[ https://issues.apache.org/jira/browse/FLINK-18744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17173766#comment-17173766 ]
Congxian Qiu(klion26) commented on FLINK-18744: ----------------------------------------------- The problem here exists, it was caused by FLINK-5763, after FLINK-5763 we assume that all the data and meta were in the same directory. But currently, state process API does not apply a deep copy of the previous data. The workaround, for now, needs to copy the data from the previous directory to the new directory. I think the problem here can be fixed by FLINK-14942. > resume from modified savepoint dirctionary: No such file or directory > --------------------------------------------------------------------- > > Key: FLINK-18744 > URL: https://issues.apache.org/jira/browse/FLINK-18744 > Project: Flink > Issue Type: Bug > Components: API / State Processor > Affects Versions: 1.11.0 > Reporter: tao wang > Priority: Major > > If I resume a job from a savepoint which is modified by state processor API, > such as loading from /savepoint-path-old and writing to /savepoint-path-new, > the job resumed with savepointpath = /savepoint-path-new while throwing an > Exception : > _*/savepoint-path-new/\{some-ui-id} (No such file or directory)*_. > I think it's an issue because of flink 1.11 use absolute path in savepoint > and checkpoint, but state processor API missed this. > The job will work well with new savepoint(which path is /savepoint-path-new) > if I copy all dictionary except `_metadata from` /savepoint-path-old to > /savepoint-path-new. -- This message was sent by Atlassian Jira (v8.3.4#803005)