Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.misc > #14140 > unrolled thread

GUI terminated suddenly and restarted..

Back to article view | Back to comp.os.linux.misc

Contents

  GUI terminated suddenly and restarted.. The Natural Philosopher <tnp@invalid.invalid> - 2015-03-18 13:38 +0000
    Re: GUI terminated suddenly and restarted.. Bobbie Sellers <bliss-sf4ever@dslextreme.com> - 2015-03-18 07:19 -0700
      Re: GUI terminated suddenly and restarted.. Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2015-03-18 17:01 +0000
    Re: GUI terminated suddenly and restarted.. Aragorn <thorongil@telenet.be.invalid> - 2015-03-18 17:24 +0100
      Re: GUI terminated suddenly and restarted.. Baho Utot <baho-utot@columbus.rr.com> - 2015-03-18 16:34 -0400
        Re: GUI terminated suddenly and restarted.. Dan Espen <despen@verizon.net> - 2015-03-18 16:52 -0400
        Re: GUI terminated suddenly and restarted.. Aragorn <thorongil@telenet.be.invalid> - 2015-03-18 22:27 +0100
          Re: GUI terminated suddenly and restarted.. The Natural Philosopher <tnp@invalid.invalid> - 2015-03-19 09:45 +0000
            Re: GUI terminated suddenly and restarted.. Aragorn <thorongil@telenet.be.invalid> - 2015-03-19 18:59 +0100
              Re: GUI terminated suddenly and restarted.. The Natural Philosopher <tnp@invalid.invalid> - 2015-03-19 18:43 +0000
            Re: GUI terminated suddenly and restarted.. faeychild <faeychild@nomail.afraid.org> - 2015-03-20 08:19 +1100
    Re: GUI terminated suddenly and restarted.. Bit Twister <BitTwister@mouse-potato.com> - 2015-03-18 18:05 +0000
      Re: GUI terminated suddenly and restarted.. The Natural Philosopher <tnp@invalid.invalid> - 2015-03-18 18:51 +0000
        Re: GUI terminated suddenly and restarted.. J G Miller <miller@yoyo.ORG> - 2015-03-18 20:41 +0000

#14140 — GUI terminated suddenly and restarted..

Mmm. Befoer I start I am not expecxting a solution to te questin, but 
noire general hints as to te class of probolem this might be.

Simply put after aroudn 7 hours of flawless operation a Mint 17.1 Mate 
computer lost the GUI completely, dropped back briefly to a text login 
and then presented me with a logon screen.

I assume the X server had crashed and restarted.

First of all, is that a valid and comprehensive assumption?

Secondly any clues as to why?

It was synchronous with trying to access a particular web page in 
Firefox, but that is all.


After relogging in the web page was just fine.



-- 
Everything you read in newspapers is absolutely true, except for the 
rare story of which you happen to have first-hand knowledge. – Erwin Knoll

[toc] | [next] | [standalone]

#14142

On 03/18/2015 06:38 AM, The Natural Philosopher wrote:
> Mmm. Befoer I start I am not expecxting a solution to te questin, but
> noire general hints as to te class of probolem this might be.
>
> Simply put after aroudn 7 hours of flawless operation a Mint 17.1 Mate
> computer lost the GUI completely, dropped back briefly to a text login
> and then presented me with a logon screen.
>
> I assume the X server had crashed and restarted.
>
> First of all, is that a valid and comprehensive assumption?
>
> Secondly any clues as to why?
>
> It was synchronous with trying to access a particular web page in
> Firefox, but that is all.
>
>
> After relogging in the web page was just fine.
>
>
>
	You might want to flush your history et al a bit more often in the 
future.  At least I begin to have problems if I do not clean out
Firefox every so often.
	Mozilla products seem to expect to be shut down every so often.
and I restart Thunderbird more often than I reboot except when testing
new Live distributions.

	bliss

[toc] | [prev] | [next] | [standalone]

#14146

On 2015-03-18, Bobbie Sellers <bliss-sf4ever@dslextreme.com> wrote:

> 	You might want to flush your history et al a bit more often in
> the future.  At least I begin to have problems if I do not clean out
> Firefox every so often.
> 	Mozilla products seem to expect to be shut down every so often.
> and I restart Thunderbird more often than I reboot except when testing
> new Live distributions.

