You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@gb-beng would it be possible to use docker stop whisper-asr-worker instead of the kill command? From what I could find, it seems like the standard approach as it first sends SIGTERM, then after 10s it sends SIGKILL. But if kubernetes handles these operations through docker kill, then I'll keep investigating how to handle it because it is taking longer than expected to find a solution
Also, using docker container kill but with the SIGINT signal doesn't work for me (I'm not sure how you tested that)
@gb-beng would it be possible to use docker stop whisper-asr-worker instead of the kill command? From what I could find, it seems like the standard approach as it first sends SIGTERM, then after 10s it sends SIGKILL.
Kubernetes doesn't use docker commands to send these signals, I just used it for testing/illustration. There is an equivalent to docker stop, but the point is that I'd like to gracefully stop the process instead of killing it.
I'll keep investigating how to handle it because it is taking longer than expected to find a solution
I wouldn't spend too much time on it now, let's focus on the distributed/scalable deployment first, as it will affect this issue as well.
Also, using docker container kill but with the SIGINT signal doesn't work for me (I'm not sure how you tested that)
It would be great if the worker/container would gracefully terminate on
SIGTERM
. Kubernetes can send it e.g. when draining nodes.To test this I used:
Note thatSIGINT
does work.The text was updated successfully, but these errors were encountered: