juicedata/juicefs · error

get %s err: %w

Error message

get %s err: %w

What it means

Wrapped by dumpQuotas in redis_bak.go:329 when HGetAll on the quota type's usedInodesKey hash fails during a metadata dump. The Redis error is wrapped with the key name so the dump operation can report exactly which hash could not be read. Any Redis connectivity/cluster error on that hash triggers it, aborting the quota section of the dump.

Source

Thrown at pkg/meta/redis_bak.go:329

	}

	acls := make([]*pb.Acl, 0, len(vals))
	for k, v := range vals {
		id, _ := strconv.ParseUint(k, 10, 32)
		acls = append(acls, &pb.Acl{
			Id:   uint32(id),
			Data: []byte(v),
		})
	}
	return dumpResult(ctx, ch, &dumpedResult{msg: &pb.Batch{Acls: acls}})
}

func (m *redisMeta) dumpQuotas(ctx Context, quotas *[]*pb.Quota, qtype uint32) error {
	quotaMap := make(map[uint64]*pb.Quota)
	quotaKeys, _ := m.getQuotaKeys(qtype)
	vals, err := m.rdb.HGetAll(ctx, quotaKeys.usedInodesKey).Result()
	if err != nil {
		return fmt.Errorf("get %s err: %w", quotaKeys.usedInodesKey, err)
	}
	for k, v := range vals {
		id, err := strconv.ParseUint(k, 10, 64)
		if err != nil {
			logger.Warnf("parse used inodes %d: %s: %v", qtype, k, err)
			continue
		}
		usedInodes, err := strconv.ParseInt(v, 10, 64)
		if err != nil {
			logger.Warnf("invalid usedInodes for %d %s: %s", qtype, k, v)
			continue
		}

		usedSpace, err := m.rdb.HGet(ctx, quotaKeys.usedSpaceKey, k).Int64()
		if err != nil && err != redis.Nil {
			return err
		}

View on GitHub (pinned to c9a67b23e8)

Solutions

1. Read the wrapped Redis error: fix connectivity (address, password, network) and retry the dump.
2. If the target is a replica returning READONLY, dump from the primary.
3. For cluster errors (MOVED), ensure the client uses the cluster mode URL so slots are followed.
4. Increase timeouts or retry during load/traffic peaks; dump during a quieter window.

Example fix

// before: dump against replica
juicefs dump redis://replica:6379/1 out.json
// after: dump against primary
juicefs dump redis://primary:6379/1 out.json

// shell: verify Redis reachable and writable before dumping
redis-cli -h primary -p 6379 ping || exit 1
redis-cli -h primary -p 6379 info replication | grep role:master || echo "warning: replica target"

if err := dumpMeta(...); err != nil {
    if strings.Contains(err.Error(), "get ") && strings.Contains(err.Error(), "usedInodes") {
        // transient Redis failure: back off and retry the dump
    }
    return err
}

When it happens

Trigger: Running `juicefs dump` against a Redis metadata engine while HGetAll(<usedInodesKey>) returns an error — connection loss, MOVED/cluster misconfiguration, READONLY replica, or timeouts.

Common situations: Dumping while Redis is failing over; dump targeting a read-only replica; cluster client pointing at wrong slots; network interruption during a large dump.

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

• get dirDataLengthKey err: %w
• get dirUsedInodesKey err: %w
• get dirUsedSpaceKey err: %w
• load user/group quotas: %w
• unknown quota type: %d

AI-assisted analysis of juicedata/juicefs@c9a67b23e8 (2026-09-06). Data as JSON: /api/errors/e282136b2162246d. Report an issue: GitHub.