[SPARK-54583][SS] Add SQLConf to enable use of OffsetMap #53365
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
sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala
Outdated
Show resolved
Hide resolved
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/checkpointing/OffsetSeq.scala
Outdated
Show resolved
Hide resolved
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/checkpointing/OffsetSeq.scala
Show resolved
Hide resolved
anishshri-db
approved these changes
Dec 8, 2025
Contributor
anishshri-db
left a comment
There was a problem hiding this comment.
lgtm pending green CI
liviazhu
reviewed
Dec 8, 2025
Contributor
There was a problem hiding this comment.
Can you be more clear about the mappings between OffsetSeqMetadata and OffsetSeq/OffsetMap? Can any version of OffsetSeqMetadata be used with OffsetSeq? Can any version of metadata be used with OffsetMap? Looking at the PR description the answer is yet, but there's only 1 conf for both. How do we control the metadata version, do we need to introduce another conf?
| object OffsetSeqLog { | ||
| private[streaming] val VERSION_1 = 1 | ||
| private[streaming] val VERSION_2 = 2 | ||
| val VERSION_1 = 1 |
| val metadata = v1.metadataOpt.get.asInstanceOf[OffsetSeqMetadata] | ||
| v1.copy(metadataOpt = Some(metadata.copy( | ||
| conf = metadata.conf + (SQLConf.SHUFFLE_PARTITIONS.key -> numPartitions.toString)))) | ||
| case _ => throw OfflineStateRepartitionErrors.unsupportedOffsetSeqVersionError( |
Contributor
There was a problem hiding this comment.
Do we need to create a TODO + add ticket to support OffsetMap here?
| */ | ||
| case class OffsetSeqV2( | ||
| offsets: Seq[Option[OffsetV2]], | ||
| metadataOpt: Option[OffsetSeqMetadataV2] = None) extends OffsetSeqBase |
Contributor
There was a problem hiding this comment.
Should we require metadata for OffsetSeqV2 (don't make it an option)?
| case class OffsetMap( | ||
| offsetsMap: Map[String, Option[OffsetV2]], | ||
| metadataOpt: Option[OffsetSeqMetadata] = None) extends OffsetSeqBase { | ||
| metadataOpt: Option[OffsetSeqMetadataBase] = None) extends OffsetSeqBase { |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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?
This PR decouples the versioning of OffsetSeqMetadata from OffsetLog, allowing them to evolve independently while maintaining backward compatibility
with existing streaming checkpoints.
Key changes:
- sourceMetadataInfo: Map of source metadata keyed by sourceId
- controlBatchInfo: Information about control batches (e.g., rewind batches)
Why are the changes needed?
Previously, OffsetSeqMetadata version was tightly coupled to the OffsetLog version. This meant that any change to metadata format would require
bumping the entire offset log version, making it difficult to evolve metadata independently.
This change enables:
Does this PR introduce any user-facing change?
No. The changes are backward compatible:
How was this patch tested?
Was this patch authored or co-authored using generative AI tooling?
No