[
https://issues.apache.org/jira/browse/FLINK-6545?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16006471#comment-16006471
]
ASF GitHub Bot commented on FLINK-6545:
---------------------------------------
Github user StefanRRichter commented on the issue:
https://github.com/apache/flink/pull/3870
@gyfora Yes, this is still a limitation, because the incremental
checkpoints are currently not (yet) externalizable. The missing piece is,
exactly as you said, familiarizing the `SavepointSerializer` with the new state
handle classes from the incremental checkpoints.
I will create a jira to track this:
https://issues.apache.org/jira/browse/FLINK-6545
> Make incremental checkpoints externalizable
> -------------------------------------------
>
> Key: FLINK-6545
> URL: https://issues.apache.org/jira/browse/FLINK-6545
> Project: Flink
> Issue Type: Sub-task
> Components: State Backends, Checkpointing
> Affects Versions: 1.3.0
> Reporter: Stefan Richter
> Priority: Blocker
> Fix For: 1.3.0
>
>
> Incremental checkpoints are currently not externalizible. The missing piece
> is familiarizing the {{SavepointSerializer}}(s) with the new state handles
> classes that we added for incremental checkpointing. Currently, some of those
> (e.g.
> {{org.apache.flink.contrib.streaming.state.RocksDBIncrementalKeyedStateHandle}})
> currently live in the contrib.rocksdb package and need to be migrated. We
> can also push them to a different abstraction level, i.e.
> {{IncrementalKeyedStateHandle}} with some private data, referenced existing
> shared data (from previous checkpoints), and (presumably) newly created
> shared data (first created by the current checkpoint).
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
