[
https://issues.apache.org/jira/browse/HDFS-5096?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13797144#comment-13797144
]
Chris Nauroth commented on HDFS-5096:
-------------------------------------
+1 for the patch, pending resolution of feedback from Andrew too. Thanks very
much, Colin!
I've had a chance to take this patch for a manual test run too in a
pseudo-distributed deployment. I created some files in a directory, and then
applied a cache directive on that directory. All of the existing files got
cached relatively quickly due to {{CacheReplicationMonitor#kick}}. Next, I
added some new files in the same directory. After the
{{dfs.namenode.path.based.cache.refresh.interval.ms}} elapsed,
{{CacheReplicationMonitor}} scanned again and cached the new files. I ran pmap
to confirm that the block files were memory-mapped into the datanode process.
I also put my namenode through a restart to confirm that we had fixed the
hanging problem I reported in HDFS-5313. I'll close that issue now.
It all looks good!
> Automatically cache new data added to a cached path
> ---------------------------------------------------
>
> Key: HDFS-5096
> URL: https://issues.apache.org/jira/browse/HDFS-5096
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: datanode, namenode
> Reporter: Andrew Wang
> Assignee: Colin Patrick McCabe
> Attachments: HDFS-5096-caching.005.patch,
> HDFS-5096-caching.006.patch, HDFS-5096-caching.009.patch,
> HDFS-5096-caching.010.patch, HDFS-5096-caching.011.patch,
> HDFS-5096-caching.012.patch
>
>
> For some applications, it's convenient to specify a path to cache, and have
> HDFS automatically cache new data added to the path without sending a new
> caching request or a manual refresh command.
> One example is new data appended to a cached file. It would be nice to
> re-cache a block at the new appended length, and cache new blocks added to
> the file.
> Another example is a cached Hive partition directory, where a user can drop
> new files directly into the partition. It would be nice if these new files
> were cached.
> In both cases, this automatic caching would happen after the file is closed,
> i.e. block replica is finalized.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
