Re: MBR -> GPT migration questions

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Wed, 04 Sep 2024 08:58:39 -0700
    From:        beaker%sdf.org@localhost (beaker)
    Message-ID:  <66d883af./Y0ZU4zyrRZebOmC%beaker%sdf.org@localhost>

  | Some questions:
  | 1) is MBR -> GPT migration as simple as running '$ gpt migrate /dev/rwd0'
  |    or is something else involved ?

As far as the drive layout is concerned, that should do it, but you'd
probably want to go and add labels to the partitions.

You also need to fix fstab, and anything else which knew the wd0a type
names for partitions, as they will turn into dkN (for some N).

  | 2) is there any benefit to doing this migration ?

You get device unit number independence, once the partitions are
labelled, you can use NAME=label almost everywhere, and nothing will
care whether it is wd0 wd1 or is now in an external USB case and has
become sd3.

But this is purely optional for drives that small - you only really
require GPT for drives above 2TiB (for 512 byte sectors - bigger if
the exposed sector size is > 512 bytes, but that is rare these days).

  | 3) does my disc look like it meets the before & aft partitions space
  |    requirements (see below) ?

Yes, the 63 sectors at the start are plenty for the primary label, which
will typically occupy 34 sectors at the start (1 for the PMBR, 1 for the
GPT header, and 32 (though that's just the default, you can adjust it up
or down) for the partition table - 32 sectors allows up to 128 partitions
(yes, 4/sector).   Then 33 sectors at the end for the backup (no PMBR sector
needed there) and you have plenty of free space for that.

kre

ps: I'm not sure I'd bother unless this drive gets moved around a lot.
If it simply sits there, for a drive that small (but yes, I used systems
when a 7MB drive was too big and needed to be partitioned so its sectors
could be addressed, 12 bit block numbers, but that was long long ago)
I'd just leave it alone.