Omigod, some former Microsofties must have joined the Mozilla staff...

-- 
/~\  cgibbs@kltpzyxm.invalid (Charlie Gibbs)
\ /  I'm really at ac.dekanfrus if you read it the right way.
 X   Top-posted messages will probably be ignored.  See RFC1855.
/ \  HTML will DEFINITELY be ignored.  Join the ASCII ribbon campaign!

[toc] | [prev] | [next] | [standalone]

#14145

On Wednesday 18 March 2015 14:38, The Natural Philosopher conveyed the 
following to comp.os.linux.misc...

> Mmm. Befoer I start I am not expecxting a solution to te questin, but
> noire general hints as to te class of probolem this might be.

Your spiel choker also appears to be b0rken. <grin>

> Simply put after aroudn 7 hours of flawless operation a Mint 17.1 Mate
> computer lost the GUI completely, dropped back briefly to a text login
> and then presented me with a logon screen.
> 
> I assume the X server had crashed and restarted.
> 
> First of all, is that a valid and comprehensive assumption?

In my experience, yes, that would be a correct analysis, provided that 
you do of course boot up to a graphical login screen in the first place.  
It would not be logical for this to happen if you start X manually from 
the command line in a character mode login console.

> Secondly any clues as to why?

That's a tricky one.  Most of X11 ─ and specifically, the desktop 
environment ─ runs quite monolithically, and it all communicates with 
your graphics adapter through the video driver.  It is possible for an 
application ─ and quite probably so with things like Adobe's Flash 
Player plugin ─ to crash and bring down the whole desktop with it.

When the standard runlevel is 5 ─ i.e. the runlevel in which the machine 
boots up to a graphical login screen ─ is active, this will then trigger 
an automatic restart of the X server (and by consequence, the desktop 
environment running on top of it).

> It was synchronous with trying to access a particular web page in
> Firefox, but that is all.

My money would be on the Flash Player plugin.


> After relogging in the web page was just fine.

It may have been a glitch ─ an unfortunate combination of circumstances, 
caused by an innocuous bug somewhere, which only gets triggered when 
certain conditions are all met. 

-- 
= Aragorn =

         http://www.linuxcounter.net - registrant #223157

[toc] | [prev] | [next] | [standalone]

#14150

On 03/18/15 12:24, Aragorn wrote:
> On Wednesday 18 March 2015 14:38, The Natural Philosopher conveyed the

[putolin]

>
> That's a tricky one.  Most of X11 ─ and specifically, the desktop
> environment ─ runs quite monolithically, and it all communicates with
> your graphics adapter through the video driver.  It is possible for an
> application ─ and quite probably so with things like Adobe's Flash
> Player plugin ─ to crash and bring down the whole desktop with it.
>
> When the standard runlevel is 5 ─ i.e. the runlevel in which the machine
> boots up to a graphical login screen ─ is active, this will then trigger
> an automatic restart of the X server (and by consequence, the desktop
> environment running on top of it).


If it was a init based system that is true.  It most likely is a systemd 
machine.  Since I don't use linux any more ( moved to freebsd ) I don't 
know how systemd would respond.  It should most likely restart X but who 
knows.

