question Re: Problem when Distcp between two HA Cluster. in Archives of Support Questions (Read Only)

Problem when Distcp between two HA Cluster.

mrdashanzi — Tue, 09 Oct 2018 10:52:00 GMT

Hi, ALL;

I have two hadoop cluster named as cluster_1 and cluster_2 with different zookeepers. Now I want to Distcp hdfs files from cluster_1 to cluster_2.

following as cluster_1's and cluster_2's information.

### cluster_1
1, active namenode: g001.server.edu.tk standby namenode: g002.server.edu.tk
2, zookeeper hosts: g003.server.edu.tk g004.server.edu.tk g005.server.edu.tk
### cluster_2
1, active namenode: d001.server.edu.tk standby namenode: d002.server.edu.tk
2, zookeeper hosts: d003.server.edu.tk d004.server.edu.tk d005.server.edu.tk<br>

1, In order to distcp data from cluster_1 to cluster_2, I copied whole hadoop configuration files from $HADOOP_HOME/etc/hadoop to /configurations/hadoop in cluster_1 and added following properties in hdfs-site.xml in g001.server.org.tk:

<property>
    <name>dfs.nameservices</name>  
    <value>cluster_1,cluster_2</value>
</property>

<property>   
    <name>dfs.client.failover.proxy.provider.cluster_2</name> 
    <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>
</property> 
<property>
     <name>dfs.ha.namenodes.cluster_2</name>    
    <value>d001.server.edu.tk,d002.server.edu.tk</value> </property>
<property>
    <name>dfs.namenode.rpc-address.cluster_2.d001.server.edu.tk</name> 
    <value>d001.server.edu.tk:8020</value>
</property> 

<property> 
    <name>dfs.namenode.servicerpc-address.cluster_2.d001.server.edu.tk</name> 
    <value>d001.server.edu.tk:54321</value> 
</property>

<property>
    <name>dfs.namenode.http-address.cluster_2.d001.server.edu.tk</name> 
    <value>d001.server.edu.tk:50070</value>
</property>

<property> 
    <name>dfs.namenode.https-address.cluster_2.d001.server.edu.tk</name> 
    <value>d001.server.edu.tk:50470</value>
</property>

<property>
    <name>dfs.namenode.rpc-address.cluster_2.d002.server.edu.tk</name> 
    <value>d002.server.edu.tk:8020</value>
</property>

<property>
    <name>dfs.namenode.servicerpc-address.cluster_2.d002.server.edu.tk</name> 
    <value>d002.server.edu.tk:54321</value>
</property>

<property>
    <name>dfs.namenode.http-address.cluster_2.d002.server.edu.tk</name>
    <value>d002.server.edu.tk:50070</value>
</property>

<property> 
    <name>dfs.namenode.https-address.cluster_2.d002.server.edu.tk</name>  
    <value>d002.server.edu.tk:50470</value>
</property>

<property>
    <name>dfs.client.failover.proxy.provider.cluster_1</name>     
    <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value> 
</property> 

<property>
     <name>dfs.ha.namenodes.cluster_1</name> 
    <value>g001.server.edu.tk,g002.server.edu.tk</value>
</property>
<property>   
    <name>dfs.namenode.rpc-address.cluster_1.g001.server.edu.tk</name> 
    <value>g001.server.edu.tk:8020</value>
</property>

<property>
    <name>dfs.namenode.servicerpc-address.cluster_1.g001.server.edu.tk</name>   
    <value>g001.server.edu.tk:54321</value>
</property>

<property>
    <name>dfs.namenode.http-address.cluster_1.g001.server.edu.tk</name> 
    <value>g001.server.edu.tk:50070</value>
</property>
<property> 
    <name>dfs.namenode.https-address.cluster_1.g001.server.edu.tk</name> 
    <value>g001.server.edu.tk:50470</value>
</property>

<property>
    <name>dfs.namenode.rpc-address.cluster_1.g002.server.edu.tk</name>
    <value>g002.server.edu.tk:8020</value>
</property>

<property>
    <name>dfs.namenode.servicerpc-address.cluster_1.g002.server.edu.tk</name> 
    <value>g002.server.edu.tk:54321</value>
</property>

<property> 
    <name>dfs.namenode.http-address.cluster_1.g002.server.edu.tk</name> 
     <value>g002.server.edu.tk:50070</value>
</property>

<property>
    <name>dfs.namenode.https-address.cluster_1.g002.server.edu.tk</name> 
    <value>g002.server.edu.tk:50470</value>
</property>

2, Then I ran the Distcp command in g001.server.org.tk:

hdfs --config /configurations/hadoop distcp -update -p hdfs://cluster_1:8020/tmp/ hdfs:/cluster_2:8020/tmp/

3, But got errors like:

