Add support for saving video files to object storage (#4290)

* Add support for saving video files to object storage

* Add support for custom url generation on s3 stored files

Uses two config keys to support url generation that doesn't directly go
to (compatible s3). Can be used to generate urls to any cache server or
CDN.

* Upload files to s3 concurrently and delete originals afterwards

* Only publish after move to object storage is complete

* Use base url instead of url template

* Fix mistyped config field

* Add rudenmentary way to download before transcode

* Implement Chocobozzz suggestions

https://github.com/Chocobozzz/PeerTube/pull/4290#issuecomment-891670478

The remarks in question:
    Try to use objectStorage prefix instead of s3 prefix for your function/variables/config names
    Prefer to use a tree for the config: s3.streaming_playlists_bucket -> object_storage.streaming_playlists.bucket
    Use uppercase for config: S3.STREAMING_PLAYLISTS_BUCKETINFO.bucket -> OBJECT_STORAGE.STREAMING_PLAYLISTS.BUCKET (maybe BUCKET_NAME instead of BUCKET)
    I suggest to rename moveJobsRunning to pendingMovingJobs (or better, create a dedicated videoJobInfo table with a pendingMove & videoId columns so we could also use this table to track pending transcoding jobs)
    https://github.com/Chocobozzz/PeerTube/pull/4290/files#diff-3e26d41ca4bda1de8e1747af70ca2af642abcc1e9e0bfb94239ff2165acfbde5R19 uses a string instead of an integer
    I think we should store the origin object storage URL in fileUrl, without base_url injection. Instead, inject the base_url at "runtime" so admins can easily change this configuration without running a script to update DB URLs

