Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
0.8.0
-
spark3.1.1
hive3.1.1
hadoop3.1.1
Description
when we acquire hive lock failed and the lock state is WATING, we should release this WATING lock; otherwise this hive lock will be locked forever。
test step:
use hive lock to control concurrent write for hudi, let‘s call this lock hive_lock
start three writers to write hudi table by using hive_lock concurrently, one of the writer will failed to acquire hive lock due to competition issues。
Exception in thread "main" org.apache.hudi.exception.HoodieLockException: Unable to acquire lock, lock object LockResponse(lockid:76, state:WAITING)
start another writer to write hudi table by using same hive_lock, then we find hive_lock is locked forever, we have no way to acquire it
Exception in thread "main" org.apache.hudi.exception.HoodieLockException: Unable to acquire lock, lock object LockResponse(lockid:87, state:WAITING)
Attachments
Issue Links
- links to