| zph1nx | is there a way to have the storage part pick the first biggest possible disk as part of the path: option when partititioning? trying to set up an automated ubuntu installation, which has the partition path set to /dev/sda, is there anyway to "wildcard" this? | 15:27 |
|---|---|---|
| zph1nx | is there a catch all for disk path under the storage key? | 15:32 |
| zph1nx | https://pastebin.com/eBAZLuRb is the user-data | 15:36 |
| dbungert | zph1nx: I think you want `match: {size: largest}` - https://canonical-subiquity.readthedocs-hosted.com/en/latest/reference/autoinstall-reference.html#disk-selection-extensions - this is more autoinstall then cloud-init though, we should move to #ubuntu-server if there is more to discuss | 15:40 |
| zph1nx | dbungert: im there, the problem here if you've noticed my user-data is that i have a very specific storage layout. and the problem atm is that autoinstall is bugging out on me as soon as /dev/sda is not a valid path | 15:56 |
| zph1nx | my layout dont really follow the preordained simple gpt -> lvm/direct layouts | 15:58 |
| dbungert | right, so the idea is that you remove the `path` declaration from the `type: disk` object and use the `match` instead | 15:59 |
| zph1nx | to * ? | 15:59 |
| dbungert | { ptable: gpt, match: {size: largest}, wipe: superblock, preserve: false, name: '', grub_device: false, type: disk, id: disk-sda } | 16:00 |
| zph1nx | so the logic would be that path would get changed to match: {size: largest} | 16:01 |
| dbungert | right | 16:02 |
| zph1nx | and type_ disk, id: largest would be the proper naming also | 16:02 |
| dbungert | the value of id is arbitrary, just needs to be consistent with the other objects referring to it. | 16:04 |
| zph1nx | so the id can be anything | 16:09 |
| zph1nx | as long as the other partitions reference it carving in lvms? | 16:09 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!