[
https://issues.apache.org/jira/browse/HDFS-9608?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15098745#comment-15098745
]
Andrew Wang commented on HDFS-9608:
-----------------------------------
Patch looks good to me, nice find! For any watchers, the issue IIUC is that the
RR chooser has a single shared iterator for all storage types.
Out of curiosity, did you notice any perf issues from the shared lock?
Some nits:
* Please add some short javadoc to curVolumes and syncLocks to explain how the
indexing and locking works
* Some of the lines are longer than 80chars, but we'll catch that and other
stuff in precommit.
> Disk IO imbalance in HDFS with heterogeneous storages
> -----------------------------------------------------
>
> Key: HDFS-9608
> URL: https://issues.apache.org/jira/browse/HDFS-9608
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 2.6.0
> Reporter: Wei Zhou
> Assignee: Wei Zhou
> Attachments: HDFS-9608.01.patch, HDFS-9608.02.patch
>
>
> Currently RoundRobinVolumeChoosingPolicy use a shared index to choose volumes
> in HDFS with heterogeneous storages, this leads to non-RR choosing mode for
> certain type of storage.
> Besides, it uses a shared lock for synchronization which limits the
> concurrency of volume choosing process. Volume choosing threads that
> operating on different storage types should be able to run concurrently.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
