I've been seeing this problem several times a day. I can't reproduce it in any reliable way and I have no idea how I can go about collecting the needed info... Anyway, here is what happens.
My most recent experience with "freeze" is not a solid one. The screen is frozen but I can still move the mouse around. Also, if there is a video playing it will freeze but the audio will continue on. I can get out of it by logging in from a terminal and doing a "killall -u myusername" which gets me back to the login screen.
I think "kwin" is the most likely suspect and have upgraded to kwin-5.4.0-4 from updates testing as the changelog included this nugget which gave me hope....
* Tue Aug 25 2015 Daniel Vrátil dvratil@redhat.com - 5.4.0-2 - add upstream patch to fix crash - make sure kwayland-integration is installed for kwin-wayland
But no... Still have issues. Any suggestion on what to look for to see if the culprit can be identified?
On 09/05/2015 07:04 AM, Ed Greshko wrote:
My most recent experience with "freeze" is not a solid one.
I suffer from the "freeze" issue as well. It's as bad as it's ever been (on average about once/day). If I'm in front of the machine and catch it quickly enough, I can drop to console and "systemctl restart sddm", but it degrades quickly. At its worst, I can ssh from another machine, but a "systemctl stop sddm" or either "systemctl isolate multi-user" just hangs, and the machine has to be rebooted.
FYI, always running the latest updates, and I've an NVidia GeForce GTX 650 on the nouveau driver.
On 09/05/15 21:13, Glenn Holmer wrote:
On 09/05/2015 07:04 AM, Ed Greshko wrote:
My most recent experience with "freeze" is not a solid one.
I suffer from the "freeze" issue as well. It's as bad as it's ever been (on average about once/day). If I'm in front of the machine and catch it quickly enough, I can drop to console and "systemctl restart sddm", but it degrades quickly. At its worst, I can ssh from another machine, but a "systemctl stop sddm" or either "systemctl isolate multi-user" just hangs, and the machine has to be rebooted.
FYI, always running the latest updates, and I've an NVidia GeForce GTX 650 on the nouveau driver.
I'm running a GeForce GTX 660 with the nVidia drivers. Sort of good to know the failure is the same with nouveau.
I suppose I should try running something other than sddm to see if that makes any difference.
Really is frustrating since one can't predict when/what triggers the "freeze". It seemed to be worse when I would be editing a video file. But that could also have just been Murphy's Law at work.
On Sat, Sep 5, 2015 at 5:04 AM, Ed Greshko ed.greshko@greshko.com wrote:
I've been seeing this problem several times a day. I can't reproduce it in any reliable way and I have no idea how I can go about collecting the needed info... Anyway, here is what happens.
I've been having this problem consistently (as in several times a day) since the move to F22. You can check out the bug report and the traces here. Currently it's assigned as an ati-driver issue, but like you seems curious that it also happens with other video (nvidia) which makes me think it's related somehow to whatever was introduced by F22 by kde... I've tried GNOME and LXQt and no problems with either of them - but I'm not the expert in this so I'm deferring to greater minds than me. ;-)
It is extremely frustrating... continues to linger. I will provide whatever type of doc required - but since I'm not the expert someone needs to advise what they need. For me, as far as I can tell it appears to be triggered by notification messages - but that could be purely anecdotal.
In the meantime I've memorized killall plasmashell ; plasmashell
Oops forgot the URL: https://bugzilla.redhat.com/show_bug.cgi?id=1217844
On Sat, Sep 5, 2015 at 8:18 AM, Gerald B. Cox gbcox@bzb.us wrote:
On Sat, Sep 5, 2015 at 5:04 AM, Ed Greshko ed.greshko@greshko.com wrote:
I've been seeing this problem several times a day. I can't reproduce it in any reliable way and I have no idea how I can go about collecting the needed info... Anyway, here is what happens.
I've been having this problem consistently (as in several times a day) since the move to F22. You can check out the bug report and the traces here. Currently it's assigned as an ati-driver issue, but like you seems curious that it also happens with other video (nvidia) which makes me think it's related somehow to whatever was introduced by F22 by kde... I've tried GNOME and LXQt and no problems with either of them - but I'm not the expert in this so I'm deferring to greater minds than me. ;-)
It is extremely frustrating... continues to linger. I will provide whatever type of doc required - but since I'm not the expert someone needs to advise what they need. For me, as far as I can tell it appears to be triggered by notification messages - but that could be purely anecdotal.
