[ https://issues.apache.org/jira/browse/FLUME-2543?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14201336#comment-14201336 ]
Joey Echeverria commented on FLUME-2543: ---------------------------------------- It will use time based or size based, though I'm not sure if we want to include size based rolls yet since we don't know the output data size, only the input data size. I can update the patch with doc fixes once we're settled on whether we want to support size based rolls yet. > Add support for Parquet datasets to the DatasetSink > --------------------------------------------------- > > Key: FLUME-2543 > URL: https://issues.apache.org/jira/browse/FLUME-2543 > Project: Flume > Issue Type: Bug > Components: Sinks+Sources > Affects Versions: v1.5.0.1 > Reporter: Joey Echeverria > Assignee: Joey Echeverria > Attachments: FLUME-2543-1.patch > > > It should be possible to support writing to Parquet datasets with > long-running transactions (1-5 minutes). -- This message was sent by Atlassian JIRA (v6.3.4#6332)