Stand alone Ubuntu orbiter

[Riggs]

Does any one know why there is no stand alone Ubuntu orbiter like we have for Windows.

[tschak909]

Because there is no package in the package database that builds one, and there needs to be a subclass of StartOrbiter written that provides a UI for Orbiter to start up.

-Thom

[Riggs]

I read through the forum and seen an old post i think it was by you back in 2007 http://forum.linuxmce.org/index.php?topic=2707.0
did you get anywhere with it.
I'm looking into it because i'm finding it really hard to get the UPS idea working  http://forum.linuxmce.org/index.php?topic=10234.msg70441;topicseen#msg70441
so i thought i'd start the basics and see if i could first control an Ubuntu PC (boot up shutdown etc), from the core on an event
but i just get no where, so that lead me to seeing if i could find an orbiter for Ubuntu, I did try PXE booting but that loops and I'ts an old distro which I don't think would be much use as dual workstation/orbiter like say windows vista with MCE

[tschak909]

Currently, remote shutdown/bootup of PC is confined to media director devices, with or without dual boot.

It is done via a combination of changing tftpboot parameters (courtesy of the general info plugin), and sending commands to the target computer to stop the computer. Wake On LAN packets are sent (again courtesy of the General Info PlugIn), to wake the computer back up.

This was utilized by the Windows Pluto Backup and Switcher, to send the relevant DCE commands, and to recieve the shutdown DCE commands from the core. All the infrastructure is there, and it works.

Sorry man, you won't be able to duct tape things together, no matter how hard you try... You'll actually need to learn the system architecture to bring features to other systems.

-Thom

[Riggs]

I wish it was more easier to understand LMCE/Linux, the last time i did any programming and had it work was with a commodore 64/Amiga 501, back in the day.
I cant even get a timed events to run a .sh script if i could then i could make the script ssh in to the PC's and run commands that way.
With LMCE everthing is just guess work with out any manual to give you hints, thats why so meny users keep post on the forum like me, and bothering the Devs

[tschak909]

Everything in LinuxMCE  is framed as passing messages back and forth of messages from one device, to another, as explained in the Developers Guide.

MessageSend is a command line program that can send raw DCE messages to the various devices.
http://wiki.linuxmce.org/index.php/MessageSend

As for a manual, there is tons of raw content on the wiki, are we supposed to stop everything we are doing and write a manual? Is that what you expect of us?

-Thom

[Riggs]

the wiki has some basic info for a none programmer to do some basic tasks that are pre defined, so if you want to do something out of the box you have to:
1, post on the forum and see if you get answer from the only people that know the system (Devs) who are over worked (annoyed pissed off, frustrated, tired) and don't have the time to give help to new comers so they give them a basic response, and in most cases you still need to know something about the system to work out what they are telling you so they will post back again for some more info on where what and how to for the response they previously got.
2, look at the wiki which does have a load of info but you need to have an understanding of the programming/and or the system to understand most of it.

I see in webadmin there are tooltips which would be brill it would solve all my info needs and would never have to post in the forum again exept for advanced stuff,
but some of them have no info or they don't have any meaning to a none Dev of this system.

for instance i say to you "I cant even get a timed events to run a .sh script if i could then i could make the script ssh in to the PC's and run commands that way"
1). in basic form i want the core at 12:00 to run shutdown.sh (i can write the shutdown.sh script which i'm working on and mostly works)

2). from this i can then go to the next thing and try to get NUT to run a script when on battery which I'm working on.

3). Look into having LMCE read responses from the output of the script from NUT.

4). Look at how 1. works and see about changing it from a timed event to when that is called when output is read from NUT.

I cant find anything in the wiki as to how to run a script .sh at a pre defind time.

[tschak909]

You have to use the App Server, to run shell scripts, but this isn't an ideal situation at all. You should write a full DCE device for what you're wanting to do.

-Thom

[Marie.O]

1). in basic form i want the core at 12:00 to run shutdown.sh (i can write the shutdown.sh script which i'm working on and mostly works)

setup a timed event that sends and off message to the core (I don't know out of the top of my head, what message it is. Look at /var/log/pluto/DCERouter.log and note the message that gets executed when you hit poweroff on the orbiter for the core.

3). Look into having LinuxMCE read responses from the output of the script from NUT.
4). Look at how 1. works and see about changing it from a timed event to when that is called when output is read from NUT.

Why would you need LinuxMCE to read responses from NUT? You only need NUT to tell the devices to shutdown (see above)

I cant find anything in the wiki as to how to run a script .sh at a pre defind time.

Because LinuxMCE does not have anything included to run arbitrary scripts. And as you can read from above, there is no real need for it, either. 

[totallymaxed]

1). in basic form i want the core at 12:00 to run shutdown.sh (i can write the shutdown.sh script which i'm working on and mostly works)

setup a timed event that sends and off message to the core (I don't know out of the top of my head, what message it is. Look at /var/log/pluto/DCERouter.log and note the message that gets executed when you hit poweroff on the orbiter for the core.

