[Solved] KDE in a new Mageia 6 installation is freezing

[mark9117]

I'm interested in either a fix or a recommendation to reinstall. I'm not sure where to go here because it could be a couple of things.
Brand new Mageia 5 installation,

Code: Select all

uname -a
Linux spike 4.9.43-desktop-1.mga6 #1 SMP Sun Aug 13 15:52:35 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux

I've got a support message here about Akonadi personal management service is not working. That is still unresolved. I am also experiencing trouble with ksmserver. On every boot, I have a notification the ksmserver couldn't start (or something to that effect). Clicking okay or cancel to get rid of that box just causes KDE to restart. On the second restart, I still get the warning box about ksmserver, but killing it does not cause a reboot and everything loads up as expected. I first noticed this when I uninstalled the Otter browser. I was trying Otter because Chromium-browser wasn't work, an issue that has since been resolve. I unstalled Otter because it opened on every boot.

When I started seeing these ksmserver warnings, I went to systemsettings > desktop sessions and selected "Start with an empty session". That worked in terms of getting a KDE session started with no applications open. Looking good, so I setup my usual session configuration (Firefox, Chromium, Thunderbird, etc.) and selected "Restore previous session" then rebooted. It did not resolve the issue - still had ksmserver warnings.

parsing journalctl for "ksmserver" yields this:

Code: Select all

Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart script  "/home/madams/.config/autostart-scripts/dropboxd"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart script  "/home/madams/.config/autostart-scripts/gkrellm_trans.sh"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting notification thread
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Kcminit phase 2 done
Aug 31 18:05:56 spike ksmserver[10057]: completeShutdownOrCheckpoint called
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/home/madams/.config/autostart/dropbox.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/home/madams/.config/autostart/KNotes.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/home/madams/.config/autostart/ownCloud.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/home/madams/.config/autostart/pia_manager.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/home/madams/.config/autostart/SpiderOak.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/home/madams/.config/autostart/XChat IRC.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/etc/xdg/autostart/mageiawelcome.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/etc/xdg/autostart/org.kde.korgac.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/etc/xdg/autostart/powerdevil.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/home/madams/.config/autostart/recollindex.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/etc/xdg/autostart/mageia-mgaonline.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/etc/xdg/autostart/net_applet.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Starting autostart service  "/etc/xdg/autostart/user-dirs-update-gtk.desktop"
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: Autostart 2 done
Aug 31 18:05:56 spike ksmserver[10057]: QObject::connect: Cannot queue arguments of type 'ObjectDescriptionType'
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: autostart service ("/bin/xdg-user-dirs-gtk-update") finished with exit code  0
Aug 31 18:05:56 spike ksmserver[10057]: ksmserver: autostart service ("/bin/mageiawelcome-launcher") finished with exit code  0
Aug 31 18:05:57 spike ksmserver[10057]: ksmserver: autostart service ("/usr/libexec/org_kde_powerdevil") finished with exit code  0
Aug 31 18:05:57 spike ksmserver[10057]: ksmserver: autostart script ("/home/madams/.config/autostart-scripts/gkrellm_trans.sh") finished with exit code  0
Aug 31 18:05:57 spike ksmserver[10057]: ksmserver: autostart service ("/bin/korgac") finished with exit code  0
Aug 31 18:05:59 spike ksmserver[10057]: ksmserver: autostart script ("/home/madams/.config/autostart-scripts/dropboxd") finished with exit code  11
Aug 31 18:06:03 spike ksmserver[10057]: ksmserver: autostart service ("/bin/knotes") finished with exit code  0
Aug 31 18:07:21 spike ksmserver[10057]: ksmserver: autostart service ("/bin/dropbox", "start", "-i") finished with exit code  0

That's the tail of the log from the last reboot. The only thing out of the ordinary I see is that dropboxd is exiting with code 11. Dropbox isn't working and my understanding is that it's some sort of trouble on their end and there's nothing to be done about it.

I searched through the form for any similar trouble and found a thread about cauldron. That thread saw a request for a journalctl -ab log that I can get, but it's insanely long, and these to pieces of information.

