13.2
There were a bunch of updates that were waiting to install in Apper, so I marked them for installation and hit Apply.
Apper never did anything after half an hour. Said it was waiting for credentials or something. I figured there was a password prompt that popped behind the other windows (like it usually does), but it was nowhere to be found. I clicked to cancel the update. It didn’t cancel. After 10 minutes, I X’d out of the window. An hour later, I rebooted.
When it came back up, it failed to boot properly.
Instead, it shows this (after briefly loading the login screen wallpaper) http://i.imgur.com/xJuXZKkh.jpg
And the keyboard is unresponsive (mouse powers off too). I googled, but didn’t find any info on what to do at this screen.
On first thought, password prompts from shell really do not give typing feedback. Just type in your password as usual and enter. It should drop you to a root shell and let you work from there.
If it does, run a “zypper ve” to try to repair the system.
Good luck!
On 2015-03-24 10:46, OhWell wrote:
>
> 13.2
> There were a bunch of updates that were waiting to install in Apper, so
> I marked them for installation and hit Apply.
> Apper never did anything after half an hour. Said it was waiting for
> credentials or something. I figured there was a password prompt that
> popped -behind- the other windows (like it usually does), but it was
> nowhere to be found. I clicked to cancel the update. It didn’t cancel.
> After 10 minutes, I X’d out of the window. An hour later, I rebooted.
> When it came back up, it failed to boot properly.
> Instead, it shows
> [image: http://i.imgur.com/xJuXZKkh.jpg]
>
> And the keyboard is unresponsive. I googled, but didn’t find any info on
> what to do at this screen.
>
> Any help is appreciated. Thanks
>
>
I mean the keyboard and mouse are not responsive at all. I have an optical mouse, and the led on it turns off. The capslock and numlock keys on the keyboard are off and will not turn on. No keys or key combinations I’ve tried result any response on the screen. Also, they are USB devices, and the error message indicates something USB related.
I’m guessing I might need to do something before it gets to this point…?
Tried moving those to another USB port or hub, and using another pointing and typing thing?
Also, there should be another option in GRUB to let you boot openSUSE in rescue mode (or was it failsafe?). It’s probably under something like “Advanced boot options”. Try booting from there. Hopefully, something should work.
Failing that, if you have another device that can make a live USB, maybe attempting to repair the system with a chroot can help?
Make a live USB, preferably with the version you have, try to mount / from your live instance, run a terminal and chroot there, then do a “zypper ve”?
Normally I’d brush off errors like that during boot, but in this case, it can be related.
On 2015-03-24 23:06, OhWell wrote:
>
> I mean the keyboard and mouse are not responsive at all. I have an
> optical mouse, and the led on it turns off. The capslock and numlock
> keys on the keyboard are off and will not turn on. No keys or key
> combinations I’ve tried result any response on the screen. Also, they
> are USB devices, and the error message indicates something USB related.
>
> I’m guessing I might need to do something before it gets to this
> point…?
>
> Thanks
>
>
That’s more or less, yes. We’ll see how it goes from there.
On 2015-03-25 05:26, OhWell wrote:
>
> I don’t get a GRUB boot menu. I must’ve disabled it or something…?
>
> Just to be clear, since I’ve never done this before:
>
>
>
> - Boot liveusb of the version of opensuse I have (13.2).
> - Mount / of my broken installation.
> - Terminal & chroot to / of my broken installation.
> - Run “zypper ve”
> - ???
> - Profit
>
>
> I’ll give that a try in a bit.
>
> Thanks
>
>
Well, I figured out part of the problem… IOMMU setting disabled in the BIOS was causing problems with USB devices.
However, opensuse still doesn’t load completely, and it drops me to a prompt. I tried “zypper ve” and it gives me
Download (curl) error for ‘http://download.opensuse.org/update/13.2-non-oss/repo...somethin…’
Error Code: Connection failed.
