[
https://issues.apache.org/jira/browse/HDFS-7496?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lei (Eddy) Xu updated HDFS-7496:
--------------------------------
Attachment: HDFS-7496.000.patch
Thanks for the suggestions [~cmccabe].
In this patch, I made the following changes:
# Add {{FsVolumeImpl#reference}}, {{FsVolumeImpl#unreference}} and
{{FsVolumeImpl#closeAndWait()}} to manage the liveness of a {{FsVolumeImpl}}
instance.
# After calling {{closeAndWait()}}, all future {{reference()}}s on the same
{{FsVolumeImpl}} will throw IOException.
# The functions that get stats on the {{FsVolume}} return 0 after close, e.g.,
{{getAverage()}}.
# The functions that requires IOs return {{IOException}}, e.g., {{checkDirs()}}
and {{getVolumeMap()}}.
# Since {{FsVolumeList#getNextVolume()}} exploits {{FsVolume}} to other
threads, we let {{BlockReceiver}} manages the reference of a volume that it is
actively writing.
> Fix FsVolume removal race conditions on the DataNode
> -----------------------------------------------------
>
> Key: HDFS-7496
> URL: https://issues.apache.org/jira/browse/HDFS-7496
> Project: Hadoop HDFS
> Issue Type: Bug
> Reporter: Colin Patrick McCabe
> Assignee: Lei (Eddy) Xu
> Attachments: HDFS-7496.000.patch
>
>
> We discussed a few FsVolume removal race conditions on the DataNode in
> HDFS-7489. We should figure out a way to make removing an FsVolume safe.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
