diff options
| author | Jeff Widman <jeff@jeffwidman.com> | 2017-10-05 22:20:42 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2017-10-05 22:20:42 -0700 |
| commit | e418b7c3b000375ec4d95c8b302cfec2e4efc85f (patch) | |
| tree | 4e4b70318210b34b80a9c80345739b588e514caa | |
| parent | ffc7caef13a120f69788bcdd43ffa01468f575f9 (diff) | |
| download | kafka-python-check-for-none-in-on-join-complete.tar.gz | |
Explicitly check for `None` rather than Falsecheck-for-none-in-on-join-complete
If the group leader somehow gets in a state that it has an empty partition assignment, then `self._assignment_snapshot` will be `{}` which evaluates to `False`. So `self._subscription.mark_for_reassignment()` will never be triggered, even if `self._assignment_snapshot != self._metadata_snapshot`.
Fixes the symptoms of https://github.com/dpkp/kafka-python/issues/1237 although I suspect there's an additional bug in that case that triggers the condition of the the group leader getting an empty partition assignment.
| -rw-r--r-- | kafka/coordinator/consumer.py | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/kafka/coordinator/consumer.py b/kafka/coordinator/consumer.py index 123699f..84c62df 100644 --- a/kafka/coordinator/consumer.py +++ b/kafka/coordinator/consumer.py @@ -190,7 +190,7 @@ class ConsumerCoordinator(BaseCoordinator): # if we were the assignor, then we need to make sure that there have # been no metadata updates since the rebalance begin. Otherwise, we # won't rebalance again until the next metadata change - if self._assignment_snapshot and self._assignment_snapshot != self._metadata_snapshot: + if self._assignment_snapshot is not None and self._assignment_snapshot != self._metadata_snapshot: self._subscription.mark_for_reassignment() return |