* Import correct function

* Support multipart upload

* Remove import of node 15.0 module stream/promises

* Extend maximum upload job length

Using the same value as for redundancy downloading seems logical

* Use dynamic part size for really large uploads

Also adds very small part size for local testing

* Fix decreasePendingMove query

* Resolve various PR comments

* Move to object storage after optimize

* Make upload size configurable and increase default

* Prune webtorrent files that are stored in object storage

* Move files after transcoding jobs

* Fix federation

* Add video path manager

* Support move to external storage job in client

* Fix live object storage tests

Co-authored-by: Chocobozzz <me@florianbigard.com>

4 files changed, 642 insertions, 0 deletions

diff --git a/server/tests/api/object-storage/index.ts b/server/tests/api/object-storage/index.ts
new file mode 100644
index 000000000..f319d6ef5
--- /dev/null
+++ b/server/tests/api/object-storage/index.ts
@@ -0,0 +1,3 @@
+export * from './live'
+export * from './video-imports'
+export * from './videos'
diff --git a/server/tests/api/object-storage/live.ts b/server/tests/api/object-storage/live.ts
new file mode 100644
index 000000000..d3e6777f2
--- /dev/null
+++ b/server/tests/api/object-storage/live.ts
@@ -0,0 +1,136 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import 'mocha'
+import * as chai from 'chai'
+import { FfmpegCommand } from 'fluent-ffmpeg'
+import {
+  areObjectStorageTestsDisabled,
+  createMultipleServers,
+  doubleFollow,
+  expectStartWith,
+  killallServers,
+  makeRawRequest,
+  ObjectStorageCommand,
+  PeerTubeServer,
+  setAccessTokensToServers,
+  setDefaultVideoChannel,
+  stopFfmpeg,
+  waitJobs,
+  waitUntilLivePublishedOnAllServers,
+  waitUntilLiveSavedOnAllServers
+} from '@shared/extra-utils'
+import { HttpStatusCode, LiveVideoCreate, VideoFile, VideoPrivacy } from '@shared/models'
+
+const expect = chai.expect
+
+async function createLive (server: PeerTubeServer) {
+  const attributes: LiveVideoCreate = {
+    channelId: server.store.channel.id,
+    privacy: VideoPrivacy.PUBLIC,
+    name: 'my super live',
+    saveReplay: true
+  }
+
+  const { uuid } = await server.live.create({ fields: attributes })
+
+  return uuid
+}
+
+async function checkFiles (files: VideoFile[]) {
+  for (const file of files) {
+    expectStartWith(file.fileUrl, ObjectStorageCommand.getPlaylistBaseUrl())
+
+    await makeRawRequest(file.fileUrl, HttpStatusCode.OK_200)
+  }
+}
+
+describe('Object storage for lives', function () {
+  if (areObjectStorageTestsDisabled()) return
+
+  let ffmpegCommand: FfmpegCommand
+  let servers: PeerTubeServer[]
+  let videoUUID: string
+
+  before(async function () {
+    this.timeout(120000)
+
+    await ObjectStorageCommand.prepareDefaultBuckets()
+
+    servers = await createMultipleServers(2, ObjectStorageCommand.getDefaultConfig())
+
+    await setAccessTokensToServers(servers)
+    await setDefaultVideoChannel(servers)
+    await doubleFollow(servers[0], servers[1])
+
+    await servers[0].config.enableTranscoding()
+  })
+
+  describe('Without live transcoding', async function () {
+
+    before(async function () {
+      await servers[0].config.enableLive({ transcoding: false })
+
+      videoUUID = await createLive(servers[0])
+    })
+
+    it('Should create a live and save the replay on object storage', async function () {
+      this.timeout(220000)
+
+      ffmpegCommand = await servers[0].live.sendRTMPStreamInVideo({ videoId: videoUUID })
+      await waitUntilLivePublishedOnAllServers(servers, videoUUID)
+
+      await stopFfmpeg(ffmpegCommand)
+
+      await waitUntilLiveSavedOnAllServers(servers, videoUUID)
+      await waitJobs(servers)
+
+      for (const server of servers) {
+        const video = await server.videos.get({ id: videoUUID })
+
+        expect(video.files).to.have.lengthOf(0)
+        expect(video.streamingPlaylists).to.have.lengthOf(1)
+
+        const files = video.streamingPlaylists[0].files
+
+        await checkFiles(files)
+      }
+    })
+  })
+
+  describe('With live transcoding', async function () {
+
+    before(async function () {
+      await servers[0].config.enableLive({ transcoding: true })
+
+      videoUUID = await createLive(servers[0])
+    })
+
+    it('Should import a video and have sent it to object storage', async function () {
+      this.timeout(240000)
+
+      ffmpegCommand = await servers[0].live.sendRTMPStreamInVideo({ videoId: videoUUID })
+      await waitUntilLivePublishedOnAllServers(servers, videoUUID)
+
+      await stopFfmpeg(ffmpegCommand)
+
+      await waitUntilLiveSavedOnAllServers(servers, videoUUID)
+      await waitJobs(servers)
+
+      for (const server of servers) {
+        const video = await server.videos.get({ id: videoUUID })
+
+        expect(video.files).to.have.lengthOf(0)
