Difference between revisions of "Clean Core"

From LinuxMCE
Jump to: navigation, search
(updated)
(Personal Notes: forgot something)
Line 33: Line 33:
 
I've tried the way bobpaul described in his post, but unfortunately working with the Start_X box didn't change anything for me. Then I tried it the "old linux way" ;-) I removed the S99kdm script from /etc/rc2.d, giving me a text login at bootup. Next step was to modify the /etc/rc.local by adding /usr/pluto/bin/StartCoreServices.sh. After next bootup the login web page was up again. This is the point I stopped testing and tried to get more information on startup before moving on...
 
I've tried the way bobpaul described in his post, but unfortunately working with the Start_X box didn't change anything for me. Then I tried it the "old linux way" ;-) I removed the S99kdm script from /etc/rc2.d, giving me a text login at bootup. Next step was to modify the /etc/rc.local by adding /usr/pluto/bin/StartCoreServices.sh. After next bootup the login web page was up again. This is the point I stopped testing and tried to get more information on startup before moving on...
 
(done with beta2 and rc1)
 
(done with beta2 and rc1)
 +
 +
-- [[User:KingCrab|KingCrab]]
  
 
----
 
----
Line 38: Line 40:
 
I looked on a dedicated core, but the Start_X box was already unticked in the web interface for the core. So I looked at "/etc/rc2.d/S99kdm" which turned out to be a symlink to a non-existant file.  In the end I just rebooted to "rescue" mode and launched "/usr/pluto/bin/StartCoreServices.sh" by hand.  After taking controll with "screen -r" I discover that the DCERouter is unable to connect to the MySQL server.
 
I looked on a dedicated core, but the Start_X box was already unticked in the web interface for the core. So I looked at "/etc/rc2.d/S99kdm" which turned out to be a symlink to a non-existant file.  In the end I just rebooted to "rescue" mode and launched "/usr/pluto/bin/StartCoreServices.sh" by hand.  After taking controll with "screen -r" I discover that the DCERouter is unable to connect to the MySQL server.
  
I reboot the machine to single usermode again to have a clean starting point, this time I decide to change to runlevel 3 with the command "telinit 3".  LinuxMCE is only fully started in runlevel 2 so there are no screen sessions to attach to except for the "VoiceMailMonitor" one (obviously started by /etc/rc3.d/S99voiceMailMonitor).  After launching "/usr/pluto/bin/StartCoreServices.sh" by hand again I see 2 additional screen sessions for "discovery" and "DCERouter", none of the sessions seem to show errors.
+
I reboot the machine to single usermode again to have a clean starting point, this time I decide to change to runlevel 3 with the command "telinit 3".  LinuxMCE is only fully started in runlevel 2 so there are no screen sessions to attach to except for the "VoiceMailMonitor" one (obviously started by /etc/rc3.d/S99voiceMailMonitor).  After launching "/usr/pluto/bin/StartCoreServices.sh" by hand again I see 2 additional screen sessions for "discovery" and "DCERouter", none of the sessions seem to show errors.  The script itself however ended with the message "ERROR: Couldn't attach to DCOP server".
  
The web-admin now appears to fully work, and so does booting a diskless Media Director.  Samba might not fully work yet though, as there are still 2 pluto related scripts (symlinked) in "/etc/rc2.d" that I haven't started yet, "StorageDevices_SambaRadar.sh" and "StorageDevices_StatusRadar.sh" but I'm not entirely sure if they normally even get started since they don't have the usual S## prefix.
+
The web-admin appears to fully work, and so does booting a diskless Media Director.  Storage might not fully work yet though, as there are still 2 pluto related scripts (symlinked) in "/etc/rc2.d" that I haven't started yet, "StorageDevices_SambaRadar.sh" and "StorageDevices_StatusRadar.sh" but I'm not entirely sure if they normally even get started since they don't have the usual S## prefix.
  