18/10/08 07:55:00 ERROR tools.Distcp: Exception encountered
java.io.IOException: org.apache.hadoop.yarn.exception.YarnException: Failed to submit application_xxx to YARN: Failed to renew token Kind: HDFS_DELEGATION_YOKEN, Service: hdfs:cluster_2, Ident: (HDFS_DELEGATION_TOKEN token 50168 for hdfs)
   at org.apache.hadoop.mapred.YARNRunner.submitJob(YARNRunner.java:306)
   at org.apache.hadoop.mapreduce.JobSubmitter.submitJobInternal(Jobsubmitter.java:240)
   at org.apache.hadoop.mapreduce.Job$10.run(Job.java:1290)
   at org.apache.hadoop.mapreduce.Job$10.run(Job.java:1287)
   at java.security.AccessController.doPrivileged(Native Method)
   at javax.security.auth.Subject.doAs(Subject;.java:422)
   at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1698)
   at org.apache.hadoop.mapreduce.Job.submit(Job.java:1287)
   at org.apache.hadoop.tools.DistCp.createAndSubmitJob(DistCp.java:183)
   at org.apache.hadoop.tools.DistCp.execute(DistCp.java:153)
   at org.apache.hadoop.tools.DistCp.run(DistCp.java:126)
   at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:70)
   at org.apache.hadoop.tools.DistCp.main(DistCp.java:430)
Caused by: org.apache.hadoop.yarn.exceptions.YarnException: Failed to submit application_xxxx to YARN: Failed to RENEW token: Kind: HDFS_DELEGATION_TOKEN, Service: hfs:cluster_2, Ident: (HDFS_DELEGATION_TOKEN token 50168 for hdfs)
   at org.apache.hadoop.yarn.client.api.impl.YarnClientImpl.submitApplication(YarnClientImpl.java:271)
   at org.apache.hadoop.mapred.ResourceMgrDelegate.submitApplication(ResourceMgrDelegate.java:290)
   at org.apache.hadoop.mapred.YARNRunner.submitJob(YARNRunner.java:291)
   ... 12 more

4, Hadoop version is 2.7.3

Re: Problem when Distcp between two HA Cluster.

pminovic — Tue, 09 Oct 2018 10:52:01 GMT

Use this, and server name and port if you are doing distcp directly to the active NN on remote cluster:

-Dmapreduce.job.hdfs-servers.token-renewal.exclude=cluster_2

Re: Problem when Distcp between two HA Cluster.

JonathanSneep — Tue, 09 Oct 2018 10:52:01 GMT

Hi @Shen Sean

It looks like you may be hitting YARN-3021 - https://issues.apache.org/jira/browse/YARN-3021

Try the same distcp operation after adding in following parameters to distcp command

-Dmapreduce.job.hdfs-servers.token-renewal.exclude=<destinationNN1>,<destinationNN2>

Re: Problem when Distcp between two HA Cluster.

mrdashanzi — Tue, 09 Oct 2018 10:52:01 GMT

4, Hadoop version is 2.7.3

Re: Problem when Distcp between two HA Cluster.

mrdashanzi — Tue, 09 Oct 2018 10:52:02 GMT

Hi, @Jonathan Sneep Thank you for your reply. I tried the method you suggest but got the same error. Below is the command what I ran:

hdfs --config /configurations/hadoop distcp -Dmapreduce.job.hdfs-servers.token-renewal.exclude=d001.server.edu.tk,d002.server.edu.tk -update -p hdfs://cluster_1:8020/tmp/ hdfs:/cluster_2:8020/tmp/

Re: Problem when Distcp between two HA Cluster.

mrdashanzi — Tue, 09 Oct 2018 10:52:10 GMT

Hi, @Predrag Minovic Thank you for your reply. I tried the method you suggest but got the same error. Below is the command what I ran:

hdfs --config /configurations/hadoop distcp -Dmapreduce.job.hdfs-servers.token-renewal.exclude=d001.server.edu.tk,d002.server.edu.tk -update -p hdfs://cluster_1:8020/tmp/ hdfs:/cluster_2:8020/tmp/

hdfs --config /configurations/hadoop distcp -Dmapreduce.job.hdfs-servers.token-renewal.exclude=d001.server.edu.tk:8020,d002.server.edu.tk:8020 -update -p hdfs://cluster_1:8020/tmp/ hdfs:/cluster_2:8020/tmp/


hdfs --config /configurations/hadoop distcp -Dmapreduce.job.hdfs-servers.token-renewal.exclude=d001.server.edu.tk,d002.server.edu.tk:8020 -update -p hdfs://cluster_1:8020/tmp/ hdfs:/cluster_2:8020/tmp/


hdfs --config /configurations/hadoop distcp -Dmapreduce.job.hdfs-servers.token-renewal.exclude=cluster_2 -update -p hdfs://cluster_1:8020/tmp/ hdfs:/cluster_2:8020/tmp/


hdfs --config /configurations/hadoop distcp -Dmapreduce.job.hdfs-servers.token-renewal.exclude=cluster_2:8020 -update -p hdfs://cluster_1:8020/tmp/ hdfs:/cluster_2:8020/tmp/

Re: Problem when Distcp between two HA Cluster.

pminovic — Tue, 09 Oct 2018 16:06:42 GMT

Sorry for hard to understand message, try this:

hdfs --config /configurations/hadoop distcp -Dmapreduce.job.hdfs-servers.token-renewal.exclude=cluster_2 -update -p hdfs://cluster_1/tmp/ hdfs:/cluster_2/tmp/

Note that you don't need port when using NN service name. Also I suggest to copy first a small file or directory in /tmp, like /tmp/mydir1, just create that dir and put a few files inside. Also remove '-update -p' during initial tests. Once it starts working you can try all that.

Re: Problem when Distcp between two HA Cluster.

mrdashanzi — Wed, 10 Oct 2018 09:07:40 GMT

Hi, @Predrag Minovic; The hadoop version is 2.7.3. It seems property 'mapreduce.job.hdfs-servers.token-renewal.exclude' is not available for hadoop version under 2.8.0. I've tried the method you provided, but got same error.