Re: F20: System doesn't resume correctly from suspend any more
On Mon, 2013-12-16 at 13:50 -0800, Adam Williamson wrote: Shell is running and alt-f2, r recovers, FWIW. Hrm, I tired 'alt f2 r' and it didn't recover for me. I'll try again. The issue Neal reported looks like the same thing? I'll try to catch X logs next time this occurs. -- Thanks, Warm regards, Ankur (FranciscoD) http://fedoraproject.org/wiki/User:Ankursinha Join Fedora! Come talk to us! http://fedoraproject.org/wiki/Fedora_Join_SIG signature.asc Description: This is a digitally signed message part -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Thu, 2013-12-19 at 13:02 +1100, Ankur Sinha wrote: On Mon, 2013-12-16 at 13:50 -0800, Adam Williamson wrote: Shell is running and alt-f2, r recovers, FWIW. Hrm, I tired 'alt f2 r' and it didn't recover for me. I'll try again. Only works for Shell, of course. The issue Neal reported looks like the same thing? It looks similar, but I'm still trying to tell for sure. I'll try to catch X logs next time this occurs. I don't see anything useful in them, just X re-evaluating the modes (which is normal, I think). No errors. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Dec 16, 2013, at 1:47 PM, nonamedotc nonamed...@gmail.com wrote: On 12/11/2013 06:57 PM, Chris Murphy wrote: On Dec 11, 2013, at 5:41 PM, Ankur Sinha sanjay.an...@gmail.com wrote: On Wed, 2013-12-11 at 17:25 -0600, nonamedotc wrote: Is anyone still seeing this? Just curious as I have also been seeing this ... I am. I can't reproduce it reliably though. It happens randomly here. Maybe need to go to more aggressive kernel parameters to get more information. debug loglevel=7 udev_log=debug Chris Murphy I might be **misunderstanding** - but with the kernel-3.12.5-301, the suspend issue seems to be gone (in my tests - 3 repetitions). Suspend *seems to work* without issues - Is this the case across the board or is this a one off occurance. I'm not sure. On 3.13.rc2 my laptop does suspend with pulsing power button as expected. On resume, I have a black screen and I can't ssh into it. So it's dead. When I reboot the logs show that it suspended OK and nothing but the reboot after that. So I have no information. This is with systemd debugging enabled. Chris Murphy -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On 12/16/2013 03:09 PM, Chris Murphy wrote: On Dec 16, 2013, at 1:47 PM, nonamedotc nonamed...@gmail.com wrote: I might be **misunderstanding** - but with the kernel-3.12.5-301, the suspend issue seems to be gone (in my tests - 3 repetitions). Suspend *seems to work* without issues - Is this the case across the board or is this a one off occurance. I'm not sure. On 3.13.rc2 my laptop does suspend with pulsing power button as expected. On resume, I have a black screen and I can't ssh into it. So it's dead. When I reboot the logs show that it suspended OK and nothing but the reboot after that. So I have no information. This is with systemd debugging enabled. Chris Murphy Thanks. Let me keep looking at this more carefully to understand what is going on. -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Mon, 2013-12-16 at 14:09 -0700, Chris Murphy wrote: On Dec 16, 2013, at 1:47 PM, nonamedotc nonamed...@gmail.com wrote: On 12/11/2013 06:57 PM, Chris Murphy wrote: On Dec 11, 2013, at 5:41 PM, Ankur Sinha sanjay.an...@gmail.com wrote: On Wed, 2013-12-11 at 17:25 -0600, nonamedotc wrote: Is anyone still seeing this? Just curious as I have also been seeing this ... I am. I can't reproduce it reliably though. It happens randomly here. Maybe need to go to more aggressive kernel parameters to get more information. debug loglevel=7 udev_log=debug Chris Murphy I might be **misunderstanding** - but with the kernel-3.12.5-301, the suspend issue seems to be gone (in my tests - 3 repetitions). Suspend *seems to work* without issues - Is this the case across the board or is this a one off occurance. I'm not sure. On 3.13.rc2 my laptop does suspend with pulsing power button as expected. On resume, I have a black screen and I can't ssh into it. So it's dead. When I reboot the logs show that it suspended OK and nothing but the reboot after that. So I have no information. This is with systemd debugging enabled. I'm consistently getting the case where I can see the bootsplash but Shell is running and alt-f2, r recovers, FWIW. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Dec 16, 2013, at 2:50 PM, Adam Williamson awill...@redhat.com wrote: On Mon, 2013-12-16 at 14:09 -0700, Chris Murphy wrote: On Dec 16, 2013, at 1:47 PM, nonamedotc nonamed...@gmail.com wrote: On 12/11/2013 06:57 PM, Chris Murphy wrote: On Dec 11, 2013, at 5:41 PM, Ankur Sinha sanjay.an...@gmail.com wrote: On Wed, 2013-12-11 at 17:25 -0600, nonamedotc wrote: Is anyone still seeing this? Just curious as I have also been seeing this ... I am. I can't reproduce it reliably though. It happens randomly here. Maybe need to go to more aggressive kernel parameters to get more information. debug loglevel=7 udev_log=debug Chris Murphy I might be **misunderstanding** - but with the kernel-3.12.5-301, the suspend issue seems to be gone (in my tests - 3 repetitions). Suspend *seems to work* without issues - Is this the case across the board or is this a one off occurance. I'm not sure. On 3.13.rc2 my laptop does suspend with pulsing power button as expected. On resume, I have a black screen and I can't ssh into it. So it's dead. When I reboot the logs show that it suspended OK and nothing but the reboot after that. So I have no information. This is with systemd debugging enabled. I'm consistently getting the case where I can see the bootsplash but Shell is running and alt-f2, r recovers, FWIW. I use nomodeset/basic graphics, so it's possible there's an issue with it and not nouveau. But I'd still expect ssh to work if services are actually up and running. Is this a use case for netconsole? Or is there something better/more recent? Chris Murphy -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On 12/10/2013 05:32 PM, Chris Murphy wrote: On Dec 10, 2013, at 4:23 PM, Chris Murphy li...@colorremedies.com wrote: Check systemd devel list, thread udev rule not applied after resume from sleep and see if that's related. http://lists.freedesktop.org/archives/systemd-devel/2013-December/015325.html Chris Murphy Is anyone still seeing this? Just curious as I have also been seeing this ... -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Dec 11, 2013, at 5:41 PM, Ankur Sinha sanjay.an...@gmail.com wrote: On Wed, 2013-12-11 at 17:25 -0600, nonamedotc wrote: Is anyone still seeing this? Just curious as I have also been seeing this ... I am. I can't reproduce it reliably though. It happens randomly here. Maybe need to go to more aggressive kernel parameters to get more information. debug loglevel=7 udev_log=debug Chris Murphy -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Tue, 2013-12-10 at 09:45 +1100, Ankur Sinha wrote: Downgrade - works Re-Upgrade to reproduce - couldn't reproduce, started working. Oh well. *sigh* it happens again. Resume brought me to the fedora bubble. No gnome-shell screen lock. Any ideas on what to look for in journalctl? -- Thanks, Warm regards, Ankur (FranciscoD) http://fedoraproject.org/wiki/User:Ankursinha Join Fedora! Come talk to us! http://fedoraproject.org/wiki/Fedora_Join_SIG signature.asc Description: This is a digitally signed message part -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Wed, 2013-12-11 at 08:39 +1100, Ankur Sinha wrote: On Tue, 2013-12-10 at 09:45 +1100, Ankur Sinha wrote: Downgrade - works Re-Upgrade to reproduce - couldn't reproduce, started working. Oh well. *sigh* it happens again. Resume brought me to the fedora bubble. No gnome-shell screen lock. Any ideas on what to look for in journalctl? I'm seeing something like this on my desktop lately, but I think it's actually suspend failing messily, not a problem on resume. Do you see something like this around the time of the attempted suspend? Dec 09 01:08:55 adam.happyassassin.net systemd[1]: Started Suspend. Dec 09 01:08:55 adam.happyassassin.net systemd[1]: Requested transaction contradicts existing jobs: File exists Dec 09 01:08:55 adam.happyassassin.net systemd[1]: Service sleep.target is not needed anymore. Stopping. