aboutsummaryrefslogtreecommitdiff
path: root/weed/mount/weedfs.go
blob: 95864ef0098aa49dd7f28b497df5be00e9e6c62b (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
package mount

import (
	"context"
	"errors"
	"math/rand/v2"
	"os"
	"path"
	"path/filepath"
	"sync"
	"sync/atomic"
	"time"

	"github.com/hanwen/go-fuse/v2/fuse"
	"google.golang.org/grpc"

	"github.com/seaweedfs/seaweedfs/weed/filer"
	"github.com/seaweedfs/seaweedfs/weed/glog"
	"github.com/seaweedfs/seaweedfs/weed/mount/meta_cache"
	"github.com/seaweedfs/seaweedfs/weed/pb"
	"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
	"github.com/seaweedfs/seaweedfs/weed/pb/mount_pb"
	"github.com/seaweedfs/seaweedfs/weed/storage/types"
	"github.com/seaweedfs/seaweedfs/weed/util"
	"github.com/seaweedfs/seaweedfs/weed/util/chunk_cache"
	"github.com/seaweedfs/seaweedfs/weed/util/grace"
	"github.com/seaweedfs/seaweedfs/weed/util/version"
	"github.com/seaweedfs/seaweedfs/weed/wdclient"

	"github.com/hanwen/go-fuse/v2/fs"
)

type Option struct {
	filerIndex         int32 // align memory for atomic read/write
	FilerAddresses     []pb.ServerAddress
	MountDirectory     string
	GrpcDialOption     grpc.DialOption
	FilerMountRootPath string
	Collection         string
	Replication        string
	TtlSec             int32
	DiskType           types.DiskType
	ChunkSizeLimit     int64
	ConcurrentWriters  int
	CacheDirForRead    string
	CacheSizeMBForRead int64
	CacheDirForWrite   string
	CacheMetaTTlSec    int
	DataCenter         string
	Umask              os.FileMode
	Quota              int64
	DisableXAttr       bool
	IsMacOs            bool

	MountUid         uint32
	MountGid         uint32
	MountMode        os.FileMode
	MountCtime       time.Time
	MountMtime       time.Time
	MountParentInode uint64

	VolumeServerAccess string // how to access volume servers
	Cipher             bool   // whether encrypt data on volume server
	UidGidMapper       *meta_cache.UidGidMapper

	// RDMA acceleration options
	RdmaEnabled       bool
	RdmaSidecarAddr   string
	RdmaFallback      bool
	RdmaReadOnly      bool
	RdmaMaxConcurrent int
	RdmaTimeoutMs     int

	uniqueCacheDirForRead  string
	uniqueCacheDirForWrite string
}

type WFS struct {
	// https://dl.acm.org/doi/fullHtml/10.1145/3310148
	// follow https://github.com/hanwen/go-fuse/blob/master/fuse/api.go
	fuse.RawFileSystem
	mount_pb.UnimplementedSeaweedMountServer
	fs.Inode
	option               *Option
	metaCache            *meta_cache.MetaCache
	stats                statsCache
	chunkCache           *chunk_cache.TieredChunkCache
	signature            int32
	concurrentWriters    *util.LimitedConcurrentExecutor
	copyBufferPool       sync.Pool
	concurrentCopiersSem chan struct{}
	inodeToPath          *InodeToPath
	fhMap                *FileHandleToInode
	dhMap                *DirectoryHandleToInode
	fuseServer           *fuse.Server
	IsOverQuota          bool
	fhLockTable          *util.LockTable[FileHandleId]
	rdmaClient           *RDMAMountClient
	FilerConf            *filer.FilerConf
}

func NewSeaweedFileSystem(option *Option) *WFS {
	wfs := &WFS{
		RawFileSystem: fuse.NewDefaultRawFileSystem(),
		option:        option,
		signature:     util.RandomInt32(),
		inodeToPath:   NewInodeToPath(util.FullPath(option.FilerMountRootPath), option.CacheMetaTTlSec),
		fhMap:         NewFileHandleToInode(),
		dhMap:         NewDirectoryHandleToInode(),
		fhLockTable:   util.NewLockTable[FileHandleId](),
	}

