在MPI集群上训练paddle报如下错误:
Created by: CruiseSun
Mon Oct 8 17:00:17 2018[1,4]:F1008 17:00:17.361272 7286 LightNetwork.cpp:395] connection refused by pserver, maybe pserver failed! Mon Oct 8 17:00:17 2018[1,4]:*** Check failure stack trace: *** Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c5091c27d google::LogMessage::Fail() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c5091fd2c google::LogMessage::SendToLog() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c5091bda3 google::LogMessage::Flush() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c5092123e google::LogMessageFatal::~LogMessageFatal() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c507793c1 paddle::SocketClient::TcpClient() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c507795a1 paddle::SocketClient::SocketClient() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c513fc9b0 paddle::ParameterClient2::init() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c50f8909d paddle::RemoteParameterUpdater::init() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c508fc1ea ParameterUpdater::init() Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c505a5f7b _wrap_ParameterUpdater_init Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b4cb9 PyEval_EvalFrameEx Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b6b28 PyEval_EvalCodeEx Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b5d10 PyEval_EvalFrameEx Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b6b28 PyEval_EvalCodeEx Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b5d10 PyEval_EvalFrameEx Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b5fb8 PyEval_EvalFrameEx Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b6b28 PyEval_EvalCodeEx Mon Oct 8 17:00:17 2018[1,4]: @ 0x4b6c52 PyEval_EvalCode Mon Oct 8 17:00:17 2018[1,4]: @ 0x4e1c7d PyRun_FileExFlags Mon Oct 8 17:00:17 2018[1,4]: @ 0x4e3501 PyRun_SimpleFileExFlags Mon Oct 8 17:00:17 2018[1,4]: @ 0x4159dd Py_Main Mon Oct 8 17:00:17 2018[1,4]: @ 0x7f4c538f7bd5 __libc_start_main Mon Oct 8 17:00:17 2018[1,4]: @ 0x414b71 (unknown) Mon Oct 8 17:00:17 2018[1,4]: @ (nil) (unknown)