[ https://issues.apache.org/jira/browse/DRILL-2512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14491928#comment-14491928 ]
Adam Gilmore commented on DRILL-2512: ------------------------------------- This is affecting me too. We query via the REST API, but only a single node is ever tasked as the foreman. The ramification is that the foreman is way overworked and we end up with a huge performance bottleneck (the foreman's CPU maxes out). I'd consider this pretty critical as it effectively prevents users from scaling by adding more drillbits as we'll always be limited by what the single foreman can handle. Can we raise the priority on this one? > Client connecting to same node through zookeeper > ------------------------------------------------ > > Key: DRILL-2512 > URL: https://issues.apache.org/jira/browse/DRILL-2512 > Project: Apache Drill > Issue Type: Bug > Components: Execution - RPC > Reporter: Ramana Inukonda Nagaraj > Assignee: Jason Altekruse > Fix For: 1.0.0 > > > When connecting to multiple drillbits and using ZK in the connection string > it looks like drill is connecting to the same drillbit everytime. -- This message was sent by Atlassian JIRA (v6.3.4#6332)