Skip to content

Conversation

@wynnw
Copy link
Contributor

@wynnw wynnw commented Sep 23, 2025

This is a fix for #2722 which is a critical bug blocking existing users from upgrading. It keeps the same logic as the original fix in #2656 (included in the 2.0.27 release) by blocking the gracefully_kill_them_all() function from finishing until all the child workers exit. BUT, it uses the WNOHANG approach which seems to allow existing signal handler logic to do the needed cleanup for the old master process to exit cleanly.

Please consider this - thanks!

@wynnw wynnw force-pushed the bugfix/gracefully-kill-them-all-nohang branch from 1fa601b to b2d2bf9 Compare September 24, 2025 14:55
…d master graceful reload

- Instead of doing a blocking waitpid call, use WNOHANG calls to avoid
  breaking existing child process signal handling which was part of the
  master fork/exit cleanup that allowed FIFO-based zero downtime reloads
  to happen. This will also allow the fix to continue on as this code
  will block until all the workers have exited.
@wynnw wynnw force-pushed the bugfix/gracefully-kill-them-all-nohang branch from b2d2bf9 to 79f00d1 Compare October 7, 2025 20:10
@xrmx xrmx merged commit 129cf47 into unbit:uwsgi-2.0 Oct 11, 2025
23 checks passed
@xrmx
Copy link
Collaborator

xrmx commented Oct 11, 2025

Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants