value must be either a float64 (interpreted as seconds) or a

Error message

value must be either a float64 (interpreted as seconds) or a valid time.Duration

What it means

Cluster.GetSrvVSchema acquires topoReadPool before reading the per-cell SrvVSchema from vtctld. If the pool acquisition fails (context cancelled or timed out waiting for a slot), the call returns this wrapped error instead of querying.

Source

Thrown at go/flagutil/float_to_duration.go:55

}

// Set parses the input and updates the duration
func (f *FloatOrDuration) Set(value string) error {
	// Try to parse as float64 first (interpreted as seconds)
	if floatVal, err := strconv.ParseFloat(value, 64); err == nil {
		f.Duration = time.Duration(floatVal * float64(time.Second))
		*f.Target = f.Duration // Update the target pointer
		return nil
	}

	// Try to parse as time.Duration
	if duration, err := time.ParseDuration(value); err == nil {
		f.Duration = duration
		*f.Target = f.Duration // Update the target pointer
		return nil
	}

	return errors.New("value must be either a float64 (interpreted as seconds) or a valid time.Duration")
}

// Type returns the type description
func (f *FloatOrDuration) Type() string {
	return "time.Duration"
}

// FloatDuration defines a flag with the specified name, default value, and usage string and binds it to a time.Duration variable.
func FloatDuration(fs *pflag.FlagSet, p *time.Duration, name string, defaultValue time.Duration, usage string) {
	*p = defaultValue
	fd := FloatOrDuration{
		Target:   p,
		Duration: defaultValue,
	}
	fs.Var(&fd, name, usage)
}

View on GitHub (pinned to 01a25a7d17)

Solutions

Retry the request
Raise the topo read pool size or lower concurrency of callers
Investigate topo server latency that keeps pool slots occupied

Defensive patterns

Strategy: retry

Try / catch

sv, err := cluster.GetSrvVSchema(ctx, cell)
if err != nil && errors.Is(err, context.DeadlineExceeded) {
	// pool contention; retry with backoff or larger deadline
	sv, err = cluster.GetSrvVSchema(ctx2, cell)
}

Prevention

Avoid bursting many concurrent per-cell topo reads
Increase topoReadPool capacity in busy deployments
Watch topo backend latency

When it happens

Trigger: Calling Cluster.GetSrvVSchema for a cell while topoReadPool is saturated and ctx expires during Acquire.

Common situations: Many concurrent vschema/topo reads; slow etcd/zookeeper causing long pool waits; aggressive client timeouts.

Related errors

AI-assisted analysis of vitessio/vitess@01a25a7d17 (2026-09-01). Data as JSON: /api/errors/c0f788fa4c3b6378. Report an issue: GitHub.