package ledis import ( "fmt" "github.com/siddontang/go/filelock" "github.com/siddontang/go/log" "github.com/siddontang/ledisdb/config" "github.com/siddontang/ledisdb/rpl" "github.com/siddontang/ledisdb/store" "io" "os" "path" "sync" "time" ) type Ledis struct { cfg *config.Config ldb *store.DB dbLock sync.Mutex dbs map[int]*DB quit chan struct{} wg sync.WaitGroup //for replication r *rpl.Replication rc chan struct{} rbatch *store.WriteBatch rwg sync.WaitGroup rhs []NewLogEventHandler wLock sync.RWMutex //allow one write at same time commitLock sync.Mutex //allow one write commit at same time lock io.Closer ttlCheckers []*ttlChecker ttlCheckerCh chan *ttlChecker } func Open(cfg *config.Config) (*Ledis, error) { if len(cfg.DataDir) == 0 { cfg.DataDir = config.DefaultDataDir } if cfg.Databases == 0 { cfg.Databases = 16 } else if cfg.Databases > MaxDatabases { cfg.Databases = MaxDatabases } os.MkdirAll(cfg.DataDir, 0755) var err error l := new(Ledis) l.cfg = cfg if l.lock, err = filelock.Lock(path.Join(cfg.DataDir, "LOCK")); err != nil { return nil, err } l.quit = make(chan struct{}) if l.ldb, err = store.Open(cfg); err != nil { return nil, err } if cfg.UseReplication { if l.r, err = rpl.NewReplication(cfg); err != nil { return nil, err } l.rc = make(chan struct{}, 1) l.rbatch = l.ldb.NewWriteBatch() l.wg.Add(1) go l.onReplication() //first we must try wait all replication ok //maybe some logs are not committed l.WaitReplication() } else { l.r = nil } l.dbs = make(map[int]*DB, 16) l.checkTTL() return l, nil } func (l *Ledis) Close() { close(l.quit) l.wg.Wait() l.ldb.Close() if l.r != nil { l.r.Close() //l.r = nil } if l.lock != nil { l.lock.Close() //l.lock = nil } } func (l *Ledis) Select(index int) (*DB, error) { if index < 0 || index >= l.cfg.Databases { return nil, fmt.Errorf("invalid db index %d, must in [0, %d]", index, l.cfg.Databases-1) } l.dbLock.Lock() defer l.dbLock.Unlock() db, ok := l.dbs[index] if ok { return db, nil } db = l.newDB(index) l.dbs[index] = db go func(db *DB) { l.ttlCheckerCh <- db.ttlChecker }(db) return db, nil } // Flush All will clear all data and replication logs func (l *Ledis) FlushAll() error { l.wLock.Lock() defer l.wLock.Unlock() return l.flushAll() } func (l *Ledis) flushAll() error { it := l.ldb.NewIterator() defer it.Close() it.SeekToFirst() w := l.ldb.NewWriteBatch() defer w.Rollback() n := 0 for ; it.Valid(); it.Next() { n++ if n == 10000 { if err := w.Commit(); err != nil { log.Fatalf("flush all commit error: %s", err.Error()) return err } n = 0 } w.Delete(it.RawKey()) } if err := w.Commit(); err != nil { log.Fatalf("flush all commit error: %s", err.Error()) return err } if l.r != nil { if err := l.r.Clear(); err != nil { log.Fatalf("flush all replication clear error: %s", err.Error()) return err } } return nil } func (l *Ledis) IsReadOnly() bool { if l.cfg.GetReadonly() { return true } else if l.r != nil { if b, _ := l.r.CommitIDBehind(); b { return true } } return false } func (l *Ledis) checkTTL() { l.ttlCheckers = make([]*ttlChecker, 0, 16) l.ttlCheckerCh = make(chan *ttlChecker, 16) if l.cfg.TTLCheckInterval == 0 { l.cfg.TTLCheckInterval = 1 } l.wg.Add(1) go func() { defer l.wg.Done() tick := time.NewTicker(time.Duration(l.cfg.TTLCheckInterval) * time.Second) defer tick.Stop() for { select { case <-tick.C: if l.IsReadOnly() { break } for _, c := range l.ttlCheckers { c.check() } case c := <-l.ttlCheckerCh: l.ttlCheckers = append(l.ttlCheckers, c) c.check() case <-l.quit: return } } }() } func (l *Ledis) StoreStat() *store.Stat { return l.ldb.Stat() } func (l *Ledis) CompactStore() error { l.wLock.Lock() defer l.wLock.Unlock() return l.ldb.Compact() }