Removing an enclosure from VC Enterprise Manager (VCEM) and retaining the profile.

[chuckk281]

Mike had a customer question:

*******************

I had a customer ask me this of VCEM and removing a enclosure form VCEM.  What is the best practice for this???

---------------------

I have been asked to remove all of our Virtual Connects from VCEM in our enclosure.  I know, the operative question is “why” would anyone want to do that, and it is a long answer, but effectively we recommended against it as the local engineering team, but were overruled and asked to do it anyway.  In the process of removing our first VC from VCEM, I can no longer get any of the blades to talk to the network.  I have recreated server profiles and done several other troubleshooting steps to figure this out and I am banging my head against the wall.

My question to you is simple.  Do you recommend a complete factory default reset and push of new server profiles when a VC is removed from VCEM?  I have created some scripts for pushing down new server profiles via the CLI so I can do that fairly quickly.  I am just wondering if there is something “stuck” in the config that makes this VC think it is still managed by VCEM and that is why it isn’t talking?

Any help you can provide is greatly appreciated.

******************

Reply from Fred:

*******************

Removing a VC domain from its VC Domain Group relinquishes VCEM control of the domain but doesn’t make any changes to the domain that would interrupt connectivity for the servers in the domain (enclosure). Here’s a snippet from the VCEM User Guide on removing a  domain from its domain group:

NOTE: Removing the VC Domain out of the VC Domain Group does not cause any changes to

the MACs or WWNs currently in use in the VC Domain. Any new MACs or WWNs assigned

through VCM after the VC Domain has been removed from the VC Domain Group will be assigned

from the ranges designated when the VC Domain was removed from the VC Domain Group.

Some other change could be causing the connectivity problems you’re experiencing.

*******************

Other comments or suggestions?

[Greg_Hooper]

Thanks for the response, Fred.  I am the customer Mike was so kind to assist with this question.  The interesting thing that is happening to me in this case is that blades with the old profiles and MAC/WWN's from VCEM are still communicating fine on the network.  When I assign a new profile using the new ranges that I defined when I removed this enclosure from VCEM, I cannot get it to communicate on the network at all.  I have verified with the network team that there is no static MAC address port security going on with the Cisco switches but they are using sticky MAC port security.  They tell me that sticky port security does not apply to trunked interfaces so there is nothing blocking on the Cisco side.  I am just trying to rule out anything on the virtual connect side and make sure I haven't missed anything there.  I am unfortunately in an environment that is fixed at firmware version 3.18, so I cannot upgrade at this time, unless I can find a definitive problem with firmware in this case.

To summarize, I have 2 ESX blade servers communicating fine with the old ranges from VCEM.  When I build out a new server and assign a new profile with the new VCM user-defined ranges, it doesn't talk.  I have crawled through the configs and made sure each vlan is properly presented so I am sure I have that piece correct.