train.log报connection错误

[sarawon]

跑的是cluster模式，执行start trainer的task时候就卡住了：
[root@192.168.30.131:8023] Executing task 'start_trainer'
[root@192.168.30.131:8023] run: cd /root/paddle/demo/recommendation; GLOG_logtostderr=0 GLOG_log_dir="./log" nohup paddle train --num_gradient_servers=2 --nics=eth0 --port=7164 --ports_num=2 --comment=paddle_process_by_paddle --pservers=192.168.30.131,192.168.30.179 --ports_num_for_sparse=2 --config=./trainer_config.py --trainer_count=4 --use_gpu=0 --num_passes=10 --save_dir=./output --log_period=50 --dot_period=10 --saving_period=1 --local=0 --trainer_id=0 > ./log/train.log 2>&1 < /dev/null &
[root@192.168.30.131:8023] out: stdin: is not a tty
[root@192.168.30.131:8023] out:

[root@192.168.30.179:8023] Executing task 'start_trainer'
[root@192.168.30.179:8023] run: cd /root/paddle/demo/recommendation; GLOG_logtostderr=0 GLOG_log_dir="./log" nohup paddle train --num_gradient_servers=2 --nics=eth0 --port=7164 --ports_num=2 --comment=paddle_process_by_paddle --pservers=192.168.30.131,192.168.30.179 --ports_num_for_sparse=2 --config=./trainer_config.py --trainer_count=4 --use_gpu=0 --num_passes=10 --save_dir=./output --log_period=50 --dot_period=10 --saving_period=1 --local=0 --trainer_id=1 > ./log/train.log 2>&1 < /dev/null &
[root@192.168.30.179:8023] out: stdin: is not a tty
[root@192.168.30.179:8023] out:

train.log的内容：
[INFO 2016-11-24 07:17:26,152 networks.py:1466] The input order is [movie_id, title, genres, user_id, gender, age, occupation, rating]
[INFO 2016-11-24 07:17:26,152 networks.py:1472] The output order is [regression_cost_0]
F1124 07:17:26.942348 352 LightNetwork.cpp:379] Check failed: connect(sockfd, (sockaddr *)&serv_addr, sizeof(serv_addr)) >= 0 ERROR connecting to 192.168.30.131: Connection refused [111]
*** Check failure stack trace: ***
@ 0x7f1604a93daa (unknown)
@ 0x7f1604a93ce4 (unknown)
@ 0x7f1604a936e6 (unknown)
@ 0x7f1604a934fb (unknown)
@ 0x7f1604a94477 (unknown)
@ 0x69552e paddle::SocketClient::TcpClient()
@ 0x696051 paddle::SocketClient::SocketClient()
@ 0x7eaa76 std::vector<>::emplace_back<>()
@ 0x7e1be3 paddle::ParameterClient2::init()
@ 0x68e2dd paddle::RemoteParameterUpdater::init()
@ 0x678de2 paddle::Trainer::init()
@ 0x5132a9 main
@ 0x7f1603c9ff45 (unknown)
@ 0x51f2a5 (unknown)
@ (nil) (unknown)
/usr/local/bin/paddle: line 109: 352 Aborted (core dumped) ${DEBUGGER} $MYDIR/../opt/paddle/bin/paddle_trainer ${@:2}

server.log的内容：
F1124 07:19:03.638399 418 SocketChannel.cpp:180] Check failed: len == sizeof(header) : Success [0]
*** Check failure stack trace: ***
@ 0x7f9fb4dbfdaa (unknown)
@ 0x7f9fb4dbfce4 (unknown)
@ 0x7f9fb4dbf6e6 (unknown)
@ 0x7f9fb4dbf4fb (unknown)
@ 0x7f9fb4dc0477 (unknown)
@ 0x667eb8 paddle::SocketChannel::readMessage()
@ 0x6657dc paddle::SocketWorker::run()
@ 0x7f9fb493ca60 (unknown)
@ 0x7f9fb5bd2184 start_thread
@ 0x7f9fb40a437d (unknown)
@ (nil) (unknown)
/usr/local/bin/paddle: line 109: 321 Aborted (core dumped) ${DEBUGGER} $MYDIR/../opt/paddle/bin/paddle_pserver_main ${@:2}

[sarawon]

请问是集群配置有问题吗？我是启动了两个docker container 分别在两台宿主机上
docker container的22端口映射到宿主机的8022端口
7164到7167映射到宿主机的7164到7167
conf.py里的hosts指定的是root@宿主机ip:8022
然后运行的cluster_train下面的run.sh

[backyes]

@sarawon 可以提供完整的日志么？

[sarawon]

还需要什么日志啊 我贴下 上面两个文件的已经都贴了全部了

[sarawon]

docker端口映射少了 现在好了