Misleading warning message about unusable space for legacy madm arrays

[alexsee75]

In mdadm/Manage.c line #805 there is no check for legacy arrays. For legacy (non-persistent) arrays the restriction to 4TB does not apply, so this warning message should be suppressed in that case (IMHO by adding ... && array->not_persistent == 0 to the condition on the if in that line - but please check, I haven't looked into the code too deeply)

It's a minor issue but likely to confuse those (probably few) people still using legacy mdadm arrays.

[mtkaczyk]

Hello @alexsee75
Just to be clear, do you mean this check:
https://github.com/md-raid-utilities/mdadm/blob/main/Manage.c#L805

For legacy (non-persistent) arrays the restriction to 4TB does not apply

You are not clarifying why it is not applying. We have super0 and this is super0 limitation, not persistent/not-persistent arrays. Could you please elaborate?

Simply, I don't know this problem :)

[alexsee75]

Hi Mariusz, I've been using a 16TB legacy raid1 array with 2 units for about a decade and when I recently had to replace one disc, I got exactly that error message shown at that line when adding a new disc. It worked with --force and I could access all 16TB of the new disc. It _might_ be that this has already been fixed. I'm using... mdadm - v4.2 - 201-12-30 - Debian 4.2-5 ... on Debian 12.6. Should I download the source code of this version from the Debian repositories and check the appropriate line or can you check this on your end? Alternatively the error message could appear somewhere else, but it was definitely a mdadm add call where I got this error message... Best, Alex Am Mon, Jan 27, 2025 at 12:53:20AM -0800 schrieb Mariusz Tkaczyk:

…

Hello @alexsee75 Just to be clear, do you mean this check: https://github.com/md-raid-utilities/mdadm/blob/main/Manage.c#L805 For legacy (non-persistent) arrays the restriction to 4TB does not apply You are not clarifying why it is not applying. We have super0 and this is super0 limitation, not persistent/not-persistent arrays. Could you please elaborate? Simply, I don't know this problem :) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.*Message ID: ***@***.***>