[
https://issues.apache.org/jira/browse/HDFS-5324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13793498#comment-13793498
]
Milind Bhandarkar commented on HDFS-5324:
-----------------------------------------
Konst,
The approach we have taken, is to get the current FSNamesystem class, and move
everything that may be reused by other namespace implementations to the
AbstractNamesystem class. The rest remains in the FSNamesystem class.
We envision two types of Namesystem implementations. One that only
re-implements the FSDirectory, and hence an AbstractDirectory is created as
well.
The other type of implementation, which completely does away with the
FSDirectory.
Currently, only the first type is considered. In future, we may want to
refactor this more to allow second type as well.
Initially, the FSNamesystem was approximately 5500 lines. Now about 5400 lines
from there are exported to AbstractNamesystem, and the remaining FSNamesystem
has about 1100 lines (including duplicate imports etc).
I am attaching AbstractNamesystem class as it stands today, but please note
that this is very much work in progress, and not a patch.
> Make Namespace implementation pluggable in the namenode
> -------------------------------------------------------
>
> Key: HDFS-5324
> URL: https://issues.apache.org/jira/browse/HDFS-5324
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: namenode
> Affects Versions: 2.1.1-beta
> Environment: All
> Reporter: Milind Bhandarkar
> Assignee: Milind Bhandarkar
> Fix For: 3.0.0
>
>
> For the last couple of months, we have been working on making Namespace
> implementation in the namenode pluggable. We have demonstrated that it can
> be done without major surgery on the namenode, and does not have noticeable
> performance impact. We would like to contribute it back to Apache HDFS.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
