Consistent gracefulReload RPC

[kenhys]

Is your feature request related to a problem? Please describe.

https://github.com/fluent/fluentd/pull/3396 tried to keep consistency even though excessive gracefulReload request is thrown (It has same problem with SIGUSR2 too) It tries to introduce blocking interval - @blocking_reload_interval parameter, but it seems that it turned out that it is too simple an approach.

• The behaviour is changed without @blocking_reload_interval = 0.
• There is a case that request is just ignored (it returns an error but it requires the user to implement retry for such a case. The breaking change)
• It depends the previous API call

Describe the solution you'd like

• Treat gracefulReload subsequently (somehow like a queue) or
• Ensure to wait reloading thread has finished

Describe alternatives you've considered

• N/A

Additional context

See https://github.com/fluent/fluentd/pull/3396

Feedbacks against #3396

• https://github.com/fluent/fluentd/pull/3396#commitcomment-51857485
• https://github.com/fluent/fluentd/pull/3396#commitcomment-51869670
• https://github.com/fluent/fluentd/pull/3396#issuecomment-856981757

[ashie]

It's for #3341

[ashie]

I confirmed that current implementation can run multiple reload threads simultaneously when it receives excessive gracefulReload request.

[ashie]

We should also take a look #3469 when we working on it.

[github-actions[bot]]

This issue has been automatically marked as stale because it has been open 90 days with no activity. Remove stale label or comment or this issue will be closed in 30 days

[github-actions[bot]]

This issue has been automatically marked as stale because it has been open 90 days with no activity. Remove stale label or comment or this issue will be closed in 30 days

[github-actions[bot]]

This issue was automatically closed because of stale in 30 days