Hadoop error: Bad connection to FS. command aborted.

运行

hadoop fs -ls

时出现错误如下：

root@ubuntu:/home/chenwq/hadoop/book/ch03/src/main/java# hadoop fs -ls
11/08/31 22:51:37 INFO security.Groups: Group mapping impl=org.apache.hadoop.security.ShellBasedUnixGroupsMapping; cacheTimeout=300000
11/08/31 22:51:38 WARN conf.Configuration: mapred.task.id is deprecated. Instead, use mapreduce.task.attempt.id
11/08/31 22:51:39 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 0 time(s).
11/08/31 22:51:40 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 1 time(s).
11/08/31 22:51:41 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 2 time(s).
11/08/31 22:51:42 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 3 time(s).
11/08/31 22:51:43 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 4 time(s).
11/08/31 22:51:44 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 5 time(s).
11/08/31 22:51:45 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 6 time(s).
11/08/31 22:51:46 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 7 time(s).
11/08/31 22:51:47 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 8 time(s).
11/08/31 22:51:48 INFO ipc.Client: Retrying connect to server: localhost/127.0.0.1:8020. Already tried 9 time(s).
Bad connection to FS. command aborted.

 

 

解决方案:

格式化namenode:

hadoop namenode -format

重新启动start-all.sh，输入

jps

 

查看后台进程

root@ubuntu:/home/chenwq/hadoop/hadoop-0.21.0/bin# jps
8734 NameNode
8934 DataNode
9137 SecondaryNameNode
9479 Jps

此时namenode启动

 

运行

hadoop fs -ls

出现：

11/08/31 23:09:37 INFO security.Groups: Group mapping impl=org.apache.hadoop.security.ShellBasedUnixGroupsMapping; cacheTimeout=300000
11/08/31 23:09:37 WARN conf.Configuration: mapred.task.id is deprecated. Instead, use mapreduce.task.attempt.id
Found 1 items
-rw-r--r--   1 root supergroup        628 2011-08-31 23:01 /user/root/books/URLCat.java

  

 

是伪分布式下的环境吗？

是的

为什么我发现网上但凡出现Datanode无法启动的问题都去格式化Namenode,就是不去找为什么启动不了的原因.难道生产环境也这么搞? 知道Namenode有多重要吗?
localhost/127.0.0.1:8020. Already tried 2 time(s).  
大多都是hdfs没有启动造成的,何必去格式化Namenode

datanode都没有启动！看日志

格式化就啥也没了。。

同求问题解决方案，我只要重启就需要格式化才能启动成功，不知道到底哪里错了。

我的问题已经查到原因，但是没有去试验
以下是我查到的资料：

每次都要重新 format namenode，有可能是您未設定 hadoop.tmp.dir 到其他路徑。
預設值是 /tmp/hadoop-${user.name}，/tmp 重開機就會容易被清空。

Hadoop的默认配置使用hadoop.tmp.dir作为本地文件系统和HDFS的临时目录，
在使用中，我们应创建自己的临时目录

$sudo mkdir -p /app/hadoop/tmp
$sudo chown chenwq:chenwq /app/hadoop/tmp