13.2 doesn't boot into X anymore

English Install/Boot/Login
1

Hi, installed 13.2 KDE with zypper dup, rebooted and it booted and worked fine at first. I may have installed updates or something while I used it, then I worked with Windows for a while (dual boot). When I tried to start openSUSE again this morning, it failed to work.

I have seen a few threads here with people that had a similar problem due to nvidia driver problems, but I don’t use that driver. I have a hybrid laptop system with the open source radeon driver, but openSUSE generally just uses the intel driver.

I took pictures of my screen, hope the most important messages are readable.

First this: https://www.dropbox.com/s/4nqutezdzhnbmje/2014-11-08%2008.49.55.jpg?dl=0

The error message seems to say:

Failed to start Load/Save Random Seed.
See "systemctl status systemd-random-seed.service" for details.

2 seconds later this: https://www.dropbox.com/s/46zrhxf4q3d802m/2014-11-08%2008.49.03.jpg?dl=0

The error message seems to say:

Failed to start Update UTMP about System Reboot/Shutdown.
See "systemctl status systemd-random-utmp.service" for details.
...
Dependency failed for Update UTMP about System Runlevel Changes.

The error message at the bottom seems to say:

Failed to start SuSEf irewa112 phase 1.
See "systemctl status SuSEf irewa112_init.service" for details.

Here, I’m not sure if I’m reading “irewa” right. It may say “irena” or even “irowa” or “irona”, but it seems to be an “ew” there.

I made sure it wasn’t a drive issue (that Windows would have blocked, as sometimes happens).

I have no idea what these messages mean or how I could possibly access the logs they are talking about. (Besides them being about systemd, of course.)

The same thing happens with different kernel versions of 3.16 and 17 (I did not install a different kernel after the upgrade, although it is possible that the updates pulled in an official update). Fall-back-mode doesn’t get me into X either.

Could anybody help me fix this? It would be much appreciated.

If possible, I’d like to avoid the hassle of putting oS13.2 on USB stick and reinstalling it, I just don’t have a lot of time.

2

This should read “SuSEfirewall2”.

Well, those messages are strange and should not appear of course.
Have you tried to run one of those “systemctl status” commands that have been suggested (in particular the first one, “systemctl status systemd-random-seed.service”)? Maybe they would provide a clue?

My first guess would be that your hard disk is full, the / partition at least. (something similar happened to me today, I got similar errors during boot…)
What does “df -h” say?
Do you maybe use btrfs? (probably not if you upgraded from 13.1)

3

Thanks for your reply!

Duh, of course it says “firewall”. :slight_smile:

I use ext4 for “/”, your guess was correct.

Hm, so I was able to run those 3 commands in a different runlevel.

I did think you may be onto something regarding disk space on “/”. I recall a large program for upgrade not downloading for apparently that reason (I may be wrong though). But then there seems there is enough left: https://www.dropbox.com/s/9gnv7h0a8opxmso/2014-11-08%2012.01.46.jpg?dl=0

Here’s a picture of the output from the first two systemd-related commands. The 3rd (firewall-related) was identical to the 2nd. Both the 2nd and the 3rd stated that the https://www.dropbox.com/s/101u451rotolvvh/2014-11-08%2012.04.39.jpg?dl=0

4

Well, my question was whether you use btrfs actually. The reason I asked is that btrfs regularly creates “snapshots” by default which can fill up your hard disk (and even will show up as free space in “df”).

I did think you may be onto something regarding disk space on “/”. I recall a large program for upgrade not downloading for apparently that reason (I may be wrong though). But then there seems there is enough left: Dropbox - File Deleted - Simplify your life

Yes. Should be plenty for booting…

Here’s a picture of the output from the first two systemd-related commands. The 3rd (firewall-related) was identical to the 2nd. Both the 2nd and the 3rd stated that the Dropbox - File Deleted - Simplify your life

Well, “systemd-random-utmp.service” does not exist. It probably should be “systemd-update-utmp.service”, I’d say… :wink:
And with the Firewall service you probably made a typing mistake as well.

Can you please run systemctl as root? It shows more information then. (I should have mentioned that immediately)

sudo systemctl status systemd-random-seed.service

The output you posted doesn’t tell more than that it failed to start, unfortunately. Not really helpful, we already know that.

And please also post the output of:

sudo systemctl status display-manager

And another thing:
Try to login as root (in text mode, the password should be the same), then run “startx” and show the output please.

5

My bad. It’s been a rough day… :slight_smile:

Here’s the output as root. Dropbox - File Deleted - Simplify your life

For status display-manager: Dropbox - File Deleted - Simplify your life
It may be noteworthy here that I use a tmpfs. (See the link from earlier for my partition table.)

