kennknowles commented on code in PR #35298:
URL: https://github.com/apache/beam/pull/35298#discussion_r2151002410
##########
sdks/python/apache_beam/runners/worker/worker_status.py:
##########
@@ -250,6 +263,13 @@ def _log_lull_in_bundle_processor(self,
bundle_process_cache):
if processor:
info = processor.state_sampler.get_info()
self._log_lull_sampler_info(info, instruction)
+ if self._restart_lull_timeout_ns and self._restart_lull(info):
+ sys.exit(1)
Review Comment:
On the design... we want to be very careful about adding `exit` calls (here
and in Java). Because eventually code ends up getting used in unexpected
contexts. It should be written without assumptions about the context it is
called in, to the extent this is possible. So this could cause some other way
of executing to exit unexpectedly. Can we raise an exception that bubbles up
and causes the top level `main` to exit? That is really the only level where we
should make any assumptions about being in charge of the process lifetime.
##########
sdks/python/apache_beam/runners/worker/worker_status.py:
##########
@@ -47,6 +48,8 @@
# 5 minutes * 60 seconds * 1000 millis * 1000 micros * 1000 nanoseconds
DEFAULT_LOG_LULL_TIMEOUT_NS = 5 * 60 * 1000 * 1000 * 1000
+DEFAULT_RESTART_LULL_TIMEOUT_NS = 10 * 60 * 1000 * 1000 * 1000
Review Comment:
This is actually an "exit/kill" timeout.
--
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]
