elastic/elasticsearch · error · IllegalArgumentException

Cannot create past TSDB backing index for replicated data st

Error message

Cannot create past TSDB backing index for replicated data stream [{}]

What it means

Thrown by validateDataStream when the data stream exists but isReplicated() returns true (a cross-cluster-replication follower data stream). Past TSDB backing index creation is a write operation that must run on the leader; follower streams are read-only replicas and cannot have new backing indices minted on the follower side.

Source

Thrown at modules/data-streams/src/main/java/org/elasticsearch/datastreams/action/TransportPastTimeSeriesIndexCreationAction.java:522

        public void init() {
            clusterService.getClusterSettings()
                .addSettingsUpdateConsumer(PAST_TSDB_INDEX_INTERVAL, tv -> indexIntervalMillis = tv.millis());
        }
    }

    record CoveredTimeWindow(long start, long end) {

        CoveredTimeWindow(Instant start, Instant end) {
            this(start.toEpochMilli(), end.toEpochMilli());
        }
    }

    private static void validateDataStream(String dataStreamName, DataStream dataStream, ProjectMetadata project) {
        if (dataStream == null) {
            throw new ResourceNotFoundException("Data stream [" + dataStreamName + "] not found");
        }
        if (dataStream.isReplicated()) {
            throw new IllegalArgumentException("Cannot create past TSDB backing index for replicated data stream [" + dataStreamName + "]");
        }
        if (dataStream.isSystem()) {
            throw new IllegalArgumentException("Cannot create past TSDB backing index for system data stream [" + dataStreamName + "]");
        }

        if (IndexMode.isTsdb(dataStream.getIndexMode()) == false) {
            throw new IllegalArgumentException(
                "Cannot create past TSDB backing index for data stream ["
                    + dataStreamName
                    + "] with mode ["
                    + dataStream.getIndexMode()
                    + "], it needs to be a time series data stream."
            );
        }
        Index writeIndex = dataStream.getWriteIndex();
        IndexMetadata writeIndexMetadata = writeIndex == null ? null : project.index(writeIndex);
        if (writeIndexMetadata == null || IndexMode.isTsdb(writeIndexMetadata.getIndexMode()) == false) {
            throw new IllegalStateException(

View on GitHub (pinned to db6a809a66)

Solutions

1. Run the past TSDB index creation request against the leader data stream, not the replicated follower.
2. Filter replicated streams out of any bulk/batch backfill job using the is_replicated flag from GET _data_stream.

if (dataStream.isReplicated()) {
  throw new IllegalArgumentException("Cannot create past TSDB backing index for replicated data stream");
}

When it happens

Trigger: Calling the past TSDB index creation API on a data stream that is a CCR follower. The check runs right after the not-found check and before the tsdb-mode check.

Common situations: Operating on the follower cluster instead of the leader; automation that targets all streams without filtering out replicated ones.

Related errors

• Cannot create past TSDB backing index for data stream [{}] b
• Data stream [{}] not found
• Cannot create past TSDB backing index for system data stream
• backing index [%s] in tsdb mode doesn't have the [%s] index
• Cannot add dynamic templates that define dimension fields on

AI-assisted analysis of elastic/elasticsearch@db6a809a66 (2026-08-12). Data as JSON: /api/errors/592e0b2dd697421b. Report an issue: GitHub.