curtin-dev team mailing list archive

[Dan Bungert <mp+426651@xxxxxxxxxxxxxxxxxx>]

Here is the input data that I used as part of this MP:
https://paste.ubuntu.com/p/TfpvN2bdVB/

> Should we preserve [last-lba and table-size] too?
This line of questioning is good, thanks for raising the topic.
I preserved only the values that were different in that dump.

last-lba: "Specify the last usable sector for GPT partitions."
sfdisk appears to be doing a good job on this, and presumably this is low risk due to the trailing 1MiB section at the end of the disk.  But suppose some software reports a lower value?  This sounds like it might be having magic data in unpartitioned space, and attempting to use the last-lba value to declare it an unused zone.  I think we should preserve last-lba.

table-length:  "Specify the maximal number of GPT partitions."
sfdisk defaults this to 128 and doesn't include this in the --dump output unless it has a non-default value.
This could be like last-lba, where another OS might decide to use a different value?  And be grumpy if different?  Safer to preserve, though I would be surprised if the value is ever different.

> Should the disk action allow setting any of the label-id/first-lba/last-lba/table-size options?
I suspect some OS are identifying the disk by label-id.  Setting a specific value might be useful but seems pretty niche.
I think the others are too esoteric to try to be proactive about.  As you say, we can add if asked.

I support including label-id as a configurable item, but suggest we not bother with the others.
-- 
https://code.launchpad.net/~dbungert/curtin/+git/curtin/+merge/426651
Your team curtin developers is subscribed to branch curtin:master.