You are not logged in.

#1 2019-09-10 14:34:22

penguinator
Member
Registered: 2019-05-13
Posts: 30

Locked in loop at logon/password prompt [RESOLVED]

I am now in Rescue Mode and the error message contains:

gtkwarning: could not open display

there was also something about 'root being locked' but that no longer appears after I managed (TTY mode) to reset the root password (actually, I reset it back to the original).  I was hoping it might resolve the problem, but obviously it did not. 

Background: I was trying to get bleachbit to run as root when I managed to instead turn-off OpenBox and got a black screen with only a cursor.  I wasn't able to see a way to get out of this (the system had become unresponsive, except for being able to move the cursor) so I used the power switch to force a system shutdown, hoping all would be well at the next reboot.

I boldfaced the above, so it's easier to pick-out from the post what I think is the likely, and major, cause of my boot-loop.  Could it have been anything else that I did, during this 'episode'?

There is another 'SOLVED' post on the forum with an (apparently) similar Subject line, but after examining the content, that user had an accident while editing etc/fstab, then doing a forced shutdown, but in my case it was solely entering an inappropriate command at terminal that blanked-out the screen.  Most likely, what solved the other problem wouldn't be good for mine?

If that command is essential to a solution, let me know as I would then have to go through my browsing history and try to recover the page where I saw it.  I think it was ssh -X, from this post, GtkWarning: could not open display.

[edit] I can now see that ssh -X was completely wrong.  Why it appeared as an answer under the heading of my error message ... ?? sad ]  [/edit]

[edit] ? What would have been a better way of regaining control of the system, after the ssh -X blanked the screen? [/edit]

Much appreciated if this provides some idea of what happened, and --
? could someone please suggest troubleshooting steps, to restore the system to normal functioning?

Last edited by penguinator (2019-09-13 15:34:19)


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#2 2019-09-11 13:15:02

rbh
Member
From: Sweden/Vasterbotten/Rusfors
Registered: 2016-08-11
Posts: 64

Re: Locked in loop at logon/password prompt [RESOLVED]

No, the command "ssh -X", could not have caused your problem!
It sholud onley returned:
"usage: ssh [-46AaCfGgKkMNnqsTtVvXxYy] [-B bind_interface] [...]"

Whatever it is impossible to guess.
Often when boot fails, it's because one or more disk devices is not accessible.
If you are still in rescuemode, can you se if all partions is mounted?

 # mount | grep sd 

lists all mounted partitions.

 # dmesg 

can give you some clues.
To get more info, you have to look in your logs in /var/log/*.

If pc could not mount partition because of trashed filesystem, use an rescue-dist (or BL-live-iso. I prefer finnix, but it is getting old...).

Often when X fails.it is possible to change to an text terminal.
Keys Ctrl + Alt + F1-F6, can give an text consol for ComandLineInterface (cli).


// Regards rbh

Offline

#3 2019-09-11 14:00:04

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

OK will do, and thanks!  Yes, this is what I was hoping for, to learn how to analyze the problem.

rbh wrote:

If you are still in rescuemode, can you see if all partions is mounted?

 # mount | grep sd 

lists all mounted partitions.

Output of above:
/dev/sda1 on / type ext4 (here, the mount parameters are listed, looks identical to what's in fstab)

This is as expected, since there's only one (very small) SSD and to give the OS enough space to load, there's just a single / partition, with everything in it.

 # dmesg 

can give you some clues.

I'll do this in sequence, but I'll have run this command again with '| more' because the output is quite massive.  Then, it will take a while to go through it line-by-line because I'm not sure if what I'm looking for will be obvious. 

Next steps will be:

To get more info, you have to look in your logs in /var/log/*.

If pc could not mount partition because of trashed filesystem, use an rescue-dist (or BL-live-iso. I prefer finnix, but it is getting old...).

Often when X fails.it is possible to change to an text terminal.
Keys Ctrl + Alt + F1-F6, can give an text consol for ComandLineInterface (cli).


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#4 2019-09-11 15:11:21

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

rbh wrote:
 # dmesg 

can give you some clues.

