Dan Hecht has posted comments on this change. ( http://gerrit.cloudera.org:8080/9053 )
Change subject: IMPALA-6416: extend Thread::Create to track instance id ...................................................................... Patch Set 4: (2 comments) http://gerrit.cloudera.org:8080/#/c/9053/4/be/src/common/thread-debug-info.h File be/src/common/thread-debug-info.h: http://gerrit.cloudera.org:8080/#/c/9053/4/be/src/common/thread-debug-info.h@115 PS4, Line 115: system_thread_id_ > I think it is possible that the parent TDI object gets destroyed by the tim Are you saying that thread_debug_info_ pointer may point to stray memory? If that's the case, maybe we shouldn't keep that pointer but instead have a way to traverse TDIs and find it if still exists? http://gerrit.cloudera.org:8080/#/c/9053/4/be/src/common/thread-debug-info.h@119 PS4, Line 119: boost::thread::id boost_thread_id_; : int64_t system_thread_id_ = 0; > boost::thread::id is less prone to being recycled, but I think this propert I'm not sure what you mean by "switch to the parent thread quickly". I don't think I fully understand the tradeoffs. Just seems confusing to have multiple IDs and I don't understand how each of these fields helps in debugging. Perhaps some comments attached to them explaining how one can use them for debugging would help? -- To view, visit http://gerrit.cloudera.org:8080/9053 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I27de2962cf0b224c17b685d77dcba3bf2e9db187 Gerrit-Change-Number: 9053 Gerrit-PatchSet: 4 Gerrit-Owner: Zoltan Borok-Nagy <borokna...@cloudera.com> Gerrit-Reviewer: Dan Hecht <dhe...@cloudera.com> Gerrit-Reviewer: Lars Volker <l...@cloudera.com> Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-Reviewer: Zoltan Borok-Nagy <borokna...@cloudera.com> Gerrit-Comment-Date: Mon, 12 Feb 2018 19:45:40 +0000 Gerrit-HasComments: Yes