khenaidoo | ab1f7bd | 2019-11-14 14:00:27 -0500 | [diff] [blame] | 1 | // Copyright 2017 The etcd Authors |
| 2 | // |
| 3 | // Licensed under the Apache License, Version 2.0 (the "License"); |
| 4 | // you may not use this file except in compliance with the License. |
| 5 | // You may obtain a copy of the License at |
| 6 | // |
| 7 | // http://www.apache.org/licenses/LICENSE-2.0 |
| 8 | // |
| 9 | // Unless required by applicable law or agreed to in writing, software |
| 10 | // distributed under the License is distributed on an "AS IS" BASIS, |
| 11 | // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 12 | // See the License for the specific language governing permissions and |
| 13 | // limitations under the License. |
| 14 | |
| 15 | package etcdserver |
| 16 | |
| 17 | import ( |
| 18 | "fmt" |
| 19 | "os" |
| 20 | "time" |
| 21 | |
| 22 | "go.etcd.io/etcd/etcdserver/api/snap" |
| 23 | "go.etcd.io/etcd/lease" |
| 24 | "go.etcd.io/etcd/mvcc" |
| 25 | "go.etcd.io/etcd/mvcc/backend" |
| 26 | "go.etcd.io/etcd/raft/raftpb" |
| 27 | |
| 28 | "go.uber.org/zap" |
| 29 | ) |
| 30 | |
| 31 | func newBackend(cfg ServerConfig) backend.Backend { |
| 32 | bcfg := backend.DefaultBackendConfig() |
| 33 | bcfg.Path = cfg.backendPath() |
| 34 | if cfg.BackendBatchLimit != 0 { |
| 35 | bcfg.BatchLimit = cfg.BackendBatchLimit |
| 36 | if cfg.Logger != nil { |
| 37 | cfg.Logger.Info("setting backend batch limit", zap.Int("batch limit", cfg.BackendBatchLimit)) |
| 38 | } |
| 39 | } |
| 40 | if cfg.BackendBatchInterval != 0 { |
| 41 | bcfg.BatchInterval = cfg.BackendBatchInterval |
| 42 | if cfg.Logger != nil { |
| 43 | cfg.Logger.Info("setting backend batch interval", zap.Duration("batch interval", cfg.BackendBatchInterval)) |
| 44 | } |
| 45 | } |
| 46 | bcfg.BackendFreelistType = cfg.BackendFreelistType |
| 47 | bcfg.Logger = cfg.Logger |
| 48 | if cfg.QuotaBackendBytes > 0 && cfg.QuotaBackendBytes != DefaultQuotaBytes { |
| 49 | // permit 10% excess over quota for disarm |
| 50 | bcfg.MmapSize = uint64(cfg.QuotaBackendBytes + cfg.QuotaBackendBytes/10) |
| 51 | } |
| 52 | return backend.New(bcfg) |
| 53 | } |
| 54 | |
| 55 | // openSnapshotBackend renames a snapshot db to the current etcd db and opens it. |
| 56 | func openSnapshotBackend(cfg ServerConfig, ss *snap.Snapshotter, snapshot raftpb.Snapshot) (backend.Backend, error) { |
| 57 | snapPath, err := ss.DBFilePath(snapshot.Metadata.Index) |
| 58 | if err != nil { |
| 59 | return nil, fmt.Errorf("failed to find database snapshot file (%v)", err) |
| 60 | } |
| 61 | if err := os.Rename(snapPath, cfg.backendPath()); err != nil { |
| 62 | return nil, fmt.Errorf("failed to rename database snapshot file (%v)", err) |
| 63 | } |
| 64 | return openBackend(cfg), nil |
| 65 | } |
| 66 | |
| 67 | // openBackend returns a backend using the current etcd db. |
| 68 | func openBackend(cfg ServerConfig) backend.Backend { |
| 69 | fn := cfg.backendPath() |
| 70 | |
| 71 | now, beOpened := time.Now(), make(chan backend.Backend) |
| 72 | go func() { |
| 73 | beOpened <- newBackend(cfg) |
| 74 | }() |
| 75 | |
| 76 | select { |
| 77 | case be := <-beOpened: |
| 78 | if cfg.Logger != nil { |
| 79 | cfg.Logger.Info("opened backend db", zap.String("path", fn), zap.Duration("took", time.Since(now))) |
| 80 | } |
| 81 | return be |
| 82 | |
| 83 | case <-time.After(10 * time.Second): |
| 84 | if cfg.Logger != nil { |
| 85 | cfg.Logger.Info( |
| 86 | "db file is flocked by another process, or taking too long", |
| 87 | zap.String("path", fn), |
| 88 | zap.Duration("took", time.Since(now)), |
| 89 | ) |
| 90 | } else { |
| 91 | plog.Warningf("another etcd process is using %q and holds the file lock, or loading backend file is taking >10 seconds", fn) |
| 92 | plog.Warningf("waiting for it to exit before starting...") |
| 93 | } |
| 94 | } |
| 95 | |
| 96 | return <-beOpened |
| 97 | } |
| 98 | |
| 99 | // recoverBackendSnapshot recovers the DB from a snapshot in case etcd crashes |
| 100 | // before updating the backend db after persisting raft snapshot to disk, |
| 101 | // violating the invariant snapshot.Metadata.Index < db.consistentIndex. In this |
| 102 | // case, replace the db with the snapshot db sent by the leader. |
| 103 | func recoverSnapshotBackend(cfg ServerConfig, oldbe backend.Backend, snapshot raftpb.Snapshot) (backend.Backend, error) { |
| 104 | var cIndex consistentIndex |
| 105 | kv := mvcc.New(cfg.Logger, oldbe, &lease.FakeLessor{}, &cIndex, mvcc.StoreConfig{CompactionBatchLimit: cfg.CompactionBatchLimit}) |
| 106 | defer kv.Close() |
| 107 | if snapshot.Metadata.Index <= kv.ConsistentIndex() { |
| 108 | return oldbe, nil |
| 109 | } |
| 110 | oldbe.Close() |
| 111 | return openSnapshotBackend(cfg, snap.New(cfg.Logger, cfg.SnapDir()), snapshot) |
| 112 | } |