[
https://issues.apache.org/jira/browse/FLINK-5480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16423983#comment-16423983
]
ASF GitHub Bot commented on FLINK-5480:
---------------------------------------
Github user zentol commented on the issue:
https://github.com/apache/flink/pull/3117
@ShashwatRastogi-Reflektion
1. For each task to load from the savepoint S1:
* Determine the existing hash by searching for the task ID in the logs/UI
* Set the uidHash for that task to that value via
`SingleOutputStreamOperator#setUidHash`
2. For each task:
* Set the uid to whatever value you wish in the future to use via
`SingleOutputStreamOperator#setUid`
3. Resume the job from the savepoint S1.
4. Create a new savepoint S2, and remove all calls to `setUidHash`
5. Resume the job from the savepoint S2.
> User-provided hashes for operators
> ----------------------------------
>
> Key: FLINK-5480
> URL: https://issues.apache.org/jira/browse/FLINK-5480
> Project: Flink
> Issue Type: Improvement
> Components: DataStream API
> Affects Versions: 1.2.0
> Reporter: Stefan Richter
> Assignee: Stefan Richter
> Priority: Major
> Fix For: 1.2.0, 1.3.0
>
>
> We could allow users to provided (alternative) hashes for operators in a
> StreamGraph. This can make migration between Flink versions easier, in case
> the automatically produced hashes between versions are incompatible. For
> example, users could just copy the old hashes from the web ui to their job.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
