[ https://issues.apache.org/jira/browse/HDFS-5776?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13878382#comment-13878382 ]
Liang Xie commented on HDFS-5776: --------------------------------- About naming, personally i'd like to keep "hedged" unchanged, it's not a big issue if HBASE-10070 use another naming, just my thought:) About increased RPC, sure, it's just a tradeoff, we all know that:) i can give a metric result against diff timeout setting, probably put into HBASE-7509 bq. doing the 3rd requests with larger timeout does not buy a lot Current patch just picks only one secondary dn, thats means there're no more than two requests under normal situation. BTW, is it possible let it go into HADOOP-2.4 if all comments are done ? i saw [~andrew.wang]'s post from mail list that there's a plan to release 2.4 in the end of this month, right ? If this patch could go in, then we can work immediately at HBASE-7509 with bumping the according dependency version to 2.4, otherwise we need to kick off once it goes into 2.5+, maybe need several weeks ? > Support 'hedged' reads in DFSClient > ----------------------------------- > > Key: HDFS-5776 > URL: https://issues.apache.org/jira/browse/HDFS-5776 > Project: Hadoop HDFS > Issue Type: Improvement > Components: hdfs-client > Affects Versions: 3.0.0 > Reporter: Liang Xie > Assignee: Liang Xie > Attachments: HDFS-5776-v2.txt, HDFS-5776-v3.txt, HDFS-5776-v4.txt, > HDFS-5776-v5.txt, HDFS-5776.txt > > > This is a placeholder of hdfs related stuff backport from > https://issues.apache.org/jira/browse/HBASE-7509 > The quorum read ability should be helpful especially to optimize read outliers > we can utilize "dfs.dfsclient.quorum.read.threshold.millis" & > "dfs.dfsclient.quorum.read.threadpool.size" to enable/disable the hedged read > ability from client side(e.g. HBase), and by using DFSQuorumReadMetrics, we > could export the interested metric valus into client system(e.g. HBase's > regionserver metric). > The core logic is in pread code path, we decide to goto the original > fetchBlockByteRange or the new introduced fetchBlockByteRangeSpeculative per > the above config items. -- This message was sent by Atlassian JIRA (v6.1.5#6160)