PeerTube/server/tests/api/transcoding/hls.ts

287 lines
8.4 KiB
TypeScript
Raw Normal View History

2020-01-31 16:56:52 +01:00
/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
2019-01-29 08:37:25 +01:00
2022-08-17 15:44:32 +02:00
import { expect } from 'chai'
2021-07-23 11:20:00 +02:00
import { basename, join } from 'path'
2019-01-29 08:37:25 +01:00
import {
checkDirectoryIsEmpty,
2020-11-04 15:31:32 +01:00
checkResolutionsInMasterPlaylist,
checkSegmentHash,
checkTmpIsEmpty,
expectStartWith,
hlsInfohashExist
} from '@server/tests/shared'
import { areObjectStorageTestsDisabled, removeFragmentedMP4Ext, uuidRegex } from '@shared/core-utils'
import { HttpStatusCode, VideoStreamingPlaylistType } from '@shared/models'
import {
cleanupTests,
2021-07-16 09:47:51 +02:00
createMultipleServers,
2021-07-16 14:27:30 +02:00
doubleFollow,
2020-01-31 16:56:52 +01:00
makeRawRequest,
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>
2021-08-17 08:26:20 +02:00
ObjectStorageCommand,
2021-07-16 09:47:51 +02:00
PeerTubeServer,
2020-01-31 16:56:52 +01:00
setAccessTokensToServers,
waitJobs,
webtorrentAdd
} from '@shared/server-commands'
2019-05-17 11:56:12 +02:00
import { DEFAULT_AUDIO_RESOLUTION } from '../../../initializers/constants'
2019-01-29 08:37:25 +01:00
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>
2021-08-17 08:26:20 +02:00
async function checkHlsPlaylist (options: {
servers: PeerTubeServer[]
videoUUID: string
hlsOnly: boolean
resolutions?: number[]
objectStorageBaseUrl: string
}) {
const { videoUUID, hlsOnly, objectStorageBaseUrl } = options
const resolutions = options.resolutions ?? [ 240, 360, 480, 720 ]
for (const server of options.servers) {
2021-07-16 09:04:35 +02:00
const videoDetails = await server.videos.get({ id: videoUUID })
const baseUrl = `http://${videoDetails.account.host}`
2019-01-29 08:37:25 +01:00
expect(videoDetails.streamingPlaylists).to.have.lengthOf(1)
const hlsPlaylist = videoDetails.streamingPlaylists.find(p => p.type === VideoStreamingPlaylistType.HLS)
expect(hlsPlaylist).to.not.be.undefined
const hlsFiles = hlsPlaylist.files
expect(hlsFiles).to.have.lengthOf(resolutions.length)
if (hlsOnly) expect(videoDetails.files).to.have.lengthOf(0)
else expect(videoDetails.files).to.have.lengthOf(resolutions.length)
2021-07-22 14:28:03 +02:00
// Check JSON files
for (const resolution of resolutions) {
const file = hlsFiles.find(f => f.resolution.id === resolution)
expect(file).to.not.be.undefined
expect(file.magnetUri).to.have.lengthOf.above(2)
2021-07-22 14:28:03 +02:00
expect(file.torrentUrl).to.match(
new RegExp(`http://${server.host}/lazy-static/torrents/${uuidRegex}-${file.resolution.id}-hls.torrent`)
)
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>
2021-08-17 08:26:20 +02:00
if (objectStorageBaseUrl) {
expectStartWith(file.fileUrl, objectStorageBaseUrl)
} else {
expect(file.fileUrl).to.match(
new RegExp(`${baseUrl}/static/streaming-playlists/hls/${videoDetails.uuid}/${uuidRegex}-${file.resolution.id}-fragmented.mp4`)
)
}
expect(file.resolution.label).to.equal(resolution + 'p')
await makeRawRequest(file.torrentUrl, HttpStatusCode.OK_200)
await makeRawRequest(file.fileUrl, HttpStatusCode.OK_200)
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('')
}
2021-07-22 14:28:03 +02:00
// Check master playlist
2019-01-29 08:37:25 +01:00
{
2021-07-09 10:21:10 +02:00
await checkResolutionsInMasterPlaylist({ server, playlistUrl: hlsPlaylist.playlistUrl, resolutions })
2019-01-29 08:37:25 +01:00
