蜂巢无法初始化类java.net.NetworkInterface [英] Hive Could not initialize class java.net.NetworkInterface
问题描述
我有一个EMR设置,应该是预先配置的Hive和Hadoop附带的.
I have an EMR Setup, which is supposed to come with Hive and Hadoop pre-configured.
到目前为止,对于需要MapReduce作业的任何命令,我都可以看到,Hive失败了.
As far, as I can see for any command which require MapReduce Job, Hive fails.
FAILED: Execution Error, return code 2 from org.apache.hadoop.hive.ql.exec.mr.MapRedTask
因此,我转到了Job Tracking URL,这是相关的输出:
So, I went to Job Tracking URL, and here is the relevant output:
Diagnostics: Could not initialize class java.net.NetworkInterface
java.lang.NoClassDefFoundError: Could not initialize class java.net.NetworkInterface
at org.apache.hadoop.net.NetUtils.isLocalAddress(NetUtils.java:690)
...
任何与该错误有关的帮助都将......
Any help with this error would be appreciated...
堆栈跟踪指向代码:
推荐答案
在尝试运行datanode时,我得到了这个确切的堆栈跟踪. I got this exact stacktrace when trying to run a datanode. The IPv4 address of the 这篇关于蜂巢无法初始化类java.net.NetworkInterface的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋! eth0
接口的IPv4地址在DNS中是可正向解析的,但不是可逆解析的(即,我可以将主机名解析为其IP,但是将IP反向查找为对应的主机名却没有)不起作用).一旦创建了从IP指向主机名的PTR记录,就可以按预期启动datanode.eth0
interface was forward-resolvable in DNS, but it wasn't reverse-resolvable (i.e. I could resolve the hostname to its IP but the reverse lookup of IP to the corresponding hostname didn't work). The datanode started as expected once the PTR record pointing from IP to hostname got created.