recovery mode gets borked after some time out

Bug #1636503 reported by mike
34
This bug affects 7 people
Affects Status Importance Assigned to Milestone
friendly-recovery (Ubuntu)
Incomplete
Undecided
Unassigned

Bug Description

(Linux mint 18 / Xenial)

after booting in recovery mode from grub.
If you just wait 30-60seconds, a new stream of messages appear and something gets messed up, you can't use recovery mode any more, you have to do a hard reboot.

You don't need to do something in particular for this to happen, simply some time out occurs.

-------------------

more precisely, one of the messages say(copied manually):

"[FAILED] Failed to start Console System on Startup logging.
see 'systemctl status console-kit-log-system-start.service' for more details.

It seams, it tries to launch a second recovery mode on top of the previous one after the error.

------------------------

I confirm, this is happening on both my PC linux mint 18 64bits, laptop linux mint 18 32bits and some one else from a forum on a ThinkPad X201 with Mint 18 KDE 64-bit.

On my systems, both are fully up to date, 32 and 64 bit. Kernels don't seam to change anything.

systemd version 229-4ubuntu11

Revision history for this message
mike (mike5346874) wrote :

affects my PC, laptop and some one else.

Changed in systemd (Ubuntu):
status: New → Confirmed
Revision history for this message
mike (mike5346874) wrote :

still borked with systemd 229-4ubuntu13(xenial)

-_- guys, isn't this a serious bug?
You are left with unusable recovery mode....

Revision history for this message
jas mann (dotmarker) wrote :

I experience the same issue as OP after a clean install of Ubuntu 16.04 4.4.0-72.

Nothing can be done in recovery mode that takes longer than one minute, otherwise some timeout fires, console spews several screens of messages indicating processes being stopped, and it loops back into the recovery mode menu; however now there is no keyboard or at least no keys respond other than ESC, thus no way to navigate the menu even if were now operable.

Hitting the ESC key at this point causes the following text to overwrite the first menu option:
  "sulogin: input overrun at /dev/tty1t"
But perhaps that is just an artifact of the unwanted and unrequested shutdown of system daemons.

Nothing more can be done at this point, a hard reboot is necessary.

This is repeatable, i.e. the same symptoms occur at each entry to recovery mode.

Revision history for this message
Paris Jones (quantera) wrote :

Can confirm, this bug affects me.

Revision history for this message
StewPedassle (anthonie-moll+bugs) wrote :

Confirmed – mostly.

Running: Mint 18 Cinnamon 64-bit
Kernel: 4.4.0-78
Laptop: Lenovo Thinkpad Yoga 12

Backstory:
I have a dual boot with windows 10, which I evidently messed up because I would get a (busybox) prompt every once in a while. Generally running a manual fsck would fix that. Last night, had problems logging in (we're talking a couple minutes to boot to login, then hours to actually get past the login screen. I went ahead and tried to go into recovery mode tonight and fix it.

First glaring issue was that I kept getting a message that I had no swap. Googled a bit found a solution to try was to reformat the swap partition and reset /etc/fstab. I was able to do this with only occasionally running into this "double screen" bug.

Great. Reboot. Same issue logging in.

When I went into recovery mode to try again, I am now getting this option almost every time. In fact, my initial boot includes "[ OK ] Started Monitoring of LVM2 mirrors, snapsots etc. using dmeventd or progress polling." across the first entry. Because the first entry is where it places the cursor, it appears there may be a timing issue.

Oooop. some crap just happened. Without touching it, it now started giving me a weird output. It starts with "[ OK ] Stopped Run anacron jobs." and has a whole misaligned lists of "[ OK ]" and a few "[FAILED]." The failed are "Failed to start Set console font and keymap" and "Failed to start Console System Startup Logging.

After all of this, I get "Welcome to emergency mode! After logging in, type "journalctl -xb" to  . . . ."

Gave root password as requested, and now we're to the second opening screen. Lighter bluew, "root@LT-A ~#" is now on the first line, but hasn't covered the "Resume normal boot" portion.

Does this help? Probably not, but I am getting extremely frustrated.

There are times when I boot and was able to work in recovery mode.

Revision history for this message
StewPedassle (anthonie-moll+bugs) wrote :

It looks like it is an error with the terminal display. I can still type and get commands to go through, but the display is garbled spacing. I can call nano /etc/fstab and get my fstab file without issue. Everything is properly spaced and I can type without issue.

Revision history for this message
mike (mike5346874) wrote :

wait a minute, it doesn't ask the root password....
That's rather bad

Revision history for this message
Dimitri John Ledkov (xnox) wrote :

please retest with 229-4ubuntu21.5

Revision history for this message
Dimitri John Ledkov (xnox) wrote :

bah

Changed in systemd (Ubuntu):
status: Confirmed → Invalid
Changed in friendly-recovery (Ubuntu):
status: New → Incomplete
Revision history for this message
Dimitri John Ledkov (xnox) wrote :

nothing to do with systemd, purely a friendly-recovery induced issue which should have been fixed across all stable releases now. Please retest, if this is still a problem for you.

Revision history for this message
mike (mike5346874) wrote :

I can confirm, that for 18.04 32/64bit it's gone.
I don't know for 16.04 .

Mathew Hodson (mhodson)
no longer affects: systemd (Ubuntu)
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.