hadoop-yarn - Hortonworks Nodemanager 启动但随后失败 : Connection refused to :8042

Question

我正在尝试解决 Hortonworks 集群上新添加的数据节点的问题。节点的 YARN namenode 管理器在启动后不久就会失败。返回以下错误消息日志:

Connection failed to http://(ipaddress):8042/ws/v1/node/info (Traceback (most recent call last):
  File "/var/lib/ambari-agent/cache/common-services/YARN/2.1.0.2.0/package/alerts/alert_nodemanager_health.py", line 166, in execute
    connection_timeout=curl_connection_timeout, kinit_timer_ms = kinit_timer_ms)
  File "/usr/lib/python2.6/site-packages/resource_management/libraries/functions/curl_krb_request.py", line 198, in curl_krb_request
    _, curl_stdout, curl_stderr = get_user_call_output(curl_command, user=user, env=kerberos_env)
  File "/usr/lib/python2.6/site-packages/resource_management/libraries/functions/get_user_call_output.py", line 61, in get_user_call_output
    raise ExecutionFailed(err_msg, code, files_output[0], files_output[1])
ExecutionFailed: Execution of 'curl --location-trusted -k --negotiate -u : -b /var/lib/ambari-agent/tmp/cookies/4268dd36-9f72-4be0-8d82-5f0a124a3a72 -c /var/lib/ambari-agent/tmp/cookies/4268dd36-9f72-4be0-8d82-5f0a124a3a72 http://gdcdrwhdb821.dir.ucb-group.com:8042/ws/v1/node/info --connect-timeout 5 --max-time 7 1>/tmp/tmp7pZrbM 2>/tmp/tmpgM4wdg' returned 7.   % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0curl: (7) Failed connect to (ipaddress):8042; Connection refused
)

这并没有真正告诉我为什么连接被拒绝，除了与端口 8042 对应的任何 Yarn 进程没有运行:

netstat -tulpn | grep 8042

我一直在寻找另一个可能包含更多信息的 nodemanager 日志，但在/var/log/hadoop-yarn 或 yarn.nodemanager.local-dirs/yarn.nodemanager.log-dirs 下找不到任何有用的信息

还有其他地方可以查找 yarn nodemanager 错误日志吗？有谁知道是什么原因造成的？

编辑:重新检查后，我在/var/log/hadoop-yarn/yarn/yarn-yarn-nodemanager-(ipaddress).log 中发现了这个有用的位

2017-04-19 14:01:14,670 FATAL nodemanager.NodeManager (NodeManager.java:initAndStartNodeManager(549)) - Error starting NodeManager
org.apache.hadoop.service.ServiceStateException: java.lang.ClassNotFoundException: org.apache.spark.network.yarn.YarnShuffleService

Answer 1

你能解决这个问题吗？

我今天遇到了类似的问题。

我在 HDP 集群中停止了 YARN，删除了/var/log/hadoop-yarn/nodemanager/recovery-state 目录，然后再次启动了 YARN。

nodemanager 正在正常运行。