Code: Select all

rpm -qa|sort|grep -iE "nvidia|kernel"
kernel-desktop-4.9.35-1.mga6-1-1.mga6
kernel-desktop-4.9.43-1.mga6-1-1.mga6
kernel-desktop-devel-4.9.35-1.mga6-1-1.mga6
kernel-desktop-devel-4.9.43-1.mga6-1-1.mga6
kernel-desktop-devel-latest-4.9.43-1.mga6
kernel-desktop-latest-4.9.43-1.mga6
kernel-firmware-20170531-1.mga6
kernel-firmware-nonfree-20170531-1.mga6.nonfree
kernel-userspace-headers-4.9.43-1.mga6
vboxadditions-kernel-4.9.35-desktop-1.mga6-5.1.22-20.mga6
vboxadditions-kernel-4.9.43-desktop-1.mga6-5.1.26-2.mga6
vboxadditions-kernel-desktop-latest-5.1.26-2.mga6
virtualbox-kernel-4.9.35-desktop-1.mga6-5.1.22-20.mga6
virtualbox-kernel-4.9.43-desktop-1.mga6-5.1.26-2.mga6
virtualbox-kernel-desktop-latest-5.1.26-2.mga6


Also:

Code: Select all

# dkms status
virtualbox, 5.1.22-1.mga6, 4.9.35-desktop-1.mga6, x86_64: installed-binary from 4.9.35-desktop-1.mga6
vboxadditions, 5.1.22-1.mga6, 4.9.35-desktop-1.mga6, x86_64: installed-binary from 4.9.35-desktop-1.mga6
virtualbox, 5.1.26-1.mga6, 4.9.43-desktop-1.mga6, x86_64: installed-binary from 4.9.43-desktop-1.mga6
vboxadditions, 5.1.26-1.mga6, 4.9.43-desktop-1.mga6, x86_64: installed-binary from 4.9.43-desktop-1.mga6


Am I chasing my tail on this, or can it be fixed more easily than a reinstall?

Thank you for any assistance.

Edit: I queried "journalctl -b -1 | grep ksmserver" and got this. The last line documents the point at which KDE restarts

Code: Select all

Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: completeShutdown: client  "requestshutdownhelper" ( 10d870696b000150401630200000270960011 )
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver:  We killed all clients. We have now clients.count()= 18
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 18
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 17
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 16
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 15
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 14
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 13
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 12
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 11
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 10
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: autostart service ("/bin/xembedsniproxy") finished with exit code  0
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 9
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: autostart service ("/usr/libexec/polkit-kde-authentication-agent-1") finished with exit code  0
Aug 29 08:18:27 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 8
Aug 29 08:18:28 spike ksmserver[27096]: ksmserver: autostart service ("/usr/bin/SpiderOak") finished with exit code  0
Aug 29 08:18:28 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 7
Aug 29 08:18:28 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 6
Aug 29 08:18:28 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 5
Aug 29 08:18:28 spike ksmserver[27096]: ksmserver: autostart service ("/bin/krunner") finished with exit code  0
Aug 29 08:18:28 spike ksmserver[27096]: ksmserver: autostart service ("/bin/plasmashell", "--shut-up") finished with exit code  0
Aug 29 08:18:28 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 4
Aug 29 08:18:29 spike ksmserver[27096]: ksmserver: KSMServer::completeKilling clients.count()= 3
Aug 29 08:18:34 spike ksmserver[27096]: The X11 connection broke (error 1). Did the X11 server die?


[doktor5000]

Am I chasing my tail on this, or can it be fixed more easily than a reinstall?

Well, if that is a fresh installation, not using a preexisting /home or migrated/restored existing settings, then unless you find out what is causing that issue the reinstallation will not help.
This is not like windows where for some issues due to missing debug/troubleshooting information (or often due to people being too lazy to find out he cause and maybe prevent that from reoccuring eventually) you have to reinstall as some issues can't be fixed otherwise.

If you want to find out why ksmserver dies you would at least have to run kdebugdialog and enable ksmserver and kdelibs-related debug information.
Also you would at least need to check the status of ksmserver when it's running normally via ps auxwww and pstree and such and in addition also the journal logs.

You should probably also disable any third-party program like the ones you list in your autostart to rule out whether those may contribute to the ksmserver issues you see.

Also try creating a new user and test with that.

[mark9117]

This is a fresh install but I am reusing my existing home directory.
When I get the chance, I'll spin up a fresh user and work under that account for awhile.

