[ https://issues.apache.org/jira/browse/HDFS-6840?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14094706#comment-14094706 ]
Jason Lowe commented on HDFS-6840: ---------------------------------- Commenting out the setSeed call in NetworkTopology on our 2.5-based build fixes the issue, so I suspect changing the param would work as well in 2.6+. Given how poorly off-rack load balancing is without the randomization, do we even want to keep the parameter added in HDFS-6701 at all? > Clients are always sent to the same datanode when read is off rack > ------------------------------------------------------------------ > > Key: HDFS-6840 > URL: https://issues.apache.org/jira/browse/HDFS-6840 > Project: Hadoop HDFS > Issue Type: Bug > Affects Versions: 2.5.0 > Reporter: Jason Lowe > Priority: Critical > > After HDFS-6268 the sorting order of block locations is deterministic for a > given block and locality level (e.g.: local, rack. off-rack), so off-rack > clients all see the same datanode for the same block. This leads to very > poor behavior in distributed cache localization and other scenarios where > many clients all want the same block data at approximately the same time. > The one datanode is crushed by the load while the other replicas only handle > local and rack-local requests. -- This message was sent by Atlassian JIRA (v6.2#6252)