timer, as the switch may come back and be reachable via de peer keep-alive but not yet ready to forward traffic?
Scenario 2: I force a manual fail-over in the vPC, i isolate the node that i want to upgrade. I shut the peer-keep alive and vPC peer-link. This way the traffic is already fail-over before i start the upgrade. (we have an oob connection into the switch).
I have not yet seen a switch version dependency inside vPC parameters? So both of them running a different version shouldn't be an issue? During the upgrade
No comments:
Post a Comment