[ https://issues.apache.org/jira/browse/HDFS-14959?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16972736#comment-16972736 ]
Hudson commented on HDFS-14959: ------------------------------- SUCCESS: Integrated in Jenkins build Hadoop-trunk-Commit #17636 (See [https://builds.apache.org/job/Hadoop-trunk-Commit/17636/]) HDFS-14959: [SBNN read] access time should be turned off (#1706) (weichiu: rev 97ec34e117af71e1a9950b8002131c45754009c7) * (edit) hadoop-hdfs-project/hadoop-hdfs/src/site/markdown/ObserverNameNode.md > [SBNN read] access time should be turned off > -------------------------------------------- > > Key: HDFS-14959 > URL: https://issues.apache.org/jira/browse/HDFS-14959 > Project: Hadoop HDFS > Issue Type: Task > Components: documentation > Reporter: Wei-Chiu Chuang > Assignee: Chao Sun > Priority: Major > Fix For: 3.3.0, 3.1.4, 3.2.2 > > > Both Uber and Didi shared that access time has to be switched off to avoid > spiky NameNode RPC process time. If access time is not off entirely, > getBlockLocations RPCs have to update access time and must access the active > NameNode. (that's my understanding. haven't checked the code) > We should record this as a best practice in our doc. > (If you are on the ASF slack, check out this thread > https://the-asf.slack.com/archives/CAD7C52Q3/p1572033324008600) -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org