In the meantime I've memorized killall plasmashell ; plasmashell
On 09/05/2015 10:18 AM, Gerald B. Cox wrote:
Oops forgot the URL: https://bugzilla.redhat.com/show_bug.cgi?id=1217844
Read over that thoroughly; whatever fixes were done hasn't resolved it for me. And I do have libxcb-1.11-8.fc22.
On Saturday 05 Sep 2015 17:09:00 Glenn Holmer wrote:
On 09/05/2015 10:18 AM, Gerald B. Cox wrote:
Oops forgot the URL: https://bugzilla.redhat.com/show_bug.cgi?id=1217844
Read over that thoroughly; whatever fixes were done hasn't resolved it for me. And I do have libxcb-1.11-8.fc22.
Just to add to this, I'm seeing something similar here on this box. IE a black screen and movable mouse pointer. However "sometimes" I have been able to switch to another virtual Desktop via CTRL Fx or bring up Yakuake F12 to then quit and restart plasma-shell.
F22, Nouveau
Colin
On 09/05/2015 06:30 PM, Colin J Thomson wrote:
On Saturday 05 Sep 2015 17:09:00 Glenn Holmer wrote:
On 09/05/2015 10:18 AM, Gerald B. Cox wrote:
Oops forgot the URL: https://bugzilla.redhat.com/show_bug.cgi?id=1217844
Read over that thoroughly; whatever fixes were done hasn't resolved it for me. And I do have libxcb-1.11-8.fc22.
Just to add to this, I'm seeing something similar here on this box.
Yeah... a person can only take so much. I have MATE/Compiz on my test box, and am spending more and more time there.
My suggestion is to create a new account in your system, then use it, if the problem still there, that means the issue from KDE Plasma, if not, that mean you should clear your kde configuration files and your cache files in your home directory.It happened with me, but now, everything is just fine.
On Friday, September 11, 2015 11:26 PM, Glenn Holmer shadowm@lyonlabs.org wrote:
On 09/05/2015 06:30 PM, Colin J Thomson wrote:
On Saturday 05 Sep 2015 17:09:00 Glenn Holmer wrote:
On 09/05/2015 10:18 AM, Gerald B. Cox wrote:
Oops forgot the URL: https://bugzilla.redhat.com/show_bug.cgi?id=1217844
Read over that thoroughly; whatever fixes were done hasn't resolved it for me. And I do have libxcb-1.11-8.fc22.
Just to add to this, I'm seeing something similar here on this box.
Yeah... a person can only take so much. I have MATE/Compiz on my test box, and am spending more and more time there.
On 09/05/15 23:18, Gerald B. Cox wrote:
On Sat, Sep 5, 2015 at 5:04 AM, Ed Greshko <ed.greshko@greshko.com mailto:ed.greshko@greshko.com> wrote:
I've been seeing this problem several times a day. I can't reproduce it in any reliable way and I have no idea how I can go about collecting the needed info... Anyway, here is what happens.
I've been having this problem consistently (as in several times a day) since the move to F22. You can check out the bug report and the traces here. Currently it's assigned as an ati-driver issue, but like you seems curious that it also happens with other video (nvidia) which makes me think it's related somehow to whatever was introduced by F22 by kde... I've tried GNOME and LXQt and no problems with either of them - but I'm not the expert in this so I'm deferring to greater minds than me. ;-)
It is extremely frustrating... continues to linger. I will provide whatever type of doc required - but since I'm not the expert someone needs to advise what they need. For me, as far as I can tell it appears to be triggered by notification messages - but that could be purely anecdotal.
In the meantime I've memorized killall plasmashell ; plasmashell
Well, for me at least, the killall plasmashell ; plasmashell is only meaningful when the entries on my systray disappear or when the background images go to black. When my system hits the "freeze" no keyboard input is possible.
On Sat, Sep 5, 2015 at 8:24 AM, Ed Greshko ed.greshko@greshko.com wrote:
Well, for me at least, the killall plasmashell ; plasmashell is only meaningful when the entries on my systray disappear or when the background images go to black. When my system hits the "freeze" no keyboard input is possible.
