Dan Hecht has posted comments on this change. ( http://gerrit.cloudera.org:8080/11337 )
Change subject: IMPALA-7335: Fix a race in HdfsScanNode ...................................................................... Patch Set 3: (1 comment) http://gerrit.cloudera.org:8080/#/c/11337/3/be/src/exec/hdfs-scan-node.cc File be/src/exec/hdfs-scan-node.cc: http://gerrit.cloudera.org:8080/#/c/11337/3/be/src/exec/hdfs-scan-node.cc@516 PS3, Line 516: before marking a range as complete. thanks, that's helpful. but why do we mark a range as complete if we get an error here, whereas if StartNextScanRange() errors, we presumably don't do that. In other words, maybe another solution to the race is to no "complete" the range if there was an error. Or, it seems that both error paths really could be common (i.e. why not mark the range complete in the case of StartNextScanRange() error). Hmm, maybe there's something subtle for abort_on_error=false, though, so maybe don't muck with that after all. Anyway, I'm fine with not trying to do too much cleanup on this right now (besides Tim's refactoring suggestion), if it will eventually be generally cleaned up. -- To view, visit http://gerrit.cloudera.org:8080/11337 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Id470818846a5c69ad28a6ff695069702982aa793 Gerrit-Change-Number: 11337 Gerrit-PatchSet: 3 Gerrit-Owner: Pooja Nilangekar <pooja.nilange...@cloudera.com> Gerrit-Reviewer: Bikramjeet Vig <bikramjeet....@cloudera.com> Gerrit-Reviewer: Dan Hecht <dhe...@cloudera.com> Gerrit-Reviewer: Impala Public Jenkins <impala-public-jenk...@cloudera.com> Gerrit-Reviewer: Pooja Nilangekar <pooja.nilange...@cloudera.com> Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-Comment-Date: Wed, 29 Aug 2018 01:02:12 +0000 Gerrit-HasComments: Yes