I did check the ksmserver process. Looks pretty straightforward:

Code: Select all

# ps auxwww |grep ksmserver
root      2472  0.0  0.0 106216   920 pts/1    S+   17:15   0:00 grep --color ksmserver
madams   10056  0.0  0.0  78492  5204 ?        S    Aug31   0:00 kwrapper5 /usr/bin/ksmserver
madams   10057  0.0  0.3 1585188 27372 ?       Sl   Aug31   0:04 /usr/bin/ksmserver


Kdebugdialog presents an issue:

Code: Select all

# kdebugdialog5
kf5.kcoreaddons.kaboutdata: Could not initialize the equivalent properties of Q*Application: no instance (yet) existing.
KUniqueApplication: Cannot find the D-Bus session server:  "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken."

KUniqueApplication: Pipe closed unexpectedly.

I've never heard of kdebugdialog before, so I'm having to pick up bits and pieces about it from the google. As near as I can tell, this just isn't working. I've verified that kcoreaddons are installed.

Still working on it.

[doktor5000]

This is a fresh install but I am reusing my existing home directory.

That is probably part of the cause of this issue. You should always explicitly mention that, as "fresh install" does usually imply a fresh user account.

Kdebugdialog presents an issue:

Code: Select all

# kdebugdialog5
kf5.kcoreaddons.kaboutdata: Could not initialize the equivalent properties of Q*Application: no instance (yet) existing.
KUniqueApplication: Cannot find the D-Bus session server:  "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken."

KUniqueApplication: Pipe closed unexpectedly.

kdebugdialog is just to enable more verbose / debug messages from the various KDE components.
If you started that as regular user then this is definitely a bigger issue, as it shows that your dbus session is broken, this would also explain why ksmserver (and probably other programs) are not working properly.
Please add the output as regular user of

Code: Select all

ps auxwww|grep -v grep|grep -i dbus
echo $DISPLAY
echo $DBUS_SESSION_BUS_ADDRESS


[mark9117]

After a night's rest and another reboot after today's lockup, kdebugdialog5 is working. I activated ksmserver debugging.

Here is the requested info:

Code: Select all

$ ps auxwww|grep -v grep|grep -i dbus
message+  1046  0.0  0.0  24668  4044 ?        Ss   Aug29   2:27 /usr/bin/dbus-daemon --system --address=systemd: --nofork --nopidfile --systemd-activation
lp        2128  0.0  0.0  72092  2316 ?        S    Aug29   0:00 /usr/lib/cups/notifier/dbus dbus://
sddm     14968  0.0  0.0  34916  2164 ?        S    14:43   0:00 dbus-launch --autolaunch 79f2231730124b1592f3e4d49b87c15b --binary-syntax --close-stderr
sddm     14969  0.0  0.0  23456  2344 ?        Ss   14:43   0:00 /usr/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session
madams   15098  0.0  0.0  34916  2016 ?        S    14:43   0:00 /usr/bin/dbus-launch --exit-with-session --sh-syntax
madams   15099  0.0  0.0  24568  3724 ?        Ss   14:43   0:00 /usr/bin/dbus-daemon --fork --print-pid 5 --print-address 15 --session
madams   16340  0.0  0.0  23456  3280 ?        S    14:43   0:00 /usr/bin/dbus-daemon --config-file=/usr/share/defaults/at-spi2/accessibility.conf --nofork --print-address 3


Code: Select all

$ echo $DISPLAY
:0


Code: Select all

$ echo $DBUS_SESSION_BUS_ADDRESS
unix:abstract=/tmp/dbus-zg1t4a149R,guid=1d183994439c245e382dcc5e59ac696c


When I got to this machine today, performance was sluggish. I found "/bin/plasmashell --shut-up" soaking up 95% of my processor cycles. I was trying to kill that when the lockup happened. Not sure exactly what the dynamic was, but I include this information because I noticed it. I have not had any trouble with that process prior to today.

Thank you for the attention.

[doktor5000]

You should probably start with a freshy booted machine, and not try to kill any arbitrary processes, and then provide a journalctl -ab log when the issue occurs.

[mark9117]

I haven't had a chance to do anything with this machine since last week. Got on it tonight and it locked up. Couldn't even access a different tty to kill Xorg which is what I usually have to do when kde freezes. I eventually ran fsck on it, fixed a long line of stuff and was able to get the machine to boot.

