Fedora CoreOS boot failure

Ask Fedora
1

Hello,
I’ve setup fedora coreos with the default partitioning scheme and luks encryption + tpm. I can’t access the services running on that system since the 8th Jan (yesterday).
The system was in “emergency mode” as I hooked it up to a monitor. It must’ve rebooted for any reason.
I tried multiple reboots and always landed back in “emergency mode”, but every line I catch says [ OK ].
The last lines I’m able to see are (static on screen)

[    9.608738] i915 0000:00:02.0: [drm] *ERROR* conflict detected with stolen region: [mem 0x2f800000-0x8f7fffff]
[  OK  ] Finished systemd-update-utmp.servi.ecord System Boot/Shutdown in UTMP.
[  OK  ] Finished audit-rules.service - Load Audit Rules.
[  OK  ] Started systemd-resolvd.service - Network Name Resolution.
[  OK  ] Reached target network.target - Network.
[  OK  ] Reached target nss-lookup.target - Host and Network Name Lookups.
         Mounting var-lib-nfs-rpc_pipefs.mount - RPC Pipe File System...
[  OK  ] Mounted var-lib-nfs-rpc_pipefs.mount - RPC Pipe File System.
[  OK  ] Reached target rpc_pipefs.target.
[  OK  ] Reached target nfs-client.target - NFS client services.
[  OK  ] Reached target remote-fs-pre.targe.reparation for Remotr File Systems.
[  OK  ] Reached target remote-cryptsetup.target - Remote Encrypted Volumes.
[  OK  ] Reached target remote-fs.target - Remote File Systems.
[  OK  ] Reached target remote-integrityset.Remote Integrity Protected Volumes.
[  OK  ] Reached target remote-veritysetup.. - Remote Verity Protected Volumes.
You are in emergency mode. After logging in, type "journalctl -xb" to view system logs, "systemctl reboot" to reboot, or "exit"
to continue bootup.

Cannot open access to console, the root account is locked.
See sulogin(8) man page for more details.

Press Enter to continue.

Please note, the above is almost the 1:1 output. Things like remotr-veritysetup.. are not my mistake.

I was able to trace back the first line of the output, to a error of the iGPU (don’t think this is the problem).
The nfs part could be critical. I did setup a nfs mount on /mnt/backup via the fstab file.

Kind regards

2

Hello @japecol457 and welcome to :fedora: !

Have you tried booting the previous OS deployment following the instructions in Temporary Rollback from Manual Rollbacks?

3

Hello Hristo Marinov,
thank you for your response.
Yes, I did try that although to no success (I selected ostree:1). I got the same result as described in my first post.
I suspect the nfs mount being the problem. I’ve got enough experience with linux to confidently unlock the luks partition and mount the filesystem via terminal. I’d just need someone to walk me through the ostree part, to remove the mount entry from the fstab. (Only if you think that is worth a try)

Kind regards

4

I don’t have much experience troubleshooting LUKS, but if no one else chimes in, I can at least try to reproduce it. Can you post the exact configuration in /etc/fstab that you think may be causing the issue?

5

Yes, I can somewhat recall it from memory.

  1. Create /mnt/backup (/var/mnt/backup),
  2. add nfs.local:/backups /mnt/backup nfs rw 0 0 to /etc/fstab,
  3. systemctl daemon-reload and reboot
6

You should be able to mount the root partition of the system from a LiveCD and then look for /etc/fstab in /sysroot/ostree/deploy/fedora-coreos/deploy/<hash of the deployment>.0/etc/fstab.

Make sure to setup all NFS mounts with nofail as option (fstab(5) - Linux manual page & https://www.freedesktop.org/software/systemd/man/latest/systemd.mount.html#nofail) to not block your system if the NFS mount point is not available.

7

That worked, thank you.

1 Like