Error message: Could not resolve host:download.opensuse.org
It won’t resolve any websites at this prompt, so I’m guessing there’s a networking/dns thing that isn’t loaded…?
I haven’t tried booting from a opensuse live usb yet, though. Should I still try that?
If your network connection is wired, you can run “dhclient -v” so it can set up networking for you. The connection might have to be brought up. If ever, “ifconfig enp5s0 up” should work, replace device with appropriate name if necessary.
If it’s wireless, I’d recommend running it with a live USB. Connecting to wireless is a lot less cumbersome with a desktop to work with.
On 2015-03-27 05:16, OhWell wrote:
>
> Well, I figured out part of the problem… IOMMU setting disabled in the
> BIOS was causing problems with USB devices.
>
> However, opensuse still doesn’t load completely, and it drops me to a
> prompt. I tried “zypper ve” and it gives me
> Download (curl) error for
> ‘http://download.opensuse.org/update/13.2-non-oss/repo...somethin…’
> Error Code: Connection failed.
> Error message: Could not resolve host:download.opensuse.org
>
> It won’t resolve any websites at this prompt, so I’m guessing there’s a
> networking/dns thing that isn’t loaded…?
>
> I haven’t tried booting from a opensuse live usb yet, though. Should I
> still try that?
>
> Thanks
>
>
I got a network connection, and ran “zypper ve”. I installed everything with the following exceptions: Firefox and Thunderbird were both being prompted to uninstall or break. I chose break for now. Also, patch #173 (something to do with glibc) would not install. Says something about it can’t be uninstalled. However, this error predates the issue at hand. I had tried installing that patch through Apper and it would fail for the same reason.
So, now I am still coming to the prompt, but now I also get this, after it tries to load the login/desktop screen:
“Could not connect to DBus server: org.freedesktop.DBus.Error.NotSupported: Unable to autolaunch a dbus-daemon without a $DISPLAY for X11”
Been searching google and haven’t yet found anything that helps me understand this error.
Tried doing a “zypper up” after to see if anything was left out?
This one might not make the desktop load, but it could give us some useful information on what’s going on:
“sudo systemctl start display-manager”
If it doesn’t start, what does “sudo systemctl status display-manager” say?
Throwing a guess, but was xorg installed properly?
As for patch #173, it looks like a few security fixes for glibc. Can’t comment much on the uninstall problem, don’t know much about that. For FF and TB, looks like it’s harmless, also getting prompted with my machine.
On 2015-03-28 23:46, OhWell wrote:
>
> I got a network connection, and ran “zypper ve”. I installed everything
> with the following exceptions: Firefox and Thunderbird were both being
> prompted to uninstall or break. I chose break for now. Also, patch #173
> (something to do with glibc) would not install. Says something about it
> can’t be uninstalled. However, this error predates the issue at hand. I
> had tried installing that patch through Apper and it would fail for the
> same reason.
>
> So, now I am still coming to the prompt, but now I also get this, after
> it tries to load the login/desktop screen:
> “Could not connect to DBus server:
> org.freedesktop.DBus.Error.NotSupported: Unable to autolaunch a
> dbus-daemon without a $DISPLAY for X11”
>
>
> Been searching google and haven’t yet found anything that helps me
> understand this error.
>
> Thanks
>
>
Does “systemctl start display-manager.service” work?
Enable it to boot to a graphical system:
systemctl enable display-manager.service
zypper ps also shows me this:
which I guess is the udev Kernel Device Manager, loaded and running per systemctl list-units.
That’s because you installed updates before. systemd-udev was already running and is still using the old files.
Reboot and this should be “fixed”.
But I’m more worried about that error when installing glibc-locale, which you actually cut off.
What error was that?
Maybe your disk is full? That would give you problems in starting KDE too.
If you are using btrfs for /, you should probably remove some snapshots with “yast snapper”.
Regarding your missing boot menu, removing the file “/boot/grub2/grubenv” should bring it back.
Nope. It does the same thing. I see the login screen wallpaper load up. Then it dumps me into emergency mode.