+        expect(video.streamingPlaylists).to.have.lengthOf(1)
+
+        const files = video.streamingPlaylists[0].files
+        expect(files).to.have.lengthOf(4)
+
+        await checkFiles(files)
+      }
+    })
+  })
+
+  after(async function () {
+    await killallServers(servers)
+  })
+})
diff --git a/server/tests/api/object-storage/video-imports.ts b/server/tests/api/object-storage/video-imports.ts
new file mode 100644
index 000000000..efc01f550
--- /dev/null
+++ b/server/tests/api/object-storage/video-imports.ts
@@ -0,0 +1,112 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import 'mocha'
+import * as chai from 'chai'
+import {
+  areObjectStorageTestsDisabled,
+  createSingleServer,
+  expectStartWith,
+  FIXTURE_URLS,
+  killallServers,
+  makeRawRequest,
+  ObjectStorageCommand,
+  PeerTubeServer,
+  setAccessTokensToServers,
+  setDefaultVideoChannel,
+  waitJobs
+} from '@shared/extra-utils'
+import { HttpStatusCode, VideoPrivacy } from '@shared/models'
+
+const expect = chai.expect
+
+async function importVideo (server: PeerTubeServer) {
+  const attributes = {
+    name: 'import 2',
+    privacy: VideoPrivacy.PUBLIC,
+    channelId: server.store.channel.id,
+    targetUrl: FIXTURE_URLS.goodVideo720
+  }
+
+  const { video: { uuid } } = await server.imports.importVideo({ attributes })
+
+  return uuid
+}
+
+describe('Object storage for video import', function () {
+  if (areObjectStorageTestsDisabled()) return
+
+  let server: PeerTubeServer
+
+  before(async function () {
+    this.timeout(120000)
+
+    await ObjectStorageCommand.prepareDefaultBuckets()
+
+    server = await createSingleServer(1, ObjectStorageCommand.getDefaultConfig())
+
+    await setAccessTokensToServers([ server ])
+    await setDefaultVideoChannel([ server ])
+
+    await server.config.enableImports()
+  })
+
+  describe('Without transcoding', async function () {
+
+    before(async function () {
+      await server.config.disableTranscoding()
+    })
+
+    it('Should import a video and have sent it to object storage', async function () {
+      this.timeout(120000)
+
+      const uuid = await importVideo(server)
+      await waitJobs(server)
+
+      const video = await server.videos.get({ id: uuid })
+
+      expect(video.files).to.have.lengthOf(1)
+      expect(video.streamingPlaylists).to.have.lengthOf(0)
+
+      const fileUrl = video.files[0].fileUrl
+      expectStartWith(fileUrl, ObjectStorageCommand.getWebTorrentBaseUrl())
+
+      await makeRawRequest(fileUrl, HttpStatusCode.OK_200)
+    })
+  })
+
+  describe('With transcoding', async function () {
+
+    before(async function () {
+      await server.config.enableTranscoding()
+    })
+
+    it('Should import a video and have sent it to object storage', async function () {
+      this.timeout(120000)
+
+      const uuid = await importVideo(server)
+      await waitJobs(server)
+
+      const video = await server.videos.get({ id: uuid })
+
+      expect(video.files).to.have.lengthOf(4)
+      expect(video.streamingPlaylists).to.have.lengthOf(1)
+      expect(video.streamingPlaylists[0].files).to.have.lengthOf(4)
+
+      for (const file of video.files) {
+        expectStartWith(file.fileUrl, ObjectStorageCommand.getWebTorrentBaseUrl())
+
+        await makeRawRequest(file.fileUrl, HttpStatusCode.OK_200)
+      }
+
+      for (const file of video.streamingPlaylists[0].files) {
+        expectStartWith(file.fileUrl, ObjectStorageCommand.getPlaylistBaseUrl())
+
+        await makeRawRequest(file.fileUrl, HttpStatusCode.OK_200)
+      }
+    })
+  })
+
+  after(async function () {
+    await killallServers([ server ])
+  })
+})
diff --git a/server/tests/api/object-storage/videos.ts b/server/tests/api/object-storage/videos.ts
new file mode 100644
index 000000000..3958bd3d7
--- /dev/null
+++ b/server/tests/api/object-storage/videos.ts
@@ -0,0 +1,391 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import 'mocha'
+import * as chai from 'chai'
+import { merge } from 'lodash'
+import {
+  areObjectStorageTestsDisabled,
+  checkTmpIsEmpty,
+  cleanupTests,
+  createMultipleServers,
+  createSingleServer,
+  doubleFollow,
+  expectStartWith,
+  killallServers,
+  makeRawRequest,
+  MockObjectStorage,
+  ObjectStorageCommand,
+  PeerTubeServer,
+  setAccessTokensToServers,
+  waitJobs,
+  webtorrentAdd
+} from '@shared/extra-utils'
+import { HttpStatusCode, VideoDetails } from '@shared/models'
+
+const expect = chai.expect
+
+async function checkFiles (options: {
+  video: VideoDetails
+
+  baseMockUrl?: string
+
+  playlistBucket: string
+  playlistPrefix?: string
+
+  webtorrentBucket: string
+  webtorrentPrefix?: string
+}) {
+  const {
+    video,
+    playlistBucket,
+    webtorrentBucket,
+    baseMockUrl,
+    playlistPrefix,
+    webtorrentPrefix
+  } = options
+
+  let allFiles = video.files
+
+  for (const file of video.files) {
+    const baseUrl = baseMockUrl