I've found I am able to usually break out with ALT-F2 - sometimes it takes 10, 15 or more times, but eventually I get the command box.
You might also try alternating with ALT-TAB, that also seems to help.
On Sat, Sep 5, 2015 at 8:46 AM, Gerald B. Cox gbcox@bzb.us wrote:
On Sat, Sep 5, 2015 at 8:24 AM, Ed Greshko ed.greshko@greshko.com wrote:
Well, for me at least, the killall plasmashell ; plasmashell is only meaningful when the entries on my systray disappear or when the background images go to black. When my system hits the "freeze" no keyboard input is possible.
I've found I am able to usually break out with ALT-F2 - sometimes it takes 10, 15 or more times, but eventually I get the command box.
On 2015-09-05 06:04, Ed Greshko wrote:
I've been seeing this problem several times a day. I can't reproduce it in any reliable way and I have no idea how I can go about collecting the needed info... Anyway, here is what happens.
My most recent experience with "freeze" is not a solid one. The screen is frozen but I can still move the mouse around. Also, if there is a video playing it will freeze but the audio will continue on. I can get out of it by logging in from a terminal and doing a "killall -u myusername" which gets me back to the login screen.
I think "kwin" is the most likely suspect and have upgraded to kwin-5.4.0-4 from updates testing as the changelog included this nugget which gave me hope....
- Tue Aug 25 2015 Daniel Vrátil dvratil@redhat.com - 5.4.0-2
- add upstream patch to fix crash
- make sure kwayland-integration is installed for kwin-wayland
But no... Still have issues. Any suggestion on what to look for to see if the culprit can be identified?
Have had similar issues on a machine and in all cases it is the kde bug reporter that has locked the session applications. Open a tty session and kill it and everything works as expected.
I cannot remember the application at this second.
Robin
On 09/15/2015 11:52 PM, Robin Laing wrote:
Have had similar issues on a machine and in all cases it is the kde bug reporter that has locked the session applications. Open a tty session and kill it and everything works as expected.
I cannot remember the application at this second.
Anybody who can? Is there a bug report against this?
On Wed, 16 Sep 2015 08:48:51 -0500, Glenn Holmer shadowm@lyonlabs.org wrote:
On 09/15/2015 11:52 PM, Robin Laing wrote:
Have had similar issues on a machine and in all cases it is the kde bug reporter that has locked the session applications. Open a tty session and kill it and everything works as expected.
I cannot remember the application at this second.
Anybody who can? Is there a bug report against this?
When I got the "near freeze" yesterday, I tried killing drkonqi. That seemed to do the trick. Robin, is that the program you mean?
On 09/17/2015 08:44 AM, George Avrunin wrote:
On Wed, 16 Sep 2015 08:48:51 -0500, Glenn Holmer shadowm@lyonlabs.org wrote:
On 09/15/2015 11:52 PM, Robin Laing wrote:
Have had similar issues on a machine and in all cases it is the kde bug reporter that has locked the session applications. Open a tty session and kill it and everything works as expected.
I cannot remember the application at this second.
Anybody who can? Is there a bug report against this?
When I got the "near freeze" yesterday, I tried killing drkonqi. That seemed to do the trick. Robin, is that the program you mean?
Dropping back to console and issuing "killall -9 drkonqi" when Plasma 5 did its daily freeze just now allowed the desktop to recover.
So this is the real problem behind the freeze? The crash is not a good thing of course, but why does drkonqi whack the machine like this?
Glenn Holmer wrote:
On 09/17/2015 08:44 AM, George Avrunin wrote:
On Wed, 16 Sep 2015 08:48:51 -0500, Glenn Holmer shadowm@lyonlabs.org wrote:
On 09/15/2015 11:52 PM, Robin Laing wrote:
Have had similar issues on a machine and in all cases it is the kde bug reporter that has locked the session applications. Open a tty session and kill it and everything works as expected.
I cannot remember the application at this second.
Anybody who can? Is there a bug report against this?
When I got the "near freeze" yesterday, I tried killing drkonqi. That seemed to do the trick. Robin, is that the program you mean?
Dropping back to console and issuing "killall -9 drkonqi" when Plasma 5 did its daily freeze just now allowed the desktop to recover.
So this is the real problem behind the freeze? The crash is not a good thing of course, but why does drkonqi whack the machine like this?
