[ https://issues.apache.org/jira/browse/CRUNCH-660?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16743122#comment-16743122 ]
Andrew Olson commented on CRUNCH-660: ------------------------------------- On a side note, even DistCp can be pretty severely afflicted by https://issues.apache.org/jira/browse/HADOOP-15281 (it writes to a temp file then does a costly S3 rename, but that extra work is parallelized across all of the map tasks). > FileTargetImpl uses Distcp vs FileUtils.copy > -------------------------------------------- > > Key: CRUNCH-660 > URL: https://issues.apache.org/jira/browse/CRUNCH-660 > Project: Crunch > Issue Type: Improvement > Components: Core > Reporter: Micah Whitacre > Assignee: Josh Wills > Priority: Major > > So for handling multiple runtimes I'm not sure there is a way to solve this > but documenting as a JIRA regardless. > If you are running in a multi-cluster environment where you might want to > read data from one cluster and then write the output on another cluster (e.g. > generating HFiles to be loaded into a separate HBase cluster), the > performance of moving files is noticeable. Specifically due to the fact that > the moving of the files happens in the launcher/driver process versus as part > of the node execution it seems.[1] > An efficient option would be to kick off a DistCp instead but that would tie > the target directly to a runtime which is not a great approach. > [1] - > https://github.com/apache/crunch/blob/5609b014378d3460a55ce25522f0c00659872807/crunch-core/src/main/java/org/apache/crunch/io/impl/FileTargetImpl.java#L157 -- This message was sent by Atlassian JIRA (v7.6.3#76005)