Description of problem: When trying to reboot my virtual machine using: reboot or shutdown -r now or poweroff or maybe other smilar commands, only "rc0: goal changed" is displayed on virtual console and this appears in /var/log/messages: Aug 26 12:59:28 dhcp2 init: tty4 main process (1214) killed by TERM signal Aug 26 12:59:28 dhcp2 init: tty5 main process (1215) killed by TERM signal Aug 26 12:59:28 dhcp2 init: tty2 main process (1216) killed by TERM signal Aug 26 12:59:28 dhcp2 init: tty3 main process (1217) killed by TERM signal Aug 26 12:59:28 dhcp2 init: tty1 main process (1218) killed by TERM signal Aug 26 12:59:28 dhcp2 init: tty6 main process (1219) killed by TERM signal Aug 26 12:59:28 dhcp2 init: rc6 main process (1326) killed by TERM signal Then serial and VGA consoles are dead, I can't type on them. When trying to shutdown again over ssh, shutdown/reboot works. Version-Release number of selected component (if applicable): upstart-0.3.11-1.fc11.i586 How reproducible: always on my server (tested multiple times) Steps to Reproduce: I am not sure, what is different on this server. It's an minimal fedora installation + dhcp server. Does not happen on workstation.
I am not able to reproduce your problem. Can you reproduce it on another system? Also try to add "--debug" to kernel grub line and follow debug output on console or log via serial
Created attachment 363313 [details] Problematic rc script. Looks like this was caused by an script, which has "stop" command in "stop)" section. This command does nothing from shell, only displays "missing job name". Curious, that stops shutdown script execution. This bug has been fixed in this script and reboot works well now. If you consider, this is not a bug, you can close this bug.
Created attachment 363323 [details] Problematic rc skript (previous one was fixed one)
the problem is that calling initctl command ("stop" is same as "initctl stop") inside running job causes that the command is applied to actual job. in this case stop is called on rc6 job, rc6 is stopped and doesn't continue anymore.
This seems more like a configuration error, than anything else. I think we should likely close this.
You could argue that its good form to unset the UPSTART_* variables from our rc* job before calling sysv scripts, but I don't think that it causes bugs like this with enough frequency to warrant that.
I agree that it's more configuration error, closing as not a bug