Lars Volker has posted comments on this change. Change subject: IMPALA-3977: TransmitData() should not block ......................................................................
Patch Set 1: (4 comments) http://gerrit.cloudera.org:8080/#/c/5491/1/be/src/runtime/data-stream-sender.cc File be/src/runtime/data-stream-sender.cc: PS1, Line 226: FLAGS_datastream_sender_timeout_ms > rather than have the sleep period be a function of the total timeout, it ma Does it make sense to add some exponential back-off here, let's say up to a maximum value? That way, if the sender out-races the receiver we will try again sooner and bound delays. http://gerrit.cloudera.org:8080/#/c/5491/1/common/thrift/generate_error_codes.py File common/thrift/generate_error_codes.py: Line 233: ("DATASTREAM_SENDER_TIMEOUT", 72, "Sender timed out waiting for receiver fragment " Should we mark this as deprecated here? PS1, Line 314: DATASTREAM_RECEIVER_EAGAIN Would it make sense to name this more general, e.g. "RPC_TRY_AGAIN" so we can re-use it in similar situations? Line 315: "Datastream receiver is not yet ready"), nit: single line? -- To view, visit http://gerrit.cloudera.org:8080/5491 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I47f9d6c1de95a72ab73471dcd2a3118ef13e7db0 Gerrit-PatchSet: 1 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Sailesh Mukil <sail...@cloudera.com> Gerrit-Reviewer: Henry Robinson <he...@cloudera.com> Gerrit-Reviewer: Lars Volker <l...@cloudera.com> Gerrit-Reviewer: Sailesh Mukil <sail...@cloudera.com> Gerrit-HasComments: Yes