Files
seaweedfs/weed/cluster/lock_manager/lock_ring.go
T
Chris Lu a3c0baa9b0 filer: cooling-off dual-read for POSIX locks during ring changes (#9672)
While the ring changed within the last snapshot interval, a fresh owner
asks the key's previous owner (LockRing.PriorOwner) whether it still
holds a conflicting lock before granting TRY_LOCK or answering GET_LK, so
it does not double-grant before re-assertion rebuilds its local state.
The probe is marked cooling_probe so the previous owner answers from
local state without recursing. PriorOwner uses the snapshot's prebuilt
ring rather than rebuilding a hash ring per call.
2026-05-25 12:34:15 -07:00

194 lines
5.5 KiB
Go

package lock_manager
import (
"sort"
"sync"
"time"
"github.com/seaweedfs/seaweedfs/weed/glog"
"github.com/seaweedfs/seaweedfs/weed/pb"
)
type LockRingSnapshot struct {
servers []pb.ServerAddress
ts time.Time
ring *HashRing // prebuilt ring for servers, so PriorOwner need not rebuild per call
}
type LockRing struct {
sync.RWMutex
snapshots []*LockRingSnapshot
lastCompactTime time.Time
snapshotInterval time.Duration
onTakeSnapshot func(snapshot []pb.ServerAddress)
cleanupWg sync.WaitGroup
Ring *HashRing // consistent hash ring
version int64 // monotonic version from master, rejects stale updates
}
func NewLockRing(snapshotInterval time.Duration) *LockRing {
return &LockRing{
snapshotInterval: snapshotInterval,
Ring: NewHashRing(DefaultVnodeCount),
}
}
func (r *LockRing) SetTakeSnapshotCallback(onTakeSnapshot func(snapshot []pb.ServerAddress)) {
r.Lock()
defer r.Unlock()
r.onTakeSnapshot = onTakeSnapshot
}
// SetSnapshot replaces the ring with a new server list from the master.
// The version must be >= the current version, otherwise the update is rejected
// (protects against reordered messages). Version 0 is always accepted (bootstrap).
func (r *LockRing) SetSnapshot(servers []pb.ServerAddress, version int64) bool {
sort.Slice(servers, func(i, j int) bool {
return servers[i] < servers[j]
})
r.Lock()
if version > 0 && version < r.version {
glog.V(0).Infof("LockRing: rejecting stale update v%d (current v%d)", version, r.version)
r.Unlock()
return false
}
r.version = version
// Update the ring while holding the lock so version and ring state
// are always consistent — prevents a concurrent SetSnapshot from
// seeing the new version but applying its servers to the old ring.
r.Ring.SetServers(servers)
// Append the snapshot under the same lock as the ring update so a concurrent
// PriorOwner always sees snapshots[0] matching r.Ring (and snapshots[1] as the
// true prior); otherwise it could pair a new ring with a stale prior snapshot.
r.addOneSnapshotLocked(servers)
r.Unlock()
r.cleanupWg.Add(1)
go func() {
defer r.cleanupWg.Done()
<-time.After(r.snapshotInterval)
r.compactSnapshots()
}()
return true
}
// Version returns the current ring version.
func (r *LockRing) Version() int64 {
r.RLock()
defer r.RUnlock()
return r.version
}
// addOneSnapshotLocked appends a new snapshot (newest at index 0). The caller
// must hold r.Lock(), so the ring update and snapshot append are one atomic step.
func (r *LockRing) addOneSnapshotLocked(servers []pb.ServerAddress) {
ts := time.Now()
ring := NewHashRing(DefaultVnodeCount)
ring.SetServers(servers)
t := &LockRingSnapshot{
servers: servers,
ts: ts,
ring: ring,
}
r.snapshots = append(r.snapshots, t)
for i := len(r.snapshots) - 2; i >= 0; i-- {
r.snapshots[i+1] = r.snapshots[i]
}
r.snapshots[0] = t
if r.onTakeSnapshot != nil {
r.onTakeSnapshot(t.servers)
}
}
func (r *LockRing) compactSnapshots() {
r.Lock()
defer r.Unlock()
ts := time.Now()
recentSnapshotIndex := 1
for ; recentSnapshotIndex < len(r.snapshots); recentSnapshotIndex++ {
if ts.Sub(r.snapshots[recentSnapshotIndex].ts) > r.snapshotInterval {
break
}
}
if recentSnapshotIndex+1 <= len(r.snapshots) {
r.snapshots = r.snapshots[:recentSnapshotIndex+1]
}
r.lastCompactTime = ts
}
func (r *LockRing) GetSnapshot() (servers []pb.ServerAddress) {
r.RLock()
defer r.RUnlock()
if len(r.snapshots) == 0 {
return
}
return r.snapshots[0].servers
}
// WaitForCleanup waits for all pending cleanup operations to complete
func (r *LockRing) WaitForCleanup() {
r.cleanupWg.Wait()
}
// GetSnapshotCount safely returns the number of snapshots for testing
func (r *LockRing) GetSnapshotCount() int {
r.RLock()
defer r.RUnlock()
return len(r.snapshots)
}
// GetPrimaryAndBackup returns the primary and backup servers for a key
// using the consistent hash ring.
func (r *LockRing) GetPrimaryAndBackup(key string) (primary, backup pb.ServerAddress) {
return r.Ring.GetPrimaryAndBackup(key)
}
// GetPrimary returns the primary server for a key using the consistent hash ring.
func (r *LockRing) GetPrimary(key string) pb.ServerAddress {
return r.Ring.GetPrimary(key)
}
// PriorOwner returns the key's owner from the previous ring snapshot, but only
// while the ring changed within the last snapshotInterval and that owner differs
// from the current primary. This is the cooling-off window in which the previous
// owner may still hold locks the new owner has not yet rebuilt — a caller can
// consult it before granting so a fresh owner does not double-grant during a
// rebalance. Returns "" outside the window or when ownership did not move. It
// uses the snapshot's prebuilt ring, so it does not rebuild a hash ring per call.
func (r *LockRing) PriorOwner(key string) pb.ServerAddress {
r.RLock()
defer r.RUnlock()
if len(r.snapshots) < 2 {
return ""
}
if time.Since(r.snapshots[0].ts) > r.snapshotInterval {
return ""
}
current := r.Ring.GetPrimary(key)
var prior pb.ServerAddress
if pr := r.snapshots[1].ring; pr != nil {
prior = pr.GetPrimary(key)
} else {
prior = hashKeyToServer(key, r.snapshots[1].servers)
}
if prior != "" && prior != current {
return prior
}
return ""
}
// hashKeyToServer uses a temporary consistent hash ring for the given server list.
func hashKeyToServer(key string, servers []pb.ServerAddress) pb.ServerAddress {
if len(servers) == 0 {
return ""
}
ring := NewHashRing(DefaultVnodeCount)
ring.SetServers(servers)
return ring.GetPrimary(key)
}