[
https://issues.apache.org/jira/browse/HBASE-4169?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13080467#comment-13080467
]
stack commented on HBASE-4169:
------------------------------
@Andrew
bq. I'm inclined to close this as invalid. It works with HDFS, which ships with
Apache Hadoop, and we are Apache HBase that runs on Apache Hadoop. Is it
incumbent on us to deal with the vagaries of XYZ vendor filesystems?
I don't have a problem making "small" changes to accomodate other FS
implementations. It could broaden our user base and experience gained atop
other FSs could help improve the home base.
@Lohit
I don't think this will work:
{code}
if (!(fs instanceof DistributedFileSystem)) {
// Change permission on file to ReadOnly.
return
}
{code}
The above will happen when running atop the local filesystem, not just for mapr
(or ceph).
I think the Todd suggestion of a mapr:// would be a better way to go though as
the lads note above, the abstraction is likely lacking when it comes to other
FS implementations (let me look at your patch).
> FSUtils LeaseRecovery for non HDFS FileSystems.
> -----------------------------------------------
>
> Key: HBASE-4169
> URL: https://issues.apache.org/jira/browse/HBASE-4169
> Project: HBase
> Issue Type: Bug
> Components: util
> Affects Versions: 0.90.3, 0.90.4
> Reporter: Lohit Vijayarenu
> Attachments: HBASE-4169.1.patch
>
>
> FSUtils.recoverFileLease uses HDFS's recoverLease method to get lease before
> splitting hlog file.
> This might not work for other filesystem implementations.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