>
>> It was synchronous with trying to access a particular web page in
>> Firefox, but that is all.
>
> My money would be on the Flash Player plugin.
>
>
>> After relogging in the web page was just fine.
>
> It may have been a glitch ─ an unfortunate combination of circumstances,
> caused by an innocuous bug somewhere, which only gets triggered when
> certain conditions are all met.
>

[toc] | [prev] | [next] | [standalone]

#14151

Baho Utot <baho-utot@columbus.rr.com> writes:

> On 03/18/15 12:24, Aragorn wrote:
>> On Wednesday 18 March 2015 14:38, The Natural Philosopher conveyed the
>
> [putolin]
>
>>
>> That's a tricky one.  Most of X11 ─ and specifically, the desktop
>> environment ─ runs quite monolithically, and it all communicates with
>> your graphics adapter through the video driver.  It is possible for an
>> application ─ and quite probably so with things like Adobe's Flash
>> Player plugin ─ to crash and bring down the whole desktop with it.
>>
>> When the standard runlevel is 5 ─ i.e. the runlevel in which the machine
>> boots up to a graphical login screen ─ is active, this will then trigger
>> an automatic restart of the X server (and by consequence, the desktop
>> environment running on top of it).
>
>
> If it was a init based system that is true.  It most likely is a
> systemd machine.  Since I don't use linux any more ( moved to freebsd
> ) I don't know how systemd would respond.  It should most likely
> restart X but who knows.

With systemd, works the same as always.

-- 
Dan Espen

[toc] | [prev] | [next] | [standalone]

#14152

On Wednesday 18 March 2015 21:34, Baho Utot conveyed the following to 
comp.os.linux.misc...

> On 03/18/15 12:24, Aragorn wrote:
>
>> On Wednesday 18 March 2015 14:38, The Natural Philosopher conveyed
>> the
> 
> [putolin]

*What* did you call me?! :p

>> That's a tricky one.  Most of X11 ─ and specifically, the desktop
>> environment ─ runs quite monolithically, and it all communicates with
>> your graphics adapter through the video driver.  It is possible for
>> an application ─ and quite probably so with things like Adobe's Flash
>> Player plugin ─ to crash and bring down the whole desktop with it.
>>
>> When the standard runlevel is 5 ─ i.e. the runlevel in which the
>> machine boots up to a graphical login screen ─ is active, this will
>> then trigger an automatic restart of the X server (and by
>> consequence, the desktop environment running on top of it).
> 
> If it was a init based system that is true.  It most likely is a
> systemd machine.  Since I don't use linux any more ( moved to freebsd
> ) I don't know how systemd would respond.  It should most likely
> restart X but who knows.

He's using Mint as his distribution, but I don't know what version.  
Mint is an offspring of Ubuntu, and Ubuntu itself is based upon Debian 
Sid (with the exception of the LTS releases).  So there is a chance that 
his init system is still Ubuntu's Upstart ─ which is a wrapper around 
sysvinit ─ or, if it is the newest (non-LMDE) release, possibly that 
he's already got systemd.

Runlevels are different in systemd, but systemd-based distributions ─ 
and especially Mint or Ubuntu ─ tend to favor a "GUI login" setup, so 
systemd would indeed normally automatically restart the X server if X 
goes down.  

The fact that systemd automatically restarts services ─ not just the 
display server but also other daemons ─ and that it has all of those 
under its management is exactly one of the "sales arguments" touted by 
the systemd developers and advocates, so I'd be thoroughly surprised if 
systemd didn't at the very least live up to its own claims in that 
regard. ;-)

That said, I'm still running a systemd-free GNU/Linux distribution here, 
and I intend to remain systemd-free for as long as I can ─ there are 
plenty of very decent alternatives, most of which are wrappers around 
sysvinit, and Slackware uses the BSD-style init (which is in fact also a 
wrapper around sysvinit, but implemented differently).  

I have my /personal/ objections against FreeBSD [*], but that would take 
us too far off-topic. ;-)


[*] I'm not saying FreeBSD is a bad operating system.  It's just that
    there are several things about it which I myself don't like, both
    technical and on account of the permissiveness of the license.

-- 
= Aragorn =

         http://www.linuxcounter.net - registrant #223157

[toc] | [prev] | [next] | [standalone]

#14156

On 18/03/15 21:27, Aragorn wrote:

>
> He's using Mint as his distribution,

17.1

On another identical software distro machine I have since had two 
instances of firefox crashing and one of the flash plugin., I note that 
both have been 'upgraded' on the last few days...