And for startx as root: Dropbox - File Deleted - Simplify your life
Again it says it could not create a lock file in /tmp/.tX0-lock .

Thank you for your help.

6

No problem.

Here’s the output as root. Dropbox - File Deleted - Simplify your life

For status display-manager: Dropbox - File Deleted - Simplify your life
It may be noteworthy here that I use a tmpfs. (See the link from earlier for my partition table.)

And for startx as root: Dropbox - File Deleted - Simplify your life
Again it says it could not create a lock file in /tmp/.tX0-lock .

Well, I guess the problem is clear now. All of them cannot create some file in /tmp, because /tmp is a “Read-only filesystem”.
And this is the reason why you do not get to graphics mode.

You say you use a tmpfs, but I do not see /tmp in the list printed by “df”. THere’s just one tmpfs mounted as Firefox cache.
Is this on purpose? Or should /tmp be on a tmpfs too?

Please post the output of “ls -ld /tmp”, I suspect the permissions are wrong. This wouldn’t matter if you mounted some tmpfs (or whatever fs) there, but as none is mounted…
The permissions should look like this:

wolfi@linux-lf90:~> ls -ld /tmp
drwxrwxrwt 1 root root 6270  9. Nov 00:40 /tmp

PS: the “read-only filesystem” could indicate a different problem as well. That your root is mounted read-only because of an error.
So please post the output of this as well:


mount
systemctl status /
7

You know what? I have no idea. I remember I tampered a little bit a few months back when I got my SSD. It’s possible that I didn’t actually use a temporary file system for /tmp, just for other parts. So, scratch that, I guess.

Permissions: Dropbox - File Deleted - Simplify your life
Not a complete match.

Mount: Dropbox - File Deleted - Simplify your life
systemctl status /: Dropbox - File Deleted - Simplify your life

8

Ok. I was just wondering.

Permissions: Dropbox - File Deleted - Simplify your life
Not a complete match.

Only the date and sizes differ (which is to be expected), the permissions are ok.

Mount: Dropbox - File Deleted - Simplify your life

There we have it:

/dev/sda4 on / type ext4 (ro,relatime,data=ordered)

It is indeed mounted read-only.

systemctl status /: Dropbox - File Deleted - Simplify your life

Unfortunately this doesn’t tell anything, because “Journal has been rotated since unit was started. Log output is incomplete or unavailable”.:stuck_out_tongue:
Well, can you please post /etc/fstab? Maybe it is set to ro there.

If not, than some error must occur. This could be a filesystem error (maybe caused by a crash/incomplete shutdown), or even a hardware error (dying disk).
In this case, try to add “errors=continue” to the mount options in /etc/fstab. This should at least get you into graphics mode if there is no severe damage.
Of course this is not really a solution, but should at least make it easier to further investigate.
Another workaround for now would be to indeed specify /tmp as tmpfs, but I think you would still have problems because of a non-writeable /var (/var/tmp in particular)…

9

/etc/fstab:
Entire file, but hard to read: https://www.dropbox.com/s/wpo9c5syfw77gq8/2014-11-09%2012.46.11.jpg?dl=0
Partitions with mount options: https://www.dropbox.com/s/1ab3wok5jg7gh8v/2014-11-09%2012.46.48.jpg?dl=0
Settings at the end of the file: https://www.dropbox.com/s/hblve38dhnhz8rm/2014-11-09%2012.47.29.jpg?dl=0

I think you’re referring to “/'s” mount option “errors-remount-ro”, aren’t you? I’m not very good when it comes to mounting, but I don’t recall adding that option, I wonder if openSUSE 13.2 did that.

If I need to edit /etc/fstab, would you be so kind and tell me how I can do that from the command line? I can open a file with cat, but I’m not sure how to edit it.

Herzlichen Dank.

10

If I need to edit /etc/fstab, would you be so kind and tell me how I can do that from the command line? I can open a file with cat, but I’m not sure how to edit it.

Herzlichen Dank.

You can use a CLI editor, run with root privileges. For example

sudo nano /etc/fstab
  • Be careful with edits though. Typos could cause boot problems.
11

Ok, great, deano_ferrari! And just to be sure: I’m supposed to replace “errors-remount-ro” with “errors=continue”?

12

That’s what wolfie323 suggested :slight_smile:

13

Yes.

I don’t think this option would be set by openSUSE (I’ve never seen it), but that doesn’t matter as it’s the default anyway. So you would have the same problem if that option wasn’t there.

Apparently there are some errors on your / and it gets remounted as read-only therefore.
“error=continue” should prevent this and get you into graphical mode.

Then please post the output of:

dmesg | grep sda4

This should show the error that causes the partition to be mounted ro I hope.

14

One small problem: I can’t edit anything on / since it’s mounted as ro.

