Re: Rebuilding 24TB Raid5 array (was btrfs corruption: parent transid verify failed + open_ctree failed)

[Marc MERLIN <marc@merlins.org>]

On Mon, Jun 13, 2022 at 02:29:30PM -0400, Josef Bacik wrote:
> Alright so we have the missing csum things, which can be fixed with
> 
> btrfs rescue init-csum-tree <device>

Took a while, but it worked now, thanks.
 
(gdb) run rescue init-csum-tree /dev/mapper/dshelf1
Starting program: /var/local/src/btrfs-progs-josefbacik/btrfs rescue init-csum-tree /dev/mapper/dshelf1
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
FS_INFO IS 0x555555652bc0
Couldn't find the last root for 8
FS_INFO AFTER IS 0x555555652bc0
processed 21904289792 of 21907435520 possible data bytes, 99%doing close???
Init csum tree finished, you can run check now
> 
> After that you can mount the fs and run a scrub, because you have
> copies of metadata that are fucked, but copies that were ok, which is
> why I disabled the block error warnings since they were confusing me.
> Scrub should go along and fix all of that up.

gargamel:/mnt/mnt# btrfs scrub start -B .
running that now, I expect it will take a while.

> The next thing is to fix the fs errors, which I imagine will cause
> other problems like not being able to find directories and files and
> such.  Once we've gotten the csum tree and the scrub done we can
> tackle the remaining fs error problems which should be less terrifying
> to mess with.  Thanks,

would that be check --repair ?

Thanks,
Marc
-- 
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
 
Home page: http://marc.merlins.org/