Index build of {} failed. Please run full index rebuild to f

Error message

Index build of {} failed. Please run full index rebuild to fix it.

What it means

logAndMarkIndexesFailed records that building the given indexes failed: it inspects the throwable for JVM-stopping decisions (e.g. OOM/WTD), logs this WARN including the exception, and marks every index as FAILED so it will not be used for queries. The message instructs the operator to run a full index rebuild because the incremental build left the index unusable.

Source

Thrown at src/java/org/apache/cassandra/index/SecondaryIndexManager.java:897

            if (DatabaseDescriptor.isDaemonInitialized())
                SystemKeyspace.setIndexRemoved(baseCfs.getKeyspaceName(), indexName);

            needsFullRebuild.add(indexName);

            if (!index.getSupportedLoadTypeOnFailure(isInitialBuild).supportsWrites() && writableIndexes.remove(indexName) != null)
                logger.info("Index [{}] became not-writable because of failed build.", indexName);

            if (!index.getSupportedLoadTypeOnFailure(isInitialBuild).supportsReads() && queryableIndexes.remove(indexName))
                logger.info("Index [{}] became not-queryable because of failed build.", indexName);
        }
    }

    private void logAndMarkIndexesFailed(Set<Index> indexes, Throwable indexBuildFailure, boolean isInitialBuild)
    {
        JVMStabilityInspector.inspectThrowable(indexBuildFailure);
        if (indexBuildFailure != null)
            logger.warn("Index build of {} failed. Please run full index rebuild to fix it.", getIndexNames(indexes), indexBuildFailure);
        else
            logger.warn("Index build of {} failed. Please run full index rebuild to fix it.", getIndexNames(indexes));
        indexes.forEach(i -> this.markIndexFailed(i, isInitialBuild));
    }

    /**
     * Marks the specified index as removed.
     *
     * @param indexName the index name
     */
    private synchronized void markIndexRemoved(String indexName)
    {
        SystemKeyspace.setIndexRemoved(baseCfs.getKeyspaceName(), indexName);
        queryableIndexes.remove(indexName);
        writableIndexes.remove(indexName);
        needsFullRebuild.remove(indexName);
        inProgressBuilds.remove(indexName);
        // remove existing indexing status

View on GitHub (pinned to 88fd0f6a0e)

Solutions

Read the attached exception in the log to identify the root cause.
Run `nodetool rebuild_index <keyspace> <table> <index>` (or ALTER TABLE ... DROP INDEX and re-CREATE) as the message suggests.
Check disk space and SSTable integrity if IOExceptions appear.
Query paths will return the index as failed/unavailable until the rebuild succeeds.

Example fix

// before
nodetool rebuild_index keyspace table index_name
// after (if repeated): drop and recreate cleanly
// DROP INDEX keyspace.index_name; CREATE INDEX ... ;

Defensive patterns

Strategy: try-catch

Try / catch

try { session.execute("CREATE INDEX IF NOT EXISTS ..."); } catch (Exception e) { // follow with rebuild
 session.execute(String.format("nodetool rebuild_index %s %s %s", ks, table, indexName)); }

Prevention

Read the attached exception before rebuilding to fix root causes (disk, corruption).
Run repair if SSTables were deleted under the build.
Schedule index builds during low-traffic periods.
Keep custom Index implementations well-tested for build-time exceptions.

When it happens

Trigger: Called from buildIndexesBlocking, buildIndex, or the asynchronous onFailure callback when an index build future fails with a non-null Throwable — e.g. IOException reading SSTables, custom Index.build throwing, or disk errors.

Common situations: Index build racing with SSTable deletion/compaction; malformed custom index implementations; disk full during index write; failed upgrades leaving incompatible index metadata on disk.

Understand the failure class

Background: Database query failed: Internal Server Error 500s wrapping SQL, Prisma, and connection failures — what to check first — this error's family across 16 libraries.

Related errors

AI-assisted analysis of apache/cassandra@88fd0f6a0e (2026-09-10). Data as JSON: /api/errors/b5be80a27850dd86. Report an issue: GitHub.