:: Re: [DNG] Some RAID1's inaccessible…
Top Page
Delete this message
Reply to this message
Author: Daniel Abrecht
Date:  
To: dng
Subject: Re: [DNG] Some RAID1's inaccessible after upgrade to beowulf from ascii
Am 2021-11-09 19:56, schrieb Hendrik Boom via Dng:
> I upgraded my server to beowulf.
>
> After rebooting, all home directories except root's are no longer
> accessible.
>
> They are all on an LVM on software RAID.
>
> The problem seems to be that two of my three RAID1 systems are not
> starting up properly. What can I do about it?


Do you still know which devices contained the raid?
I'd first check out the metadata on them:
```
mdadm --query --examine /dev/sdXY
```

Ideally, the ones for the same arrays should have the same Version,
Array UUID, and a number for Events which doesn't differ much.

I'm not much of an export with raid myself. In fact, I already hosed one
of mine once (I followed some tutorial which suggested using "mdadm
--create" for a certain thing, but it was wrong and I should have used
--grow instead).

Anyway maybe try explicitly assembling the devices as a new raid device.
There is also a --readonly option that should avoid it doing anything to
the array, and a --readwrite option, which could be helpful.
```
mdadm --assemble /dev/md5 --readonly /dev/sdXY /dev/sdXY
```

If that works & the raid device works fine (maybe try mounting it), you
can make it readwrite using
```
mdadm --readwrite /dev/md5
```

Alternatively, you could also try --re-add to add devices to an existing
raid again.

I'd avoid doing anything with --create. I think that would be
destructive.

If it does stuff like resyncing after that, you can already use the
device, but maybe wait with a reboot until it finishes.
If you get it working, try doing the `mdadm --examine --scan` thing
again / recreate /etc/mdadm/mdadm.conf. Maybe also recreate the
initramfs using `update-initramfs -u`, just in case.
And if all that works out, you may try rebooting to see if it detects it
again.

If things don't work out, maybe check for error messages / look at
dmesg.

Regards,
Daniel Abrecht