Joe McDonnell has posted comments on this change. ( http://gerrit.cloudera.org:8080/20801 )
Change subject: IMPALA-11909: Use absolute path when calling resolve_minidumps.py ...................................................................... Patch Set 1: (1 comment) http://gerrit.cloudera.org:8080/#/c/20801/1//COMMIT_MSG Commit Message: http://gerrit.cloudera.org:8080/#/c/20801/1//COMMIT_MSG@14 PS1, Line 14: On some platforms (Centos 7), resolve_minidump.py's call to > Is it a prerequisite for the parallel run? I'm not clear what about these c Basically, I had merged something to use resolve_minidumps.py in finalize.sh, but the absolute path issue prevents it from actually running. When testing the fix for the absolute path issue, I found out that resolve_minidumps.py goes crazy with high memory usage on Centos 7. I didn't want to fix the absolute path issue without having a workaround for the memory issue. That's what I mean by fixing the memory issue being a prerequisite. Basically, resolve_minidumps.py is dumping symbols from all the system libraries, which is something we don't do for other minidump resolution methods. One of those libraries has corrupt symbol information (or something like that) and when those symbols are present they cause the stackwalker to go haywire. Maybe it is an infinite loop. It is not a prerequisite for the parallel symbol dumping, and I could split that out. -- To view, visit http://gerrit.cloudera.org:8080/20801 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I063843554b52d3e8ed79ee32d9fd4c90d059c482 Gerrit-Change-Number: 20801 Gerrit-PatchSet: 1 Gerrit-Owner: Joe McDonnell <joemcdonn...@cloudera.com> Gerrit-Reviewer: Impala Public Jenkins <impala-public-jenk...@cloudera.com> Gerrit-Reviewer: Joe McDonnell <joemcdonn...@cloudera.com> Gerrit-Reviewer: Michael Smith <michael.sm...@cloudera.com> Gerrit-Comment-Date: Fri, 15 Dec 2023 22:44:56 +0000 Gerrit-HasComments: Yes
