Re: BTRFS errors following bad SATA connection

[Roman Mamedov <rm@romanrm.net>]

On Fri, 3 Jan 2025 22:24:19 +0100
Victor Banon <banon.victor@gmail.com> wrote:

> It just occurred to me that I might have a way to have enough space to 
> retrieve most of the un-backed up files. The 4th disk of the RAID was a 
> recent addition (that's probably when the SATA issues appeared), so I 
> have closer to 24 TB of data than 36 TB. So I'm thinking I may be able 
> to leverage up to 2 extra disks. Can I get your input on how dumb the 
> following ideas are:
> 
> A. Delete some files then shrink the array (4x12TB RAID5 aka 36 usable 
> TB -> 3x12 TB RAID5 aka 24 usable TB). Remove 1 drive, use it for extra 
> storage.
> 
> B. Degrade the array (4x12TB RAID5 aka 36 usable TB -> degraded 3x12 TB 
> RAID5 aka 36 usable TB). Remove 1 drive, use it for extra storage.
> 
> C. Do both (4x12TB RAID5 aka 36 usable TB -> degraded 2x12 TB RAID5 aka 
> 24 usable TB) . Remove 2 drives, use them for extra storage

If gaining 1 free drive is enough, I'd avoid reshaping and put it in degraded
(--fail, --remove in mdadm), so option B. I remember you said it was not super
important data anyway, and in this state you'd have like a RAID0 of 3
remaining drives, reliability-wise.

> It doesn't seem smart to do any of this while the array is in this 
> precarious state, but I feel like I don't have a whole lot to lose by 
> trying this-- even if it fails catastrophically, the alternative is to 
> do nothing and reformat anyway. But I'd like to have some idea of what 
> to expect before I do that... Is it even possible to shrink or degrade 
> the array?

You'd need to shrink the Btrfs first, and I am not sure if it would let you,
or run into the same transid errors as during other operations.

-- 
With respect,
Roman