Automatically clean disks

[josecastillolema]

There is an automatedCleanMode field on the BMH that is normally set to disabled for assisted integration but can be set to metadata to wipe partition tables.

[josecastillolema]

cc @akrzos @jtaleric @rsevilla87 @venkataanil @ebattat

[akrzos]

We don't have a bmh object in this context, what is being suggested here?

[josecastillolema]

So the issue that we are trying to solve is that as you know QUADS only wipes the primary disk of each server, so sometimes the leftovers from previous allocations can get in the way.
Yesterday you showed me this is possible to do via ignition files, i.e.: https://github.com/redhat-performance/jetlag/blob/main/ansible/roles/configure-local-storage/templates/worker.ign.j2#L13
The BMH approach was suggested by the Assisted Installer team, which one do you feel is best suited for jetlag?

[akrzos]

Well since we do not have a full cluster here with all the typically components to provision clusters, we can not use a bmh object to clean disks, perhaps the assisted team didn't completely understand the context of your original question. I also do not think cleaning the disks is really within the scope of jetlag since hardware changes in the lab and do we really want to support building a robust script that wipes all disks and has the potentially to accidently earse someone's results if they do not fully understand what they are doing with it. We really need the lab team to hand off hardware in a reproducible and consistent manner. I have spoken to them about this issue before and I believe they stuck by only cleaning the install disks. What I have already added was due to absolutely necessity to achieve the testing I am performing and is rather specific to that testing. Alternatively perhaps a separate project dedicated to solving this issue would be more appropriate since in my opinion cleaning disks is far from jetlag's original scope.