Reboot after update ends in "Emergency mode" - Logs of no help...

Hi again!

Have a TW 64bit install (plain vanilla with multimedia) doing fine for more than 2 years, was updated 2-3 weeks ago without problems, but after yesterdays “zypper dup” the reboot ends always in “emergency mode” (after seeing the arrows hunt on the infinity symbol, normally the mouse cursor arrow would appear next and then the login screen, so apparently the graphical system crashes).

What I have done so far:

  • Tried to boot in GRUB2 the older kernel and/or recovery mode of both kernels shown. Always ends in “emergency mode”.

  • Updated a second TW install on the same computer (different HDD), worked fine, so no hardware issue, apparently. - Mounted the HDD with the problematic install on another TW computer, without any problems I could access all partitions, so apparently no problem with the HDD (dual-boot form this HDD with Win7 also working fine).

  • Screened the log (systemctl …, as proposed at the emergency promp), ends with lots of “start-up result is RESULT”, but no apparent errors. . “journalctl -b | grep -i “fail” or “error” finds no hint in my opinion, as “fail” is only for ASPM and for “usr/bin/setfont” (exit status 71) and “error” for ASPM, PIO_UNIMAPPCLR: Input/output error” and for setting fonts.

I have no idea how to debug this, where to start…

Hi,

I usually encounter the emergency mode when something’s wrong with mounts. Do you have any network dependent mounts in /etc/fstab? Try to mount only the devices required by the system at boot time and then mount the rest manually, see if that works. Are you using LVM? Maybe something’s wrong there?

Regards

Just checked (again): No, nothing in fstab besides swap, / and /home…

No LVM in use here…

Is there any way to read the systemd logs (/var/lg/journal…) when the HDD is mounted on a different computer?

PS: Fogot to mention: / is EXT4, /home is XFS… So no BTRFS involved here.

The system always “says” something when entering emergency mode. Like “one or more filesystems could not be mounted” or something like that. Can you show that message?

Hi!

To my total surprise there is no such error message:

You are in emergency mode. After logging in, type “journalctl -xb” to view
system logs, “systemctl reboot” to reboot, “systemctl default” or “exit” to boot into default mode.
Give root password for maintenance
(or press Control-D to continue):

I have real problems to extract some info from the booted system in such a “vegetative state”… Could take a photo, but how to insert here? Have no pic hoster currently in use.

PS: If I try to boot to default mode, the infinity symbol hunting the arrows appear, after LONG time, the arrows dissapear, the infinity symbol changes colour from white to blue and “Tumbleweed” appears underneath the infinity symbol. I let it sit there for hours but nothing will happen in this state.

I can’t switch to any text console, btw…

Thanks for reading…

paste.opensuse.org , no need to login, the top right has an Image button, click it, upload your image, and post the link here.
A thing to try if possible: connect the drive to a different SATA port. The drive itself seems OK, since mounting on another machine works.
Another option to try, is to run YaST’s bootloader module in single user mode, if you can login / access the system that way.

Works! SUSE Paste

I tried different SATA adaptors and ports (eSATA, USB,…) makes no difference. The HDD (WD red 1 TB 2.5") is only 2.5 years old and only booted from time to time. SMART throws no errors or signs of deprecation

https://www.cyberciti.biz/faq/grub-boot-into-single-user-mode/ …would this “Append “S” to the kernel parameters” work with TW and Grub2 to enter single user mode?

Yes. You can also use " 1 " the number. If you don’t manage that way, it can still be done from an install medium

Booting to single user mode results in the same emergency mode… Any ideas where to look for relevant error messages? I don’t have the slightest idea what is going on here… Would it be an option to reinstall TW to the root partition and spare the home partition? As a final solution…

And this emergency mode doesn’t ask for the root password? Could you post a photo of the screen on paste.opensuse.org ?

Absolutely identical to first photo, including the question for root password, tried with “S” and “1” added to the “linux…” line in the grub2 entry… Or is this wrong?

Try this, once in emergency mode:


sudo systemctl mask systemd-vconsole-setup.service 


, then reboot. From the messages that looks like the culprit.

…tried it, got a reply, something with “…dev null” in the end, but the next (and the over next…) reboot also ends in emergency mode…

And the other options the output suggests, like


systemctl default

don’t work ?

Nope, after systemctl default the infinity symbol starts hunting the arrows, then changes colour from white to blue and the “Tumbleweed” is added underneath the infinity symbol, but then nothing happens… I let it sit there for more than half an hour now, but nothing changes.

Provide full output of “journalctl -b”.

Have no idea how to get it out of this box! That’s why I asked if it is possible to mount the HDD on another TW and read the systemd journal :slight_smile:

Just copy it on USB stick. Of course you will need to manually mount it.

That’s why I asked if it is possible to mount the HDD on another TW and read the systemd journal :slight_smile:

Yes, if you read “man journalctl” you may notice “-D DIR” option.