Alexey Serbin has uploaded a new change for review. http://gerrit.cloudera.org:8080/6683
Change subject: [rpc-test] fixed TestConnectionKeepalive flakiness ...................................................................... [rpc-test] fixed TestConnectionKeepalive flakiness After introducing TLS option for Kudu RPC, the TestConnectionKeepalive became a bit flaky, failing time to time with output like the following: [ RUN ] OptionalSSL/TestRpc.TestConnectionKeepalive/1 I0419 01:31:19.019090 4863 rpc-test.cc:223] Connecting to 0.0.0.0:52055 src/kudu/rpc/rpc-test.cc:225: Failure Failed Bad status: Network error: Recv() got EOF from remote (error 108) [ RUN ] OptionalSSL/TestRpc.TestConnectionKeepalive/1 I0419 01:31:25.609285 370 rpc-test.cc:223] Connecting to 0.0.0.0:41418 W0419 01:31:25.853305 398 connection.cc:462] client connection to \ 0.0.0.0:41418 recv error: Network error: recv error: \ Connection reset by peer (error 104) src/kudu/rpc/rpc-test.cc:225: Failure Failed Bad status: Network error: recv error: Connection reset by peer (error 104) It seems the TLS connection establishment phase in some cases takes too long, so by the time of test RPC call might close. Bumping the keepalive interval helped to address the issue. Prior to this fix, if running with --stress_cpu_threads=32 option, usually 15 out of 1024 runs failed. After the fix 0 out of 4096 runs failed. Change-Id: I8fb2dec953c1047ea43fe9e28d470ae6566077c5 --- M src/kudu/rpc/rpc-test.cc 1 file changed, 2 insertions(+), 2 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/83/6683/1 -- To view, visit http://gerrit.cloudera.org:8080/6683 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newchange Gerrit-Change-Id: I8fb2dec953c1047ea43fe9e28d470ae6566077c5 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Alexey Serbin <aser...@cloudera.com>