Based on these findings (excellent detective work!), I've opened a bug to track it at
https://bugzilla.redhat.com/show_bug.cgi?id=1264486
-- rex
On 09/18/15 23:02, Rex Dieter wrote:
Glenn Holmer wrote:
On 09/17/2015 08:44 AM, George Avrunin wrote:
On Wed, 16 Sep 2015 08:48:51 -0500, Glenn Holmer shadowm@lyonlabs.org wrote:
On 09/15/2015 11:52 PM, Robin Laing wrote:
Have had similar issues on a machine and in all cases it is the kde bug reporter that has locked the session applications. Open a tty session and kill it and everything works as expected.
I cannot remember the application at this second.
Anybody who can? Is there a bug report against this?
When I got the "near freeze" yesterday, I tried killing drkonqi. That seemed to do the trick. Robin, is that the program you mean?
Dropping back to console and issuing "killall -9 drkonqi" when Plasma 5 did its daily freeze just now allowed the desktop to recover.
So this is the real problem behind the freeze? The crash is not a good thing of course, but why does drkonqi whack the machine like this?
Based on these findings (excellent detective work!), I've opened a bug to track it at
I'll follow the BZ but just FYI I've not encountered a "near freeze" in at least a week.
On 09/19/15 07:19, Ed Greshko wrote:
I'll follow the BZ but just FYI I've not encountered a "near freeze" in at least a week.
The best way to get a "near freeze" is to tell everyone that you've not had one in at least a week. Seriously, less than one hour after writing this I got a "near freeze" and was able to confirm that "killall -9 drkonqi" got me going again.
Thanks Glenn.
On 09/18/2015 06:49 PM, Ed Greshko wrote:
On 09/19/15 07:19, Ed Greshko wrote: The best way to get a "near freeze" is to tell everyone that you've not had one in at least a week. Seriously, less than one hour after writing this I got a "near freeze" and was able to confirm that "killall -9 drkonqi" got me going again.
I've reached my limit with the Plasma 5 freezes. Speaking to users who've tried alpha builds: do you think there's any hope of this situation improving in Fedora 23? I see we're scheduled for a beta this week. If the cause of the freezes hasn't been found yet, I'll start making other plans for this machine.
On Sunday 20 Sep 2015 18:36:46 Glenn Holmer wrote:
On 09/18/2015 06:49 PM, Ed Greshko wrote:
On 09/19/15 07:19, Ed Greshko wrote: The best way to get a "near freeze" is to tell everyone that you've not had one in at least a week. Seriously, less than one hour after writing this I got a "near freeze" and was able to confirm that "killall -9 drkonqi" got me going again.
I've reached my limit with the Plasma 5 freezes. Speaking to users who've tried alpha builds: do you think there's any hope of this situation improving in Fedora 23? I see we're scheduled for a beta this week. If the cause of the freezes hasn't been found yet, I'll start making other plans for this machine.
Well I've been running 5.4.1 for a while now, which I think it has been pushed to stable and I have not seen any actual "freeze" at all on this box.
Nouveau drivers FWIW
I get do the odd crash of plasma (some days it is fine) but I can always switch to another virtual desktop and restart plasma. Which is strange to say the least. I have not yet found the trigger of this and Drkonq never shows up when it happens. Bit of a weird one.
My only real gripe with P5 now is the amount of RAM it eats over the course of a day.. up to almost 5 gig now and hitting swap! (8 gig on this box) uptime ~ 17 hours. So kquitapp plasmashell resets it nicely.
Still searching kde.bugs for this one.
Not much help but thought I would share this.
Colin
On Monday 21 Sep 2015 00:58:37 Colin J Thomson wrote:
My only real gripe with P5 now is the amount of RAM it eats over the course of a day.. up to almost 5 gig now and hitting swap! (8 gig on this box) uptime ~ 17 hours. So kquitapp plasmashell resets it nicely.
I have tried removing various plasmoids and tray icons but still get the high CPU usage, so...
Still searching kde.bugs for this one.
I have found found an interesting thread on the kde forum and a link to a bug report that points to kactivitymanager being to source of the high CPU usage and maybe the huge amount of RAM P5 eats up over a period of time.
https://forum.kde.org/viewtopic.php?f=289&t=121533&start=30
Colin