Warnings, or errors, seen in dmesg:

ACPI Warning: SystemIO range 0x00...0400-0x00...041F conflicts with OpRegion 0x00...0400-0x00...041F (\SMRG) (20160831/utaddress-247)
ACPI Warning: SystemIO range 0x00...0400-0x00...041F conflicts with OpRegion 0x00...0400-0x00...041F (\_SB.PCI0.SBRG.SMRG) (20160831/utaddress-247)

EXT4-fs (sda1): re-mounted. Opts: discard,errors=remount-ro

Four consecutive ACPI 'Warning' messages:
SystemIO range (range specified) conflicts with OpRegion (range specified) all for (20160831/utaddress-247), each 'Warning' followed by:
ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver

I think I should do a fsck on /dev/sda1. 

Would this be the best way to do this from where I am now, in Recovery Mode?

To check the root file system, you should
touch /forcefsck
and reboot. That will force a fsck before the root file system is mounted.

Or, is it necessary to power-down and do it from a Live CD/DVD?

Will try the fast way, first, and then post back, with the result.

[edit] No luck. The loop condition persists. [/edit]

Last edited by penguinator (2019-09-11 15:19:15)


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#5 2019-09-11 15:30:21

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

Additional information:

On the log-in screen, there is a 'Tool Icon' that opens two check boxes, Default Xsession and Openbox.
On the first boot following entry of the touch-fsck command (which looped, again) I looked and Default Xsession was selected.

I changed it back to Openbox, but after the next boot, which was still looping, I checked the setting again, and the choice had, once again, reverted to Default Xsession.

Is this something that needs troubleshooting?

Last edited by penguinator (2019-09-11 15:31:53)


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#6 2019-09-11 16:07:54

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

rbh wrote:

To get more info, you have to look in your logs in /var/log/*.

Possibly relevant lines from syslog:

X10:  fatal IO error 11 (Resource temporarily unavailable) on X server ":0"
Registry[736]:     after 53 requests (53 known processed) with 0 events remaining.

More ACPI warnings, similar to those already reported.

Nothing else apparently significant.  If one of the other logs would be more productive, which one, or ones should I open?

Question:
Is fsck sensitive as to the distro from which it comes, relative to the distro on which it's used? I already have a live Lubuntu disk that (I think?) loads more quickly than the BL Linux DVD, and has lots of utilities.  But, could this cause damage if used to 'fsck' a Debian directory?

Last edited by penguinator (2019-09-11 16:18:41)


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#7 2019-09-11 19:39:38

rbh
Member
From: Sweden/Vasterbotten/Rusfors
Registered: 2016-08-11
Posts: 64

Re: Locked in loop at logon/password prompt [RESOLVED]

Forgive me, I often forget debian has migrated to systemd. Then you can read systemlog for last boot, easy with for example:

# journalctl -xb 
# touch /forcefsck

is as good as booting rescue CD. When rebooted and the computer finds the file, it wont try to mount the partition before doing fsck. But, the diskcheck goes fast, hard to se messeges. If your grub.cfg, passes "quiet" to your grub, you will not se anything.. "fsck -f /dev/sdxy", from an terminal when booted from live/rescue-cd, gives more information.
Sometimes when I have had big problems, I`ve worked from live-cd booted X. Then I can use the compouter and work on an solution.

It does not matter if you use rescue cd from another dist. The important thing is the age of the tools. If you use an to old Lubuntu-cd, it will protest when you execute fsck and ask you to use a newer version.

The ACPI Warnings, have you had them before? Lookin old logs.


// Regards rbh

Offline

#8 2019-09-11 22:25:47

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

I am looking at the line that says, "remounted. ... remount-ro"

penguinator wrote:

...

...
EXT4-fs (sda1): re-mounted. Opts: discard,errors=remount-ro
...

I think I should do a fsck on /dev/sda1. 

Would this be the best way to do this from where I am now, in Recovery Mode?
...

Is this a transient condition that happened because the filesystem needed a 'check' or is it a continuing problem that requires some other kind of attention to prevent its recurrence?  Is it saying the device (sda1) was mounted as read-only?

In any event, a fsck was supposedly done upon the next boot, yet the looping continues.

I am wondering why this continues, if the fsck (should have?) resolved the problem causing the file system to mount as read-only.  Am I correct or incorrect in describing the situation in this way?

I'm going to try that command "fsck -f /dev/sdxy" to see if more details turn up as to my /dev/sda1's integrity (or lack thereof!).  Will post with details, as needed.

Other than that, are there any other commands might I run from the command line, to examine what's 'right' and what's 'going wrong'?

