Orbiter not starting on Core Hybrid

[rafik24]

Hi All,

 I upgraded my core hybrid  3 days ago via apt-get upgrade. Since then when my core starts it stays at Loading LinuxMCE screen and just sits there for hours.

 I investigated the logs and found the following in LaunchManager.log:

05      08/27/09 16:06:23.227           ^[[33;1mConnect() failed, Error Code 111 (Connection refused))^[[0m <0xb7a34b90>

I tried using the webadmin interface and i works fine, so i assume that the database access is granted.

I also tried lauching Startup_Core-Hybrid.sh manually which gives the same results.

Could you please help.

Rafik

[jimbodude]

Sound similar to:
http://forum.linuxmce.org/index.php?topic=8720.msg59031#msg59031

[WhateverFits]

I was just going to say that. Try deleting those files (I can't remember where but it ends with /pluto/lock) and rebooting. That fixed it for me.

[jimbodude]

Please keep track of what files you need to delete.  This should be reported as a bug.

[rafik24]

 Thank you guys for your amazingly fast answers.

 Removing the files in /usr/pluto/locks did the trick.

 Know dcerouter is downloading stuff but it got further.

 All the best,

Rafik

[dlewis]

to jimbodude's point, we need to figure out how to fix this instead of manually deleting files...

[jimbodude]

So to follow up with the people who are having this issue - please confirm that this is what is actually going on:
1) The core is powered off without going through proper shutdown (hard power-off)
2) When the core is powered on, the DCERouter does not start
--) There are no new entries in /var/log/pluto/DCERouter.log
--) The LaunchManager is unable to connect to the router, and loops forever (see /var/log/pluto/LaunchManager.progress.log)
3) The directory /usr/pluto/locks is manually cleared by the user (and NO other manual actions are taken)
4) The core is rebooted, and the DCERouter starts normally

I also cleared some locks in /var/run I think - not sure if that was needed or not.  Are there any circumstances other than a hard power-off that trigger this issue?  Can you reproduce it easily?

[rafik24]

 It happened to me after my machine got down due to a mains cutdown in our building

[dlewis]

hmm... so you're saying a power failure caused this issue? Sort of hard to believe, but I'm not one to rule anything out. Maybe a process was terminated...? Were you in the middle of running a process?

[jimbodude]

In my case, it was a power failure.  Numerous reboots did not help.  After I removed the lock files and rebooted again, everything was fine.

[rafik24]

 Same here, unfortunately i was not in front of my machine so it would be hard for me to tell if there was any process ongoing

[WhateverFits]

Jimbodude, that sounds just like my scenario. Just a little more info on it, my system was completely unresponsive and appeared to be in a swap of death. My hard drive light was stuck on and I could not log in to it from any where. After about 20 minutes of trying, I hard powered off the machine. That is when these problems started. This also happened on the same machine on the 7.10 series and is actually the reason why I switched to the beta. What I am going to suggest is that this folder get cleared early on in startup. Has anyone entered a ticket yet? If not, I will. If so, post it here.

[schaferj]

Hello,

There may be two issues if I'm reading this correctly and it may help to clarify.  I have the issue described by this topic - that is my core hybrid runs and DCErouter is running, but the orbiter does not launch.

I mentioned it in a discusson of another issue and recieved some advice here:
http://forum.linuxmce.org/index.php?topic=8705.msg59255#msg59255

The issue that DCErouter won't start may be a different issue as described in another thread:
http://forum.linuxmce.org/index.php?topic=8720

If the dcerouter doesn't start, then it may be certain that the orbiter can't start, but the reverse is not true.  A core can be running as a core, which means that dcerouter is running.  And the core may or may not be running an orbiter; a core running an orbiter makes it a hybrid.  

If it helps, I don't have any old locks in /usr/pluto/locks (they are all timestamped after a reboot earlier today) and have not experienced any crashes since I installed from scratch a few days ago on the 24th of August.

I checked several logs and didn't find anything helpful

madmin@dcerouter:/var/log/pluto$ less DCERouter.log
madmin@dcerouter:/var/log/pluto$ less LaunchManager.log
madmin@dcerouter:/var/log/pluto$ tail LaunchManager.progress.log

Thanks and appreciate any insights,
joseph

[jimbodude]

schaferj - I don't think anyone on this thread has had the DCERouter running, which causes Orbiter to not start.

WhateverFits - This does not sound like the same issue.  I am able to ssh in and do whatever I want with the system.  Web Admin also works fine.

I was able to reproduce this issue again today thanks to a certain contractor who has no respect for power cords.  I've entered ticket #316 describing my findings.
http://svn.linuxmce.org/trac.cgi/ticket/316