[
https://issues.apache.org/jira/browse/FLINK-6364?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15993343#comment-15993343
]
ASF GitHub Bot commented on FLINK-6364:
---------------------------------------
Github user StefanRRichter commented on a diff in the pull request:
https://github.com/apache/flink/pull/3801#discussion_r114368960
--- Diff:
flink-contrib/flink-statebackend-rocksdb/src/main/java/org/apache/flink/contrib/streaming/state/RocksDBKeyedStateBackend.java
---
@@ -137,6 +156,14 @@
/** Number of bytes required to prefix the key groups. */
private final int keyGroupPrefixBytes;
+ /** The sst files materialized in pending checkpoints */
+ private final SortedMap<Long, Map<String, StreamStateHandle>>
materializedSstFiles = new TreeMap<>();
+
+ /** The identifier of the last completed checkpoint */
+ private final long lastCompletedCheckpointId = -1;
--- End diff --
Currently, this value `lastCompletedCheckpointId` is not maintained at all,
and also the `materializedSstFiles` is ever-growing. I think the whole feedback
from the checkpoint coordinator about completed checkpoints is still missing.
Are you planning to do this in another PR?
> 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)
