Dan Burkert has posted comments on this change. Change subject: catalog_manager: improve IsAlterInProgress performance ......................................................................
Patch Set 2: (3 comments) What's your thoughts on test coverage, are you assuming we have enough coverage of this already? Only thing I might like to see is more debug-enabled 'sanity checks' that the versions map is consistent in certain scenarios. E.g. maybe add a check that if a tablet is dropped and the # of tablets is 0, the versions map is empty. http://gerrit.cloudera.org:8080/#/c/7950/2/src/kudu/master/catalog_manager.cc File src/kudu/master/catalog_manager.cc: Line 4381: if (it->first == TabletInfo::NOT_YET_REPORTED) { Can't this check be skipped? Line 4389: return it->first < version; I recommend casting version to an int64_t explicitly. It's ambiguous which one is getting cast right now, and I think it's brittle to have it cast to uint32_t, even though 'it->first' should be in range given the check above. http://gerrit.cloudera.org:8080/#/c/7950/2/src/kudu/master/catalog_manager.h File src/kudu/master/catalog_manager.h: Line 161: // Simple accessor for reported_schema_version_. Should this doc line be moved to be aove reported_schema_version()? -- To view, visit http://gerrit.cloudera.org:8080/7950 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Id8c1f48c0febad038833edd555ee88f1db83249d Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo <[email protected]> Gerrit-Reviewer: Dan Burkert <[email protected]> Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon <[email protected]> Gerrit-HasComments: Yes