I think that’s related to my inability to install update opensuse-2015-173, as mentioned earlier.
I am using btfrs. Baobab shows I’m only using 16 out of 68 GB when it looks at the whole partition (doing this through IceWM). When I look at the breakdown by folder, it told me the .snapshots folder was 150 GB, which is larger than the entire disk. I deleted several snapshots that were made the day the problems began (already tried snapping back to these. didn’t help), and that brought .snapshots down to 97 GB, which is still larger than the entire partition. But after a reboot, I still get dumped into emergency mode after it fails to load the display manager.
I deleted /boot/grub2/grubenv and got my grub menu back.
Tried to boot into rescue mode, but it does the same thing.
Tried to boot to a few different snapshots made well before the issue, and either it dumps me to emergency mode, or doesn’t boot at all.
I thought you already “fixed” that, and are booting to text mode now.
Does it still say “Welcome to emergency mode”?
If your system boots to emergency mode, this is unrelated to the display-manager.service.
Can you please post the output of “systemctl get-default”?
And check which service is failing to cause emergency mode to be started in the first place (for how, read the message on the screen).
systemctl --failed
should show.
Does the boot continue if you enter “systemctl default”?
I think that’s related to my inability to install update opensuse-2015-173, as mentioned earlier.
Hm, I don’t see that being mentioned at all in this thread.
openSUSE-2015-173 is a glibc update. But why are you unable to install it? What does the error message say?
It might be related to the failing boot.
I am using btfrs. Baobab shows I’m only using 16 out of 68 GB when it looks at the whole partition (doing this through IceWM). When I look at the breakdown by folder, it told me the .snapshots folder was 150 GB, which is larger than the entire disk.
Well, snapshots don’t really take space on the disk when the files haven’t been changed. They are no backups.
But baobab just naively sums up the file “sizes”, so reaches a value that’s bigger than your hard disk.
I deleted several snapshots that were made the day the problems began (already tried snapping back to these. didn’t help), and that brought .snapshots down to 97 GB, which is still larger than the entire partition. But after a reboot, I still get dumped into emergency mode after it fails to load the display manager.
The system shouldn’t drop to emergency mode if it fails to load the display-manager service. I rather think it’s the opposite: it doesn’t load the display-manager.service because it drops to emergency mode before that.
No, the same thing happens. It goes back to emergency mode.
Sorry, that was in post #9.
I ran zypper up again in order to replicate the issue. But this time it seems that glibc-locale installed correctly. There were several other things that failed to install, though.
Not sure if this is related, but right before the “Welcome to emergency mode!” line, I have these two lines:
[19.907765] xhci_hcd 0000:02:00.0: can’t setup: -110
[19.908361] xhci_hcd 0000:02:00.0: init 0000:02:00.0 fail, -110
I think this indicates a USB 3.0 driver issue. XHCI and EHCI are enabled in my BIOS (pretty sure they need to be).
I ran zypper up again in order to replicate the issue. But this time it seems that glibc-locale installed correctly. There were several other things that failed to install, though.
And what failed to install now? If random things fail to install, that doesn’t really speak for the healthiness of your system.
Maybe try to run a fsck, check the kernel log (dmesg) whether there are mentions of bad sectors/filesystem errors, and have a look at the smartctl output.
Not sure if this is related, but right before the “Welcome to emergency mode!” line, I have these two lines:
[19.907765] xhci_hcd 0000:02:00.0: can’t setup: -110
[19.908361] xhci_hcd 0000:02:00.0: init 0000:02:00.0 fail, -110
I think this indicates a USB 3.0 driver issue. XHCI and EHCI are enabled in my BIOS (pretty sure they need to be).
Yes this does indicate a USB 3.0 issue.
But I don’t think that would cause emergency mode, unless you have some partitions on USB drives in your fstab.
Again, please run “systemctl --failed” to see what service exactly failed to start, causing the system to go into emergency mode.
