I performed the change recently, and this is how I did it.
It didn't go as well as expected, but it ended up working. The first step was to de-configure everything using the VC-Enet 1/10 module in bay 1, this was more than just removing all the external ethernet links from the vNet configuration, this also meant we had to edit all the server profiles to set network cards 1 and 3 to have "Unassigned" as their network. This is because the LOM NICs are assigned to the VC-Enet in bay 1 (way to go HP for poor redundancy design!). I have firmware 2.3, so that allowed me to change the Server Profiles without rebooting the servers. I also ensured that I updated the firmware on the Flex-10 modules before I started this process. I then performed a failover to the standby module (Tools, Reset Virtual Connect Manager), and then I could physically unplug the VC-Enet module in bay 1. Then a few seconds later, I could navigate to the bottom of the screen in Hardware Overview to "Remove from Domain".
Then I could insert the Flex-10 module - but it was reported as "unsupported". I re-configured the VC-Enet 1/10 module in bay 2 to have no connections, and then removed it from the chassis.
Then this is where I came unstuck. The Domain Controller that performs the LDAP authentication was running as a virtual machine on the blades. It could not be contacted to let me log in to the Flex-10 module in Bay 1. When I inserted the next Flex-10 module into bay 2, the Virtual Connect Manager logged me out and I could not get back in. I was able to log in with the default administrator log in from the iLo type ticket on the module (the default password), and then I was presented with the first time wizard.
I was given two options - restore from backup (no thanks, everything is different), or import settings. It took a good 10-15 minutes to import, and then it took me through the rest of the wizard to set up Ethernet networks and SAN settings.
It took me a while, and the virtual machines that had been running on the blades were mostly OK - some had rebooted, but most were still online. We also had some firewall maintenance going on, so most machines were doing nothing, and only around 5% reported problems writing to the disk (SAN connections were down). I had powered down 5 of the blades, and the problem was that these all got different WWNs for their SAN connection (the running blades were unchanged), but we re-presented the LUNs from our EVA 8100 SAN and all came right. I don't know if these blades that were powered off also got new MACs.
It may have been impossible to do this (even if I had non-LDAP login to the Virtual Connect modules) because it appears that when the "incompatible" Flex-10 module was plugged in to bay 1, it did not get the Virtual Connect Domain settings applied to it. The blades remained up and running, and if the SAN configuration had remained, then for me it would have been a minimal impact outage (because I scheduled it at the same time as a network outage), but as it turns out it was not as bad as I thought (powering down all servers would have been impossible).
