Yida Wu has posted comments on this change. ( http://gerrit.cloudera.org:8080/23094 )
Change subject: IMPALA-12057: Track removed coordinators to reject queued queries early ...................................................................... Patch Set 8: (2 comments) http://gerrit.cloudera.org:8080/#/c/23094/7/be/src/scheduling/cluster-membership-mgr.cc File be/src/scheduling/cluster-membership-mgr.cc: http://gerrit.cloudera.org:8080/#/c/23094/7/be/src/scheduling/cluster-membership-mgr.cc@195 PS7, Line 195: _markCoordinatorAsRemoved > addToRemoved sounds conflicting to me. Perhaps MarkCoordinatorAsMissing is Changed the function name to _markCoordinatorAsRemoved. http://gerrit.cloudera.org:8080/#/c/23094/7/be/src/scheduling/cluster-membership-mgr.cc@221 PS7, Line 221: _markCoordinatorAsRemoved(state, be, *actual_be); > Under intermittent network partition, is it possible that same coordinator Yeah, I think it's possible a coordinator marked as removed could come back with the same backend id. If that happens, admissiond would pick it up in the current backends and doens't need to check the removed list. Even if it briefly flaps and a request is rejected, it's not critical, the coordinator can retry. The current design keeps things simple while remaining safe, so we keep the removed coord entry as-is -- To view, visit http://gerrit.cloudera.org:8080/23094 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I1e0f270299f8c20975d7895c17f4e2791c3360e0 Gerrit-Change-Number: 23094 Gerrit-PatchSet: 8 Gerrit-Owner: Yida Wu <[email protected]> Gerrit-Reviewer: Abhishek Rawat <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Riza Suminto <[email protected]> Gerrit-Reviewer: Wenzhe Zhou <[email protected]> Gerrit-Reviewer: Yida Wu <[email protected]> Gerrit-Comment-Date: Thu, 10 Jul 2025 17:20:41 +0000 Gerrit-HasComments: Yes
