"Stop" hang on "stopping" state when there are more than one workers in distributed mode.

[xllalitha]

Observed an issue after clicking "stop" load test from web console of master node. "stop" status change to "stopping" status and hang there without reaching "stopped" state. When there is only 1 worker, it reaches properly to "stopped" state.. I had to restart locust master again to restart load test when it was hung on "stopping" state. Issue repeats when I have more than one worker.

I experienced this issue first when running master and workers on AWS fargate containers .Interestingly, there were no errors logged in master and worker containers after pressing stop button. With just single worker container, Stop status properly move to "stopped" on web console master. Having more than 1 worker cause this issue.

Later I could reproduce this issue on local machine without even docker.

When I run locust master and 2 workers from my local MacBook itself, I am experiencing this issue.
There are no errors thrown after clicking "stop". With 1 worker, it work fine and move to "stopped" status.

How to reproduce in local MacBook

1. Start Locust master from one terminal with below command

locust -f my_locustfile.py --master

2. Start 2 locust workers from another 2 terminals with below command

locust -f my_locustfile.py --worker.

(Ignored "--master-host" parameter here as it is local run. In AWS Fargate setup, we have that parameter)

After this master web console(http://0.0.0.0:8089/) shows 2 workers connected, also load test work fine. When click on "stop" it changes to "stopping" and never come to "stopped".