Login loop

Questions about Grub, UEFI,the liveCD and the installer
Forum rules
Before you post please read how to get help
Post Reply
alice2wondr
Level 1
Level 1
Posts: 7
Joined: Tue Sep 03, 2019 6:51 pm

Login loop

Post by alice2wondr » Tue Sep 03, 2019 7:05 pm

Hi there,
I too, have this viewtopic.php?f=46&t=301050 problem as of today: since I updated a kernel update that has been sitting around in the Update Manager for some time now. I cannot even log in as a guest. If there is a way to get around this without my having to loose all the information in my 18.3 Sylvia, that would be wonderful!

Here is the info:

Code: Select all

mint@mint ~ $ inxi -Fxzd
System:    Host: mint Kernel: 4.10.0-38-generic x86_64 (64 bit gcc: 5.4.0) Desktop: Xfce 4.12.3 (Gtk 2.24.28)
           Distro: Linux Mint 18.3 Sylvia
Machine:   System: Dell (portable) product: Latitude E6420 v: 01
           Mobo: Dell model: 0K0DNP v: A02 Bios: Dell v: A04 date: 05/11/2011
CPU:       Dual core Intel Core i5-2520M (-HT-MCP-) cache: 3072 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 9977
           clock speeds: max: 3200 MHz 1: 1495 MHz 2: 2801 MHz 3: 2964 MHz 4: 3065 MHz
Graphics:  Card: Intel 2nd Generation Core Processor Family Integrated Graphics Controller bus-ID: 00:02.0
           Display Server: X.Org 1.18.4 drivers: intel (unloaded: fbdev,vesa) Resolution: 1366x768@59.99hz
           GLX Renderer: Mesa DRI Intel Sandybridge Mobile GLX Version: 3.0 Mesa 17.0.7 Direct Rendering: Yes
Audio:     Card Intel 6 Series/C200 Series Family High Definition Audio Controller
           driver: snd_hda_intel bus-ID: 00:1b.0
           Sound: Advanced Linux Sound Architecture v: k4.10.0-38-generic
Network:   Card-1: Intel 82579LM Gigabit Network Connection (Lewisville)
           driver: e1000e v: 3.2.6-k port: 4080 bus-ID: 00:19.0
           IF: eno1 state: up speed: 1000 Mbps duplex: full mac: <filter>
           Card-2: Intel Centrino Advanced-N 6205 [Taylor Peak] driver: iwlwifi bus-ID: 02:00.0
           IF: wlp2s0 state: down mac: <filter>
Drives:    HDD Total Size: 1036.4GB (1.3% used) ID-1: /dev/sda model: TOSHIBA_MQ02ABD1 size: 1000.2GB temp: 28C
           ID-2: USB /dev/sdb model: Flash_Drive size: 32.1GB temp: 0C
           ID-3: USB /dev/sdc model: U3_Cruzer_Micro size: 4.1GB temp: 0C
           Optical-1: /dev/sr0 model: TEAC DVD-ROM DV18SA rev: 3.0D dev-links: cdrom,dvd
           Features: speed: 24x multisession: yes audio: yes dvd: yes rw: none state: running
           Optical-2: /dev/sr1 model: N/A rev: N/A dev-links: N/A
           Features: speed: 40x multisession: yes audio: yes dvd: no rw: cd-r state: N/A
Partition: ID-1: /home size: 22G used: 248M (2%) fs: ext4 dev: /dev/sdb2
           ID-2: swap-1 size: 7.00GB used: 0.00GB (0%) fs: swap dev: /dev/sda6
RAID:      No RAID devices: /proc/mdstat, md_mod kernel module present
Sensors:   System Temperatures: cpu: 53.0C mobo: N/A
           Fan Speeds (in rpm): cpu: N/A
Info:      Processes: 214 Uptime: 1 min Memory: 420.6/5845.0MB Init: systemd runlevel: 5 Gcc sys: 5.4.0
           Client: Shell (bash 4.3.481) inxi: 2.2.35

