You are not logged in.
I’m running Arch (64 bit) on a Thinkpad x201.
The ssd is encrypted with dm-crypt (file system: btrfs), and this setup has been stable for a couple of years without reinstall, with daily updates.
I use Xmonad as my only window manager.
About a week ago, the following problem occurred for the first time:
In any application, or in urxvt, Xmonad will get unresponsive. That means, I can’t switch between workspaces, the status bar (trayer/xmobar) will not show any new information, I can’t change the window arrangement or open new terminals. None of the xmonad keyboard shortcuts will then work.
Basically, I’m stuck within the window I was working in at the moment the error occurs. This application will however work flawlessly, unless it tries to spawn a new window (e.g. the “save as” dialogue in libreoffice).
Ctrl+alt+F1-6 still works, so I can kill xmonad from another tty (only with "kill -9").
Sometimes it’s an hour until the error occurs, but mostly it’s seconds to minutes.
At first I suspected it had to do with libreoffice or chromium, but the error occurs even without any graphical application running.
I’m using urxvt-unicode-patched from aur, so I tried urxvt-unicode from the community repository, but it didn’t help.
Then I thought my Xmonad configuration was incompatible with later versions, but the same error occurs without trayer/xmobar and a very clean config (from ArchWiki).
Next I feared my RAM wasn’t working properly, but I ran memtest86+ for two hours without any errors being detected.
For now, I just make sure I have backups of my data and hope to solve the issue. Do you think it could be the ssd failing or the file system being corrupt?
Do you have an idea what could cause this behaviour?
edit: There are no error messages in xmonad or xorg log files. CPU and RAM usage seem to be normal.
Last edited by xiliyi (2016-08-25 09:53:05)
Offline
Have you looked which packages were upgraded around the time this first occurred?
I have a similar issue in i3 with linux 4.7 and 4.7.1 but not with 4.6.5 or earlier.
Offline
Have you looked which packages were upgraded around the time this first occurred?
I have a similar issue in i3 with linux 4.7 and 4.7.1 but not with 4.6.5 or earlier.
I’m not quite sure, it started about two weeks ago now, and that was the time of the update from 4.6-5 to 4.7-1.
edit: I have to say I was away for some time in August. Did pacman -Syyu on the 14th. During this update, linux was updated (4.6.5-1 to 4.7.1), but also the Intel video driver, xmonad, xmonad-contrib and xmobar (among many other packages).
The kernel received another update (4.7.1 to 4.7.1-1) on the 18th.
Last edited by xiliyi (2016-08-27 07:20:14)
Offline
Reverting linux, linux-headers (and tp_smapi for dependency reasons) didn’t help, same error still occurs. I’ll try going back to a previous version of xmonad.
edit: I updated linux, linux-headers and tp_smapi back to the latest version and then downgraded the display manager:
Reverted to xmobar-0.23.1-9, xmonad-0.12-7 and xmonad-contrib-0.12-6. After 10 minutes of switching between workspaces and opening lots of windows, the error hasn’t occurred! I’ll try for a little longer and get back to you.
Last edited by xiliyi (2016-08-27 07:57:53)
Offline
Too happy too soon. Error still occurs. Updating to the latest versions again.
I’m starting to fear it might be a problem with btrfs, because I found some weird looking directories in /. (Like described here: https://bbs.archlinux.org/viewtopic.php?id=212657)
Offline
Heya,
Started having the same problem today after a reboot (laptop hardly gets restarted - so I'm not sure which set updates it started borking on). As a side note; the only reason it got powered off, was because it ran outta juice (don't think anything got corrupted - it was sitting idle for quite a while).
Dont think its btrfs related; as this machine doesnt use btrfs anywhere.
Let me know if need any info from me; or if you find a solution.
Cheers
Offline
Dont think its btrfs related; as this machine doesnt use btrfs anywhere.
Let me know if need any info from me; or if you find a solution.
Cheers
Hi,
I haven’t been able to find a solution yet. Do you happen to use a non standard keyboard layout, or one that needed special attention somehow? Maybe the problem has nothing to do with xmonad crashing (since there are no error messages in the logs), but with the shortcuts not working properly? Just a thought.
I ruled out the kernel version, intel drivers, memory failure, xmonad/xmobar/trayer configs, xorg configuration and apps like chromium/libreoffice. What’s left?
Also: after killing xmonad from a different tty, and after restarting it, the pid gets significantly higher every time. Like, from around 800 (first time), to 3000, to 5000, up to 20000 and higher. Why would that be?
Last edited by xiliyi (2016-09-02 05:37:42)
Offline
Nothing non-standard really - just a Lenovo X220. Have Caps Lock mapped to Control via setxkbmap in .xinitrc, and a few brightness keys etc hooked up in my xmonad config.
Seems to be another post here with a similar issue. I had the lockup happen again a couple of times again today. The last time, I happened to be inside a terminal window and tried the
pkill -CONT xmonad-x86_64-lcommand. Everything came back to life after that, and it hasn't frozen again yet (only been a couple of hours since, but fingers crossed).
I guess you could try that; I'm not 100% sure but you could even try it from another TTY
Ctrl+Alt+F2and see if that helps. Still doesn't solve the issue; but at least you don't have to kill xmonad and lose all your current work.
Offline
For reference, I wasn’t able to solve the issue with xmonad, but (after 5 years of using xmonad) switched to i3.
Offline
Sad to hear.
I've had pretty good luck with issuing the pkill directly in to a terminal after xmonad fires up (no freezes since I started doing it).
Offline
I just switched to arch linux (from an ubuntu 16.10 setup) on a Thinkpad P50 and have started seeing this issue since I changed. I'm using the same XMonad setup as before. I noticed there's an entry for this in the the xmonad FAQ but the suggested solution given there didn't help. It sounds like it could be something similar though.
My arch installation is very standard, except that I am using xorg-server 1.19 (from testing). I thought that might be causing a problem but given that I'm not alone, that seems unlikely now. I will try using pkill next time it happens.
Offline
Heya, Tekul.
I posted a issue on the xmonad github: https://github.com/xmonad/xmonad/issues/55
From what I can make out, the its got to do with one of the log pipes locking up, and its Thinkpad specific.
I might try removing my battery stats from xmobar over the next week and see if I get anymore lockups - have a feeling its change in the kernel somewhere along the line that stops the battery stats from getting reporting. Mainly due to the fact I don't get xmonad locking up too much anymore, but if I do, its within a minute of me resuming from a suspend.
Offline
Hi illis.
Thanks for your reply. I also found this thread which seems likely to be the cause in my case, and maybe yours too if you are using dbus-launch. When I did get the problem, xmonad was indeed stopped (along with other processes) and 'pkill -CONT' from another login was enough to restart them.
I posted the info on github too 'cos it seems like something other xmonad users will likely get bitten by and it's confusing since it has the same symptoms as the FAQ entry.
Offline