We have discussed the need for implementing functionality that would allow a seamless switch to backup hardware with no interruption to a running show if a hardware problem or software crash were to occur.
We are still discussing the details of this implementation, so I can't estimate a timeline for when it will happen.
The two obvious considerations are:
1) How do we handle a console failure? 2) How do we handle a node failure (DP2000 or IOP)?
My answers would be:
1) Run a server on the backup console that only listens for changes and keeps a current copy of the show file. If the primary console fails, allow this server to begin communicating with nodes.
2) Allow a node to run in "backup mode", whereby it has the same net number as an active node and begin processing DMX and providing feedback if the primary active node with that net number goes offline.
These are both high-level answers. There are a lot of details that need to be taken into consideration for this to work properly.
How would you like this functionality to work? I can make sure you're opinion gets logged so that we can implement something that functions as our users would want it to.
I think we do 1 and 2.l We have two hardwares the Node and the console. So no hardware alone can do anything in a case of fire to get the show work perfect.
I agree. I think both suggestions should be implemented.
Only question I have is what happens if the original console/DP comes back to life. Are we left with a scenario where we have 2 servers or nodes conflicting?
Tom, you said "1) Run a server on the backup console that only listens for changes and keeps a current copy of the show file. If the primary console fails, allow this server to begin communicating with nodes."
You used to be able to do this. Several years ago this was possible. For a time, you could run two servers on separate consoles and they would cleanly track each other. I was able to take either console on and offline with no conflicts. The only difference from what you suggested, is that both consoles would talk to the DP's at all times. I'd love to see this ability come back.
Currently when I do a backup/tracking situation I set up it as a server/client situation. My server desk just sits there holding the show. I do all programming and playback on the client desk. This keeps most of the load off the server and let's it concentrate on running the show. This setup is not the best because you are still reliant on the server desk, and are in a world of hurt if it goes sideways. Having said that, the two desks will track each other when set up this way...plus you can do independent playback on both desks if wanted.
I'm all for the tracking backup. That's been on my top 5 list since the beginning.
the problem on this is that this solution is not nearly a BACKUP in a worse case of fire. We need very very soon a full tracked backup. So my ongoing think is that I link two console together with a midi cable. So I`f programming ends a load the show into the second III and it will track my cuelist in playback. So with this setup you have a full second system for backup. As you do this I have not good minds about it ( for BACKUP ) for others it is very very fine. So often I had crashes with server / client and all came down.
I also run things in this way, however the problem comes when desk A goes offline the DP's reload the entire patch and go through their startup routine on connection to Desk B. They also do not hold the last dmx state all the way through the changeover process. Furthermore, the system is not truly backed up until the DP's have a spare also.
Yes I agree total. What I do next for a small cost backup solution is: When Hog III PC and the console have the same software on it ( 1.3.8 <> 1.4 ) then I conect a superwidget to my pc version and drive them. Cheaper then 2 consoles ; also for my client. I know that ´has a limit for 7 universes.
I've opened feature request #9212 for the fully tracking backup functionality that has been discussed here. If you have any great ideas or think of anything that's been missed, continue to post those thoughts to this thread and I'll continue adding comments to this bug.
We don't have a publicly searchable database of our request or bug lists. You can take a look through the Requests & Enhancements forum to see what people have requested there.
As I mentioned above, we have enhancement #9212 open for the fully tracking backup functionality. Please post any additional ideas you have for that feature here and we can get them added to the bug.
Are you saying that you want to be able to manually switch which console is actively running the server?
I can see how this might be useful. If I had to relocate my consoles from my tech position to my show position I could move the backup, manually switch to make that console the server, then move the other desk.