Code: Select all

mint@mint ~ $ df -m
Filesystem     1M-blocks  Used Available Use% Mounted on
udev                2908     0      2908   0% /dev
tmpfs                585     9       576   2% /run
/dev/sdb1           8177  5835      2342  72% /cdrom
/dev/loop1          1692  1692         0 100% /rofs
aufs                3961  2842       916  76% /
/dev/sdb2          21920   248     20537   2% /home
tmpfs               2923     1      2923   1% /dev/shm
tmpfs                  5     1         5   1% /run/lock
tmpfs               2923     0      2923   0% /sys/fs/cgroup
tmpfs               2923     1      2923   1% /tmp
Hope this helps!!
Last edited by WharfRat on Tue Sep 03, 2019 9:18 pm, edited 1 time in total.
Reason: Split from another topic and added link to it.

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Tue Sep 03, 2019 9:44 pm

boot to previous kernel.

pull up the grub menu and select e advanced to get the previous choice of kernel.

after that you can uninstall this latest one.

maybe you get in the login loop and hop to a tty with ctrl + alt + F1 and type dkms status and see if you have any custom drivers that are listed for previous kernel but not listed for this latest kernel update, then you know where the problem is(driver module was not correctly built for the new kernel. -often a gpu module could fail on a new kernel and that can cause this type of problem).

or just use previous and uninstall the scary one and skip it until the next kernel update to try that one instead.

tell us what you find out.

alice2wondr
Level 1
Level 1
Posts: 7
Joined: Tue Sep 03, 2019 6:51 pm

Re: Login loop

Post by alice2wondr » Wed Sep 04, 2019 10:22 am

Hello zcot,
Thanks for a quick reply!
...
I had to ctrl + alt +F1 on login screen....here is the result:
===================================
..../dkms: line 428 echo:write error:No space left on device
dkms:conf Error!
No dest_name directive
No package name
No package version
Bad conf file does not rep.dkms.conf.file
===================================

???
The advanced boot option menu offers GNU ....4.15.0-60-generic, 58, and one other ; 'e' is to edit these options (?)...

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Wed Sep 04, 2019 12:55 pm

so it seems like /boot partition might have run out of space, so the -60 kernel is not fully installed correctly.

you can select 4.15.0-58-generic and boot that previous kernel, you don't need to edit anything.

Then see in Update Manager, "View" menu -> Linux kernels. And if you have a lot of kernels, you need to uninstall one or more of the oldest ones to make space on the partition. But you will keep the currently running -58. And you could keep the one before that, but no need to keep more than 2 or 3 total. The 4.10 versions are all end of life and unpatched from the Meltdown/Spectre exploits so get rid of all of those.

And the -60 version will have to be corrected manually.

then copy/paste the info from command: ls /boot


gm10 wrote:ppa
Does the custom kernel manager have a clean removal function? And that works on 18.3? ;)

gm10
Level 18
Level 18
Posts: 8738
Joined: Thu Jun 21, 2018 5:11 pm

Re: Login loop

Post by gm10 » Wed Sep 04, 2019 2:58 pm

zcot wrote:
Wed Sep 04, 2019 12:55 pm
gm10 wrote:ppa
Does the custom kernel manager have a clean removal function? And that works on 18.3? ;)
I'm not sure what you are asking, what custom kernel manager? Also what PPA? Mine? Not compatible with 18.3.
Tune up your LM 19.x: ppa:gm10/linuxmint-tools

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Wed Sep 04, 2019 4:50 pm

gm10 wrote:
Wed Sep 04, 2019 2:58 pm
zcot wrote:
Wed Sep 04, 2019 12:55 pm
gm10 wrote:ppa
Does the custom kernel manager have a clean removal function? And that works on 18.3? ;)
I'm not sure what you are asking, what custom kernel manager? Also what PPA? Mine? Not compatible with 18.3.
Ahh, noted. :(

