Open makman88 opened 1 year ago
Thanks for raising the issue, will try and figure out a way to shut the process down gracefully
Is it possible to trap the SIGINT in algorun and call a exec a script that executes a kill or killall of the goal in the container? Not sure if that is best, just a thought. In any case, thanks!
when an algorun user uses the -w flag for continuing status, and then control-C's out of it, the goal process that was spawned inside the container continues to live. if the user makes repeated runs of "algorun goal node status -w" , these goal processes continue to accumulate.
In looking at my particular container, each residual goal process is consuming about 39 MB (seen via "top" RES) and holding an established socket connection to port 8080, which i see using netstat. So if the user uses -w a lot, the resource load starts to add up pretty fast.