2021-07-16 09:04:35 +02:00
const masterPlaylist = await server.streamingPlaylists.get({ url: hlsPlaylist.playlistUrl })
2019-01-29 08:37:25 +01:00
2021-09-07 15:16:26 +02:00
let i = 0
2019-01-29 08:37:25 +01:00
for (const resolution of resolutions) {
2019-11-26 16:25:36 +01:00
expect(masterPlaylist).to.contain(`${resolution}.m3u8`)
2019-01-29 08:37:25 +01:00
expect(masterPlaylist).to.contain(`${resolution}.m3u8`)
2021-09-07 15:16:26 +02:00
const url = 'http://' + videoDetails.account.host
await hlsInfohashExist(url, hlsPlaylist.playlistUrl, i)
i++
2019-01-29 08:37:25 +01:00
}
}
2021-07-22 14:28:03 +02:00
// Check resolution playlists
2019-01-29 08:37:25 +01:00
{
for (const resolution of resolutions) {
2021-07-23 11:20:00 +02:00
const file = hlsFiles.find(f => f.resolution.id === resolution)
const playlistName = removeFragmentedMP4Ext(basename(file.fileUrl)) + '.m3u8'
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>
2021-08-17 08:26:20 +02:00
const url = objectStorageBaseUrl
2021-08-18 10:00:57 +02:00
? `${objectStorageBaseUrl}hls/${videoUUID}/${playlistName}`
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>
2021-08-17 08:26:20 +02:00
: `${baseUrl}/static/streaming-playlists/hls/${videoUUID}/${playlistName}`
const subPlaylist = await server.streamingPlaylists.get({ url })
2019-01-29 08:37:25 +01:00
2021-07-22 14:28:03 +02:00
expect(subPlaylist).to.match(new RegExp(`${uuidRegex}-${resolution}-fragmented.mp4`))
expect(subPlaylist).to.contain(basename(file.fileUrl))
2019-01-29 08:37:25 +01:00
}
}
{
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>
2021-08-17 08:26:20 +02:00
const baseUrlAndPath = objectStorageBaseUrl
2021-08-18 10:00:57 +02:00
? objectStorageBaseUrl + 'hls/' + videoUUID
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>
2021-08-17 08:26:20 +02:00
: baseUrl + '/static/streaming-playlists/hls/' + videoUUID
2019-01-29 08:37:25 +01:00
for (const resolution of resolutions) {
2021-07-09 10:21:10 +02:00
await checkSegmentHash({
server,
baseUrlPlaylist: baseUrlAndPath,
baseUrlSegment: baseUrlAndPath,
resolution,
hlsPlaylist
})
2019-01-29 08:37:25 +01:00
}
}
}
}
describe('Test HLS videos', function () {
2021-07-16 09:47:51 +02:00
let servers: PeerTubeServer[] = []
2019-01-29 08:37:25 +01:00
let videoUUID = ''
2019-05-17 11:56:12 +02:00
let videoAudioUUID = ''
2019-01-29 08:37:25 +01:00
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>
2021-08-17 08:26:20 +02:00
function runTestSuite (hlsOnly: boolean, objectStorageBaseUrl?: string) {
2021-07-09 16:23:01 +02:00
it('Should upload a video and transcode it to HLS', async function () {
this.timeout(120000)
2019-01-29 08:37:25 +01:00
2021-07-16 09:04:35 +02:00
const { uuid } = await servers[0].videos.upload({ attributes: { name: 'video 1', fixture: 'video_short.webm' } })
2021-07-15 10:02:54 +02:00
videoUUID = uuid
2019-01-29 08:37:25 +01:00
await waitJobs(servers)
2019-01-29 08:37:25 +01:00
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>
2021-08-17 08:26:20 +02:00
await checkHlsPlaylist({ servers, videoUUID, hlsOnly, objectStorageBaseUrl })
})
2019-01-29 08:37:25 +01:00
it('Should upload an audio file and transcode it to HLS', async function () {
this.timeout(120000)
2019-01-29 08:37:25 +01:00
2021-07-16 09:04:35 +02:00
const { uuid } = await servers[0].videos.upload({ attributes: { name: 'video audio', fixture: 'sample.ogg' } })
2021-07-15 10:02:54 +02:00
videoAudioUUID = uuid
2019-01-29 08:37:25 +01:00
await waitJobs(servers)
