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
I use socket.io-redis with my node js cluster app. Instead of just exiting on SIGTERM and SIGINT, I gracefully exit and close all connections. However, using the socket.io-redis my child processes never exit, problaby redis connection is persisted, leaving a dead process that has to be forced killed using kill -9. Here is a quick reproduction script:
After running this you should see Worker <PID> exiting gracefully, but if you look at your process manager you will still see that worker running with the master process dead. In this particular case sending SIGTERM or SIGINT will trigger a channel closed error but in my actual code the program will just return leaving the worker running in the background. Commenting out the redis adapter line will resolve the problem.
Thanks for taking a look at this!
The text was updated successfully, but these errors were encountered:
I use socket.io-redis with my node js cluster app. Instead of just exiting on SIGTERM and SIGINT, I gracefully exit and close all connections. However, using the socket.io-redis my child processes never exit, problaby redis connection is persisted, leaving a dead process that has to be forced killed using
kill -9
. Here is a quick reproduction script:After running this you should see
Worker <PID> exiting gracefully
, but if you look at your process manager you will still see that worker running with the master process dead. In this particular case sending SIGTERM or SIGINT will trigger a channel closed error but in my actual code the program will just return leaving the worker running in the background. Commenting out the redis adapter line will resolve the problem.Thanks for taking a look at this!
The text was updated successfully, but these errors were encountered: