[ https://issues.apache.org/jira/browse/FLINK-5229?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15715405#comment-15715405 ]
Till Rohrmann commented on FLINK-5229: -------------------------------------- Thanks for the pointer [~aljoscha]. The PR I've opened is for the release-1.1 branch. I've checked the PR and it looks as if some of the work has been done. However, it is not yet exhaustive (e.g. the non partitionable state handles are not discarded properly and the error reporting can be improved). I'll base my work in this branch then. So please don't split the commit up. > Cleanup StreamTaskStates if a checkpoint operation of a subsequent operator > fails > ---------------------------------------------------------------------------------- > > Key: FLINK-5229 > URL: https://issues.apache.org/jira/browse/FLINK-5229 > Project: Flink > Issue Type: Bug > Components: State Backends, Checkpointing, TaskManager > Affects Versions: 1.2.0, 1.1.3 > Reporter: Till Rohrmann > Assignee: Till Rohrmann > Fix For: 1.2.0, 1.1.4 > > > Due to chaining, a {{StreamTask}} needs to checkpoint multiple operators. If > the first operators succeed in creating a checkpoint but a subsequent > operator in the chain fails, the {{StreamTask}} has to clean up the already > completed checkpoints. Otherwise we might end up with orphaned state data. -- This message was sent by Atlassian JIRA (v6.3.4#6332)