Physically move stacked enclosure

[Robert Cook_6]

I need to move a c7000 stacked enclosure to a different rack.  There are 3, c7000's in domain.  I will have to disconnect all the cables.  There are 4 uplink connections in this enclosure.  Is there anything special I need to do other than the obvious of shutting down all the systems in the enclosure?

Regards,

Bob

[Hongjun Ma]

will OA and VC keep the same IPs?

[shocko]

If you are moving all the OA/VC modules and reconnecting to the same management networks then no IP changes are needed.

Note: Backup all your configs regardless.

[Psychonaut]

Make sure to label you connections - I just swapped out two chassis and didn't do that (only have 1GbE for these two).  Then I had problems telling which connection was OA and which was VC, luckily they are all grey cables.... 

[Robert Cook_6]

Thanks for the input.  I'm physically moving the entire enclosure from one rack to a different rack so all the addressing will stay the same.  My main concern is the stacked link (FLEX-10).   I will have to unplug the stacked links.  The I/O should fail over to the redundant ports on the other enclosures, as long as the configuration is set up correctly.  I'm concerned that the servers in the other enclosures will stay on the network. I'm hoping someone had done this successfully!

Regards,

Bob

[Psychonaut]

Yes, as long as you have the OS setup for redundancy and have Smartlink enabled on the vNets.  I have tested FlexFabric in 2 and 4-wide with Windows and ESX - failing the links and the VC with the active connections.  I've also failed individual VC's in the domains and the servers continued to stay up.

[Jim_SAN_DR]

If you use the OA feature to identify the "Rack" or labeled cabinet, you will want to update that confguration item too.

[Robert Cook_6]

I did the move yesterday. The configuration was there were uplinks to the network on enclosure 1 (D1) and enclosure 3 (D2).  I was moving enclosure 1.   Shutdown all the blades then unplugged the enclosure.  Pinged servers on the other 2 enclosures and all were still on the network.  Unplugged the Link cables and still could ping other servers.  Moved the enclosure.  Reseated all the modules in the enclosure.  Plugged in all the cables then plugged in the enclosure power.  Virtual Connect Manager was running on the enclosure I moved.  It took about 10 minutes before I could log into Virtual Connect Manager.  It showed all links were up.  Powered up the blades one at a time.  All came back up. 

The key to doing this is to make sure that all servers are configured with redundant network links.  That is to say all servers had network connections through D1 and D2.

Regards,

Bob