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Tue, 2013-12-10 at 14:04 -0800, Adam Williamson wrote: I'm seeing something like this on my desktop lately, but I think it's actually suspend failing messily, not a problem on resume. Do you see something like this around the time of the attempted suspend? I haven't been able to find any errors relating to suspend. It shows these lines for each suspend I attempted: [asinha@ankur-laptop ~]$ sudo journalctl --no-pager --since=2013-12-10 00:01 | egrep suspend Dec 10 08:27:04 ankur-laptop kernel: Suspending console(s) (use no_console_suspend to debug) Dec 10 08:27:04 ankur-laptop kernel: nouveau [ DRM] suspending fbcon... Dec 10 08:27:04 ankur-laptop kernel: nouveau [ DRM] suspending display... Dec 10 08:27:04 ankur-laptop kernel: mei_me :00:16.0: suspend Dec 10 08:27:04 ankur-laptop kernel: nouveau [ DRM] suspending client object trees... Dec 10 08:27:04 ankur-laptop kernel: nouveau [ DRM] suspending kernel object tree... Dec 10 08:27:04 ankur-laptop kernel: PM: suspend of devices complete after 2224.750 msecs Dec 10 08:27:04 ankur-laptop kernel: PM: late suspend of devices complete after 0.232 msecs Dec 10 08:27:04 ankur-laptop kernel: PM: noirq suspend of devices complete after 59.778 msecs Resume does show some errors, but it spews System resumed after, so I'm not sure about this again: [asinha@ankur-laptop ~]$ sudo journalctl --no-pager --since=2013-12-10 00:01 | egrep systemd | egrep suspend|resume Dec 10 08:27:00 ankur-laptop systemd-sleep[20600]: System resumed. Dec 10 08:46:35 ankur-laptop systemd-udevd[159]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 10 08:46:35 ankur-laptop systemd-udevd[159]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' Dec 10 09:14:01 ankur-laptop systemd-udevd[160]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 10 09:14:01 ankur-laptop systemd-udevd[160]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' Dec 10 09:16:42 ankur-laptop systemd-sleep[2495]: System resumed. Dec 10 09:19:51 ankur-laptop systemd-udevd[159]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 10 09:19:51 ankur-laptop systemd-udevd[159]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' Dec 10 09:22:30 ankur-laptop systemd-sleep[2466]: System resumed. Dec 10 09:23:20 ankur-laptop systemd-sleep[2855]: System resumed. Dec 11 08:12:57 ankur-laptop systemd-sleep[21685]: System resumed. Dec 11 08:17:56 ankur-laptop systemd-udevd[159]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 11 08:17:56 ankur-laptop systemd-udevd[159]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' [asinha@ankur-laptop ~]$ There is no /etc/udev/rules.d/99-resume.rules file: [asinha@ankur-laptop ~]$ ls /etc/udev/rules.d/ 98-kexec.rules Here's something else, from dracut: [asinha@ankur-laptop ~]$ sudo journalctl --no-pager --since=2013-12-11 08:00 | egrep dracut Dec 11 08:17:56 ankur-laptop systemd[1]: Starting dracut cmdline hook... Dec 11 08:17:56 ankur-laptop dracut-cmdline[94]: dracut-20 (Heisenbug) dracut-034-64.git20131205.fc20 Dec 11 08:17:56 ankur-laptop dracut-cmdline[94]: warn Cancelling resume operation. Device not found.; Dec 11 08:17:56 ankur-laptop systemd[1]: Started dracut cmdline hook. Dec 11 08:17:56 ankur-laptop systemd[1]: Started dracut pre-udev hook. I haven't been able to get any more info on what device this is referring to though. At the moment, I can only think of downgrading systemd as a way to confirm it's the culprit here. -- Thanks, Warm regards, Ankur (FranciscoD) http://fedoraproject.org/wiki/User:Ankursinha Join Fedora! Come talk to us! http://fedoraproject.org/wiki/Fedora_Join_SIG signature.asc Description: This is a digitally signed message part -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Dec 10, 2013, at 3:22 PM, Ankur Sinha sanjay.an...