Zakelly commented on code in PR #23679:
URL: https://github.com/apache/flink/pull/23679#discussion_r1387572052
##########
docs/content.zh/docs/deployment/cli.md:
##########
@@ -153,6 +153,35 @@ $ ./bin/flink savepoint \
Triggering the savepoint disposal through the `savepoint` action does not only
remove the data from
the storage but makes Flink clean up the savepoint-related metadata as well.
+### Creating a Checkpoint
+[Checkpoints]({{< ref "docs/ops/state/checkpoints" >}}) can also be manually
created to save the
+current state. To get the difference between checkpoint and savepoint, please
refer to
+[Checkpoints vs. Savepoints]({{< ref
"docs/ops/state/checkpoints_vs_savepoints" >}}). All that's
+needed to trigger a checkpoint manually is the JobID:
+```bash
+$ ./bin/flink checkpoint \
+ $JOB_ID
+```
+```
+Triggering checkpoint for job 99c59fead08c613763944f533bf90c0f.
+Waiting for response...
+Checkpoint(CONFIGURED) 26 for job 99c59fead08c613763944f533bf90c0f completed.
+You can resume your program from this checkpoint with the run command.
+```
+If you want to trigger a full checkpoint while the job periodically triggering
incremental checkpoints,
+please use the `--full` option.
Review Comment:
After some digging, I found when a job is configured to run with rocksdb and
incremental checkpoint disabled, the `--incremental` flag could not take
effect. That is because the `RocksNativeFullSnapshotStrategy` does not take
`CheckpointOptions` into account when taking async snapshots. So since only
`--full` flag works when incremental checkpoint enabled, I only describe this
usage here. But I'm ok with your phrasing.
Also, we should reconsider the incremental file-sharing chain in this
senario. How does a manually triggered incremental checkpoint reuse the files
from previous full checkpoints, which exist in `chk-x` instead of `shared`
directory? I think we need a more flexible checkpoint directory layout. WDYT?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