I conclude that something is breaking firefox/flash on both setups with 
unpredictable results.


-- 
Everything you read in newspapers is absolutely true, except for the 
rare story of which you happen to have first-hand knowledge. – Erwin Knoll

[toc] | [prev] | [next] | [standalone]

#14157

On Thursday 19 March 2015 10:45, The Natural Philosopher conveyed the 
following to comp.os.linux.misc...

> On 18/03/15 21:27, Aragorn wrote:
>
>> He's using Mint as his distribution,
> 
> 17.1
> 
> On another identical software distro machine I have since had two
> instances of firefox crashing and one of the flash plugin., I note
> that both have been 'upgraded' on the last few days...
> 
> I conclude that something is breaking firefox/flash on both setups
> with unpredictable results.

It is not uncommon for the Adobe Flash Player plugin to conflict with 
the graphics driver, so that would be my first guess.  And it is also 
not uncommon for that Flash Player plugin to freeze or bring down 
Firefox, which itself is a GTK+ application and thus may tie in with the 
GTK+ libraries used by your desktop environment of choice and form 
another point of weakness. [*]


[*] I don't remember exactly which one you said you were using, but
    I'm pretty sure it isn't KDE, which is based upon the Qt widget
    libraries, and most of the other desktop environments supported
    in Mint are all GTK-based ─ XFCE, MATE, Cinnamon, LXDE, GNOME...

-- 
= Aragorn =

         http://www.linuxcounter.net - registrant #223157

[toc] | [prev] | [next] | [standalone]

#14158

On 19/03/15 17:59, Aragorn wrote:
> On Thursday 19 March 2015 10:45, The Natural Philosopher conveyed the
> following to comp.os.linux.misc...
>
>> On 18/03/15 21:27, Aragorn wrote:
>>
>>> He's using Mint as his distribution,
>>
>> 17.1
>>
>> On another identical software distro machine I have since had two
>> instances of firefox crashing and one of the flash plugin., I note
>> that both have been 'upgraded' on the last few days...
>>
>> I conclude that something is breaking firefox/flash on both setups
>> with unpredictable results.
>
> It is not uncommon for the Adobe Flash Player plugin to conflict with
> the graphics driver, so that would be my first guess.  And it is also
> not uncommon for that Flash Player plugin to freeze or bring down
> Firefox, which itself is a GTK+ application and thus may tie in with the
> GTK+ libraries used by your desktop environment of choice and form
> another point of weakness. [*]
>
>

well had a flash crash today and another firefox crash yesterday. So it 
looks like latest firefox+flash is bugged

> [*] I don't remember exactly which one you said you were using, but
>      I'm pretty sure it isn't KDE, which is based upon the Qt widget
>      libraries, and most of the other desktop environments supported
>      in Mint are all GTK-based ─ XFCE, MATE, Cinnamon, LXDE, GNOME...
>

MATE is my desktop of choice. Nice compromise between pretty, and 
functional.


-- 
Everything you read in newspapers is absolutely true, except for the 
rare story of which you happen to have first-hand knowledge. – Erwin Knoll

[toc] | [prev] | [next] | [standalone]

#14159

The Natural Philosopher wrote:

> On 18/03/15 21:27, Aragorn wrote:
> 
>>
>> He's using Mint as his distribution,
> 
> 17.1
> 
> On another identical software distro machine I have since had two
> instances of firefox crashing and one of the flash plugin., I note that
> both have been 'upgraded' on the last few days...
> 
> 
> I conclude that something is breaking firefox/flash on both setups with
> unpredictable results.
> 
> 

I was about to register my two cents worth this morning about Firefox.

For the last couple of days the online TV guide page has been displaced to 
the right and can't be fully dragged back. This may have coincided with the 
flash update.

Konqueror displays correctly.

I was all fired up to go this morning, but the page is displaying correctly.

