Some options to do that were proposed in #7740 but they got (rightly) shut down, so it ended up with a much coarser workaround. Maybe the webservers of tomorrow will offer synchronous reload APIs instead of working on signals :D.
Or simply have the reload process wait until all the old worker processes have been superseded by a new one. Or did you mean the same thing?
That's a good idea. nginx already resolved its own portability problems, it might not be that hard to write nginx -s reload-and-wait
.
As far as I can see in the nginx code, internally it also works kinda asynchronous..
Thanks man! That indeed helped
Increasing the number of seconds to 6 already made the command pass in most cases. I will use this as a workaround for now.
It would be great though if there is a better native solution used by default. If you have any idea how to accomplish that, please create a feature request and link it here. I will try to help if I can (I just contacted Nginx support with a question about the synchronous reload).
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.