hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Scott Carey (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HDFS-1547) Improve decommission mechanism
Date Fri, 07 Jan 2011 09:50:47 GMT

    [ https://issues.apache.org/jira/browse/HDFS-1547?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12978710#action_12978710
] 

Scott Carey commented on HDFS-1547:
-----------------------------------

I like Todd's proposal to have only one file, that lists each node at most once, and do not
see any explanation why it won't work.

A node has only one state from the administrator POV, and what should be shown in the UI (dead,
decomission in progress, etc) can be derived from that.

Why have 3 files when one will do?  Its only more confusing.

Yes, the current two file format has issues because the meaning is overloaded or the names
are bas.  But a single file with a format like Todd suggests seems like it would work.   Possible
format:

{noformat}
node1=active
node2=decommission
node3=exclude
{noformat}

When an administrator wants to decommission a node, the part after the = in the file for that
node is changed from active to decommission.  Nodes in the decommission state are allowed
to talk to the NN and register with it, but will shut down after successful decommission.
  Nodes marked exclude are not allowed to talk to the NN.  Nodes marked active are tracked
and compared to what is regisered (along with decommission marked nodes) to identify dead
nodes.

In short, all three files in this proposal could be combined into one.

> Improve decommission mechanism
> ------------------------------
>
>                 Key: HDFS-1547
>                 URL: https://issues.apache.org/jira/browse/HDFS-1547
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: name-node
>    Affects Versions: 0.23.0
>            Reporter: Suresh Srinivas
>            Assignee: Suresh Srinivas
>             Fix For: 0.23.0
>
>
> Current decommission mechanism driven using exclude file has several issues. This bug
proposes some changes in the mechanism for better manageability. See the proposal in the next
comment for more details.

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


Mime
View raw message