[SPARK-53751][SDP] Explicit Versioned Checkpoint Location #52487
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
JiaqiWang18
commented
Sep 30, 2025
| val resolvedGraph = resolveGraph() | ||
| if (context.fullRefreshTables.nonEmpty) { | ||
| State.reset(resolvedGraph, context) | ||
| } |
There was a problem hiding this comment.
with explicit storage location for checkpoint, we shouldn't need to create the tables and obtain its path beforehand. resolvedGraph should suffice.
JiaqiWang18
commented
Sep 30, 2025
sql/pipelines/src/main/scala/org/apache/spark/sql/pipelines/graph/DatasetManager.scala
Outdated
Show resolved
Hide resolved
JiaqiWang18
changed the title
JiaqiWang18
force-pushed
the
SPARK-53751-explicit-versioned-multiflow-checkpoint
branch
from
f3533ff to
60cfada
Compare
JiaqiWang18
commented
Oct 6, 2025
| } | ||
|
|
||
| override def afterEach(): Unit = { | ||
| protected override def afterEach(): Unit = { |
There was a problem hiding this comment.
keep consistency with signature of BeforeAndAfterEach and fix an inheritance error when introducing StorageRootMixin
JiaqiWang18
commented
Oct 7, 2025
| * The path to the temporary directory is available via the `storageRoot` variable. | ||
| */ | ||
| trait StorageRootMixin extends BeforeAndAfterEach { self: Suite => | ||
|
|
There was a problem hiding this comment.
extracting this out from PipelineTest so spark connect pipeline test can also extend this
JiaqiWang18
force-pushed
the
SPARK-53751-explicit-versioned-multiflow-checkpoint
branch
4 times, most recently
from
7ecc4dd to
22a4c49
Compare
|
@sryza diff is a bit large but tried to include only the necessary changes for checkpoints |
sryza
approved these changes
Oct 7, 2025
|
|
||
| /** | ||
| * Resets the checkpoint for the given flow by creating the next consecutive directory. Also | ||
| * clears out batch append state if it exists. |
There was a problem hiding this comment.
I don't think this clears out batch append state, right?
There was a problem hiding this comment.
yeah stale comment, let me remove
JiaqiWang18
force-pushed
the
SPARK-53751-explicit-versioned-multiflow-checkpoint
branch
from
9413fd8 to
dcad414
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
Add a
storagefield in pipeline spec to allow users specify locations of metadata such as streaming checkpoints.Below is the structure of the directory, which offer supports for multi-flow and versioned directory where version number is incremented after a full refresh.
Why are the changes needed?
Currently, SDP stores streaming flow ckpts in the table path. This does not allow support for versioned checkpoints and does not work for sinks.
Does this PR introduce any user-facing change?
How was this patch tested?
New and existing tests
Was this patch authored or co-authored using generative AI tooling?
No