----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/58225/#review171175 -----------------------------------------------------------
geode-core/src/main/java/org/apache/geode/internal/cache/DistributedRegion.java Line 2129 (original), 2129 (patched) <https://reviews.apache.org/r/58225/#comment244057> just get rid of the "final" instead of commenting it out geode-core/src/main/java/org/apache/geode/internal/cache/SearchLoadAndWriteProcessor.java Line 220 (original), 222 (patched) <https://reviews.apache.org/r/58225/#comment244059> did you look into using this existing "pendingResponders" instead of adding the new "departedMembers" list? If you get a response message from a member that is not in pendingResponders then ignore it. Hmmm... incomingResponse already checks pendingResponders so perhaps that part of the fix is not needed? The only problem I see is that when netSearchForBlob is doing the replicates one at a time it does not set pendingResponders geode-core/src/main/java/org/apache/geode/internal/cache/SearchLoadAndWriteProcessor.java Lines 229 (patched) <https://reviews.apache.org/r/58225/#comment244058> I think "id" would be better here than "selectedNode". Either would be correct since you just made sure they are equal but I think id is clearer since it was passed to memberDeparted as the id of the member that departed. geode-core/src/main/java/org/apache/geode/internal/cache/SearchLoadAndWriteProcessor.java Lines 510 (patched) <https://reviews.apache.org/r/58225/#comment244061> I think this synchronization is too complicated. Just sync here and hold it until line 532. Both sendValueRequest and waitForObject2 are synchronized methods so it really does not make any sense to keep releasing the sync and reacquiring it. - Darrel Schneider On April 5, 2017, 3:20 p.m., Eric Shu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/58225/ > ----------------------------------------------------------- > > (Updated April 5, 2017, 3:20 p.m.) > > > Review request for geode, anilkumar gingade and Darrel Schneider. > > > Bugs: GEODE-2757 > https://issues.apache.org/jira/browse/GEODE-2757 > > > Repository: geode > > > Description > ------- > > Do not process netsearch reply from a departed node that membership listener > already detected. > > > Diffs > ----- > > > geode-core/src/main/java/org/apache/geode/internal/cache/DistributedRegion.java > fa02574 > > geode-core/src/main/java/org/apache/geode/internal/cache/SearchLoadAndWriteProcessor.java > 3d969f9 > > geode-core/src/test/java/org/apache/geode/internal/cache/SearchLoadAndWriteProcessorTest.java > bfe78b0 > > > Diff: https://reviews.apache.org/r/58225/diff/1/ > > > Testing > ------- > > precheckin. > > > Thanks, > > Eric Shu > >
