Reboot Interface Feature Request!


#1

Howdy folks!

I would like to submit a new feature request. Currently, the Sandbox stack_manager has the ability to be Stopped/Started/Reboot from the Windows Tray. This makes it really easy for me to remote administer an a simple instance without the need for command-line ninja-fu. (better to have it and not need it, than to need it and not have it… :wink: )

The system expectations of Interface are quite high. My solution to allow new team members the ability to test things within interface is to allow them to remotely access a very high-end machine. To do this; interface must be started on-site prior to their engagement. Most of you that have suffered through 2014-15 alpha years already see where this is going.

So what happens if the interface crashes? (it does still happen btw…)

Well, the current solution is it needs to be restarted in Windows somehow…

But what if the Stack_Manager could reboot Interface on my behalf ???

I can’t make it into the Friday meetups anymore to share my little suggestions so I’m putting this one here. I need a way to Reboot Interface such that if a remote user of interface closes or crashes, it can startup on it’s own without failing due to Remote Desktop graphics. (anyone that has attempted to run Substance Designer or Interface remotely knows exactly what I’m talking about)

Any other recommendations welcome, @OmegaHeron, @Cracker.Hax, @Menithal possibly a windows service could be authored that simply sniffs the status and reboots if needed? Could this be a Windows problem that is currently unsolvable?

image

cc: @Jess


#2

I think it may be possible to do that to a degree with some behind the scenes command-line stuff. I know in Linux you can kill processes by name and ID, which is what I use to start and stop my servers (don’t judge me). I think Windows has a similar thing which can be used to at least stop Interface, be it on crash or otherwise. This leaves starting, which is more or less trivial. Get both working and you can ‘restart’ a program like a service.

Just my 2 cents worth of an idea behind a possible solution.


#3

Not a programmer so I can’t help, but I can say that this is something definitely possible on Windows.

I used to run a private server for an old MMO. The software was buggy and crashed frequently, so someone made a tool that did exactly what you describe, for Windows. It was an extra process that restarted the main server executable automatically when it detected it wasn’t running anymore.

Granted this was like 10 years ago, on Windows XP. But it should still work.


#4

Yes. Keepalive functions are (or should be) available in most OSes.


#5

From a remote user standpoint; before I go mucking around on a shared workstation… do you think this will enable the program to run (around) the OpenGL 4.1 requirement? Sorry if I haven’t described it well enough, unfortunately my English isn’t to blame but describing the fail_to_restart remotely is just a tough concept for me to explain.


#6

I’ve been keeping this idea in mind in writing a rather simple Watchdog (both a JS script that runs with interface and keeps chatting with the watchdog app), but one problem that I know will be the bane of my existence is the confirmation dialog that pops up when the client crashes. While possible to just kill interface and start it back up, the problem will be in knowing how to deal with that dialog.