[ https://issues.apache.org/jira/browse/HDFS-4953?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13741623#comment-13741623 ]
Andrew Wang commented on HDFS-4953: ----------------------------------- Great, I think we're very close. bq. * comapre the genStamp here. That way, we will not return a shorter typo {code} return new ClientMmapManager(conf.getInt(DFS_CLIENT_MMAP_CACHE_SIZE, DFS_CLIENT_MMAP_CACHE_SIZE_DEFAULT), conf.getLong(DFS_CLIENT_MMAP_CACHE_TIMEOUT_MS, DFS_CLIENT_MMAP_CACHE_TIMEOUT_MS_DEFAULT)); {code} weird spacing * I think the CacheCleaner delay and period could still be improved. An initial delay of timeout still makes sense to me, and maybe we could introduce a conf parameter that affects the period. I doubt that this will be twiddled all that much, so let's set good defaults. A more aggressive default period of {{timeout/4}} might make more sense. * javac warnings still need to be addressed (looks like via the ignore file) * I'd like to see yet more class javadoc for ZeroCopyCursor (man page level). I bet many HDFS users will be eager to use this more efficient interface but won't necessarily understand e.g. ownership of the fallback buffer, and how to handle short vs. not short reads. It's also important since there isn't a design doc. We can defer this to a follow-on JIRA though if you prefer. > enable HDFS local reads via mmap > -------------------------------- > > Key: HDFS-4953 > URL: https://issues.apache.org/jira/browse/HDFS-4953 > Project: Hadoop HDFS > Issue Type: New Feature > Affects Versions: 2.3.0 > Reporter: Colin Patrick McCabe > Assignee: Colin Patrick McCabe > Attachments: benchmark.png, HDFS-4953.001.patch, HDFS-4953.002.patch, > HDFS-4953.003.patch, HDFS-4953.004.patch, HDFS-4953.005.patch, > HDFS-4953.006.patch, HDFS-4953.007.patch > > > Currently, the short-circuit local read pathway allows HDFS clients to access > files directly without going through the DataNode. However, all of these > reads involve a copy at the operating system level, since they rely on the > read() / pread() / etc family of kernel interfaces. > We would like to enable HDFS to read local files via mmap. This would enable > truly zero-copy reads. > In the initial implementation, zero-copy reads will only be performed when > checksums were disabled. Later, we can use the DataNode's cache awareness to > only perform zero-copy reads when we know that checksum has already been > verified. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira