During a recent code update (1.4_3m to 1.4_3y), one of our chassis’s was in a funky state. After some research & consulting with Cisco Advanced Services (yes, we are that lucky), recommendation was to re-acknowledge the problem chassis. Here are the steps we took to properly remove a chassis from UCS, and re-acknowledge it:
- Power off all of the servers in the problem chassis
- Disassociate service profiles for all of the aforementioned blades
- Unplug the power from the problem chassis
- Remove the chassis from UCS (right click, Remove)
- Once the chassis is gone from UCS, unconfigure the server ports for the problem chassis (the ports will disappear from your list of server ports)
- Power on the problem chassis
- Reconfigure the server ports for the problem chassis on each FI, this will cause the chassis to reappear
- Right click on the problem chassis, and acknowledge the chassis
- Once properly acknowledged, start associating service profiles (FSM gives great details on status, as does the Service profile>Server Details>Status details drop down view)
- The problem chassis should no longer be a problem chassis. 🙂
This may seem a little verbose, but this thorough method is proven to work properly. Thanks!
How long does it take to acknowledge the chassis? I have a change in now and this would be my first acknowledgement.
Sorry for slow reply. Empty chassis takes like 3-5 minutes. Can get a little scary as errors can crank up and clear when chassis is discovered.
Hope it went well.