You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 42 Next »

A producer ID (PID) is a 64-bit identifier assigned by the broker to each idempotent or transactional producer. It has three key properties:

  1. Unique: uniquely identifies a producer for idempotent deduplication and transaction tracking.
  2. Stable: once assigned, a PID persists across producer sessions (for transactional producers) or until expiration.
  3. Non-negative: valid PIDs are >= 0. The value -1 (NO_PRODUCER_ID) marks non-idempotent batches.

Without PID mapping, two independent clusters can assign the same producer ID to different producers. When records from both source clusters are mirrored into the same destination partition, the ProducerStateManager (PSM) sees two unrelated producers sharing one PID.

Current Approach: Stateless Transformation

The following PID transformation is applied before appending mirrored data to the destination cluster:

-(PID + 2)

This is a simple negation that maps all non-negative PIDs into the negative space. The +2 offset avoids mapping PID 0 to 0 and keeps PID -1 (non-idempotent) untouched. 

There are a couple of problems with this approach that are evident when looking at the chained mirroring use case.

Non-idempotent transformation

When B mirrors to C, PIDs already negative from A get re-transformed: -((-7) + 2) = 5 , which restores the original PID and collides with local producers on C.

 A           B           C           D
-1 -------> -1 -------> -1 -------> -1
 5 -------> -7 --------> 5 -------> -7              
                         5 -------> -7 # collision

PID collision with local producers

Even if we make the mapping idempotent by skipping negative PIDs, when A has local PID 5 and B also has local PID 5, both map to -7 on any downstream cluster. These are different producers, but they become indistinguishable. The PSM cache stores the transformed PID with no awareness of its origin, so a collision silently overwrites the previous entry breaking txn consistency within the log.

A           B           C
5 -------> -7 -------> -7
            5 -------> -7 # collision

We identified the following scenarios caused by interleaving records from a PID collision:

  1. Same epoch, wrong sequence: No OutOfOrderSequenceException. Batches are silently accepted under the same PID as they are coming from the leader (append origin == REPLICATION). The PSM cache entry is updated with whatever arrives last. Silent data corruption with zero signals, not even a warning.
  2. Different epochs: No fencing exception. Lower epoch batch is accepted with a warning log. Both producers coexist under the same PID. Silent corruption, only a WARN log line as a hint.
  3. Transactional interleaving: Commit/abort markers from one producer close the other's transaction. No exception. Silent transaction corruption.

New Approach: Barrier Control Batch

Rather than transforming PIDs at write time, this approach proactively expires stale producer state on failover. The key insight is that during mirroring, the destination partition is read-only: no local producers exist, so all PSM entries originate from mirrored data. When mirroring stops, all PSM entries are stale and can be safely expired. Records from the source are stored as-is on the destination, with no PID modification, which otherwise would require a checksum recalculation. A MIRROR_PID_RESET control batch (type 7) is written to each destination partition's log during the STOPPING state transition, after the fetcher has been removed and truncation to last stable offset is complete, but before the partition becomes writable.

The mirror partition state transitions are:

  • STOPPING: remove fetchers, truncate to LSO, persist last mirrored offsets, write MIRROR_PID_RESET barrier
  • STOPPED:  partition is writable (terminal state, no actions)

The key follows the standard control record format (version=0, type=7). The value uses the MirrorPidResetRecord schema:

{
  "type": "data",
  "name": "MirrorPidResetRecord",
  "validVersions": "0",
  "flexibleVersions": "0+",
  "fields": [
    { "name": "Version", "type": "int16", "versions": "0",
      "about": "The version of the mirror PID reset record."},
    { "name": "SourceClusterId", "type": "string", "versions": "0",
      "about": "The source cluster UUID for verification."}
  ]
}

The SourceClusterId field records which source cluster the mirrored data came from, enabling future validation (e.g. detecting unexpected source cluster changes) and data provenance tracing from the log itself.

When the barrier batch is encountered during append or during log recovery, all producer entries are removed from the PSM. This ensures leaders, followers, and recovery all handle the barrier consistently. Because the partition is read-only during mirroring, all PSM entries originate from mirrored data. Expiring all entries is safe: no local producer state exists to preserve. Control batches are filtered out by the consumer fetcher via isControlBatch checks. The barrier is invisible to application consumers, just like transaction markers (commit/abort). The log dump tool is enhanced to deserialize and display MIRROR_PID_RESET records.

Supported Topologies

The barrier approach works correctly with all practical mirroring topologies:                                                                                                                                                                                                            

  • Active-passive (A to B): B mirrors from A, stores records as-is. On failover, barrier expires all PSM entries. Local producers get fresh PIDs from the coordinator with no collision risk.                                                                                            

  • Failback (A to B, then B to A): After failover, B becomes writable. Later, A starts mirroring from B, stores B's records as─is. B's barrier record is included in the fetched data and appended to A's log, but it does not trigger PSM expiration on A. This is because A is actively mirroring, not transitioning. Its PSM state reflects the mirrored data it is currently tracking, and wiping it mid─stream would break producer state consistency for the ongoing mirroring operation. When A later stops mirroring from B, A writes its own barrier, which is the correct trigger for PSM expiration on A, producing a clean slate before A becomes writable again. The cycle can repeat safely in either direction because each failover produces clean PSM state via the barrier, and each new mirroring session starts from a truncated, offset─aligned log.

  • Fan-out (A to B, A to C): B and C mirror independently from A, each with its own PSM per partition. On failover, each writes its own barrier independently.
  • Fan-in (A to C, B to C, different topics): Each topic's partitions have independent PSMs. The barrier is written per partition during the STOPPING transition of each mirror.
  • Chain (A to B to C): B mirrors from A, stores records as-is. C mirrors from B, stores records as-is. On failover at any point in the chain, the barrier expires all PSM entries on the stopping node. Longer chains work inductively by the same principle.


  • No labels