Yida Wu has posted comments on this change. ( http://gerrit.cloudera.org:8080/23873 )
Change subject: IMPALA-14682: Use centralized async cleanup for admission state cleanup ...................................................................... Patch Set 3: (3 comments) http://gerrit.cloudera.org:8080/#/c/23873/2//COMMIT_MSG Commit Message: http://gerrit.cloudera.org:8080/#/c/23873/2//COMMIT_MSG@9 PS2, Line 9: adde > nit: added Done http://gerrit.cloudera.org:8080/#/c/23873/2//COMMIT_MSG@10 PS2, Line 10: This patch refactors all admission state deletions : to use this centralized async method > Any concerns with keeping AdmissionState in memory longer with async cleanu I had the same concern, so I ran a stress test. The cleanup thread can handle about 380k cleanup per second under contention from 5 producer threads (cleanup 200k items in ~0.5s). Our normal query rate is much lower (hundreds/sec), so the queue should stay near empty and AdmissionState objects only live for a very short time http://gerrit.cloudera.org:8080/#/c/23873/2/tests/custom_cluster/test_admission_controller.py File tests/custom_cluster/test_admission_controller.py: http://gerrit.cloudera.org:8080/#/c/23873/2/tests/custom_cluster/test_admission_controller.py@2409 PS2, Line 2409: admission_state_siz > Maybe 'admission_state_size' is clearer? Done -- To view, visit http://gerrit.cloudera.org:8080/23873 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I04f46f2e42ec5e50f4dcccb6b73a34a376615ab0 Gerrit-Change-Number: 23873 Gerrit-PatchSet: 3 Gerrit-Owner: Yida Wu <[email protected]> Gerrit-Reviewer: Abhishek Rawat <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Yida Wu <[email protected]> Gerrit-Comment-Date: Tue, 20 Jan 2026 08:27:26 +0000 Gerrit-HasComments: Yes
