[
https://issues.apache.org/jira/browse/HBASE-10052?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13834194#comment-13834194
]
Andrew Purtell edited comment on HBASE-10052 at 11/27/13 9:49 PM:
------------------------------------------------------------------
Looks like we could do this easily with a bit of reflection.
Edit: Hit enter or something, oops. Not sure what you mean about compacted
files not being read again, though. We will open readers on the new file
(eventually). In fact we might want to preload the blocks (HBASE-9857 - just an
example). Maybe you meant drop the data of the old discarded HFiles from
blockcache, but the API you point out is for DataOutputStream?
was (Author: apurtell):
Looks like we could do this easily with a bit of reflection.
> use HDFS advisory caching to avoid caching HFiles that are not going to be
> read again (because they are being compacted)
> ------------------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-10052
> URL: https://issues.apache.org/jira/browse/HBASE-10052
> Project: HBase
> Issue Type: Improvement
> Reporter: Colin Patrick McCabe
>
> HBase can benefit from doing dropbehind during compaction since compacted
> files are not read again. HDFS advisory caching, introduced in HDFS-4817,
> can help here. The right API here is {{DataOutputStream#setDropBehind}}.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
