Bharath Vissapragada has posted comments on this change. ( http://gerrit.cloudera.org:8080/11871 )
Change subject: IMPALA-6436: exit instead of abort for catalog startup failure ...................................................................... Patch Set 1: (1 comment) What is the criterion that you used to decide between clean_exit or abort? I feel that most aborts() can be can be converted to clean exits. Like - Anything that runs in the process setup phase like FindClass()/GetMethodID() etc, the stack trace should be helpful enough to figure out the problem. - All New*() methods (like NewString()) return a NULL / throw exception. We catch the exception anyway (like OOM) and dump it (but interestingly we are not doing a null check anywhere) Wondering where exactly we might need core dump for debugging. http://gerrit.cloudera.org:8080/#/c/11871/1/be/src/util/jni-util.h File be/src/util/jni-util.h: http://gerrit.cloudera.org:8080/#/c/11871/1/be/src/util/jni-util.h@89 PS1, Line 89: log the backtrace at FATAL level and abort the : // process nit: I think LOG(FATAL) calls the abort. Maybe clarify that? -- To view, visit http://gerrit.cloudera.org:8080/11871 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I4026dccb39843b847426112fc0fe9ba897e48dcc Gerrit-Change-Number: 11871 Gerrit-PatchSet: 1 Gerrit-Owner: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-Reviewer: Bharath Vissapragada <bhara...@cloudera.com> Gerrit-Reviewer: Impala Public Jenkins <impala-public-jenk...@cloudera.com> Gerrit-Comment-Date: Sat, 03 Nov 2018 00:18:27 +0000 Gerrit-HasComments: Yes