I thought I recalled you mentioning your custom update manager having a broken-kernel removal function. I guess I could've just checked the launchpad to realize it wasn't for 18.x though. :oops:

gm10
Level 18
Level 18
Posts: 8738
Joined: Thu Jun 21, 2018 5:11 pm

Re: Login loop

Post by gm10 » Wed Sep 04, 2019 5:05 pm

zcot wrote:
Wed Sep 04, 2019 4:50 pm
I thought I recalled you mentioning your custom update manager having a broken-kernel removal function. I guess I could've just checked the launchpad to realize it wasn't for 18.x though. :oops:
No such thing even on 19.x, my Update Manager removes kernels pretty much the same way as the official one by letting the package management system remove the packages. No reason to have anything fancy there. Except that I have improved handling of meta-packages, those can lead to removal issues with the official version. That aside, with the automatic maintenance service that I also added to the official version we can prevent filled up boot partitions in the first place, so hopefully these issues will go away.

And yes, I'm not backwards compatible, my versions use Python 3.6+ and GTK 3.20+ features, both not present in 18.x, so I couldn't build it for that if I wanted to. I'm upwards compatible though, tested and working on upcoming Ubuntu 19.10. ;)
Tune up your LM 19.x: ppa:gm10/linuxmint-tools

alice2wondr
Level 1
Level 1
Posts: 7
Joined: Tue Sep 03, 2019 6:51 pm

Re: Login loop

Post by alice2wondr » Thu Sep 05, 2019 10:03 am

To zcot:see below line...

so it seems like /boot partition might have run out of space, so the -60 kernel is not fully installed correctly.

you can select 4.15.0-58-generic and boot that previous kernel, you don't need to edit anything.

Then see in Update Manager, "View" menu -> Linux kernels. And if you have a lot of kernels, you need to uninstall one or more of the oldest ones to make space on the partition. But you will keep the currently running -58. And you could keep the one before that, but no need to keep more than 2 or 3 total. The 4.10 versions are all end of life and unpatched from the Meltdown/Spectre exploits so get rid of all of those.

And the -60 version will have to be corrected manually.

then copy/paste the info from command: ls /boot
====================================================================
I log in and it gives me:
30. 176159 ] [drm:drm_atomic_helper_wait_for_flip_done [drm_kms_helper]] *Error* [CRTC:37: pipe A] flip_done timed out

etc...down to 153.
Then, it will loop back to login screen.:-/

-Alice

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Thu Sep 05, 2019 12:02 pm

Ok. That looks like a kernel bug. A lot of reports and workarounds.

But did you boot to the previous kernel instead? which version?

Are you able to see the /boot directory of the broken file system when using the live session installer media?

alice2wondr
Level 1
Level 1
Posts: 7
Joined: Tue Sep 03, 2019 6:51 pm

Re: Login loop

Post by alice2wondr » Fri Sep 06, 2019 10:12 am

Hi zcot,

I booted from 4.15- 58 generic with last reply.

But, after I booted the live session, it has available in boot file:
Screenshot_2019-09-06_09-20-57.jpg
Screenshot_2019-09-06_10-03-31.jpg
So it looks like the live that I have is the 4.10...but the updates took it to 4.15..? so the kernel 'update' regressed it...?

-A

gm10
Level 18
Level 18
Posts: 8738
Joined: Thu Jun 21, 2018 5:11 pm

Re: Login loop

Post by gm10 » Fri Sep 06, 2019 10:19 am

No, you've got no other kernel than the old one installed. That bug was fixed a year ago.

edit: nvm, had been confused by the screenshot from the live environment. as zcot said below, choose the kernel to boot from the GRUB boot menu.
Last edited by gm10 on Fri Sep 06, 2019 11:50 am, edited 1 time in total.
Tune up your LM 19.x: ppa:gm10/linuxmint-tools

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Fri Sep 06, 2019 11:43 am

Since /boot partition is full, you will not be able to use the system until you clear space in /boot.