2019-01-29 08:37:25 +01:00
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>
2021-08-17 08:26:20 +02:00
await checkHlsPlaylist({
servers,
videoUUID: videoAudioUUID,
hlsOnly,
resolutions: [ DEFAULT_AUDIO_RESOLUTION, 360, 240 ],
objectStorageBaseUrl
})
})
2019-01-29 08:37:25 +01:00
it('Should update the video', async function () {
2021-08-18 10:00:57 +02:00
this.timeout(30000)
2019-05-17 11:56:12 +02:00
2021-07-16 09:04:35 +02:00
await servers[0].videos.update({ id: videoUUID, attributes: { name: 'video 1 updated' } })
2019-05-17 11:56:12 +02:00
await waitJobs(servers)
2019-05-17 11:56:12 +02:00
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>
2021-08-17 08:26:20 +02:00
await checkHlsPlaylist({ servers, videoUUID, hlsOnly, objectStorageBaseUrl })
})
2019-05-17 11:56:12 +02:00
it('Should delete videos', async function () {
this.timeout(10000)
2019-05-17 15:51:42 +02:00
2021-07-16 09:04:35 +02:00
await servers[0].videos.remove({ id: videoUUID })
await servers[0].videos.remove({ id: videoAudioUUID })
2019-01-29 08:37:25 +01:00
await waitJobs(servers)
2019-01-29 08:37:25 +01:00
for (const server of servers) {
2021-07-16 09:04:35 +02:00
await server.videos.get({ id: videoUUID, expectedStatus: HttpStatusCode.NOT_FOUND_404 })
await server.videos.get({ id: videoAudioUUID, expectedStatus: HttpStatusCode.NOT_FOUND_404 })
}
})
2019-01-29 08:37:25 +01:00
it('Should have the playlists/segment deleted from the disk', async function () {
for (const server of servers) {
await checkDirectoryIsEmpty(server, 'videos')
await checkDirectoryIsEmpty(server, join('streaming-playlists', 'hls'))
}
})
2019-05-17 15:51:42 +02:00
it('Should have an empty tmp directory', async function () {
for (const server of servers) {
await checkTmpIsEmpty(server)
}
})
}
2019-01-29 08:37:25 +01:00
before(async function () {
this.timeout(120000)
2019-01-29 08:37:25 +01:00
const configOverride = {
transcoding: {
enabled: true,
allow_audio_files: true,
hls: {
enabled: true
}
}
2019-01-29 08:37:25 +01:00
}
2021-07-16 09:47:51 +02:00
servers = await createMultipleServers(2, configOverride)
// Get the access tokens
await setAccessTokensToServers(servers)
// Server 1 and server 2 follow each other
await doubleFollow(servers[0], servers[1])
2019-01-29 08:37:25 +01:00
})
describe('With WebTorrent & HLS enabled', function () {
runTestSuite(false)
2019-01-29 08:37:25 +01:00
})
describe('With only HLS enabled', function () {
before(async function () {
2021-07-16 09:04:35 +02:00
await servers[0].config.updateCustomSubConfig({
2021-07-07 11:51:09 +02:00
newConfig: {
transcoding: {
enabled: true,
allowAudioFiles: true,
resolutions: {
'144p': false,
2021-07-07 11:51:09 +02:00
'240p': true,
'360p': true,
'480p': true,
'720p': true,
'1080p': true,
'1440p': true,
'2160p': true
},
hls: {
enabled: true
},
webtorrent: {
enabled: false
}
}
}
})
})
runTestSuite(true)
2019-01-29 08:37:25 +01:00
})
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>
2021-08-17 08:26:20 +02:00
describe('With object storage enabled', function () {
if (areObjectStorageTestsDisabled()) return
before(async function () {
this.timeout(120000)
const configOverride = ObjectStorageCommand.getDefaultConfig()
await ObjectStorageCommand.prepareDefaultBuckets()
await servers[0].kill()
await servers[0].run(configOverride)
})
runTestSuite(true, ObjectStorageCommand.getPlaylistBaseUrl())
})
2019-04-24 15:10:37 +02:00
after(async function () {
await cleanupTests(servers)
2019-01-29 08:37:25 +01:00
})
})