[ 
https://issues.apache.org/jira/browse/BIGTOP-1174?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13878219#comment-13878219
 ] 

Evans Ye commented on BIGTOP-1174:
----------------------------------

[~mackrorysd]
To clarify my original thought when providing the patch, the code snippet 
{code}
Hadoop::Namenode<||> -> Hadoop::Datanode<||> -> Exec<| title == "init hdfs" |>
{code}
chain 
[collectors|http://docs.puppetlabs.com/puppet/2.7/reference/lang_collectors.html]
 in the following [executing 
order|http://docs.puppetlabs.com/puppet/2.7/reference/lang_relationships.html#chaining-arrows]:
first, run all the resources found in Hadoop::Namenode namespace
second, run all the resources found in Hadoop::Datanode namespace
thrid, run the Exec resource which has title equal to "init hdfs"

Just In case if you would like to leave any comments on the patch:)

> missing dependency setting on init-hdfs.sh in puppet recipes
> ------------------------------------------------------------
>
>                 Key: BIGTOP-1174
>                 URL: https://issues.apache.org/jira/browse/BIGTOP-1174
>             Project: Bigtop
>          Issue Type: Bug
>          Components: Deployment
>    Affects Versions: 0.7.0
>            Reporter: Evans Ye
>            Assignee: Sean Mackrory
>             Fix For: 0.8.0
>
>         Attachments: BIGTOP-1174.1.patch
>
>
> There's no dependency setting to explicitly define {{init-hdfs.sh}} to be 
> executed after hdfs get ready, so when running the puppet recipes to deploy a 
> hadoop environment, the {{init-hdfs.sh}} script failed to setup folders and 
> permissions on hdfs since it will be executed before namenode and datanode 
> has been started.
> Sample output:
> {code}
> notice: 
> /Stage[main]/Hadoop_worker_node/Hadoop::Datanode[datanode]/File[/data/1/hdfs]/mode:
>  mode changed '0700' to '0755'
> notice: /Stage[main]/Hadoop_head_node/Exec[init hdfs]/returns: executed 
> successfully
> notice: 
> /Stage[main]/Hadoop_worker_node/Hadoop::Datanode[datanode]/File[/data/2/hdfs]/mode:
>  mode changed '0700' to '0755'
> notice: 
> /Stage[main]/Hadoop_worker_node/Hadoop::Datanode[datanode]/Service[hadoop-hdfs-datanode]/ensure:
>  ensure changed 'stopped' to 'running'
> notice: 
> /Stage[main]/Hadoop_head_node/Hadoop::Namenode[namenode]/Service[hadoop-hdfs-namenode]/ensure:
>  ensure changed 'stopped' to 'running'
> notice: /Stage[main]/Hadoop_head_node/Hadoop-hbase::Master[hbase 
> master]/Service[hbase-master]/ensure: ensure changed 'stopped' to 'running'
> notice: Finished catalog run in 135.44 seconds
> {code}



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

Reply via email to