/etc/init/failsafe.conf doesn't actually stop on runlevel
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
upstart (Ubuntu) |
Fix Released
|
High
|
Clint Byrum |
Bug Description
The failsafe job is meant to inform the user that the system is waiting for the network to come up, and start the runlevel if a timeout is reached.
However, the 'stop on runlevel' does *not* work as designed, because all the activity here is done in the pre-start script which upstart will not kill when the job target changes to stop:
$ sudo start failsafe&
[1] 3979
vorlon@
failsafe start/pre-start, process 3981
vorlon@
[2] 3985
vorlon@
failsafe stop/pre-start, process 3981
[1]- Done sudo start failsafe
So the pre-start script continues sleeping, and continues sending messages to plymouth.
If you happen to reboot the system within two minutes of startup, this results in the messages being printed on the splash screen at *shutdown*, which is rather confusing.
To avoid this, I think the failsafe job should be switched to do its work in 'script' instead of 'pre-start script', and rc-sysinit changed to listen for an event emitted by the failsafe job at the end of this script instead of on 'started failsafe'.
ProblemType: Bug
DistroRelease: Ubuntu 11.10
Package: upstart 1.3-0ubuntu9
ProcVersionSign
Uname: Linux 3.0.0-12-generic x86_64
NonfreeKernelMo
ApportVersion: 1.23-0ubuntu2
Architecture: amd64
Date: Sat Oct 1 00:02:54 2011
InstallationMedia: Ubuntu 10.04 "Lucid Lynx" - Beta amd64 (20100406.1)
ProcEnviron:
LANG=en_US.UTF-8
SHELL=/bin/bash
SourcePackage: upstart
UpgradeStatus: Upgraded to oneiric on 2011-09-23 (7 days ago)
modified.
mtime.conffile.
Related branches
Changed in upstart (Ubuntu): | |
assignee: | nobody → Clint Byrum (clint-fewbar) |
Changed in upstart (Ubuntu): | |
status: | Triaged → Fix Committed |
Raised to High, this also causes odd problems with plymouth in boot, and shows the message if lightdm takes more than 20 seconds.