Subsequent to that, it locked up again. I killed Xorg and got kde back up. The log entry is too long to post here, so I've got a postbin link:

https://pastebin.com/8LefJ8NB

It's late. I should have been in bed an hour ago.

I appreciate the help. I'm happy to provide any more info.

[mark9117]

This afternoon I reinstalled the Mageia 6 on this hardware. I did not replace my /home directory. While it appears that the Akonadi server is running (see viewtopic.php?f=7&t=11964 for more), KDE crashed again while I was editing a photo in the Gimp.

This may be a problem with my hardware, specifically, my motherboard. In a last ditch effort to avoid having to apply my hard-pressed budget to new hardware, i filed a bug at KDE: https://bugs.kde.org/show_bug.cgi?id=384579

If it's hardware, nobody else will have this trouble. On the off chance that there's something not quite right about the KDE install I'm working on, or some other software trouble, perhaps it will get tracked down.

Just wanted to update the record.

[doktor5000]

This afternoon I reinstalled the Mageia 6 on this hardware. I did not replace my /home directory.

Meaning you are still using existing settings, and did not try with a freshly created user account yet?

[mark9117]

I did create a account and worked with it for a while. It experienced a lockup, so I reinstalled. Since the reinstall I have experienced one lockup, but that was early on and none since. I also haven't seen a lockup when I wasn't using the system.

Oddly, the lockups I was seeing when the system wasn't attended tended to happen when the second hand of the analog clock applet on the desktop was at the very top of the hour - directly over the "12'. I checked cron jobs and didn't find anything that should have been running at that time. Just kind of puzzling.

Edit: Meant to me that I've had limited time to work on it since the reinstall. I'm off tomorrow and Friday. Will try to test it more thoroughly.

[mark9117]

Tonight I decided to do a legitimate fresh install, home directory and all. I fired up my live usb installer for Mageia 6. I spent the next 2 hours attempting to reinstall this OS. It seemed like it was fighting me every step of the way. In the space of 2 hours or so, I attempted 5 installs, none of which got beyond the network configuration screen.

The first trouble I ran into was presaged by the bootloader failing to install correctly. I was installing grub2 with graphical interface, but grub2 text kept showing up. I'm not sure exactly what that was all about, but it shouldn't have happened and when it did I had further trouble.

I kept rebooting and getting one of a few failures, then reinstalling. The mouse and keyboard failing to respond at the network configuration dialog happened about every other install. Rebooting from that resulted in either a failure to get past an ACPI error:



When it did get past that error, it was likely to stop a bit further downstream:



Eventually, I just formatted every partition that wasn't either a data partition or /usr/local. That actually worked and I was able to get the OS installed. I reinstalled all my applets, applications and important configurations from scratch. I've copied over a lot of the more benign data from backup files. That's going to take some time, but so far things look good.

More as it develops.

[mark9117]

Things seem to be stable and 99% back. I'm marking this one "solved".

[doktor5000]

The two screenshots show bogus errors. Also the grub2 issue showing a text menu is also expected - it shouldn't be like that, but unless you add vga= boot option then it will only show text menu. See e.g. https://bugs.mageia.org/show_bug.cgi?id=19642
Simply add the vga= option to your /etc/default/grub and it's graphical boot again.

[mark9117]

Yeah, none of what I was seeing looked right. I'll file that info away, but that installation was just plain broken. I blew away all the old configurations, as you kept suggesting, and the whole world started to spin again.

Next, I'm doing the same to my HTPC. It's having trouble booting and can't seem to find its splash screen. I may open up a new thread on it to document the experience, so that's another kettle of fish.

Thanks again.

[Myles]

Seeing that this problem is marked as [SOLVED], should I open a new issue?

I too have been having a few niggling problems with #6 install using my existing /home and existing /var. (already added my 2c worth about KNotes not working - but found it was only with my existing /home drive; ok with a new user - + now getting all these msgs at boot about write cache on my USB drives and some other stuff).
Anyway, I too have experienced total system lockup - one just about an hour ago - more than 10 times just in the past week. I had 3 in an hour the other day! I don't know how to go about investigating system logs/messages properly, or even what I'm looking for relating to the lockups but if someone could help me out that would be great. At least I can have a look at the logs and see what might be causing this. I just hope I don't have to do what mark9117 did and wipe my /home drive and start again.

