[
https://issues.apache.org/jira/browse/FLINK-6364?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15989920#comment-15989920
]
ASF GitHub Bot commented on FLINK-6364:
---------------------------------------
GitHub user shixiaogang opened a pull request:
https://github.com/apache/flink/pull/3801
[FLINK-6364] [checkpoints] Implement incremental checkpointing in
RocksDBKeyedStateBackend
This is the initial implementation of incremental checkpointing in
RocksDBKeyedStateBackend. Changes include
1. Add a new `CheckpointType` for incremental checkpoints.
2. Call the `restore()` method for all `KeyedStateBackend` if the restore
state handle is null or empty.
3. Implement `RocksDBIncrementalSnapshotOperation` and
`RocksDBIncrementalRestoreOperation` which supports incremental snapshotting
and restoring with/without parallelism changes, respectively.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/shixiaogang/flink flink-6364
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/3801.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #3801
----
commit c5340a2186ecfbc48c46048db5adb11af83db511
Author: xiaogang.sxg <[email protected]>
Date: 2017-04-29T15:44:36Z
Implement incremental checkpointing in RocksDBKeyedStateBackend
----
> Implement incremental checkpointing in RocksDBStateBackend
> ----------------------------------------------------------
>
> Key: FLINK-6364
> URL: https://issues.apache.org/jira/browse/FLINK-6364
> Project: Flink
> Issue Type: Sub-task
> Components: State Backends, Checkpointing
> Reporter: Xiaogang Shi
> Assignee: Xiaogang Shi
>
> {{RocksDBStateBackend}} is well suited for incremental checkpointing because
> RocksDB is base on LSM trees, which record updates in new sst files and all
> sst files are immutable. By only materializing those new sst files, we can
> significantly improve the performance of checkpointing.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