As posde says use a timed event; Select 'Timed Event' from the 'Wizard' side panel in Web Admin. Now choose the basic options for your event like 'Interval based' or 'Day of the week based' etc etc. Then use the 'Advanced Wizard' options and drop-down at the bottom of the page to select the event you want to be triggered. In your case you would select 'The Core/Hybrid' from the Device drop-down then select 'Off' from the Command drop-down.

After you have configured the above do a reload router and you should be in business :-)

All the best

Andrew

[totallymaxed]

Does any one know why there is no stand alone Ubuntu orbiter like we have for Windows.

We have done some work to take the Orbiter to Ubuntu 9.10 this was based on the code used on the Nokia N800/810. This runs well on the DT362 touch tablet which is based on an Atom processor. You can see it running on the DT362 here http://wiki.linuxmce.org/index.php/WebDT362. This gives you full access to the 9.10 desktop alongside the Orbiter app...there are many improvements that could be made. But essentially this provides a much larger and easier to use screen than the N800/810. Battery life is not that wonderful but acceptable if you have a charging cradle handy ;-)

I am sure this could be ported to 10.04 without any problems at all. We still sell the DT362 but have not updated to 10.04 as these units are beaten badly on price and form factor by iPad and other tablets. But any suitable Atom based hardware would be suitable and increasingly this is true of Arm based devices too.

I will see if we can get our DT362 code uploaded to the SVN next week so you can have a look at it.

All the best

Andrew

[Riggs]

The reason i want LinuxMCE to read responses from NUT, I want to use LinuxMCE as a whole house solution the brain if you must, have it take care of devices and report to me what happens and when.
Yes i can have NUT and many types of software talk to the other computers directly but then i would have so many things to keep an i on, if LinuxMCE was the goto like i see it as, (dcerouter) then i only have one thing to take care of.

yes scripts may not be the best way to do these things but i can do scripts i can research them and write them, but i can't do that with LinuxMCE it is programming that i cant get my head round and there is only the wiki to go on.

with my script I can shutdown/control any devices that i can ssh into, all device are logged to LinuxMCE in DHCP now that would be brill for me if instead of listing all devices with an IP/MAC it would list them with a device number like it does everything els, (like when you PXE boot that device is given a number).

Lets think of a computer like a bluetooth phone, when linux detects it, it will ask what device it is then send software (orbiter) to the device etc, it picks up the IP of the computer then knows the OS and sends the orbiter.
the orbiter would include SIP and messaging, and ability for the core to control it like shutdown/reboot, in linux run update manager etc, the core would also be able to use the input devices on said devices.

Now i'm sure you could agree with the above being a really good idea, and also tell me it would be a nightmare to program so i am attempting to try and do some of this by means of script because I can't program and i'm not gonna ask you to do it for me.

to you it may seem crazy to say duct tape it but to me, its away to learn and get good at scripting I'm thinking i might need to use python for some scripts. (I AM A NEWBE)

[tschak909]

This is an extremely advanced topic,

In order to do this, you'll have to literally learn to take the system apart and study it.

What you want to do is more than doable once you understand how the system is actually put together.

You'll need to dig into parts of the C++ code to understand how and why things are done, and spend time reading the developers guide and asking questions. This is what the IRC channels are for. and you will need to exercise a great deal of patience. Given your current knowledge, I estimate that this would be a year long project for you, learning the system and then implementing the design (which will be a lot less time, than learning the system.)

-Thom

[Riggs]

Ok, so lets say i do it your way.

first i guess i would install kubuntu 10.04 on a VBOX just to keep life separate i do all programming in the VBOX.

2. I guess i need to start looking at source code so i need to learn how to use SVN and how to install it

what els, think basic minded for now to get me started, I guess i'll make it a hobby of learning LinuxMCE instead of learning scripts

PS. does it sound like a good idea though LOL.

as for the IRC i'm looking it up

[tschak909]

You have two options, C++ or Ruby...

For C++, on your existing 0810 core:

* check out a copy of the LinuxMCE-0810 branch from SVN into a folder in /home
* copy the contents of /usr/pluto/lib into its src/lib
* add the following lines to your ~/.bashrc

Code: [Select]

export SNR_CPPFLAGS="-DKDE_LMCE"
export SNR_LDFLAGS=" "

install the following bits,

Code: [Select]

apt-get install pluto-dcegen
apt-get install pluto-sql2cpp

and you'll have everything you need for C++ development.

If you choose to develop in Ruby/GSD, then you can do all your work in the Web Admin. Read the "GSD" page in the wiki, it is much simpler, but also not as flexible (meaning, you can only talk to a single TCP or RS232 port.), but if that is enough, it is substantially simpler to write a device.

-Thom