Details
-
Bug
-
Status: Resolved
-
Normal
-
Resolution: Fixed
-
None
-
Normal
Description
AsyncOneResponse has a bug where it uses the initial timeout value instead of the adjustedTimeout. Combined with passing in the wrong TimeUnit, it leads to a shorter timeout than expected. This can have unintended consequences for example, in StorageService::sendReplicationNotification instead of waiting 10 seconds (request_timeout_in_ms), we wait for 10000 Nano Seconds.
Attachments
Attachments
Issue Links
- is duplicated by
-
CASSANDRA-14510 Flaky uTest: RemoveTest.testRemoveHostId
- Resolved
- relates to
-
CASSANDRA-14514 StorageService::sendReplicationNotification has potential to end up in an infinite loop causing JVM to GC and die
- Open