[
https://issues.apache.org/jira/browse/HDFS-11548?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15943933#comment-15943933
]
Anu Engineer commented on HDFS-11548:
-------------------------------------
Thank you for updating the patch. Some very minor comments below.
* Could we please rewrite this pattern ?
{code}
try {
lock.writeLock().lock();
{code}
* Replace this pattern with
{code}
lock.writeLock().lock();
try {
...
} finally {
lock.writeLock().unlock();
}
That seems to be the Java idiom.
https://docs.oracle.com/javase/7/docs/api/java/util/concurrent/locks/Lock.html
{code}
* nit: Preconditions.CheckNotNull(pool);
{code}
public void addNode(final String pool, final DatanodeID node)
{code}
Same feedback in removeNode, checkNotNull on node.
* Hard coded nodeCount in {{testDefaultNodePoolReload}}
> OZone: SCM: Add node pool management API
> ----------------------------------------
>
> Key: HDFS-11548
> URL: https://issues.apache.org/jira/browse/HDFS-11548
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Xiaoyu Yao
> Assignee: Xiaoyu Yao
> Attachments: HDFS-11548-HDFS-7240.001.patch,
> HDFS-11548-HDFS-7240.002.patch, HDFS-11548-HDFS-7240.003.patch
>
>
> The idea is to group registered nodes into pools of fixed size (say 24 nodes
> per pool) so that the container allocation and report can all be handled
> independently on a pool basis by SCM.
> The initial patch will implement the following Node Pool API that
> 1) add node to a node pool
> 2) remove a node from a pool
> 3) get the pool name that a node belongs to
> 4) get all the pool names
> 5) get all nodes of a pool
> The integration with SCM container allocation can be all nodes in a single
> default pool upon registration. We will provide a CLI to manage multiple
> pools and support for pool definition file later.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
