Yifan Zhang has posted comments on this change. ( http://gerrit.cloudera.org:8080/16319 )
Change subject: KUDU-3180: prioritize larger mem-stores in time-based flusing ...................................................................... Patch Set 1: (1 comment) http://gerrit.cloudera.org:8080/#/c/16319/1/src/kudu/tablet/tablet_replica_mm_ops.cc File src/kudu/tablet/tablet_replica_mm_ops.cc: http://gerrit.cloudera.org:8080/#/c/16319/1/src/kudu/tablet/tablet_replica_mm_ops.cc@120 PS1, Line 120: std::max(1.0, extra_mb) > Clever! We want to distinguish between mem-stores that above the threshold and that doesn't reach the threshold, now if there is a mem-store over threshold less than 1MB, it will have the same perf_improvement with a small mem-store that has been there for a while. This change ensures over threshold mem-stores always have a higher perf_improvement than those below threshold. With this change I think we could reduce memory usage and avoid generating too much small diskrowsets. However I didn't test this on the real production cluster and it's difficult to reproduce actual workloads in a test cluster. Of course enable tablet throttler could also help our cluster I think. -- To view, visit http://gerrit.cloudera.org:8080/16319 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: kudu Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I0a826643709a4990e40b0a49f89f4ea34f14163b Gerrit-Change-Number: 16319 Gerrit-PatchSet: 1 Gerrit-Owner: Yifan Zhang <[email protected]> Gerrit-Reviewer: Andrew Wong <[email protected]> Gerrit-Reviewer: Kudu Jenkins (120) Gerrit-Reviewer: Yifan Zhang <[email protected]> Gerrit-Comment-Date: Tue, 11 Aug 2020 11:37:27 +0000 Gerrit-HasComments: Yes