Delete all(4) of the 4.15.0-60 files:
Image

Then you must boot the system into the 4.10 or 4.15.0-58 kernel. Do you know how to boot to a previous kernel? select Mint, press e, select 4.10 or not the -60 kernel?

You need to gain free space in /boot first. And then get into the system so you can fix the rest of it because the package system will show an error until you clean up the 4.15.0-60 problem.

Can you get there?

alice2wondr
Level 1
Level 1
Posts: 7
Joined: Tue Sep 03, 2019 6:51 pm

Re: Login loop

Post by alice2wondr » Fri Sep 06, 2019 8:36 pm

Hi zcot,

The 'delete' button does not work on these files.;-/ These articles expresses a pre-emptive concern I have:
https://unix.stackexchange.com/question ... ry-ubuntu and
https://askubuntu.com/questions/711801/ ... boot-linux

If memory serves, the kernel update(s) are from Ubuntu-right?

if I delete the 4.15.0 60 files, and am able to reboot to 4.15.0 58, and get in, will Update Manager retrieve packages for upgrading to 4.15.0 60 and install them correctly?

Would an uninstall function differently than deleting? If so, is there a way to uninstall 4.15.0 60s instead of deleting them?

-A

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Sat Sep 07, 2019 12:10 pm

alice2wondr wrote:
Fri Sep 06, 2019 8:36 pm
a pre-emptive concern I have:
That's good to be concerned.

Actually the stackexchange article is relevant, but not as much the askubuntu one.

Let me explain.

It's different than both of those articles. It looks like in both of those cases they deleted 1 or more kernel files that are marked as the current kernel, thus breaking the current grub default kernel(and possibly they could've just booted to previous kernel manually -depends on what random files they were deleting). In your case that's already broken, -you can't boot the latest entry which would be that -60 kernel since it was just installed during the last update, except that you ran into an "out of space" problem, which means that -60 did not succeed, even though you didn't get any other indication that it was unsuccessful(it probably did say so in the terminal window of the update manager but you would've had to open that and scroll through to see it). If you have a "no space" error on that /boot partition(which you did get when running the dkms command), and you had just attempted a kernel install, then that install failed. And the pic does show 0 bytes free.

To me it means the -60 failed.

So, to take it further, you can check it out. Do a quick file size comparison, you can just boot into the live session and look. Check each of the sets of files and compare the file size to each other. they should be very close to the same size(the 4.15's anyway). Or in tty: ls -l /boot

Further, in live session, open the broken filesystem and look at the file /var/log/apt/term.log. At the top of the file you will see some "log started" date, and at some point later you will see "log ended". So that's the last updates that took place and you can find info about the kernel package installation somewhere in there. In tty: cat /var/log/apt/term.log (of course you can use path/file autocomplete with the TAB key at any time) and then page with shift+pgup/pgdn or pipe the command with less(for using up/dn arrows or paging), so add | less at the end.

alice2wondr wrote:
Fri Sep 06, 2019 8:36 pm
If memory serves, the kernel update(s) are from Ubuntu-right?

if I delete the 4.15.0 60 files, and am able to reboot to 4.15.0 58, and get in, will Update Manager retrieve packages for upgrading to 4.15.0 60 and install them correctly?

Would an uninstall function differently than deleting? If so, is there a way to uninstall 4.15.0 60s instead of deleting them?

-A
yes they are Ubuntu kernel packages.