+      ? `${baseMockUrl}/${webtorrentBucket}/`
+      : `http://${webtorrentBucket}.${ObjectStorageCommand.getEndpointHost()}/`
+
+    const prefix = webtorrentPrefix || ''
+    const start = baseUrl + prefix
+
+    expectStartWith(file.fileUrl, start)
+
+    const res = await makeRawRequest(file.fileDownloadUrl, HttpStatusCode.FOUND_302)
+    const location = res.headers['location']
+    expectStartWith(location, start)
+
+    await makeRawRequest(location, HttpStatusCode.OK_200)
+  }
+
+  const hls = video.streamingPlaylists[0]
+
+  if (hls) {
+    allFiles = allFiles.concat(hls.files)
+
+    const baseUrl = baseMockUrl
+      ? `${baseMockUrl}/${playlistBucket}/`
+      : `http://${playlistBucket}.${ObjectStorageCommand.getEndpointHost()}/`
+
+    const prefix = playlistPrefix || ''
+    const start = baseUrl + prefix
+
+    expectStartWith(hls.playlistUrl, start)
+    expectStartWith(hls.segmentsSha256Url, start)
+
+    await makeRawRequest(hls.playlistUrl, HttpStatusCode.OK_200)
+
+    const resSha = await makeRawRequest(hls.segmentsSha256Url, HttpStatusCode.OK_200)
+    expect(JSON.stringify(resSha.body)).to.not.throw
+
+    for (const file of hls.files) {
+      expectStartWith(file.fileUrl, start)
+
+      const res = await makeRawRequest(file.fileDownloadUrl, HttpStatusCode.FOUND_302)
+      const location = res.headers['location']
+      expectStartWith(location, start)
+
+      await makeRawRequest(location, HttpStatusCode.OK_200)
+    }
+  }
+
+  for (const file of allFiles) {
+    const torrent = await webtorrentAdd(file.magnetUri, true)
+
+    expect(torrent.files).to.be.an('array')
+    expect(torrent.files.length).to.equal(1)
+    expect(torrent.files[0].path).to.exist.and.to.not.equal('')
+
+    const res = await makeRawRequest(file.fileUrl, HttpStatusCode.OK_200)
+    expect(res.body).to.have.length.above(100)
+  }
+
+  return allFiles.map(f => f.fileUrl)
+}
+
+function runTestSuite (options: {
+  playlistBucket: string
+  playlistPrefix?: string
+
+  webtorrentBucket: string
+  webtorrentPrefix?: string
+
+  useMockBaseUrl?: boolean
+
+  maxUploadPart?: string
+}) {
+  const mockObjectStorage = new MockObjectStorage()
+  let baseMockUrl: string
+
+  let servers: PeerTubeServer[]
+
+  let keptUrls: string[] = []
+
+  const uuidsToDelete: string[] = []
+  let deletedUrls: string[] = []
+
+  before(async function () {
+    this.timeout(120000)
+
+    const port = await mockObjectStorage.initialize()
+    baseMockUrl = options.useMockBaseUrl ? `http://localhost:${port}` : undefined
+
+    await ObjectStorageCommand.createBucket(options.playlistBucket)
+    await ObjectStorageCommand.createBucket(options.webtorrentBucket)
+
+    const config = {
+      object_storage: {
+        enabled: true,
+        endpoint: 'http://' + ObjectStorageCommand.getEndpointHost(),
+        region: ObjectStorageCommand.getRegion(),
+
+        credentials: ObjectStorageCommand.getCredentialsConfig(),
+
+        max_upload_part: options.maxUploadPart || '2MB',
+
+        streaming_playlists: {
+          bucket_name: options.playlistBucket,
+          prefix: options.playlistPrefix,
+          base_url: baseMockUrl
+            ? `${baseMockUrl}/${options.playlistBucket}`
+            : undefined
+        },
+
+        videos: {
+          bucket_name: options.webtorrentBucket,
+          prefix: options.webtorrentPrefix,
+          base_url: baseMockUrl
+            ? `${baseMockUrl}/${options.webtorrentBucket}`
+            : undefined
+        }
+      }
+    }
+
+    servers = await createMultipleServers(2, config)
+
+    await setAccessTokensToServers(servers)
+    await doubleFollow(servers[0], servers[1])
+
+    for (const server of servers) {
+      const { uuid } = await server.videos.quickUpload({ name: 'video to keep' })
+      await waitJobs(servers)
+
+      const files = await server.videos.listFiles({ id: uuid })
+      keptUrls = keptUrls.concat(files.map(f => f.fileUrl))
+    }
+  })
+
+  it('Should upload a video and move it to the object storage without transcoding', async function () {
+    this.timeout(20000)
+
+    const { uuid } = await servers[0].videos.quickUpload({ name: 'video 1' })
+    uuidsToDelete.push(uuid)
+
+    await waitJobs(servers)
+
+    for (const server of servers) {
+      const video = await server.videos.get({ id: uuid })
+      const files = await checkFiles({ ...options, video, baseMockUrl })
+
+      deletedUrls = deletedUrls.concat(files)
+    }
+  })
+
+  it('Should upload a video and move it to the object storage with transcoding', async function () {
+    this.timeout(40000)
+
+    const { uuid } = await servers[1].videos.quickUpload({ name: 'video 2' })
+    uuidsToDelete.push(uuid)
+
+    await waitJobs(servers)
+
+    for (const server of servers) {
+      const video = await server.videos.get({ id: uuid })