Last edited by penguinator (2019-09-11 22:44:45)


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#9 2019-09-11 22:49:13

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

You are correct, "fsck -f /dev/sda1" gave a better report on the device than I've ever seen.  Five passes of checks, and it passed all of them.

I'm going to get out of the live environment, power down, and reboot tomorrow.  If it's still remounting as read-only, and then looping, I'll continue the thread.

Incidentally, that -f flag is a good tip.  I read fsck's man-page several times and didn't see it.  kiss Maybe it will appear tomorrow morning?


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#10 2019-09-11 22:54:28

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

rbh wrote:

The ACPI Warnings, have you had them before? Lookin old logs.

I'll wait with that until the 'boot-loop' has been stopped.  I suspect that's not the cause of the current problem, but more of a point of interest and curiosity ... I wonder if it's correctable, given the netbook's hardware dates back about 12 years?


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#11 2019-09-12 08:41:47

rbh
Member
From: Sweden/Vasterbotten/Rusfors
Registered: 2016-08-11
Posts: 64

Re: Locked in loop at logon/password prompt [RESOLVED]

penguinator wrote:

I am looking at the line that says, "remounted. ... remount-ro"

Yes, from my fstab:

# / was on /dev/sda3 during installation
UUID=4e3b85e8-42b0-4e00-84c3-d30397504a54 / ext4    errors=remount-ro 0  2

"errors=remount-ro". If the boot process can't mount / (rot) read-write (rw), / will be mounted read only.

Is this a transient condition that happened because the filesystem needed a 'check' or is it a continuing problem that requires some other kind of attention to prevent its recurrence?

Why it's happening, is still a mystery. When the cause is fixed, / will be mounted rw again.

Other than that, are there any other commands might I run from the command line, to examine what's 'right' and what's 'going wrong'?

Nothing else intresting when you read the journal from rescue-mode with "journalctl -xb"?

Have you tried to continue out of rescue-mode?
Try remount / rw:

# mount -o remount,rw /dev/sda1 /

Exit rescue-modewith "exit" or Ctrl+D.


// Regards rbh

Offline

#12 2019-09-12 10:33:08

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

rbh wrote:

Why it's happening, is still a mystery. When the cause is fixed, / will be mounted rw again.
...
Nothing else intresting when you read the journal from rescue-mode with "journalctl -xb"?

Nothing that would point to a cause. There is, in red letters, Failure to execute /ub/udev/mtp-probe (copied by hand). As each boot begins there's a briefly-displayed message in the upper-left corner of the screen, saying 'root: clean'.

I do have a Timeshift backup of the entire /dev/sda1 drive, taken shortly before the boot-loop began, and given the 'mystery' of what's causing the SSD to repeatedly mount as read-only, I would like to now use that.

Running Timeshift from terminal isn't working, and that may explain why the developer pulled the documentation for that feature.  I suspect it works on some installations, but not on others.

I have a backported Timeshift package for a different, but still, stretch-based, disto, but Gdebi is failing to install that into a live instance of BL Linux, giving a missing dependency error: libgee (it wants version >=0.8.3). It seems Gdebi doesn't, at least in this case, make any attempts to locate or install this, but instead just 'aborts'.

Any advice on how to locate and install a package for this Debian-Stretch dependency would be appreciated.  My searching hasn't turned up anything that works.

Last edited by penguinator (2019-09-12 10:41:37)


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

#13 2019-09-12 20:50:53

penguinator
Member
Registered: 2019-05-13
Posts: 30

Re: Locked in loop at logon/password prompt [RESOLVED]

Update on restoring operating system:

I was able to restore the backup and the netbook is once again booting normally.

Some of the folders on /dev/sda1 were 'restored' but some others were being 're-created' from the backup, according to Timeshift report.  So the damage went farther than just filesystem problems, that could be repaired by fsck.

The dependency problem with installing Timeshift into a live instance of BL was resolved when I used the larger .iso, the one that fits only on a DVD.  The one that failed was from the version on the CD ROM.

Many thanks for your help!  I did learn a lot that will come in handy the next time.  smile

Last edited by penguinator (2019-09-12 20:53:14)


System Info:
ASUS EeePC 701 netbook, 4GB SSD
BL Helium

Offline

Board footer

Powered by FluxBB