[ https://issues.apache.org/jira/browse/NIFI-1588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15194041#comment-15194041 ]
Joseph Witt commented on NIFI-1588: ----------------------------------- muchhh better. Verified can clear state cleanly and it only keeps a couple small values like ListFile and it appears to work quite well for staying in sync with the underlying HDFS. +1 merged to master. > ListHDFS to retaining wrong state in zookeeper > ---------------------------------------------- > > Key: NIFI-1588 > URL: https://issues.apache.org/jira/browse/NIFI-1588 > Project: Apache NiFi > Issue Type: Improvement > Components: Core Framework > Affects Versions: 0.4.1 > Environment: latest 0.5.1 > Reporter: Matthew Clarke > Assignee: Mark Payne > Fix For: 0.6.0 > > Attachments: > 0001-NIFI-1588-Reworked-how-ListHDFS-store-state-so-that-.patch, > 0001-NIFI-1588-Reworked-how-ListHDFS-store-state-so-that-.patch > > > The expected state that should be retained by listHDFS processor is the last > modified timestamp. The processor instead is retaining the filename and path > of every file it lists. This results in excessive disk usage by zookeeper to > retain this filename based state. -- This message was sent by Atlassian JIRA (v6.3.4#6332)