fix exec_* startup delay if nofile is too high #1750
Draft
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Currently we rely on a for loop to close all possible file descriptors, and when nofile is set to a large number (a very common case in Nginx deployments), the worker process might eat up a single core for a few minutes(!) after
fork
and beforeexec
. This results in unexpected server performance impacts, and causesexec_push
andexec_pull
to be delayed a lot (e.g. you can't do low-latency realtime transcode anymore). Others have pointed out this problem but lowering nofiles is not a long term solution.This patch resolves the performance issue, and have been steadily runnning in my environment for a while. But it requires Linux 5.9 + glibc 2.34, or FreeBSD 8.0+. I'm not sure if we can detect the existence of
close_range
inconfigure
as I'm not really familiar of Nginx's build system conventions. I'd like to refine its compatibility if possible.