.....Is it Firefox or the page was fixed ( it didn't rattle Konqueror)
..... Is it just unfortunate coincidence.

Only the Natural Philosopher and myself got tackled.




-- 
faeychild
Running kde on 3.14.32-desktop-1.mga4 kernel.
Mageia release 4 (Official) for x86_64

[toc] | [prev] | [next] | [standalone]

#14147

On Wed, 18 Mar 2015 13:38:34 +0000, The Natural Philosopher wrote:
>
> Simply put after aroudn 7 hours of flawless operation a Mint 17.1 Mate 
> computer lost the GUI completely, dropped back briefly to a text login 
> and then presented me with a logon screen.
>
> I assume the X server had crashed and restarted.
> First of all, is that a valid and comprehensive assumption?

I'll +1 on that assumption.

> Secondly any clues as to why?

Not really at this point.

> It was synchronous with trying to access a particular web page in 
> Firefox, but that is all.

Based on the inordinate amount of information provided, my WAG is that
the web page attempted a malware exploit. The exploit was a
deliberate DOS attack, or it just screwed up and corrupted the
software stack with the resulting JOW.

As for clues, the only possible places I would know to look would be in
/var/log/Xorg.0.log*, ~/xsession-errors and a core.

As far as I know, default installs do not have core enabled.
Each gui login causes /var/log/Xorg.0.log.old to be over written
so you better get a copy for forensic study.

I have found it helpful to get a copy of .xsession-errors just after
starting the Desktop Manager. That way I know how far down to start
looking around for something having to do with an application.

> After relogging in the web page was just fine.

That does not surprise me. The better malware will blacklist any ip
address it has made an attempt at infecting. That is to help prevent
analysis by anti-virus software.

[toc] | [prev] | [next] | [standalone]

#14148

On 18/03/15 18:05, Bit Twister wrote:
> On Wed, 18 Mar 2015 13:38:34 +0000, The Natural Philosopher wrote:
>>
>> Simply put after aroudn 7 hours of flawless operation a Mint 17.1 Mate
>> computer lost the GUI completely, dropped back briefly to a text login
>> and then presented me with a logon screen.
>>
>> I assume the X server had crashed and restarted.
>> First of all, is that a valid and comprehensive assumption?
>
> I'll +1 on that assumption.
>
>> Secondly any clues as to why?
>
> Not really at this point.
>
>> It was synchronous with trying to access a particular web page in
>> Firefox, but that is all.
>
> Based on the inordinate amount of information provided, my WAG is that
> the web page attempted a malware exploit. The exploit was a
> deliberate DOS attack, or it just screwed up and corrupted the
> software stack with the resulting JOW.
>
> As for clues, the only possible places I would know to look would be in
> /var/log/Xorg.0.log*, ~/xsession-errors and a core.
>
> As far as I know, default installs do not have core enabled.
> Each gui login causes /var/log/Xorg.0.log.old to be over written
> so you better get a copy for forensic study.
>
> I have found it helpful to get a copy of .xsession-errors just after
> starting the Desktop Manager. That way I know how far down to start
> looking around for something having to do with an application.
>
>> After relogging in the web page was just fine.
>
> That does not surprise me. The better malware will blacklist any ip
> address it has made an attempt at infecting. That is to help prevent
> analysis by anti-virus software.
>
It wasn't malware. I know the person whose site it was and he is highly 
technical and runs a tight ship.


Sadly the x log file is gone now..


-- 
Everything you read in newspapers is absolutely true, except for the 
rare story of which you happen to have first-hand knowledge. – Erwin Knoll

[toc] | [prev] | [next] | [standalone]

#14149

On Wednesday, March 18th, 2015 at 18:51:07 +0000, The Natural Philosopher explained:

> Sadly the x log file is gone now..

Where you running a display manager which created a log file eg kdm.log, lxdm.log?

If so, do you see anything similar to

   EQ overflowing.  Additional events will be discarded
                    until existing events are processed.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.linux.misc

csiph-web