dheerajturaga opened a new pull request, #65838:
URL: https://github.com/apache/airflow/pull/65838
A wedged task could hold an edge3 worker in drain indefinitely after a
shutdown request, blocking restarts and administrative drain sequences.
Add two new [edge] configs, both opt-in:
drain_timeout_sec - max wait for jobs to finish after drain starts
(0 = wait forever, preserves legacy behavior)
drain_kill_grace_sec - grace after SIGTERM before SIGKILL (default 30)
When enabled, the worker loop enforces the deadline: SIGTERM at
timeout, SIGKILL after grace, exit regardless of remaining jobs.
Drain-start timestamp is recorded on any drain path (SIGINT, SIGTERM,
SHUTDOWN_REQUEST heartbeat response, version mismatch).
Also stop emitting Shutdown requested! on every heartbeat while the
server keeps returning SHUTDOWN_REQUEST during a drain; log only on the
transition into drain.
The SIGINT drain log now includes the configured timeout and grace
values so operators can see what the worker will actually do.
##### Was generative AI tooling used to co-author this PR?
<!--
If generative AI tooling has been used in the process of authoring this PR,
please
change below checkbox to `[X]` followed by the name of the tool, uncomment
the "Generated-by".
-->
- [ ] Yes (please specify the tool below)
ClaudeCode Opus 4.7
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