	wfs.option.filerIndex = int32(rand.IntN(len(option.FilerAddresses)))
	wfs.option.setupUniqueCacheDirectory()
	if option.CacheSizeMBForRead > 0 {
		wfs.chunkCache = chunk_cache.NewTieredChunkCache(256, option.getUniqueCacheDirForRead(), option.CacheSizeMBForRead, 1024*1024)
	}

	wfs.metaCache = meta_cache.NewMetaCache(path.Join(option.getUniqueCacheDirForRead(), "meta"), option.UidGidMapper,
		util.FullPath(option.FilerMountRootPath),
		func(path util.FullPath) {
			wfs.inodeToPath.MarkChildrenCached(path)
		}, func(path util.FullPath) bool {
			return wfs.inodeToPath.IsChildrenCached(path)
		}, func(filePath util.FullPath, entry *filer_pb.Entry) {
			// Find inode if it is not a deleted path
			if inode, inodeFound := wfs.inodeToPath.GetInode(filePath); inodeFound {
				// Find open file handle
				if fh, fhFound := wfs.fhMap.FindFileHandle(inode); fhFound {
					fhActiveLock := fh.wfs.fhLockTable.AcquireLock("invalidateFunc", fh.fh, util.ExclusiveLock)
					defer fh.wfs.fhLockTable.ReleaseLock(fh.fh, fhActiveLock)

					// Recreate dirty pages
					fh.dirtyPages.Destroy()
					fh.dirtyPages = newPageWriter(fh, wfs.option.ChunkSizeLimit)

					// Update handle entry
					newEntry, status := wfs.maybeLoadEntry(filePath)
					if status == fuse.OK {
						if fh.GetEntry().GetEntry() != newEntry {
							fh.SetEntry(newEntry)
						}
					}
				}
			}
		})
	grace.OnInterrupt(func() {
		wfs.metaCache.Shutdown()
		os.RemoveAll(option.getUniqueCacheDirForWrite())
		os.RemoveAll(option.getUniqueCacheDirForRead())
		if wfs.rdmaClient != nil {
			wfs.rdmaClient.Close()
		}
	})

	// Initialize RDMA client if enabled
	if option.RdmaEnabled && option.RdmaSidecarAddr != "" {
		rdmaClient, err := NewRDMAMountClient(
			option.RdmaSidecarAddr,
			wfs.LookupFn(),
			option.RdmaMaxConcurrent,
			option.RdmaTimeoutMs,
		)
		if err != nil {
			glog.Warningf("Failed to initialize RDMA client: %v", err)
		} else {
			wfs.rdmaClient = rdmaClient
			glog.Infof("RDMA acceleration enabled: sidecar=%s, maxConcurrent=%d, timeout=%dms",
				option.RdmaSidecarAddr, option.RdmaMaxConcurrent, option.RdmaTimeoutMs)
		}
	}

	if wfs.option.ConcurrentWriters > 0 {
		wfs.concurrentWriters = util.NewLimitedConcurrentExecutor(wfs.option.ConcurrentWriters)
		wfs.concurrentCopiersSem = make(chan struct{}, wfs.option.ConcurrentWriters)
	}
	wfs.copyBufferPool.New = func() any {
		return make([]byte, option.ChunkSizeLimit)
	}
	return wfs
}

func (wfs *WFS) StartBackgroundTasks() error {
	follower, err := wfs.subscribeFilerConfEvents()
	if err != nil {
		return err
	}

	startTime := time.Now()
	go meta_cache.SubscribeMetaEvents(wfs.metaCache, wfs.signature, wfs, wfs.option.FilerMountRootPath, startTime.UnixNano(), follower)
	go wfs.loopCheckQuota()