 
I tested this using a dedicated core installation of LMCE-1.1B2 running under VMWare as described in: [[Testing with VMWare]], which I still had lying around.
 
I tested this using a dedicated core installation of LMCE-1.1B2 running under VMWare as described in: [[Testing with VMWare]], which I still had lying around.
  
-- [[User:Zaerc|Zaerc]] 15:34, 2 August 2007 (MST)
+
-- [[User:Zaerc|Zaerc]] 15:53, 2 August 2007 (MST)
  
 
----
 
----

Revision as of 23:53, 2 August 2007

Introduction

This page is meant to document the efforts to turn the LMCE core into one or more "services" that are automaticly started at boot time. In other words we would like to get rid of the Launch Manager application, and it's dependency uppon a graphic environment, at least for starting a (dedicated) core machine.

Due to the lack of documentation we will have to figure a lot of things out on our own. Please don't be shy to add any related knowledge you have, we don't really know what we're doing, so why should you? At least by adding stuff here we'll be able to build uppon each others experience without having to reverse-engineer the wheel every time. And who knows, one day we might even achieve our common goal.

Known Information

you'll find all of the LinuxMCE scripts in /usr/pluto/bin. 
Two very interesting scripts are Start_X.sh (which I assume you already found?) and StartCoreServices.sh

My guess would be you could disable the Start_X script. 
This can be done from the web interface and goto Advanced: Software: Boot Sequence. 
Untick the box for Start_X. 
Then, manually launch /usr/pluto/bin/StartCoreServices.sh from some init script. 
I think that should do what you originally asked for, but I haven't tried it myself.
  • A request for more information on the forum was posted by KingCrab: How does a core system start?. If you don't want to sign up for the wiki or feel uncomfortable editing this page you could leave a message there.
  • More to come...


Version Specific Notes

This seems like a good place to add information specificly related to a certain version of LMCE.

LMCE 1.1 RC1

The current test version.

Personal Notes

So... what did you try, and how did it work out?


I've tried the way bobpaul described in his post, but unfortunately working with the Start_X box didn't change anything for me. Then I tried it the "old linux way" ;-) I removed the S99kdm script from /etc/rc2.d, giving me a text login at bootup. Next step was to modify the /etc/rc.local by adding /usr/pluto/bin/StartCoreServices.sh. After next bootup the login web page was up again. This is the point I stopped testing and tried to get more information on startup before moving on... (done with beta2 and rc1)

-- KingCrab


I looked on a dedicated core, but the Start_X box was already unticked in the web interface for the core. So I looked at "/etc/rc2.d/S99kdm" which turned out to be a symlink to a non-existant file. In the end I just rebooted to "rescue" mode and launched "/usr/pluto/bin/StartCoreServices.sh" by hand. After taking controll with "screen -r" I discover that the DCERouter is unable to connect to the MySQL server.

I reboot the machine to single usermode again to have a clean starting point, this time I decide to change to runlevel 3 with the command "telinit 3". LinuxMCE is only fully started in runlevel 2 so there are no screen sessions to attach to except for the "VoiceMailMonitor" one (obviously started by /etc/rc3.d/S99voiceMailMonitor). After launching "/usr/pluto/bin/StartCoreServices.sh" by hand again I see 2 additional screen sessions for "discovery" and "DCERouter", none of the sessions seem to show errors. The script itself however ended with the message "ERROR: Couldn't attach to DCOP server".

The web-admin appears to fully work, and so does booting a diskless Media Director. Storage might not fully work yet though, as there are still 2 pluto related scripts (symlinked) in "/etc/rc2.d" that I haven't started yet, "StorageDevices_SambaRadar.sh" and "StorageDevices_StatusRadar.sh" but I'm not entirely sure if they normally even get started since they don't have the usual S## prefix.

I tested this using a dedicated core installation of LMCE-1.1B2 running under VMWare as described in: Testing with VMWare, which I still had lying around.

-- Zaerc 15:53, 2 August 2007 (MST)