jaegertracing/jaeger · error

failed to Exec query '%s': %w

Error message

failed to Exec query '%s': %w

What it means

Table.Exec wraps a failed Cassandra CQL write (query.Exec()) with the query text and the underlying driver error. It exists so callers get metrics emitted plus a single error that names exactly which statement failed. The %w wrapping preserves the original driver error for errors.Is/As inspection.

Source

Thrown at internal/storage/cassandra/metrics/table.go:40

// NewTable takes a metrics scope and creates a table metrics struct
func NewTable(factory metrics.Factory, tableName string) *Table {
	t := spanstoremetrics.WriteMetrics{}
	metrics.Init(&t, factory.Namespace(metrics.NSOptions{Name: "", Tags: map[string]string{"table": tableName}}), nil)
	return &Table{t}
}

// Exec executes an update query and reports metrics/logs about it.
func (t *Table) Exec(query cassandra.UpdateQuery, logger *zap.Logger) error {
	start := time.Now()
	err := query.Exec()
	t.Emit(err, time.Since(start))
	if err != nil {
		queryString := query.String()
		if logger != nil {
			logger.Error("Failed to exec query", zap.String("query", queryString), zap.Error(err))
		}
		return fmt.Errorf("failed to Exec query '%s': %w", queryString, err)
	}
	return nil
}

View on GitHub (pinned to 806f444784)

Solutions

1. Read the wrapped driver error after 'failed to Exec query ...' to get the root cause (use %v of the full chain or errors.Unwrap).
2. Verify Cassandra connectivity and that the target keyspace/table exist (run jaeger's init/schema job).
3. Check the logged 'Failed to exec query' zap entry for the exact CQL statement and fix the statement or data.
4. Inspect Cassandra server logs / nodetool status for cluster health if errors are transient.
5. Upgrade driver/retry policies if timeouts occur under load; add retries with backoff around Exec.

Example fix

// before
err := table.Exec(query, logger)
if err != nil {
    return err // opaque
}
// after
err := table.Exec(query, logger)
if err != nil {
    var cerr *gocql.Error
    if errors.As(err, &cerr) {
        logger.Warn("cassandra write failed", zap.Int32("code", cerr.Code), zap.Error(cerr))
    }
    return fmt.Errorf("writing spans: %w", err)
}

// precondition: verify Cassandra is reachable before writes
if err := session.Ping(); err != nil {
    return fmt.Errorf("cassandra unavailable before write: %w", err)
}

if err := table.Exec(query, logger); err != nil {
    var driverErr error
    if unwrapped := errors.Unwrap(err); unwrapped != nil {
        driverErr = unwrapped
    }
    log.Printf("cassandra write failed: %v (driver: %v)", err, driverErr)
    // retry transient errors only
    return retryTransient(driverErr)
}

When it happens

Trigger: Any Insert/Update/Upsert statement executed through metrics.Table.Exec fails at the Cassandra driver level: connection loss, cluster down, timeout, query syntax error, schema/table missing, or write consistency issues.

Common situations: Cassandra node restart or network partition during a span/dependency write; keyspace or table dropped; invalid CQL after a schema migration; Cassandra unreachable at startup; tombstone/timeout errors under heavy write load.

Related errors

• error reading throughput from storage: %w
• error reading probabilities from storage: %w
• invalid version
• failed to acquire resource lock due to cassandra error: %w
• unknown column for position: %q

AI-assisted analysis of jaegertracing/jaeger@806f444784 (2026-09-01). Data as JSON: /api/errors/594ebdea05402161. Report an issue: GitHub.