Hadoop Datanode slave未连接到我的主服务器

Question

由于许多错误,我无法弄清楚为什么在没有将datanode slave vm连接到我的主vm时发生这种情况.任何建议都是受欢迎的,所以我可以尝试一下.首先,其中一个是我的slave vm日志中的错误:

WARN org.apache.hadoop.hdfs.server.datanode.DataNode: Problem connecting to server: ubuntu-378e53c1-3e1f-4f6e-904d-00ef078fe3f8:9000

因此,我无法在我的主vm中运行我想要的工作:

hadoop jar ./share/hadoop/mapreduce/hadoop-mapreduce-examples-2.2.0.jar pi 2 5

这给了我这个错误

org.apache.hadoop.ipc.RemoteException(java.io.IOException): File /user/ubuntu/QuasiMonteCarlo_1386793331690_1605707775/in/part0 could only be replicated to 0 nodes instead of minReplication (=1).  There are 0 datanode(s) running and no node(s) are excluded in this operation.

即便如此,hdfs dfsadmin -report(在主vm)给我全部0

Configured Capacity: 0 (0 B)
Present Capacity: 0 (0 B)
DFS Remaining: 0 (0 B)
DFS Used: 0 (0 B)
DFS Used%: NaN%
Under replicated blocks: 0
Blocks with corrupt replicas: 0
Missing blocks: 0
Datanodes available: 0 (0 total, 0 dead)

为此,我建立了openstack 3 vms ubuntu,一个用于master和其他奴隶.在主人,它的建立etc/hosts

127.0.0.1 localhost
50.50.1.9 ubuntu-378e53c1-3e1f-4f6e-904d-00ef078fe3f8
50.50.1.8 slave1
50.50.1.4 slave2

核心的site.xml

<name>fs.default.name</name>
<value>hdfs://ubuntu-378e53c1-3e1f-4f6e-904d-00ef078fe3f8:9000</value>
<name>hadoop.tmp.dir</name>
<value>/home/ubuntu/hadoop-2.2.0/tmp</value>

HDFS-site.xml中

<name>dfs.replication</name>
<value>3</value>
<name>dfs.namenode.name.dir</name>
<value>file:/home/ubuntu/hadoop-2.2.0/etc/hdfs/namenode</value>
<name>dfs.datanode.data.dir</name>
<value>file:/home/ubuntu/hadoop-2.2.0/etc/hdfs/datanode</value>
<name>dfs.permissions</name>
<value>false</value>

mapred-site.xml中

<name>mapreduce.framework.name</name>
<value>yarn</value>

我的slave vm文件包含每一行:slave1和slave2.

来自master vm的所有日志都没有错误,但是当我使用slave vm时,它会给出连接错误.并且nodemanager在日志中也给出了错误:

Error starting NodeManager org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.net.ConnectException: Call From ubuntu-e6df65dc-bf95-45ca-bad5-f8ddcc272b76/50.50.1.8 to 0.0.0.0:8031 failed on connection exception: java.net.ConnectException: Connection refused;

从我的Slave Machine:core-site.xml

<name>fs.default.name</name>
<value>hdfs://ubuntu-378e53c1-3e1f-4f6e-904d-00ef078fe3f8:9000</value>
<name>hadoop.tmp.dir</name>
<value>/home/ubuntu/hadoop-2.2.0/tmp</value>

HDFS-site.xml中

<name>dfs.namenode.name.dir</name>
<value>file:/home/ubuntu/hadoop-2.2.0/etc/hdfs/namenode</value>
</property>
<property>
<name>dfs.datanode.data.dir</name>
<value>file:/home/ubuntu/hadoop-2.2.0/etc/hdfs/datanode</value>

在我的/ etc/hosts上

127.0.0.1 localhost
50.50.1.8 ubuntu-e6df65dc-bf95-45ca-bad5-f8ddcc272b76
50.50.1.9 ubuntu-378e53c1-3e1f-4f6e-904d-00ef078fe3f8

JPS大师

15863 ResourceManager
15205 SecondaryNameNode
14967 NameNode
16194 Jps

奴隶

1988 Jps
1365 DataNode
1894 NodeManager

Answer 1

导致所有错误显示的原因是，以下错误是主站无法连接到从站的主要原因：

\n\n

Error starting NodeManager org.apache.hadoop.yarn.exceptions.YarnRuntimeException: java.net.ConnectException: Call From ubuntu-e6df65dc-bf95-45ca-bad5-f8ddcc272b76/50.50.1.8 to 0.0.0.0:8031 failed on connection exception: java.net.ConnectException: Connection refused;\n

\n\n

基本上，0.0.0.0:8031是港口yarn.resourcemanager.resource-tracker.address，所以我使用 lsof -i :8031 检查，该端口未启用/打开/允许。由于我使用的是 OpenStack（云），因此添加了 8031 和其他显示错误的端口，voil\xc3\xa1，按预期工作。

\n