Sailesh Mukil has posted comments on this change. ( http://gerrit.cloudera.org:8080/9158 )
Change subject: IMPALA-6449: Use CLOCK_MONOTONIC in ConditionVariable ...................................................................... Patch Set 1: (1 comment) http://gerrit.cloudera.org:8080/#/c/9158/1//COMMIT_MSG Commit Message: http://gerrit.cloudera.org:8080/#/c/9158/1//COMMIT_MSG@13 PS1, Line 13: CLOCK_MONOTONIC Reading the man pages, it basically says that even CLOCK_MONOTONIC can be affected by sources like NTP. However, the difference is that with CLOCK_MONOTONIC, the time will not jump backwards, but rather experience "time slew", which means that the frequency of clock ticks can change momentarily to reflect adjustments in NTP. CLOCK_MONOTONIC_RAW on the other hand shows the absolute wall clock time that's not affected by NTP. But it could be affected by things like temperature, environment, etc. that will not be corrected (CLOCK_MONOTONIC would have that corrected apparently) It seems a bit confusing which would be considered "right" while measuring time like we do. I'm okay with both, as long as we've considered both the options. Whatever you think is the right option, I think we can just add to the commit message that we've considered the other one and give a reason as to why we didn't choose that. -- To view, visit http://gerrit.cloudera.org:8080/9158 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I81611cfd5e7c5347203fe7fa6b0f615602257f87 Gerrit-Change-Number: 9158 Gerrit-PatchSet: 1 Gerrit-Owner: Michael Ho <k...@cloudera.com> Gerrit-Reviewer: Jim Apple <jbapple-imp...@apache.org> Gerrit-Reviewer: Michael Ho <k...@cloudera.com> Gerrit-Reviewer: Sailesh Mukil <sail...@cloudera.com> Gerrit-Comment-Date: Fri, 02 Feb 2018 22:03:42 +0000 Gerrit-HasComments: Yes