Till Westmann has posted comments on this change. Change subject: ASTERIXDB-1791, ASTERIXDB-1796: fix failure handling in runtime operators. ......................................................................
Patch Set 11: Code-Review+2 (12 comments) Looks like a big improvement to me. I have a few questions, but nothing that should keep this from being merged. https://asterix-gerrit.ics.uci.edu/#/c/1513/11/asterixdb/asterix-app/src/test/java/org/apache/asterix/test/runtime/LangExecutionUtil.java File asterixdb/asterix-app/src/test/java/org/apache/asterix/test/runtime/LangExecutionUtil.java: PS11, Line 132: IS_OS_WINDOWS Let's file an issue for windows. https://asterix-gerrit.ics.uci.edu/#/c/1513/11/asterixdb/asterix-app/src/test/resources/runtimets/queries_sqlpp/failure/q18_large_volume_customer_failure/q18_large_volume_customer_failure.2.update.sqlpp File asterixdb/asterix-app/src/test/resources/runtimets/queries_sqlpp/failure/q18_large_volume_customer_failure/q18_large_volume_customer_failure.2.update.sqlpp: PS11, Line 22: (`delimiter`=`|`)); linebreak? https://asterix-gerrit.ics.uci.edu/#/c/1513/11/asterixdb/asterix-external-data/src/main/java/org/apache/asterix/external/operators/FeedIntakeOperatorNodePushable.java File asterixdb/asterix-external-data/src/main/java/org/apache/asterix/external/operators/FeedIntakeOperatorNodePushable.java: PS11, Line 79: another fix line length? https://asterix-gerrit.ics.uci.edu/#/c/1513/11/hyracks-fullstack/hyracks/hyracks-dataflow-common/src/main/java/org/apache/hyracks/dataflow/common/io/RunFileWriter.java File hyracks-fullstack/hyracks/hyracks-dataflow-common/src/main/java/org/apache/hyracks/dataflow/common/io/RunFileWriter.java: PS11, Line 74: delete Wondering if we should null-out the file reference to check that we don't access it again. https://asterix-gerrit.ics.uci.edu/#/c/1513/11/hyracks-fullstack/hyracks/hyracks-dataflow-std/src/main/java/org/apache/hyracks/dataflow/std/join/HybridHashJoinOperatorDescriptor.java File hyracks-fullstack/hyracks/hyracks-dataflow-std/src/main/java/org/apache/hyracks/dataflow/std/join/HybridHashJoinOperatorDescriptor.java: PS11, Line 532: closeJoin Why don't we need to close the join if the while loop throws? https://asterix-gerrit.ics.uci.edu/#/c/1513/11/hyracks-fullstack/hyracks/hyracks-dataflow-std/src/main/java/org/apache/hyracks/dataflow/std/join/OptimizedHybridHashJoin.java File hyracks-fullstack/hyracks/hyracks-dataflow-std/src/main/java/org/apache/hyracks/dataflow/std/join/OptimizedHybridHashJoin.java: PS11, Line 444: return If we exit here, doesn't the finally block below close r again? Is that safe? https://asterix-gerrit.ics.uci.edu/#/c/1513/11/hyracks-fullstack/hyracks/hyracks-dataflow-std/src/main/java/org/apache/hyracks/dataflow/std/join/OptimizedHybridHashJoinOperatorDescriptor.java File hyracks-fullstack/hyracks/hyracks-dataflow-std/src/main/java/org/apache/hyracks/dataflow/std/join/OptimizedHybridHashJoinOperatorDescriptor.java: PS11, Line 458: logProbeComplete If we fail before we get to close(), we logProbeComplete. But when we fail in here, we don'r logProbeComplete. Is that intended? PS11, Line 574: setIsReversed Move this before the "try"? That would make the "open" the first statement in the try-block. PS11, Line 581: always Only if we don't throw. PS11, Line 593: closeProbe Same as above, not necessarily closed - is that ok? PS11, Line 747: close This could throw as well. Is that a problem? PS11, Line 773: closeCache This could throw as well. Is that a problem? -- To view, visit https://asterix-gerrit.ics.uci.edu/1513 To unsubscribe, visit https://asterix-gerrit.ics.uci.edu/settings Gerrit-MessageType: comment Gerrit-Change-Id: I203168171e6dac16b57d2eda960823e3810e22a3 Gerrit-PatchSet: 11 Gerrit-Project: asterixdb Gerrit-Branch: master Gerrit-Owner: Yingyi Bu <[email protected]> Gerrit-Reviewer: Jenkins <[email protected]> Gerrit-Reviewer: Taewoo Kim <[email protected]> Gerrit-Reviewer: Till Westmann <[email protected]> Gerrit-HasComments: Yes
