Closed daipom closed 2 years ago
The behavior of start_worker_delay is a little weird.
start_worker_delay
While not a major problem, there seems to be a minor mistake in the implementation.
Although Fluentd doesn't use this option, if we use the option as follows, then this works as follows.
--- a/lib/fluent/supervisor.rb +++ b/lib/fluent/supervisor.rb @@ -494,7 +494,8 @@ module Fluent main_cmd: params['main_cmd'], signame: params['signame'], - disable_shared_socket: params['disable_shared_socket'] + disable_shared_socket: params['disable_shared_socket'], + start_worker_delay: 3 } if daemonize se_config[:pid_path] = pid_path
The current behavior
2022-06-10 16:02:08 +0900 [info]: using configuration file: <ROOT> <system> workers 6 </system> <source> @type forward port 24224 bind "0.0.0.0" </source> <match test.**> @type stdout </match> </ROOT> 2022-06-10 16:02:08 +0900 [info]: starting fluentd-1.14.6 pid=353261 ruby="2.7.5" 2022-06-10 16:02:08 +0900 [info]: #0 fluentd worker is now running worker=0 2022-06-10 16:02:11 +0900 [info]: #2 fluentd worker is now running worker=2 2022-06-10 16:02:11 +0900 [info]: #1 fluentd worker is now running worker=1 2022-06-10 16:02:14 +0900 [info]: #4 fluentd worker is now running worker=4 2022-06-10 16:02:14 +0900 [info]: #3 fluentd worker is now running worker=3 2022-06-10 16:02:17 +0900 [info]: #5 fluentd worker is now running worker=5
The behavior with this fix
2022-06-10 15:56:47 +0900 [info]: using configuration file: <ROOT> <system> workers 6 </system> <source> @type forward port 24224 bind "0.0.0.0" </source> <match test.**> @type stdout </match> </ROOT> 2022-06-10 15:56:47 +0900 [info]: starting fluentd-1.14.6 pid=352530 ruby="2.7.5" 2022-06-10 15:56:47 +0900 [info]: #0 fluentd worker is now running worker=0 2022-06-10 15:56:50 +0900 [info]: #1 fluentd worker is now running worker=1 2022-06-10 15:56:53 +0900 [info]: #2 fluentd worker is now running worker=2 2022-06-10 15:56:56 +0900 [info]: #3 fluentd worker is now running worker=3 2022-06-10 15:56:59 +0900 [info]: #4 fluentd worker is now running worker=4 2022-06-10 15:57:02 +0900 [info]: #5 fluentd worker is now running worker=5
I rebased this for checking tests.
:+1:
Thanks for the merge!
The behavior of
start_worker_delay
is a little weird.While not a major problem, there seems to be a minor mistake in the implementation.
Although Fluentd doesn't use this option, if we use the option as follows, then this works as follows.
The current behavior
The behavior with this fix