Young-Seok Kim has posted comments on this change. Change subject: Deadlock-free locking protocol is enabled ......................................................................
Patch Set 6: (1 comment) https://asterix-gerrit.ics.uci.edu/#/c/825/6/asterixdb/asterix-runtime/src/main/java/org/apache/asterix/runtime/operators/AsterixLSMPrimaryUpsertOperatorNodePushable.java File asterixdb/asterix-runtime/src/main/java/org/apache/asterix/runtime/operators/AsterixLSMPrimaryUpsertOperatorNodePushable.java: Line 233: } > Why didn't we take care of the upsert use case too? It has been taken care of as is done in AsterixLSMInsertDeleteOperatorNodePushable by flushing partial frame to the next operator and eventually to commit operator. However, there is a difference between upsert and nonupsert node pushable. In upsertNodePushable writes tuples one at a time in the nextFrame() method. (This is how you implemented this upsert op). In contrast, insertXXXOpNodePushable writes in a batch manner when the frame is pushed to the nextOp. So, logically, they are doing the same thing. -- To view, visit https://asterix-gerrit.ics.uci.edu/825 To unsubscribe, visit https://asterix-gerrit.ics.uci.edu/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ie58ae2f519baa53599e99b51bd61ea5f8366dafd Gerrit-PatchSet: 6 Gerrit-Project: asterixdb Gerrit-Branch: master Gerrit-Owner: Young-Seok Kim <kiss...@gmail.com> Gerrit-Reviewer: Ian Maxon <ima...@apache.org> Gerrit-Reviewer: Jenkins <jenk...@fulliautomatix.ics.uci.edu> Gerrit-Reviewer: Murtadha Hubail <hubail...@gmail.com> Gerrit-Reviewer: Till Westmann <ti...@apache.org> Gerrit-Reviewer: Yingyi Bu <buyin...@gmail.com> Gerrit-Reviewer: Young-Seok Kim <kiss...@gmail.com> Gerrit-Reviewer: abdullah alamoudi <bamou...@gmail.com> Gerrit-HasComments: Yes
