[lustre-discuss] EC degraded read after OSS down

[Andreas Dilger]

On Apr 23, 2026, at 07:27, Stepan Beskrovnyy via lustre-discuss <lustre-discuss at lists.lustre.org> wrote:
> 
> Hi everyone! 
> 
> Up existing thread, because no I see no replies but the solution is still needed.  
> 
> Tried  https://review.whamcloud.com/64970 and resync not working at all on this commit(See pictures). 

It looks like that patch is now abandoned, so it was probably outdated?
The current patch series is in flux, with fixup patches being merged
into their parents then abandoned, reordering, etc. so it is hard to
get everything in one place at the moment.  I expect that the core EC functionality will be landed to master in 3-4 weeks.

It looks like patch https://review.whamcloud.com/62757 ("LU-12669 ec: recover data from parity") is at least able to recover EC and passing testing, so may be a suitable place for you to test.  This still does not include later (important) patches to reduce wait time for EC reads to timeout, faster rebuild IO and optimized CPU usage, but it _should_ at least be functional.

Cheers, Andreas
---
Andreas Dilger
Principal Lustre Architect
adilger at thelustrecollective.com