And if the last kernel package failed, which you should confirm as above, then using some typical package management command is going to fail too, on the -60 package which is not in a good state. Remove command is fine, but that will possibly fail because dpkg cache. The only command I see possibly working is apt install -f after you have cleared space in /boot to finally finish the -60 install correctly. But that's a paradox. You can temporarily remove the kernel files from one of the unused kernels to get enough space to run that fix command against the -60 package and maybe that works. But it leaves an issue for later if you want to use the GUI or package commands to deal with something later, which you will have to do since it appears you can only hold 2 kernels in that space. So eventually deleting one unused kernel with the GUI, putting the temporarily removed kernel files back and using the GUI on that one too. Then you're down to 1 kernel and no kernel package issues. It's a paradox of sorts, and a bit twisted. :lol:

So anyway you have to clear space at the least. And then fix the cache if that's going to be a problem.


Let us know where you stand, and someone can come up with a good way to move forward. ;)

alice2wondr
Level 1
Level 1
Posts: 7
Joined: Tue Sep 03, 2019 6:51 pm

Re: Login loop

Post by alice2wondr » Sun Sep 08, 2019 12:05 pm

Thank you very much for your reply! :D The -.15 -58 and -.15-60 are very close in size.
in live session, open the broken filesystem and look at the file /var/log/apt/term.log
The only -60 file I can open is the config file.

also:
mint@mint ~ $ /var/log/apt/term.log
bash: /var/log/apt/term.log: Permission denied

So if I delete the -60 files and am able to boot into 4.15.0 -58, I should be able to get my stuff out of there and maybe install something else, due to paradoxical issues-right..? Should 4.10.0 be left alone for now?

How do I delete -60 files? Right-click menu delete is inactive...

-A

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Sun Sep 08, 2019 2:48 pm

please share info.

use live session.

copy/paste command then share url link here: cat /media/mint/587bb63a-4044-4206-a9cb-0599dfbba465/var/log/apt/term.log | nc termbin.com 9999

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Sun Sep 08, 2019 2:59 pm

I'm trying to figure out exactly what the problem is, because it appears that you just did updates and a kernel was one of the updates and then now you can not boot. This does tend to imply that the kernel update was a problem. And also seeing messages about no space left on /boot that does seem like the root of the problem. So I have been trying to pay attention directly towards that.

However, in the first post you said "In the login screen I enter my password and the screen gets black for a few seconds and returns to the login screen.", and also "I cannot even log in as a guest.".

So, I just want to clearly understand what the problem is, to make sure that this is a problem with a kernel update, or maybe something different.

Let us know the answer to these questions please:
1. can you boot normally which should use the 4.15.0-60 kernel? (I realize the answer is no). do you get to see the login screen and enter your password?
2. can you boot to the previous kernel using the grub selection menu for the 4.15.0-58 kernel? do you see the mint logo? what about the login screen?
3. can you boot the pc from grub menu using the 4.10.0-38 kernel?

alice2wondr
Level 1
Level 1
Posts: 7
Joined: Tue Sep 03, 2019 6:51 pm

Re: Login loop

Post by alice2wondr » Mon Sep 09, 2019 10:00 pm

Here are the answers to your questions;
1:
-Shut down live 18.3;removed live media
-Booting 4.15.0 -60:
I just booted and logged in successfully. I tried this multiple times at the very beginning of the post with no success.
-Received a popup message stating that there is low disk space ~59.7mb available.
-Transferring numerous files to external media; deleted nonsense files...
-In Disk Usage Analyzer, an uninstalled file remains…

2:-yes; yes;yes ; successful login. Low disk space: 59.7mb

3:-Yes; logged in fine; File system root has only 59.4mb remaining…

Should one or more kernels be deleted?

User avatar
zcot
Level 5
Level 5
Posts: 501
Joined: Wed Oct 19, 2016 6:08 pm

Re: Login loop

Post by zcot » Mon Sep 09, 2019 11:23 pm

Ah! Okay, great!

You are fine then.

You can easily delete some files from ~/Downloads

and Empty the Trash.

Further than that you could get technical and do some things as in this thread: viewtopic.php?f=42&t=266175

Post Reply

Return to “Installation & Boot”