You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Nigel Daley (JIRA)" <ji...@apache.org> on 2008/02/18 03:48:37 UTC

[jira] Commented: (HADOOP-2850) [HOD] Split the operations of allocation from provisioning in HOD.

    [ https://issues.apache.org/jira/browse/HADOOP-2850?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12569774#action_12569774 ] 

Nigel Daley commented on HADOOP-2850:
-------------------------------------

+1 for this issue.   This would effectively make the allocate a pure front for qsub.  Of course, for most users, allocate and bringing up the daemons in a single step is what they'll want.  That should still be made easy for them (a single command line arg perhaps).

> [HOD] Split the operations of allocation from provisioning in HOD.
> ------------------------------------------------------------------
>
>                 Key: HADOOP-2850
>                 URL: https://issues.apache.org/jira/browse/HADOOP-2850
>             Project: Hadoop Core
>          Issue Type: Bug
>          Components: contrib/hod
>    Affects Versions: 0.16.0
>            Reporter: Hemanth Yamijala
>            Assignee: Hemanth Yamijala
>             Fix For: 0.17.0
>
>
> Currently, the HOD allocation operation does the following distinct steps - allocate a requested number of nodes, [optionally] transfer a hadoop tarball and install it, then provision hadoop by bringing up the appropriate daemons. It would be nice to separate these layers so each of them could be done independently.
> This would lead to a very great flexibility in users using the cluster. For e.g. one could allocate a certain number of nodes, then have hod bring up one version, then bring it down, then repeat this again and so on.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.