[Bug 1154040] [NEW] Upgrading to 1.7 on raring caused system to crash

[Jonathan Davies]

Public bug reported:

Upgrading Upstart from 1.6.1-1ubuntu3 → 1.7-0ubuntu1 caused a system to
crash (happened twice to me now). The system required a hard reboot to
start again.

There was nothing on the console, and the last thing I saw via SSH was:

"Install new configuration file /etc/dbus-1/system.d/Upstart.conf..."

Brief IRC chat about this:

< jpds> Because that's twice it's happened to me.
< Laney> killing how?
< jpds> Laney: As in, it needs a hard reboot to come back.
< jpds> Last thing I see is: "Install new configuration file /etc/dbus-1/system.d/Upstart.conf..."
< infinity> Re-execing it is ending in sadness?
< jpds> infinity: Well, this was done over SSH, and everything just stopped.
< infinity> jpds: Over SSH makes things a bit scarier.  I was about to ask you to try running "telinit u" to see if it does the same thing.  But the same thing sounds undesirable. :P
< jpds> infinity: Well, the first time it happened (on my laptop), I couldn't get any response out of it.
< jpds> Now the same thing happened on my desktop.
< infinity> jpds: Right, I gathered that.  I was asking if you could run "telinit u" on one of the affected systems after a reboot.
< infinity> jpds: To isolate if it's a consistent issue with your machines/setup, or if it's isolated to only happening during upgrade (which is still bad, but different).
< jpds> infinity: Nothing happened running "telinit u" on my laptop.
< infinity> (as root)
< jpds> Yep.
< infinity> And "initctl version" confirms that it's 1.7ish?
< jpds> Yes.
< infinity> Alright.  Then upstart's postinst may be doing something curiously silly.
< jpds> Upstart went from 1.6.1-1ubuntu3 → 1.7-0ubuntu1.
< infinity> My assumption is that your locking up was actually a kernel panic because init died.
< infinity> The kernel hates init going away.
< infinity> But, of course, that's really not meant to happen.

** Affects: upstart (Ubuntu)
     Importance: High
         Status: New

** Description changed:

  Upgrading Upstart from 1.6.1-1ubuntu3 → 1.7-0ubuntu1 caused a system to
  crash (happened twice to me now). The system required a hard reboot to
  start again.
  
  There was nothing on the console, and the last thing I saw via SSH was:
  
  "Install new configuration file /etc/dbus-1/system.d/Upstart.conf..."
  
  Brief IRC chat about this:
  
