SOS - Where is mdadm.conf?! I really really need to edit mdadm.conf on dom0

[brick]

Long story short I had a drive failure, now all my RAID arrays incorrectly show up as "raid0 inactive". Apparently one way to fix this is to manually change the arrays to the correct levels in mdadm.conf, but I can't seem to find that in my dom0 with the `locate` command.

Please help. I really need these arrays back. My damn fedora-34 template is there so I can't even use sys-net

[Mike Keehan]

Try "find / -name mdadm.conf -print

It's in /usr/lib/tmpfiles.d/ on my laptop, but I don't have raid.

Mike.

[brick]

EDIT: And no I can't just rebuild the arrays. I need to recover the data if it's still there, and something tells me it just might be.

[brick]

I created my own mdadm.conf and put it in /etc/mdadm.conf as well as in /etc/mdadm/mdadm.conf but Qubes doesn't seem to be reading it from either of those places. Does Xen do things differently or something?

On Thursday, April 13, 2023 at 1:09:03 PM UTC-5 brick wrote:

[brick]

I don't think that's it. It's only one line long and has no reference to my arrays. I tried to put mine there anyway and rebooted but my arrays are still all raid0

[brick]

Actually yeah that's not it at all. That's a `systemd-tmpfiles` file whatever the fudge that is, nothing at all to do with mdadm

[brick]

Any Qubes devs/experts please...? I asked on linuxquestions as well and according to them, mdadm will try to load whatever file it listed under `man mdadm.conf`, which in the case of dom0 is at /etc/mdadm.conf... but that file did not exist until I put it there so this is Qubes specific. Where does Qubes OS's mdadm get its settings from at boot? I need to change those settings because they are corrupted... I literally have nowhere else to go but this mailing list. PLEASE HELP

[brick]

I fixed the problem. Turns out there's a --config option. I just forced it to use my config once which was enough to repair the array that was broken. I rebooted and all my stuff is back :)