Question

在h2o实例（版本3.24.0.2（Hadoop））上运行了较长的过程（约2个小时）后，python脚本似乎失去了连接，并显示错误消息

...
File "/home/osboxes/projects/ml1c/venv/local/lib/python2.7/site-packages/h2o/h2o.py", line 274, in init
    raise H2OConnectionError('Can only start H2O launcher if IP address is localhost.')
H2OConnectionError: Can only start H2O launcher if IP address is localhost.

检查运行脚本之前启动的h2o实例，按enter后显示终端输出

...
H2O node 172.18.4.62:54321 reports H2O cluster size 5 [leader is 172.18.4.62:54321]
H2O node 172.18.4.66:54321 reports H2O cluster size 5 [leader is 172.18.4.62:54321]
H2O node 172.18.4.65:54321 reports H2O cluster size 5 [leader is 172.18.4.62:54321]
H2O cluster (5 nodes) is up
(Note: Use the -disown option to exit the driver after cluster formation)

Open H2O Flow in your web browser: http://172.18.4.62:54321

(Press Ctrl-C to kill the cluster)
Blocking until the H2O cluster shuts down...
packet_write_wait: Connection to 172.18.4.62 port 22: Broken pipe

有人知道该怎么做才能解决此问题吗？

对于上下文，我正在按行分块处理一个大熊猫数据帧，将其转换为h2o数据帧，并将其与行绑定h2o.rbind放在一起。

H2O群集意外关闭； packet_write_wait：连接到xxx.xxx.xx.xx端口22：管道断开

0 个答案: