System compositor fails to start - Failed to set the current VT mode: Input/output error (5)

[Fernando Correa Neto]

After following the instructions on http://www.olli-ries.com/running-mir/, all I could see was a balck screen with a pointer in the middle. After rebootin, I could verify that X was running because instead of the pointer, there was the classical "X" pointer.

The hardware is an Lenovo X1 Carbon. Let me know what info you need so I can provide more details about it.

I'm attaching lightdm files.

[Robert Ancell (robert-ancell)]

The specific problem here is in unity-system-compositor.log - Mir fails to control the VT and throws an exception:

ERROR: /build/buildd/mir-0.0.5bzr774saucy0/src/server/graphics/gbm/linux_virtual_terminal.cpp(137): Throw in function virtual void mir::graphics::gbm::LinuxVirtualTerminal::register_switch_handlers(mir::MainLoop&, const std::function<bool()>&, const std::function<bool()>&)
Dynamic exception type: boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> >
std::exception::what: Failed to set the current VT mode
[boost::errinfo_errno_*] = 5, "Input/output error"

[Robert Ancell (robert-ancell)]

Annoyingly for some reason the log in comment 1 seems corrupt. The lightdm.log from comment 19 in bug 1109957 doesn't have corruption.

[Robert Ancell (robert-ancell)]

Also note the log shows that LightDM tried to run standard X after the compositor fails, but in both Fernando and Mark's case it doesn't seem to have worked correctly.

[Mark Shuttleworth (sabdfl)]

Oddly, this morning my laptop booted straight into Mir with no problems. Either the issue is fixed for me, or it's racy and I just (finally!) got lucky.

[Ubuntu QA Website (ubuntuqa)]

This bug has been reported on the Ubuntu Package testing tracker.

A list of all reports related to this bug can be found here:
http://packages.qa.ubuntu.com/qatracker/reports/bugs/1195509

[Fernando Correa Neto (fcorrea)]

I've tried this morning again and it's working now. I could see the two pointers and also the unity-system-compositor process running.
However, while updating this bug my session ended abruptly and I was kicked out back into the lightdm login screen. After login, no more crashes.

Attaching the new log files

[kevin gunn (kgunn72)]

I'm suggesting we close this, there has been a handful of bugs fixes/development which has occured in recent weeks to address these boot issues. It appears we're fairly stable/reliable for boot. Please reopen if your experience is still failing to boot on the system-compositor-testing ppa circa first week of July or later.

[Daniel van Vugt (vanvugt)]

Reopened from duplicate bug 1201793.

[PS Jenkins bot (ps-jenkins)]

Fix committed into lp:mir at revision None, scheduled for release in mir, milestone 0.0.9

[kevin gunn (kgunn72)]

these are the 2 relevant competing bugs...
https://bugs.launchpad.net/xmir/+bug/1206508 "Xorg crash on xmir punts user back out to greeter" which is the ctl+c issue on sw rendering - which this mp https://code.launchpad.net/~thomas-voss/mir/revert-process-group-leader-patch/+merge/177645 was intending to fix

and this bug which is fixed by https://code.launchpad.net/~mlankhorst/mir/setsid/+merge/176676

[Chris Gagnon (chris.gagnon)]

I ran in to this again on intel machine.

ERROR: /build/buildd/mir-0.0.8+13.10.20130730bzr898saucy0/src/server/graphics/gbm/linux_virtual_terminal.cpp(137): Throw in function virtual void mir::graphics::gbm::LinuxVirtualTerminal::register_switch_handlers(mir::MainLoop&, const std::function<bool()>&, const std::function<bool()>&)
Dynamic exception type: boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> >
std::exception::what: Failed to set the current VT mode
[boost::errinfo_errno_*] = 5, "Input/output error"

[Chris Gagnon (chris.gagnon)]

apt-cache policy unity-system-compositorunity-system-compositor:
  Installed: 0.0.1bzr39saucy0.242
  Candidate: 0.0.1bzr39saucy0.242
  Version table:
 *** 0.0.1bzr39saucy0.242 0
       1002 http://ppa.launchpad.net/mir-team/system-compositor-testing/ubuntu/ saucy/main amd64 Packages
        100 /var/lib/dpkg/status
apt-cache policy mir-demos
mir-demos:
  Installed: 0.0.8+13.10.20130807.1bzr938saucy0
  Candidate: 0.0.8+13.10.20130807.1bzr938saucy0
  Version table:
 *** 0.0.8+13.10.20130807.1bzr938saucy0 0
       1002 http://ppa.launchpad.net/mir-team/system-compositor-testing/ubuntu/ saucy/main amd64 Packages
        100 /var/lib/dpkg/status
     0.0.8+13.10.20130807.1-0ubuntu1 0
        500 http://archive.ubuntu.com/ubuntu/ saucy/main amd64 Packages

[PS Jenkins bot (ps-jenkins)]

Fix committed into lp:mir at revision 942, scheduled for release in mir, milestone 0.0.9

[Launchpad Janitor (janitor)]

This bug was fixed in the package mir - 0.0.8+13.10.20130808.1-0ubuntu1

---------------
mir (0.0.8+13.10.20130808.1-0ubuntu1) saucy; urgency=low

  [ Robert Ancell ]
  * VT switch on alt+ctrl+Fn.
  * Add missing iostream import. This was accidentally removed in
    revision 939. Local builds and CI builds both seem to have treated
    this as a warning; debian package builds as an error. Fixes
    packaging builds.

  [ Robert Carr ]
  * Add an interface by which the shell may be notified of and interfere
    with surface configuration requests.

  [ Alexandros Frantzis ]
  * Bring back revision 931 with a fix for the crash it caused (LP:
    #1209053). (LP: #1209053)
  * examples: Quit the example servers and standalone programs when
    Ctrl+Alt+Backspace is pressed.

  [ Maarten Lankhorst ]
  * Re-introduce console support, and ignore control characters. (LP:
    #1102756, #1195509)

  [ Ubuntu daily release ]
  * Automatic snapshot from revision 944
 -- Ubuntu daily release <email address hidden> Thu, 08 Aug 2013 05:03:26 +0000