What causes "database redis://%s is not empty" in juicedata/juicefs?

Use a dedicated empty Redis DB: pass a distinct DB index (redis://host:6379/1) or a fresh instance.

database redis://%s is not empty

Error message

database redis://%s is not empty

What it means

prepareLoad guards `juicefs load` against overwriting live data: for standalone Redis/Sentinel it checks DBSize and refuses to proceed if the database holds any keys. Loading into a non-empty Redis would corrupt or interleave metadata.

Source

Thrown at pkg/meta/redis_bak.go:995

	return execPipe(ctx, pipe)
}

func (m *redisMeta) prepareLoad(ctx Context, opt *LoadOption) error {
	opt.check()
	if _, ok := m.rdb.(*redis.ClusterClient); ok {
		err := m.scan(ctx, "*", func(keys []string) error {
			return fmt.Errorf("found key with same prefix: %s", keys[0])
		})
		if err != nil {
			return err
		}
	} else {
		dbsize, err := m.rdb.DBSize(ctx).Result()
		if err != nil {
			return err
		}
		if dbsize > 0 {
			return fmt.Errorf("database redis://%s is not empty", m.addr)
		}
	}
	return nil
}

View on GitHub (pinned to c9a67b23e8)

Solutions

Use a dedicated empty Redis DB: pass a distinct DB index (redis://host:6379/1) or a fresh instance.
Flush the target: redis-cli -n <db> FLUSHDB (confirm nothing else uses it first).
If the leftover keys are from a failed prior load, delete only that volume's prefix keys.
Verify the --meta URL points at the intended host/port/db before loading.

Example fix

// before
juicefs load -i backup.json redis://10.0.0.5:6379/0   # DB 0 has keys
// after
redis-cli -h 10.0.0.5 -n 0 DBSIZE   # verify, then flush if safe
redis-cli -h 10.0.0.5 -n 0 FLUSHDB
juicefs load -i backup.json redis://10.0.0.5:6379/0

Defensive patterns

Strategy: validation

Validate before calling

n, err := rdb.DBSize(ctx).Result()
if err != nil { return err }
if n > 0 { return fmt.Errorf("abort: target DB has %d keys", n) }

Try / catch

if err := load(ctx, meta, f); err != nil {
    if strings.Contains(err.Error(), "is not empty") {
        log.Fatalf("refusing to load into non-empty DB; flush or choose another DB index: %v", err)
    }
    return err
}

Prevention

Dedicate a Redis DB index (or instance) to each restore.
Verify DBSize == 0 before every load.
Do not share a Redis instance between JuiceFS volumes without distinct DB indexes/prefixes.
Confirm the meta URL (host, port, db index) before running load.

When it happens

Trigger: Running `juicefs load -i dump.json redis://host:port/vol` when the target DB already contains keys — a previous volume, another JuiceFS instance, or leftover keys from an aborted load.

Common situations: Restoring a backup over an existing volume by mistake; reusing a Redis instance shared with other applications; DB index selection (e.g. redis://host:6379/0 vs /1) pointing at an occupied database.

Understand the failure class

Background: "already exists" / EEXIST / FileAlreadyExistsException: what the 'file already exists' error means and how to fix it — this error's family across 37 libraries.

Related errors

AI-assisted analysis of juicedata/juicefs@c9a67b23e8 (2026-09-06). Data as JSON: /api/errors/01f61564490c154d. Report an issue: GitHub.