Is there currently a way to keep track of the PID of the active applications started by the slave (max, nuke, mayabatch etc)?
We are running into some issues with stalled slaves, that after the initial stall is detected are not able to start max any longer. Would be great if we could identify whether the active process is still the stalled one, or belongs to the new task
You could print out the ids of slave’s managed processes in the plugin using docs.thinkboxsoftware.com/produc … ccf7097dd0. Other than that, we do not track that at all. If the issue is that the slave stalled and the processes persist, you could check the slave process’ start time vs the processes’ to determine if the were started after the current slave instance or if they preceded it.
We were able to figure it out based on the command line of the process (for max jobs it happens to have the job id in it). Might come up with a more generic solution though, that would require patching the plugins (say, touching a file that has the pid in it, or something like that), so that external applications can also figure out which process belongs to an active deadline job.