- 11:47:36 < jpds> Because that's twice it's happened to me.
- 11:47:58 < Laney> killing how?
- 11:48:10 < jpds> Laney: As in, it needs a hard reboot to come back.
- 11:48:25 < jpds> Last thing I see is: "Install new configuration file /etc/dbus-1/system.d/Upstart.conf..."
- 11:48:32 < infinity> Re-execing it is ending in sadness?
- 11:49:02 < jpds> infinity: Well, this was done over SSH, and everything just stopped.
- 11:49:33 < infinity> jpds: Over SSH makes things a bit scarier.  I was about to ask you to try running "telinit u" to see if it does the same thing.  But the same thing sounds undesirable. :P
- 11:50:10 < jpds> infinity: Well, the first time it happened (on my laptop), I couldn't get any response out of it.
- 11:50:40 < jpds> Now the same thing happened on my desktop.
- 11:50:54 < infinity> jpds: Right, I gathered that.  I was asking if you could run "telinit u" on one of the affected systems after a reboot.
- 11:51:25 < infinity> jpds: To isolate if it's a consistent issue with your machines/setup, or if it's isolated to only happening during upgrade (which is still bad, but different).
- 11:52:05 < jpds> infinity: Nothing happened running "telinit u" on my laptop.
- 11:52:09 < infinity> (as root)
- 11:52:23 < jpds> Yep.
- 11:53:04 < infinity> And "initctl version" confirms that it's 1.7ish?
- 11:53:13 < jpds> Yes.
- 11:53:36 < infinity> Alright.  Then upstart's postinst may be doing something curiously silly.
- 11:53:43 < jpds> Upstart went from 1.6.1-1ubuntu3 → 1.7-0ubuntu1.
- 11:53:48 < infinity> My assumption is that your locking up was actually a kernel panic because init died.
- 11:53:51 < infinity> The kernel hates init going away.
- 11:53:58 < infinity> But, of course, that's really not meant to happen.
+ < jpds> Because that's twice it's happened to me.
+ < Laney> killing how?
+ < jpds> Laney: As in, it needs a hard reboot to come back.
+ < jpds> Last thing I see is: "Install new configuration file /etc/dbus-1/system.d/Upstart.conf..."
+ < infinity> Re-execing it is ending in sadness?
+ < jpds> infinity: Well, this was done over SSH, and everything just stopped.
+ < infinity> jpds: Over SSH makes things a bit scarier.  I was about to ask you to try running "telinit u" to see if it does the same thing.  But the same thing sounds undesirable. :P
+ < jpds> infinity: Well, the first time it happened (on my laptop), I couldn't get any response out of it.
+ < jpds> Now the same thing happened on my desktop.
+ < infinity> jpds: Right, I gathered that.  I was asking if you could run "telinit u" on one of the affected systems after a reboot.
+ < infinity> jpds: To isolate if it's a consistent issue with your machines/setup, or if it's isolated to only happening during upgrade (which is still bad, but different).
+ < jpds> infinity: Nothing happened running "telinit u" on my laptop.
+ < infinity> (as root)
+ < jpds> Yep.
+ < infinity> And "initctl version" confirms that it's 1.7ish?
+ < jpds> Yes.
+ < infinity> Alright.  Then upstart's postinst may be doing something curiously silly.
+ < jpds> Upstart went from 1.6.1-1ubuntu3 → 1.7-0ubuntu1.
+ < infinity> My assumption is that your locking up was actually a kernel panic because init died.
+ < infinity> The kernel hates init going away.
+ < infinity> But, of course, that's really not meant to happen.

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to upstart in Ubuntu.
https://bugs.launchpad.net/bugs/1154040

Title:
  Upgrading to 1.7 on raring caused system to crash

Status in “upstart” package in Ubuntu:
  New

Bug description:
  Upgrading Upstart from 1.6.1-1ubuntu3 → 1.7-0ubuntu1 caused a system
  to crash (happened twice to me now). The system required a hard reboot
  to start again.

  There was nothing on the console, and the last thing I saw via SSH
  was:

  "Install new configuration file /etc/dbus-1/system.d/Upstart.conf..."

  Brief IRC chat about this:

  < jpds> Because that's twice it's happened to me.
  < Laney> killing how?
  < jpds> Laney: As in, it needs a hard reboot to come back.
  < jpds> Last thing I see is: "Install new configuration file /etc/dbus-1/system.d/Upstart.conf..."
  < infinity> Re-execing it is ending in sadness?
  < jpds> infinity: Well, this was done over SSH, and everything just stopped.
  < infinity> jpds: Over SSH makes things a bit scarier.  I was about to ask you to try running "telinit u" to see if it does the same thing.  But the same thing sounds undesirable. :P
  < jpds> infinity: Well, the first time it happened (on my laptop), I couldn't get any response out of it.
  < jpds> Now the same thing happened on my desktop.
  < infinity> jpds: Right, I gathered that.  I was asking if you could run "telinit u" on one of the affected systems after a reboot.
  < infinity> jpds: To isolate if it's a consistent issue with your machines/setup, or if it's isolated to only happening during upgrade (which is still bad, but different).
  < jpds> infinity: Nothing happened running "telinit u" on my laptop.
  < infinity> (as root)
  < jpds> Yep.
  < infinity> And "initctl version" confirms that it's 1.7ish?
  < jpds> Yes.
  < infinity> Alright.  Then upstart's postinst may be doing something curiously silly.
  < jpds> Upstart went from 1.6.1-1ubuntu3 → 1.7-0ubuntu1.
  < infinity> My assumption is that your locking up was actually a kernel panic because init died.
  < infinity> The kernel hates init going away.
  < infinity> But, of course, that's really not meant to happen.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/1154040/+subscriptions