I would love a hint as to how to get around this. :slight_smile:

Thanks bunches.

15

Do you have a Live distro to boot from? You could take the approach as for disaster recovery, grub configuration etc

This should be enough to give you the basic idea…

16

Nevermind, I googled it. Don’t want to bother you guys with a question a search engine can answer. It’s

mount -o remount rw /

So I did that and edited /etc/fstab (and verified it after a reboot), but it didn’t help, I still got the same error messages. But that gave me an idea.

In the running session, I tried the above command (with su) and then tried to startx, first as myself and when that didn’t work, as root. Here’s what I got and it may point to a problem with /tmp after all:

startx as myself: https://www.dropbox.com/s/xvehoao69b7wpfi/2014-11-09%2015.36.37.jpg?dl=0
Fatal error: cannot open /dev/tty0

startx as root gave me a screensaver for just the blink of an eye. Then this little message popped up and I was able to use my mouse until I confirmed it: https://www.dropbox.com/s/vuhah87363wmm9a/2014-11-09%2015.37.40.jpg?dl=0

Right after, I got this output: https://www.dropbox.com/s/0r4rpw3irijzipc/2014-11-09%2015.38.29.jpg?dl=0
Apparently, root couldn’t initiate something KDE-related.

Any clues what this means?

17

Excuse the double posting, I can’t seem to edit my last post.

So used this magical and totally new tool called “Google” on both messages. The latter has something to do with permissions on home directory. On the former, I found that that you can no longer startx without root privileges, and that you should do it on runlevel 5. All very useful to know! I’d been using runlevel 6 pretty randomly, I know next to nothing about runlevels. I bet both errors where owed to that.

Anyway, I tried the same thing as above (remounting /, then starting X) in runlevel 5 AND was able to startx as root. But I was taken straight to Plasma (where I was logged in as root, which made me uncomfortable :wink: ) and not first to KDM. I didn’t want to connect to the internet, either. Is there any way I can start X and be taken to the display manager instead?

Here is the command that wolfi323 requested:

# dmesg | grep sda4
    1.950693]  sda: sda1 sda2 sda3 sda4 sda5 sda6 sda7 sda8
    6.853767] EXT4-fs (sda4): mounted filesystem with ordered data mode. Opts: (null)
   65.630408] EXT4-fs (sda4): re-mounted. Opts: (null)
18

Once in the desktop log out

19

This won’t help when he started X with “startx”. He’ll just get back to text mode then… :wink:

20

Yes, that’s restricted to 10 minutes after posting.

So used this magical and totally new tool called “Google” on both messages. The latter has something to do with permissions on home directory. On the former, I found that that you can no longer startx without root privileges, and that you should do it on runlevel 5. All very useful to know! I’d been using runlevel 6 pretty randomly, I know next to nothing about runlevels. I bet both errors where owed to that.

Well, actually runlevels are a thing of the past. systemd uses “targets” which are more general, but it offers the old runlevels (and changing them via “init X” for compatibility).

Anyway, I tried the same thing as above (remounting /, then starting X) in runlevel 5 AND was able to startx as root. But I was taken straight to Plasma (where I was logged in as root, which made me uncomfortable :wink: ) and not first to KDM. I didn’t want to connect to the internet, either.

Yes. That’s why you shouldn’t use startx. You won’t get all necessary permissions as user.

Is there any way I can start X and be taken to the display manager instead?

“init 5”, or “systemctl start display-manager”.
Probably you’d have to switch to runlevel 3 first (“init 3”), as you are actually in runlevel 5 already, so switching to it will do nothing.

Here is the command that wolfi323 requested:

# dmesg | grep sda4
    1.950693]  sda: sda1 sda2 sda3 sda4 sda5 sda6 sda7 sda8
    6.853767] EXT4-fs (sda4): mounted filesystem with ordered data mode. Opts: (null)
   65.630408] EXT4-fs (sda4): re-mounted. Opts: (null)

Hm, this doesn’t tell anything unfortunately.

So let’s recap the current situation:
You have changed /etc/fstab to have “error=continue” for /? And removed “error=remount-ro”?
But when you boot you are still stuck in text mode and / is mounted read-only?

As I said, try to run “fsck -f /dev/sda4”.
With a read-only mounted filesystem, it should be possible to check it.
It won’t repair it of course because it is mounted, and you cannot unmount it for obvious reasons.
For that you would have to boot from a LiveCD, or add this to the boot options to stop the boot before the system switches to the mounted /:

rd.break

You should get to text mode then, run “umount /dev/sda4” and “fsck -f /dev/sda4”.

And have a look at “smartctl -a /dev/sda”. This shows whether there are hardware errors on the drive (i.e. bad sectors).