Googling for that error message led me to this: Seagate external 2Tb USB 3 drive [Archive] - FedoraForum.org
So do you mount an USB drive in your fstab? Comment out that entry (or at least add “nofail” to the mount options) and your system should boot normally again.
If in doubt, please post /etc/fstab.
It was moving too fast for me to capture.
Now if I run zypper up, it says “Warning: Repository ‘openSUSE:13.2:Update’ appears to be outdated” then “The following 150 package updates will NOT be installed:” lists a bunch of things, and then says “Nothing to do.”
zypper v shows 44 packages will be installed, and 5 that will be removed. I *think *some (maybe all) of the 5 listed for removal were ones that had errors installing as previously mentioned.
google-chrome-stable, kdelibs4, kdepim4, kdepim4-runtime, kdepimlibs4. But I’m not sure. I do remember that the ones that failed said “applydeltarpm check failed.”
fsck reports it’s unable to resolve the last one. And even if I disconnect all drives except the one containing / and reboot, it still drops me to emergency mode.
dmesg I don’t see anything about bad sectors or other errors except this:
AMD-Vi: Event logged [IO_PAGE_FAULT device=02:00.0 domain=0x0016 address=0x00000000ce9f9880 flags=0x0010]
which I’ve had for months, and never got around to looking into until now. Based on this (I have the same mobo mentioned), I can fix that by disabling IOMMU in the bios and enabling *soft *IOMMU in the OS. Not sure if that’s a viable solution; it seems to cause other problems. And I don’t think it’s related to this issue.
smartctl reports no SMART errors, passes health check.
Unrecoverable file system error can cause this kind of thing ie the system is fried. Restore from back up or reinstall. Essential if you have problems that fsck won’t fix then not much you can do. All the update problems point to a serious system problem. The only good part is that smart says the drive is ok. Might check lost&found se if any part of files landed there if so it is near impossible to stitch them back together unless they are pure text
If the drives that you removed have partition in the fstab and does not have nofail option then any inability to find them will cause a drop to emergency mode. So that tells us nothing. I notice what appears to be a couple of USB device are these in the fstab?
You should probably run “zypper ref”. Do you have auto-refresh turned off for the update repo? I wouldn’t recommend that.
zypper v shows 44 packages will be installed, and 5 that will be removed. I *think *some (maybe all) of the 5 listed for removal were ones that had errors installing as previously mentioned.
If “zypper ve” wants to install 44 packages, then something is definitely wrong with your installation.
google-chrome-stable, kdelibs4, kdepim4, kdepim4-runtime, kdepimlibs4. But I’m not sure. I do remember that the ones that failed said “applydeltarpm check failed.”
delta rpms are only used for the updates from the update repo.
If “applydeltarpm check failed”, either your filesystem or some files are corrupted or you have bad RAM I’d say.
plymouth-start.service loaded failed failed
Hm, shouldn’t really drop you to emergency mode I think, but would still be interesting why it failed.
fsck reports it’s unable to resolve the last one.
Because it is not connected/accessable.
Remove the entry from the fstab or add “nofail” to the mount options.
And the one before seems to be an external drive as well. Add “nofail” there too.
And even if I disconnect all drives except the one containing / and reboot, it still drops me to emergency mode.
Of course.
The system drops to emergency mode if it cannot mount a required (i.e. without “nofail”) partition from the fstab.
If you disconnect all drives, you basically make sure that the partitions definitely cannot be mounted, and therefore even force emergency mode.
smartctl reports no SMART errors, passes health check.
You should have a look at the “offline uncorrectable”, “current pending sector”, or similar values.
Even if the health check is passed, there still might be bad sectors.
And you should boot from a LiveCD or similar and run fsck on the / partition. This cannot be done in the running system when / is mounted, not even in recovery mode.
An alternative would be to stop the boot before the system switches to / on the hard disk, by adding “rd.break” to the boot options.
You can then unmount the / partition via “umount /sysroot” as it is not used yet, and run fsck.
