risingwavelabs/risingwave · error · HummockError

bounded compaction is not supported for copy-on-write tasks

Error message

bounded compaction is not supported for copy-on-write tasks

What it means

create_task_execution rejects bounded iceberg compaction (a request carrying max_file_sequence_number) when the resolved compaction kind is copy-on-write, or when the task is pk-index coordinated. Bounded compaction relies on append/sequence-number semantics that COW publishes cannot honor, since each COW plan rewrites and publishes whole table state.

Source

Thrown at src/storage/src/hummock/compactor/iceberg_compaction/iceberg_compactor_runner.rs:946

    let iceberg_config = IcebergConfig::from_btreemap(BTreeMap::from_iter(props))
        .map_err(|e| HummockError::compaction_executor(e.as_report()))?;

    let catalog = iceberg_config
        .create_catalog()
        .await
        .map_err(|e| HummockError::compaction_executor(e.as_report()))?;

    let table_ident = iceberg_config
        .full_table_name()
        .map_err(|e| HummockError::compaction_executor(e.as_report()))?;

    let parsed_task_type = TaskType::try_from(task_type)
        .map_err(|e| HummockError::compaction_executor(e.as_report()))?;
    let compaction_kind = IcebergCompactionKind::resolve(parsed_task_type, &iceberg_config)?;
    if max_file_sequence_number.is_some()
        && (pk_index_coordinated || compaction_kind.is_copy_on_write())
    {
        return Err(HummockError::compaction_executor(anyhow::anyhow!(
            "bounded compaction is not supported for copy-on-write tasks"
        )));
    }

    let branch = commit_branch(iceberg_config.r#type.as_str(), iceberg_config.write_mode);

    let table = catalog
        .load_table(&table_ident)
        .await
        .map_err(|e| HummockError::compaction_executor(e.as_report()))?;

    if let Some(boundary) = max_file_sequence_number {
        // An empty bounded plan is reported as `Drained`, so fail closed unless
        // the loaded branch can prove that this fixed boundary is meaningful.
        if table.metadata().format_version() < FormatVersion::V2 {
            return Err(HummockError::compaction_executor(anyhow::anyhow!(
                "bounded compaction requires Iceberg format V2 or V3"
            )));

View on GitHub (pinned to 6469eb736d)

Solutions

  1. Remove max_file_sequence_number from the compaction request so it runs as an unbounded task.
  2. Switch the iceberg table write_mode from copy-on-write to merge-on-read if bounded (incremental) compaction is required.
  3. Re-issue the request without a sequence boundary when using pk-index coordinated compaction.
  4. Align the client/tooling generating the request with the table's current write mode.

Example fix

// before
// request with max_file_sequence_number = 42 against write_mode = "copy-on-write"

// after: either drop the boundary
// max_file_sequence_number = null
// or change table config
[iceberg_compaction]
write_mode = "merge-on-read"
Defensive patterns

Strategy: validation

Validate before calling

fn is_bounded(req: &CompactionRequest) -> bool { req.max_file_sequence_number.is_some() }
// reject before submitting:
if is_bounded(req) && write_mode == WriteMode::CopyOnWrite { return Err("bounded compaction unsupported for COW"); }

Prevention

When it happens

Trigger: start_iceberg_compactor receives a task with max_file_sequence_number set while the resolved IcebergCompactionKind is CopyOnWrite/CopyOnWriteAuto (COW write mode), or pk_index_coordinated is true with a sequence-number boundary.

Common situations: A user or tooling submits a bounded/incremental compaction request against an iceberg sink configured with write_mode = copy-on-write; meta sends a sequence-number-bounded task to a COW table after a config switch mid-flight.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors


AI-assisted analysis of risingwavelabs/risingwave@6469eb736d (2026-09-11). Data as JSON: /api/errors/63b08d18daf8185e. Report an issue: GitHub.