[
https://issues.apache.org/jira/browse/HDDS-116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16495415#comment-16495415
]
Xiaoyu Yao commented on HDDS-116:
---------------------------------
{quote}I think we can instead use AtomicLong? Thoughts?
{quote}
Yes, we should use AtomicLong.
{quote}Even if we grab the VolumeSet lock here, the Container to which this
volume is returned might still write to a removed volume. We should have the
calling function grab a lock the VolumeSet when passing the volumeList to
RRVolumeChoosingPolicy.
{quote}
Agree. Maybe we use composite pattern and make VolumeChoosingPolicy part of
VolumeSet? This way, we don't have to expose the lock.
HDDS currently uses a ContainerStorageLocation/StorageLocation and starts DU
thread per location(volume) to get the usage info. From the design spec, it
seems that we are going to use VolumeSet/VolumeInfo to replace
ContainerStorageLocation/StorageLocation. However, the current VolumeInfo does
not have ability to get the usage information itself like
ContainerStorageLocation. We don't have to address it now if you plan to add
it later with a different JIRA?
> Implement VolumeSet to manage disk volumes
> ------------------------------------------
>
> Key: HDDS-116
> URL: https://issues.apache.org/jira/browse/HDDS-116
> Project: Hadoop Distributed Data Store
> Issue Type: Sub-task
> Reporter: Hanisha Koneru
> Assignee: Hanisha Koneru
> Priority: Major
> Labels: ContainerIO
> Fix For: 0.2.1
>
> Attachments: HDDS-116-HDDS-48.001.patch, HDDS-116-HDDS-48.002.patch,
> HDDS-116-HDDS-48.003.patch
>
>
> VolumeSet would be responsible for managing volumes in the Datanode. Some of
> its functions are:
> # Initialize volumes on startup
> # Provide APIs to add/ remove volumes
> # Choose and return volume to calling service based on the volume choosing
> policy (currently implemented Round Robin choosing policy)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