	return nil
}

func (wfs *WFS) String() string {
	return "seaweedfs"
}

func (wfs *WFS) Init(server *fuse.Server) {
	wfs.fuseServer = server
}

func (wfs *WFS) maybeReadEntry(inode uint64) (path util.FullPath, fh *FileHandle, entry *filer_pb.Entry, status fuse.Status) {
	path, status = wfs.inodeToPath.GetPath(inode)
	if status != fuse.OK {
		return
	}
	var found bool
	if fh, found = wfs.fhMap.FindFileHandle(inode); found {
		entry = fh.UpdateEntry(func(entry *filer_pb.Entry) {
			if entry != nil && fh.entry.Attributes == nil {
				entry.Attributes = &filer_pb.FuseAttributes{}
			}
		})
	} else {
		entry, status = wfs.maybeLoadEntry(path)
	}
	return
}

func (wfs *WFS) maybeLoadEntry(fullpath util.FullPath) (*filer_pb.Entry, fuse.Status) {
	// glog.V(3).Infof("read entry cache miss %s", fullpath)
	dir, name := fullpath.DirAndName()

	// return a valid entry for the mount root
	if string(fullpath) == wfs.option.FilerMountRootPath {
		return &filer_pb.Entry{
			Name:        name,
			IsDirectory: true,
			Attributes: &filer_pb.FuseAttributes{
				Mtime:    wfs.option.MountMtime.Unix(),
				FileMode: uint32(wfs.option.MountMode),
				Uid:      wfs.option.MountUid,
				Gid:      wfs.option.MountGid,
				Crtime:   wfs.option.MountCtime.Unix(),
			},
		}, fuse.OK
	}

	// read from async meta cache
	meta_cache.EnsureVisited(wfs.metaCache, wfs, util.FullPath(dir))
	cachedEntry, cacheErr := wfs.metaCache.FindEntry(context.Background(), fullpath)
	if errors.Is(cacheErr, filer_pb.ErrNotFound) {
		return nil, fuse.ENOENT
	}
	return cachedEntry.ToProtoEntry(), fuse.OK
}

func (wfs *WFS) LookupFn() wdclient.LookupFileIdFunctionType {
	if wfs.option.VolumeServerAccess == "filerProxy" {
		return func(ctx context.Context, fileId string) (targetUrls []string, err error) {
			return []string{"http://" + wfs.getCurrentFiler().ToHttpAddress() + "/?proxyChunkId=" + fileId}, nil
		}
	}
	return filer.LookupFn(wfs)
}

func (wfs *WFS) getCurrentFiler() pb.ServerAddress {
	i := atomic.LoadInt32(&wfs.option.filerIndex)
	return wfs.option.FilerAddresses[i]
}

func (wfs *WFS) ClearCacheDir() {
	wfs.metaCache.Shutdown()
	os.RemoveAll(wfs.option.getUniqueCacheDirForWrite())
	os.RemoveAll(wfs.option.getUniqueCacheDirForRead())
}

func (option *Option) setupUniqueCacheDirectory() {
	cacheUniqueId := util.Md5String([]byte(option.MountDirectory + string(option.FilerAddresses[0]) + option.FilerMountRootPath + version.Version()))[0:8]
	option.uniqueCacheDirForRead = path.Join(option.CacheDirForRead, cacheUniqueId)
	os.MkdirAll(option.uniqueCacheDirForRead, os.FileMode(0777)&^option.Umask)
	option.uniqueCacheDirForWrite = filepath.Join(path.Join(option.CacheDirForWrite, cacheUniqueId), "swap")
	os.MkdirAll(option.uniqueCacheDirForWrite, os.FileMode(0777)&^option.Umask)
}

func (option *Option) getUniqueCacheDirForWrite() string {
	return option.uniqueCacheDirForWrite
}

func (option *Option) getUniqueCacheDirForRead() string {
	return option.uniqueCacheDirForRead
}