Versions Compared

Old Version 4

changes.mady.by.user Manan Gupta

Saved on

compared with

New Version Current

changes.mady.by.user Manan Gupta

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Discussion thread: here 

JIRA: here [Change the link from KAFKA-1 to your own ticket]

Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).

...

Today, kafka-reassign-partitions.sh --execute submits all partition reassignments from the JSON in oneAdminClient#alterPartitionReassignments call (unless the user manually splits work). For large clusters this can create large bursts of replication traffic and controller work.

This KIP proposes tool-only pacing controls:

  1. --reassignment-batch-size — caps how many topic partitions are submitted per step (semantics differ slightly depending on mode; see below). 0 preserves legacy behavior (single request for the entire plan).
  2. --incremental — optional mode used with--reassignment-batch-size > 0 to keep at most Npartition reassignments in flight for this execution, submitting the next partition from a deterministic queue when a slot frees up.
  3. --reassignment-poll-interval-ms — how long the tool sleeps between polls while waiting for batch completion (non-incremental, between batches) and while driving incremental submissions; default 500 ms.

Ordering for both modes is (topic name, partition index), not the order of entries in the JSON file.

No broker protocol, controller, or metadata changes are required; pacing is implemented entirely in the reassignment tool using existing Admin APIs (alterPartitionReassignments, listPartitionReassignments, metadata reads).

...

  1. Operational risk: Submitting hundreds or thousands of partition reassignments in one RPC can stress replication links, disk, and the controller in ways that are hard to predict during maintenance windows.
  2. Limited operator control: Throttling (--throttle) limits bandwidth but does not limit how many partitions are simultaneously moving; operators often want serialised waves or bounded concurrency without hand-splitting JSON files.
  3. Manual workarounds: Teams split reassignment JSON by hand or wrap the tool in scripts that call alterPartitionReassignments in chunks — error-prone and inconsistent across deployments.

...

CLI (kafka-reassign-partitions.sh)

Option

Applies to

Semantics

--reassignment-batch-size <int>

--execute only

  • Default 0: legacy — entire plan in one alterPartitionReassignments request; no extra waits in the tool.
  • > 0 without --incremental: split the plan into contiguous batches of at most N partitions (sorted by topic, then partition id). After each batch except the last, the tool blocks until every partition in that batch reports complete (current ISR matches target and reassignment not active) before submitting the next batch.
  • > 0 with --incremental: N is the maximum number of partition reassignments from this JSON that may be active at once; when one completes, the tool submits the next partition from the sorted queue.

--incremental

--execute only

Requires --reassignment-batch-size > 0. Mutually exclusive interpretation of batch size as in-flight cap (see above).

--reassignment-poll-interval-ms <long>

--execute only

Milliseconds to sleep between progress polls when the tool is waiting on reassignment state. Default 500. 

Validation and compatibility

  • --reassignment-batch-size must be ≥ 0. Negative values are rejected.
  • --incremental without --reassignment-batch-size > 0 is rejected at argument validation time.
  • --reassignment-batch-size, --incremental and --incrementalreassignment-poll-interval-ms are not permitted with --list, --generate, --verify, or --cancel (same pattern as other execute-only options).
  • --reassignment-poll-interval-ms must be > 0. Negative values are rejected.

Programmatic API

ReassignPartitionsCommand.executeAssignment gains parameters:

  • int reassignmentBatchSize
  • boolean incremental
  • long reassignmentPollIntervalMs

Existing callers that omit pacing pass 0 and false to preserve legacy behaviour.

...

  1. Sort partitions deterministically (compareTopicPartitions).
  2. Maintain a pending deque and an in-flight map (submissions for this execute only).
  3. Loop until pending is empty: remove completed partitions from in-flight (using the same completion predicate as non-incremental wait paths), then submit new partitions up to the N in-flight cap.
  4. Poll interval between iterations when work remains: INCREMENTAL_REASSIGNMENT_POLL_INTERVAL_MS (500 ms in the default implementation).

...

Non-incremental wait between batches

  • Poll interval: BATCH_REASSIGNMENT_POLL_INTERVAL_MSreassignmentPollIntervalMs (500 ms in the reference implementation).
  • Completion uses existing findPartitionReassignmentStates / PartitionReassignmentState logic; inconsistent terminal states produce TerseException (same class of errors as today’s verify path).

...

  • Default CLI: unchanged legacy path (reassignment-batch-size defaults to 0; incremental absent).
  • Brokers / ZK / KRaft: no change.
  • Wire protocol: unchanged (same Admin APIs).

...

  1. No timeout on batch-completion waits in the reference implementation; stuck reassignments can poll indefinitely until operator intervention (same class of risk as long-running admin operations without deadlines).
  2. Incremental + --additional: total cluster in-flight work can exceed N when multiple tool processes run overlapping plans; documentation / optional warnings are recommended.

Rejected alternatives

  1. Only documentation: “Split your JSON manually” — does not scale and yields inconsistent operations.
  2. New broker-side “max concurrent reassignments” quota: much larger scope; tool-side pacing addresses the common case without protocol work.
  3. Order = JSON file order only: rejected in favour of deterministic sorted order so behaviour is reproducible and independent of file editing.

Test plan

  • Unit tests: batch splitting; incremental ordering; failure on second batch; completion predicate shared between wait and incremental removal.
  • Args tests: invalid combinations; execute-only restriction for new flags.
  • Integration / cluster tests: execute with small batch sizes (including batch size 1) in KRaft and ZK modes where the project already runs ClusterTest.
  • Manual: large plan with --reassignment-batch-size and --list / --verify to observe waves and completion.

Documentation

  • Extend the kafka-reassign-partitions section of the Kafka documentation / ops guides: semantics of 0, non-incremental batching, incremental, --additional, and verification with --verify.