I checked *suspend to DISK with fatal results.
BAD MOVE: after my success at suspending and resuming from Sleep I decided to test Hibernate. It appeared to work but won’t recover from hibernation. Now I can’t even get the GRUB screen. Just a load of messages which stop at “Reached target Graphical Interface”. If I touch the On/Off button, it moves a bit further **and then powers off."
******I mean when I press the On-button my laptop keeps trying to recover the Hibernate image and there is no way to boot 13.2. The messages I mentioned just scroll up the screen and I can’t stop them. I created a new Knoppix CD and I can see the content of the SSD/HD and found the partition with the system.
In the meantime I appear to have zapped the MBR because when I try to boot from the HD I get the Lenovo screen and then nothing but a blank screen with a flashing underline cursor. This cursor does respond to the keyboard.
grubenv contain:
#GRUB Environment Block
saved_entry=0
next_entry=Advanced options for openSUSE>openSUSE, with Linux 3.16.6-2-desktop(recovery mode)
########################
No, it doesn’t.
It tries to boot the set entry over and over again.
Remove /boot/grub2/grubenv to get the boot menu back.
In the meantime I appear to have zapped the MBR because when I try to boot from the HD I get the Lenovo screen and then nothing but a blank screen with a flashing underline cursor. This cursor does respond to the keyboard.
Please note that “mount /proc” and “mount /sys” will not work on a default 13.2 installation any more, use this instead:
mount -t proc proc /proc
mount -t sysfs sysfs /sys
grubenv contain:
#GRUB Environment Block
saved_entry=0
next_entry=Advanced options for openSUSE>openSUSE, with Linux 3.16.6-2-desktop(recovery mode)
########################
I.e. You booted an older kernel via “Advanced Options” and then hibernated. The hibernate script set the wrong boot entry, namely “recovery mode” which doesn’t resume at all, because it has the “noresume” parameter.
.e. You booted an older kernel via “Advanced Options” and then hibernated. The hibernate script set the wrong boot entry, namely “recovery mode” which doesn’t resume at all, because it has the “noresume” parameter.
Yes. This is what probably happened:
after checking that resume from Sleep was working, an update arrived via apper
I installed the update and it asked me to reboot
I was about to do that, got distracted and thought of testing hibernation first. [can’t be sure because my memory is sieve]
In retrospect that was stupid.
While I was trying to get out of that problem I saw a advice on the web for removing a “hibernation snapshot” by holding down the Shift key while trying to boot. That’s probably what zapped the MBR.
Doesn’t matter.
This file is loaded dynamically during boot. It’s not used at all when (re)installing GRUB2.
But with that grubenv present, grub2 will not show any menu at all because that entry tells it to directly boot Linux 3.16.6(recovery mode).
Normally you shouldn’t even have to do anything else than to remove that file, but as your system doesn’t boot/load grub at all now, a reinstallation of grub seems necessary.
You could try to just delete grubenv as a test first though.
Note that even with that entry present in that file your system should boot, just directly into “Linux 3.16.6(recovery mode)”.
That booting to recovery mode doesn’t work apparently is a different problem, unrelated to GRUB2.
Could be caused by having a proprietary video driver installed (nvidia, fglrx).
Yes. This is what probably happened:
after checking that resume from Sleep was working, an update arrived via apper
I installed the update and it asked me to reboot
I was about to do that, got distracted and thought of testing hibernation first. [can’t be sure because my memory is sieve]
In retrospect that was stupid.
Ok, so apparently you didn’t explicitely boot an older kernel, but you hibernated after you installed the kernel update.
The same problem in the end though, the currently booted kernel was one of the “Advanced Options” sub menu (the menu gets updated when a kernel package is installed e.g.), and the hibernate script set the wrong one in grubenv, i.e. the “(recovery mode)” flavour which does no resume at all.
While I was trying to get out of that problem I saw a advice on the web for removing a “hibernation snapshot” by holding down the Shift key while trying to boot. That’s probably what zapped the MBR.
Holding Shift doesn’t have any effect at all AFAIK (with the default settings in openSUSE at least).
At least it should not “zap” the MBR in any case (unless your BIOS maybe has such a function…).
Maybe you broke the boot loader in another way while trying to “fix” your problem?
I really appreciate your help. And I have good news. The Lenovo has recovered after deleting grubenv and after opening the BIOS to switch from “Legacy Mode” to EFI. After that Grub appeared and loaded. I don’t want to try “suspend to disk” again. Nonetheless I want to check if everything is OK.
I received and installed a couple of updates and looked at the Software Updater" messages. They mentioned “A transaction failed to complete successfully”.
Furthermore dmesg has a line in bold that ends with "
**Please run fsck" **referring to sda1, which is FAT-fs - fsck doesn’t exist so I ran
xfs_repair -V -n sda1
That failed for reasons I don’t understand. I didn’t set up the partitions and their formats.
However the Updater now says my system is up to date.
Is there anything else I can check? If not, this issue is:
SOLVED
.