@gmail.com wrote: [asinha@ankur-laptop ~]$ sudo journalctl --no-pager --since=2013-12-10 00:01 | egrep systemd | egrep suspend|resume Dec 10 08:27:00 ankur-laptop systemd-sleep[20600]: System resumed. Dec 10 08:46:35 ankur-laptop systemd-udevd[159]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 10 08:46:35 ankur-laptop systemd-udevd[159]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' Dec 10 09:14:01 ankur-laptop systemd-udevd[160]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 10 09:14:01 ankur-laptop systemd-udevd[160]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' Dec 10 09:16:42 ankur-laptop systemd-sleep[2495]: System resumed. Dec 10 09:19:51 ankur-laptop systemd-udevd[159]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 10 09:19:51 ankur-laptop systemd-udevd[159]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' Dec 10 09:22:30 ankur-laptop systemd-sleep[2466]: System resumed. Dec 10 09:23:20 ankur-laptop systemd-sleep[2855]: System resumed. Dec 11 08:12:57 ankur-laptop systemd-sleep[21685]: System resumed. Dec 11 08:17:56 ankur-laptop systemd-udevd[159]: unknown key '\nSYMLINK' in /etc/udev/rules.d/99-resume.rules:1 Dec 11 08:17:56 ankur-laptop systemd-udevd[159]: invalid rule '/etc/udev/rules.d/99-resume.rules:1' [asinha@ankur-laptop ~]$ There is no /etc/udev/rules.d/99-resume.rules file: From reading the systemd devel list my understanding is that udev rules only apply when hardware appears or disappears, but not on suspend/resume. So if the device in question, which ought to be the device rootfs is on, is going away and returning on suspend/resume then udev applies. Otherwise it's the kernel that needs to restore state on resume. You might look at 'udevadm monitor' before and after suspending. Also I'm confused about systemd reporting suspend when the computer indicates it's definitely not asleep. It's hibernated. And anaconda's partition code for suggested swap for hibernate support, isn't being employed (apparently at all for anyone). On my laptops, suggested swap is always the same as RAM. But the anaconda swap code for hibernation says it should be anywhere from 3x memory (for small amounts of RAM) up to 1.5x memory if you have gobs of it, for my two laptops the code suggests it should be 2x memory but suggested swap as computed by anaconda isn't 2x. I have to manually create swaps of that size. At the moment, I can only think of downgrading systemd as a way to confirm it's the culprit here. Check systemd devel list, thread udev rule not applied after resume from sleep and see if that's related. Chris Murphy -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Dec 10, 2013, at 4:23 PM, Chris Murphy li...@colorremedies.com wrote: Check systemd devel list, thread udev rule not applied after resume from sleep and see if that's related. http://lists.freedesktop.org/archives/systemd-devel/2013-December/015325.html Chris Murphy -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
Re: F20: System doesn't resume correctly from suspend any more
On Tue, 2013-12-10 at 09:07 +1100, Ankur Sinha wrote: I'm going to try downgrading the package to see if that fixes it. Downgrade - works Re-Upgrade to reproduce - couldn't reproduce, started working. Oh well. -- Thanks, Warm regards, Ankur (FranciscoD) http://fedoraproject.org/wiki/User:Ankursinha Join Fedora! Come talk to us! http://fedoraproject.org/wiki/Fedora_Join_SIG signature.asc Description: This is a digitally signed message part -- test mailing list test@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