+      const files = await checkFiles({ ...options, video, baseMockUrl })
+
+      deletedUrls = deletedUrls.concat(files)
+    }
+  })
+
+  it('Should correctly delete the files', async function () {
+    await servers[0].videos.remove({ id: uuidsToDelete[0] })
+    await servers[1].videos.remove({ id: uuidsToDelete[1] })
+
+    await waitJobs(servers)
+
+    for (const url of deletedUrls) {
+      await makeRawRequest(url, HttpStatusCode.NOT_FOUND_404)
+    }
+  })
+
+  it('Should have kept other files', async function () {
+    for (const url of keptUrls) {
+      await makeRawRequest(url, HttpStatusCode.OK_200)
+    }
+  })
+
+  it('Should have an empty tmp directory', async function () {
+    for (const server of servers) {
+      await checkTmpIsEmpty(server)
+    }
+  })
+
+  after(async function () {
+    mockObjectStorage.terminate()
+
+    await cleanupTests(servers)
+  })
+}
+
+describe('Object storage for videos', function () {
+  if (areObjectStorageTestsDisabled()) return
+
+  describe('Test config', function () {
+    let server: PeerTubeServer
+
+    const baseConfig = {
+      object_storage: {
+        enabled: true,
+        endpoint: 'http://' + ObjectStorageCommand.getEndpointHost(),
+        region: ObjectStorageCommand.getRegion(),
+
+        credentials: ObjectStorageCommand.getCredentialsConfig(),
+
+        streaming_playlists: {
+          bucket_name: ObjectStorageCommand.DEFAULT_PLAYLIST_BUCKET
+        },
+
+        videos: {
+          bucket_name: ObjectStorageCommand.DEFAULT_WEBTORRENT_BUCKET
+        }
+      }
+    }
+
+    const badCredentials = {
+      access_key_id: 'AKIAIOSFODNN7EXAMPLE',
+      secret_access_key: 'aJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY'
+    }
+
+    it('Should fail with same bucket names without prefix', function (done) {
+      const config = merge({}, baseConfig, {
+        object_storage: {
+          streaming_playlists: {
+            bucket_name: 'aaa'
+          },
+
+          videos: {
+            bucket_name: 'aaa'
+          }
+        }
+      })
+
+      createSingleServer(1, config)
+        .then(() => done(new Error('Did not throw')))
+        .catch(() => done())
+    })
+
+    it('Should fail with bad credentials', async function () {
+      this.timeout(60000)
+
+      await ObjectStorageCommand.prepareDefaultBuckets()
+
+      const config = merge({}, baseConfig, {
+        object_storage: {
+          credentials: badCredentials
+        }
+      })
+
+      server = await createSingleServer(1, config)
+      await setAccessTokensToServers([ server ])
+
+      const { uuid } = await server.videos.quickUpload({ name: 'video' })
+
+      await waitJobs([ server ], true)
+      const video = await server.videos.get({ id: uuid })
+
+      expectStartWith(video.files[0].fileUrl, server.url)
+
+      await killallServers([ server ])
+    })
+
+    it('Should succeed with credentials from env', async function () {
+      this.timeout(60000)
+
+      await ObjectStorageCommand.prepareDefaultBuckets()
+
+      const config = merge({}, baseConfig, {
+        object_storage: {
+          credentials: {
+            access_key_id: '',
+            secret_access_key: ''
+          }
+        }
+      })
+
+      const goodCredentials = ObjectStorageCommand.getCredentialsConfig()
+
+      server = await createSingleServer(1, config, {
+        env: {
+          AWS_ACCESS_KEY_ID: goodCredentials.access_key_id,
+          AWS_SECRET_ACCESS_KEY: goodCredentials.secret_access_key
+        }
+      })
+
+      await setAccessTokensToServers([ server ])
+
+      const { uuid } = await server.videos.quickUpload({ name: 'video' })
+
+      await waitJobs([ server ], true)
+      const video = await server.videos.get({ id: uuid })
+
+      expectStartWith(video.files[0].fileUrl, ObjectStorageCommand.getWebTorrentBaseUrl())
+    })
+
+    after(async function () {
+      await killallServers([ server ])
+    })
+  })
+
+  describe('Test simple object storage', function () {
+    runTestSuite({
+      playlistBucket: 'streaming-playlists',
+      webtorrentBucket: 'videos'
+    })
+  })
+
+  describe('Test object storage with prefix', function () {
+    runTestSuite({
+      playlistBucket: 'mybucket',
+      webtorrentBucket: 'mybucket',
+
+      playlistPrefix: 'streaming-playlists_',
+      webtorrentPrefix: 'webtorrent_'
+    })
+  })
+
+  describe('Test object storage with prefix and base URL', function () {
+    runTestSuite({
+      playlistBucket: 'mybucket',
+      webtorrentBucket: 'mybucket',
+
+      playlistPrefix: 'streaming-playlists_',
+      webtorrentPrefix: 'webtorrent_',
+
+      useMockBaseUrl: true
+    })
+  })
+
+  describe('Test object storage with small upload part', function () {
+    runTestSuite({
+      playlistBucket: 'streaming-playlists',
+      webtorrentBucket: 'videos',
+
+      maxUploadPart: '5KB'
+    })
+  })
+})