Zoltan Martonka has posted comments on this change. ( http://gerrit.cloudera.org:8080/20725 )
Change subject: KUDU-3527 Fix block manager test when using 64k container block alignment ...................................................................... Patch Set 14: (9 comments) http://gerrit.cloudera.org:8080/#/c/20725/14//COMMIT_MSG Commit Message: http://gerrit.cloudera.org:8080/#/c/20725/14//COMMIT_MSG@21 PS14, Line 21: RHEL 8.8 ARM 64k core > Did you mean "RHEL 8.8 ARM with 64K page size"? Done http://gerrit.cloudera.org:8080/#/c/20725/14//COMMIT_MSG@36 PS14, Line 36: 64k core > nit: 64K page size ? Done http://gerrit.cloudera.org:8080/#/c/20725/14//COMMIT_MSG@51 PS14, Line 51: Added a jira to find a solution > nit: Filed a JIRA issue to track the issue Done http://gerrit.cloudera.org:8080/#/c/20725/14/src/kudu/fs/block_manager-test.cc File src/kudu/fs/block_manager-test.cc: http://gerrit.cloudera.org:8080/#/c/20725/14/src/kudu/fs/block_manager-test.cc@161 PS14, Line 161: CreateABlock > style nit: why not just CreateBlock() ? If you want to emphasize it's a si Thank you. I just used the original lambda name (create_a_block), did not gave it much thought. I will rename it to CreateBlock() http://gerrit.cloudera.org:8080/#/c/20725/14/src/kudu/fs/block_manager-test.cc@942 PS14, Line 942: is > are Done http://gerrit.cloudera.org:8080/#/c/20725/14/src/kudu/fs/block_manager-test.cc@949 PS14, Line 949: resistent > resistant Done http://gerrit.cloudera.org:8080/#/c/20725/14/src/kudu/fs/block_manager-test.cc@981 PS14, Line 981: // The chance that we delete the : // .metadata but fail to delete the .data file is (1-x) * x > I guess that's a bit simplistic given that the deletion of these two files Look at the last 2 lines of LogBlockContainerNativeMeta::~LogBlockContainerNativeMeta(). There is no error handling there. You can verify that the delete indeed happens when this smart pointers are reset: + Put a breakpoint to line 995 (CommitDeletedBlocks) in block_manager-test.cc and run the new test (use --gtest_also_run_disabled_tests) + When the breakpoint is hit, put a breakpoint into "Status DeleteFile(...)" in env_posix.cc. Resume the debugging. You will see a .data then a .metadata delete. you can check the debug callstack to see they are indeed called from ~LogBlockContainerNativeMeta() So the deletion error chances are independent as long as it is our artificial error and not a real disk error. http://gerrit.cloudera.org:8080/#/c/20725/14/src/kudu/fs/block_manager-test.cc@1022 PS14, Line 1022: , > nit: remove the comma? Done http://gerrit.cloudera.org:8080/#/c/20725/14/src/kudu/fs/block_manager-test.cc@1028 PS14, Line 1028: " > nit: remove the stray quote? Done -- To view, visit http://gerrit.cloudera.org:8080/20725 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: kudu Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I7e325bde502b7d7f39dd17fa84cb7eb42a3d7c20 Gerrit-Change-Number: 20725 Gerrit-PatchSet: 14 Gerrit-Owner: Zoltan Martonka <zmarto...@cloudera.com> Gerrit-Reviewer: Abhishek Chennaka <achenn...@cloudera.com> Gerrit-Reviewer: Alexey Serbin <ale...@apache.org> Gerrit-Reviewer: Ashwani Raina <ara...@cloudera.com> Gerrit-Reviewer: Kudu Jenkins (120) Gerrit-Reviewer: Mahesh Reddy <mre...@cloudera.com> Gerrit-Reviewer: Wang Xixu <1450306...@qq.com> Gerrit-Reviewer: Zoltan Martonka <zmarto...@cloudera.com> Gerrit-Comment-Date: Mon, 29 Jan 2024 16:26:25 +0000 Gerrit-HasComments: Yes