You are not logged in.
Hi have been having issues with my new desktop machine and a new installation of Arch. Occassionally if I leave my machine for a while and the monitor goes to sleep I am unable to wake the machine up again, it appears to completely lock up, unable to hit the power button to make the machine shutdown, nor able to ssh in.
Looking at journalctl I noticed that upower had been activated shortly before my machine locked up. I have disabled and masked upower, though this causes problems with cinnamon, at first I noticed that my usual icon set wasn't being used, running dmesg show this.
cinnamon-settin[449]: segfault at 0 ip 00007f9368853709 sp 00007ffc26da95a0 error 4 in libupower-glib.so.3.0.1[7f9368843000+22000]
I am still uncertain as to whether disabling upower will fix my problem as it would occur perhaps once a day or so.
I have noticed in dmesg this
[ 1.612233] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
I didn't think it was particularly important though, I'm not sure how to find out information about ACPI on my system in order to try to find an ACPI driver for it. I do have ASPM disabled in the UEFI bios though.
Using a live ubuntu sessison it seemed that ubuntu didn't issue the warning about the ACPI driver, so it's possible that there is an available driver and I just need to find which one I need... Though it's also possible that in ubuntu they've just supressed that warning message instead.
Last edited by Xanomead (2015-06-02 10:18:02)
Offline
With upower disabled my monitor was still turning off, but I have had no problems with the system freezing up.
I'm not exactly sure what is controlling the monitor sleep, though I suspect that it might have been the Xorg.conf file set up by aticonfig --initial as I'm using the proprietary drivers, which I realise are unsupported. I will be surprised if that was causing any kind of conflict with upower leading to the system freezes though. I have disabled DPMS in the Xorg.conf file and re-enabled upower in order to test if it does help though.
Offline
I'm not exactly sure what is controlling the monitor sleep, though I suspect that it might have been the Xorg.conf file set up by aticonfig --initial as I'm using the proprietary drivers, which I realise are unsupported.
I'm not sure what is causing the freezes on monitor switchoff but as for what controls that behaviour, I would have thought that would be down to the cinnamon-settings-daemon. You might want to have a poke around in dconf-editor under org/cinnamon and see if you can find the relevant setting.
Offline
I think I might have solved the problem now, as I'm using the proprietary drivers that was one area I thought could be a possible issue, I'm using the packages from the unofficial repos as on the wiki. I ran catalyst_build_module all to rebuild the drivers and rebooted, so far I haven't been able to recreate the freeze. And yes I could adjust things in cinnamon settings, just telling the screen to not turn off and manually turning it off, but that would mean if I ever forgot it'd stay on till I came back. Anyway, hopefully the issue is now fixed.
Offline
Well left the machine running when I went to bed and in the morning it wouldn't respond again.
Offline
You should be able to turn off the monitor with
xset dpms force offDo you experience the freeze then?
Offline
I have had a similar issue twice recently, now that you mention it. Didnt have time to poke around and find out what went wrong.
upower is enabled on mine as well.
will do some more tests and report.
and I dont think it is your video drivers - I have onboard intel, using the xf86-intel-video driver
Last edited by surfatwork (2015-06-01 21:57:14)
Offline
You should be able to turn off the monitor with
xset dpms force offDo you experience the freeze then?
No freeze with that as far as I can tell. When I disabled upower the screen would go blank after a period of idleness which I assume to have been DPMS taking over in place of the disabled cinnamon power settings.
Offline
I have had a similar issue twice recently, now that you mention it. Didnt have time to poke around and find out what went wrong.
upower is enabled on mine as well.
will do some more tests and report.
and I dont think it is your video drivers - I have onboard intel, using the xf86-intel-video driver
Thanks, I've changed the title to something a bit more informative, maybe we can get some extra help on this. My motherboard has an inbuilt Intel card, but it's disabled in the bios and lsmod doesn't show any modules for Intel graphics being used.
Offline
I found that this happens only when left unattended, with the screen off, for an extended period of time. A few minutes don't seem to cause a problem. In any case, the machine isn't dead though - I could blind switch to another tty, blind log in and blind type "sudo reboot", which did reboot the machine cleanly.
I have acpi_backlight=vendor in the boot params, without which the brightness keys dont work. Not sure if that makes a difference.
Probably upowerd is powering down other bits and then finding itself unable to turn them back on. I'll disable upower and see if it makes a difference.
Offline
I'll need to try switching to another tty and doing that, though I'm pretty sure that my machine was locked up and wouldn't respond to anything the last time I tried.
Offline
Well it locked up again and I tried switching to another tty and rebooting, nothing happened. Tried to ping the machine and there was no response, so hard rebooted it, and just to make sure pinged it and got a response.
This is very frustrating.
Offline
I just saw this old news update which applies to my machine and might apply to yours.
https://wiki.archlinux.org/index.php/Mi … de_Updates
Offline
So far I've gone two days without my machine locking up. Not long enough to convince me that the problem is resolved, but enough to give me some hope.
If it is fixed though, I'm not really sure what did it. According to the archwiki the output I get when running 'dmesg | grep microcode' indicates that there aren't any microcode updates for my CPU which is a fairly recent Haswell i3. It could also be the updates for Cinnamon that have fixed the problem. Of course it's possible that the issue isn't fixed and I've just been a bit lucky for the past two days.
Offline
Cinnamon was updated to 2.6 five days ago. That might be what fixed it (if indeed the problem is fixed).
Last edited by Chazza (2015-06-06 18:43:49)
Offline
It froze up again last night unfortunately. At this point I'm just going to wait for the newer proprietary Radeon drivers to be made available and see if that helps at all.
Offline