As always, any help greatly appreciated.

[mark9117]

I wish I could help. I'd say if you wiped everything including your /home directory and got a good install but still had trouble, open a new thread. I would also report that my desktop is working well, though I was no longer able to get knotes to load (again) and last night I had another KDE lockup. Other than that, I'm happy enough to not want to scrap hardware or reinstall for the sixth time.

Best of luck with yours. I'll be watching.

[Myles]

mark9117 thanks for that. I may have to do just that - had another lockup in the sometime in the past 90 mins or so (was upstairs), came back to - what I thought was - a hibernated screen (yet again, has happened multiple times prior) only to find I had to reset my PC and reboot - again!
I had no issues with #5.1 at all (I have re-installed #5.1 multiple times in the past 2 months or so using my existing /home and /var with no problems - except getting Plex working again (after I stuffed up the initial #6 upgrade and lots of things didn't work so re-installed #5.1).
If this happens again, I'll have to try a re-install and new /home and new /var and see how that fares! But even though you have done that you still had a lockup! When mine locks I can't get to a screen to restart KDE (if that is the problem is), the entire system is frozen.

[mark9117]

I have seen one KDE lockup since the last install and update. Not sure how much weight to give that lockup, but I am vigilant. The thing that's chafing me now is that I can't get Knotes to start. I'm trying to like Tomboy, but I used Knotes a lot to keep track of scraps of stuff.

If I'm honest, it looks to me as if this KDE implementation is not quite ready for prime time. Not sure what to make of that either, but I'm having all sorts of trouble -- large and small. Best of luck with yours. I'm curious to follow what happens.

[Myles]

Well, another lockup. I hear you about KNotes - it didn't work for me - it did start but I couldn't create a note. but I found it was something in my existing /home drive (from #5.1 that I hadn't formatted when re-installing #6) causes the problem. If I create a new user, I am able to create a note with KNotes.
Also, I agree about this implementation of KDE, as I said, I am suffering from a few niggling problems - the major one being these lockups (and not knowing how to go about finding out why they're happening)!
After the previous lockup an hour ago, I re-booted to be met with a text GRUB menu instead of the normal GRUB2 graphic menu screen. I had to reload the bootloader from the DVD.
I'm getting pretty tired of all this mucking around. I would rather just re-install and stay on #5.1 for the time being but it's probably out of support now - they only support the previous version for 90 days I think.

[Myles]

mark9117,
rant

Update. What follows illustrates just how bad this release of Mageia #6 is - for me anyway!
Well, after my last post, I had yet another lockup/freeze! So I decided to try your approach and start with a completely new /home drive. So I finally got the system fully installed with any updates that had been released since the DVD was released. Booted system and it worked. By this time it was after 9:30 pm (28th) so I wasn't going to start mucking around re-installing everything. So due to the late hour, I just locked the screen and left it.
Yesterday morning I came Into my study to start the arduous task of re-installing my programs, only to find I couldn't wake the system up - it was completely frozen yet again!!!!!!
This was a stock-standard install - with nothing else installed and a brand new /home drive - and it still froze!!!!!! I mean what is going on???

Anyway, I re-booted and persevered with re-installing all of my stuff - only to have it freeze again soon after starting that exercise! I have had 2 more since - one when I was using it early last night, so after re-booting I thought I'd wouldn't take the chance of a freeze and selected "Hibernate" instead. Well, system powered back on and re-booted, but instead of being presented with the "lock screen" as expected - I had a completely black screen - and the system was useless - again! So re-boot, again!
This is ridiculous, I never once had this problem with any previous versions, so I am going to open a new thread in the hope that someone can help me with how to look at the logs (which logs) and what I might be looking for as I have no idea.
/rant

[katnatek]

Sorry to post so latter
Did you have a intel graphic card?
If the answer is yes you could try to add video=SVIDEO-1:d to kernel options

Related links:
https://ubuntuforums.org/showthread.php ... st13664911
https://bbs.archlinux.org/viewtopic.php ... 4#p1689914

[Myles]

Thanks katnatek, but I have a Gigabyte nVidia GeForce GT710 2Gb video card (relatively new - 24th July this year - as the other nVidia GeForce GT610 1Gb died)