0

Mint 16 Cinnamon x86_64 fully updated (kernel 3.11.0-14-generic).

Drive is encrypted as well as user's home dir.

I had been working for a few days with a new install of Mint 16 and have suspended/restarted without issues until today. Now the boot process hangs, but I can log in via "single" and X works with startx.

The following are the symptoms that I have noticed:

  • When the boot process is hung, if I try to log into an alternate tty the login appears to succeed but immediately exits back to a login prompt
  • Last message in the hung boot process is:

    Starting startpar bridge for notification of upstart job start/stop
    

    , of which there are several.

  • The last message in syslog from hung boot is:

    dbus[1191]: [system] Successfully activated service 'org.freedesktop.ConsoleKit'
    
  • Last line in kern.log is:

    kernel: [   48.127483] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
    
  • Things that I recall doing directly before the breakage: installed wine, installed some i386 libs to run a prog in wine, installed clamav, installed ttf-mscorefonts-installer, installed oracle-java7-installer

UPDATE:

If I start single, and then run "service rcS start" I get the same effect as starting normal. In fact. I have moved all of the S* symlinks out of /etc/rcS.d/ and still experience the issue when starting the rcS service.

If I manually start all of the services in rcS.d, I get no errors. I started them using "service ... start". I'm thinking now that this might be related to upstart?

0

Is there any chance that you might have installed LXC ?

Because same thing happened to me and I realised after I installed Android Dev. Bundle and Eclipse (and naturally Java) my system stuck on startpar in different stages. Removing ADT or Java didn't help so looked at the boot log and realized lxc was failing, chrooted and removed it and everything was back on track.

If that's not the case, try reinstalling the grub, upstart and dkpg, using a live CD and chroot.

Let me know if that's the case for you too.

  • I did reinstall upstart and dpkg but that did not help; I didn't have lxc installed. I have since reinstalled the entire OS and since I had kept detailed logs of my setup, I was able to quickly recreate my installation. No failures yet, so I'm not sure I'll ever know what went wrong. – Ethereal Jan 7 '14 at 20:20
  • Ok it happened again... at least it is reproducible. No LXC installed. – Ethereal Jan 7 '14 at 23:03
0

It turns out that I had a script in /etc/profile.d that was mucking things up, on two levels.

  • I had inserted "xhost +", which caused the graphical display to fail to come up
  • I had "exit" in there to avoid running part of the script, which was exiting my shells every time I tried to log in to a different terminal.

I'm thinking I should be able to use "return" instead of exit -- although since the scripts are executed, not sourced, I don't see why this is happening. Not sure why "xhost +" breaks things, as this works fine in Centos 6. I do get "upstart/mdm.log:xhost: unable to open display "" " in my logfiles but that is not a fatal error.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.