[Next: get Sleep to work with the lid; install Bumblebee (had it working on this laptop under 13.1) because I need the nVidia GPU very little and without it the tiny battery will last longer; etc.]
Good to hear!
Did you maybe install in EFI mode and switch to “Legacy Mode” later?
I don’t want to try “suspend to disk” again.
Unless you boot the older kernel it should work.
If hibernate/resume itself fails for some reason and you’re stuck without a boot menu, you’ll know what to do now anyway: Remove /boot/grub2/grubenv.
Nonetheless I want to check if everything is OK.
I received and installed a couple of updates and looked at the Software Updater" messages. They mentioned “A transaction failed to complete successfully”.
Well, PackageKit is not too good in solving problems/conflicts.
Maybe try to run “zypper up” to see what’s wrong.
Furthermore dmesg has a line in bold that ends with **“****Please run fsck” **referring to sda1, which is FAT-fs
fsck doesn’t exist so I ran
xfs_repair -V -n sda1
That failed for reasons I don’t understand. I didn’t set up the partitions and their formats.
???
You say sda1 is FAT-fs, why are you running xfs_repair? That’s for XFS filesystems.
Use dosfsck instead.
But fsck should exist as well. You’ll have to be root to be able to run it though.
Is there anything else I can check?
If the system boots up fine, everything should be ok.
The worst I think that could have happened would be an inconsistent filesystem. But fsck should be run automatically on boot to fix that (for the / and /home partitions at least).
Yes, I did. I switched to Legacy because my old Knoppix wouldn’t boot.
???
You say sda1 is FAT-fs, why are you running xfs_repair? That’s for XFS filesystems.
Use dosfsck instead.
But fsck should exist as well. You’ll have to be root to be able to run it though.
I became root and issued the fsck command as requested by dmesg. The Konsole said it didn’t exist and to use xfs_repair. I thought it pretty weird. I’ll make a note of dosfsck.
Thank again. I’ll be around for a while. Season Greetings! I’ll be moving accounting and my development servers on to the Lenovo.
Well, EFI boots completely different than legacy mode (MBR), so no wonder that grub2 couldn’t be loaded afterwards.
I became root and issued the fsck command as requested by dmesg. The Konsole said it didn’t exist and to use xfs_repair.
Sorry, I don’t really understand that.
The Konsole should not suggest xfs_repair in any case, even if fsck is missing.
Can you please try to run fsck again, and post the exact message you get?
If fsck is really missing, you should probably re-install the package util-linux. You might run into problems sooner or later without fsck, as it is run automatically on boot.
> - after checking that resume from Sleep was working, an update arrived
> via apper
> - I installed the update and it asked me to reboot
> - I was about to do that, got distracted and thought of testing
> hibernation first. [can’t be sure because my memory is sieve]
NEVER! When it tells you to reboot, after a kernel update, just reboot.
Don’t do anything else, specially not any testing, and never, under
death penalty, hibernate! Just reboot and don’t delay it.
> - In retrospect that was stupid.
> - While I was trying to get out of that problem I saw a advice on the
> web for removing a “hibernation snapshot” by holding down the Shift
> key while trying to boot. That’s probably what zapped the MBR.
I doubt it. Anyway, I have not seen that shift thing anywhere.
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
> - I received and installed a couple of updates and looked at the
> Software Updater" messages. They mentioned “A transaction failed to
> complete successfully”.
I don’t like apper/packagekit. I prefer yast/zypper.
> - Furthermore dmesg has a line in bold that ends with *"**Please run
> fsck" *referring to sda1, which is FAT-fs
> - fsck doesn’t exist so I ran
Impossible. fsck must exist, it is crucial. Your system would not boot.
> Code:
> --------------------
> xfs_repair -V -n sda1
> --------------------
> That failed for reasons I don’t understand.
Fortunately it failed.
Why do you attempt and XFS repair on a FAT partition? If it
attempted to work on it, it would destroy your partition filesystem
beyond repair!
Please, please, go slow and don’t go rushing. Don’t try two things on a
single step, like update a kernel then hibernate, or repair a partition
with the wrong tool because you don’t find the right one. Slow down, and
if you had doubts, ask.
–
Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)
> Code:
> --------------------
>
> 8.405316] FAT-fs (sda1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.
> 8.405931] XFS (sda4): Mounting V5 Filesystem
> 9.101122] XFS (sda4): Starting recovery (logdev: internal)
> --------------------
>
> As you can see, it still says to run fsck. All I did was to issue
> Code:
> --------------------
> fsck --help
> --------------------
> to find out what it did. The output (which I cannot recreate) said “You
> probably want xsf_repair”.
Please post here the output of “fsck --help”. If you get output, then
the “fsck” command is installed.
fsck can be called as that, fsck, or fsck.something:
Sorry for delay in responding (blame Xmas). Yes. I am sure you’re right about what happened. fsck does exist, of course. Anyway the important thing is that Sleep is fixed and working fine. I’ll ignore the dmesg report for now. Thanks to everyone for their help and Happy New Year from a frozen part of Spain.
