Hi Jan, Jan Damborsky wrote: > Jens Deppe wrote: > > [...] > >> >> Yes, 6488 would solve this problem. However there may well still be a >> need to manipulate the root disk before installation starts. Another >> use case is the creation and setup of partitions which the installer >> does not support. > > Hi Jens, > > do you have particular partition setup scenario at hand ? > I am asking, since in general we are trying to address > common use cases and we would like to make sure we are > not missing the important ones.
Sure. Although I think already addressed we originally needed the option to be able to create a FAT partition as the first partition on the disk. Another use case is that on some systems one has to leave a small amount of disk space free for the hardware RAID controller to write it's config data. This space has to be outside of any fdisk partitions and has to be at the end of the disk. Perhaps an RFE might be for the partition size to be given as a negative number where the value indicates how much space to leave free. > > I agree that there could be marginal cases which are > likely not to be addressed, as they are not worth the > potential effort (80/20 principle). > > That said, at this point there doesn't seem to be viable > way to obtain the information about default target disk > without running the installer. The algorithm itself is > straightforward, but it is currently integral part of AI > engine and processes internal data structures filled in > by the Orchestrator with the information collected by > Target Discovery service. > If it's not too onerous, would you mind describing the algorithm, or pointing me to any documentation which does so? > Is this something which you would need to solve immediately > or for a longer term ? If I can replicate the current algorithm I can work around it for now. > > If latter is the case, please feel free to take a look at Sarah's > CUD proposal, I think it could give you some hints how things are > to be designed, especially dry-run and checkpointing capabilities > might be potential points of interests.' Thanks, I'll do so. --Jens > > Thank you, > Jan >
