Hello devs! How is everybody doing?!
We have a big mission in front of us and I would like to ask for some guidance.
In our environment we have 2 Viprion chassis with one B2100 each. The chassis run TMOS v12.1.1 HF1 and they are configured as a device cluster (HA).
I read two articles while researching for this:
The idea of our mission is to perform an as-is migration, so no code upgrades whatsoever. I believe both these articles DO NOT apply to my case since our code is 12.1.1 and so we do not have the platform-migrate flag.
The first article I pointed out talks about "Loading a UCS From the Source Platform Directly on the New Platform" and after says that as a con it "Requires hand editing of configuration files". What exactly would I need to edit?
So the question is: Is there an article or guide that I could follow in this particular case?
so to be sure you now have configuration (LTM, ASM, ...?) on a viprion which you want to move to a vCMP guest on another viprion?
then that remains a different platform, even though the hardware at the bottom is the same the vCMP guest will have for example a different interface setup then the vCMP host.
so that is your main issue. when you load the UCS from a current BIG-IP to another it might fail due to the config on mainly interface level being different. if you have the time easiest might be to just try, load it, see where it fails and change that in the config file, then try to load again.
"so to be sure you now have configuration (LTM, ASM, ...?) on a viprion which you want to move to a vCMP guest on another viprion?"
Yes! We will disable the standby chassis, enable vCMP on it, provision a new vCMP guest and then we would like to restore the old configuration on this vCMP guest. We're moving from physical to virtual. That's it.
clear, well as pointed out above that could be your approach.
set it up, deploy the the vCMP guest, restore the UCS and see what happens. then modify step by step.
if you have an F5 partner to reach out that would probably be helpful if they done modifying UCS archives before.
Quick follow-up question if I may...
How come when using the platform-migrate flag you have to change the master-key beforehand but if it's a replacement unit (RMA) you just load the UCS with the no-license and it's all good?! This makes no sense to me...
that is also not the case, if you restore an UCS you need to set the master key before to make sure everything encrypted is available correctly.
only exception might be RMA in HA cluster, then you can sync most config to it eventually.