Improve streaming server with cluster (#1970)

.env.production.sample

@@ -81,3 +81,7 @@ SMTP_FROM_ADDRESS=notifications@example.com
 # Advanced settings
 # If you need to use pgBouncer, you need to disable prepared statements:
 # PREPARED_STATEMENTS=false
+
+# Cluster number setting for streaming API server.
+# If you comment out following line, cluster number will be `numOfCpuCores - 1`.
+STREAMING_CLUSTER_NUM=1

streaming/index.js

@@ -1,3 +1,5 @@
+import os from 'os';
+import cluster from 'cluster';
 import dotenv from 'dotenv'
 import express from 'express'
 import http from 'http'
@@ -14,300 +16,318 @@ dotenv.config({
   path: env === 'production' ? '.env.production' : '.env'
 })
 
-const pgConfigs = {
+if (cluster.isMaster) {
-  development: {
+  // cluster master
-    database: 'mastodon_development',
-    host:     '/var/run/postgresql',
-    max:      10
-  },
 
-  production: {
+  const core = +process.env.STREAMING_CLUSTER_NUM || (env === 'development' ? 1 : os.cpus().length - 1)
-    user:     process.env.DB_USER || 'mastodon',
+  const fork = () => {
-    password: process.env.DB_PASS || '',
+    const worker = cluster.fork();
-    database: process.env.DB_NAME || 'mastodon_production',
+    worker.on('exit', (code, signal) => {
-    host:     process.env.DB_HOST || 'localhost',
+      log.error(`Worker died with exit code ${code}, signal ${signal} received.`);
-    port:     process.env.DB_PORT || 5432,
+      setTimeout(() => fork(), 0);
-    max:      10
+    });
-  }
+  };
-}
+  for (let i = 0; i < core; i++) fork();
+  log.info(`Starting streaming API server master with ${core} workers`)
 
-const app    = express()
+} else {
-const pgPool = new pg.Pool(pgConfigs[env])
+  // cluster worker
-const server = http.createServer(app)
-const wss    = new WebSocket.Server({ server })
 
-const redisClient = redis.createClient({
+  const pgConfigs = {
-  host:     process.env.REDIS_HOST     || '127.0.0.1',
+    development: {
-  port:     process.env.REDIS_PORT     || 6379,
+      database: 'mastodon_development',
-  password: process.env.REDIS_PASSWORD
+      host:     '/var/run/postgresql',
-})
+      max:      10
+    },
 
-const subs = {}
+    production: {
-
+      user:     process.env.DB_USER || 'mastodon',
-redisClient.on('pmessage', (_, channel, message) => {
+      password: process.env.DB_PASS || '',
-  const callbacks = subs[channel]
+      database: process.env.DB_NAME || 'mastodon_production',
-
+      host:     process.env.DB_HOST || 'localhost',
-  log.silly(`New message on channel ${channel}`)
+      port:     process.env.DB_PORT || 5432,
-
+      max:      10
-  if (!callbacks) {
+    }
-    return
   }
 
-  callbacks.forEach(callback => callback(message))
+  const app    = express()
-})
+  const pgPool = new pg.Pool(pgConfigs[env])
+  const server = http.createServer(app)
+  const wss    = new WebSocket.Server({ server })
 
-redisClient.psubscribe('timeline:*')
+  const redisClient = redis.createClient({
+    host:     process.env.REDIS_HOST     || '127.0.0.1',
+    port:     process.env.REDIS_PORT     || 6379,
+    password: process.env.REDIS_PASSWORD
+  })
 
-const subscribe = (channel, callback) => {
+  const subs = {}
-  log.silly(`Adding listener for ${channel}`)
-  subs[channel] = subs[channel] || []
-  subs[channel].push(callback)
-}
 
-const unsubscribe = (channel, callback) => {
+  redisClient.on('pmessage', (_, channel, message) => {
-  log.silly(`Removing listener for ${channel}`)
+    const callbacks = subs[channel]
-  subs[channel] = subs[channel].filter(item => item !== callback)
-}
 
-const allowCrossDomain = (req, res, next) => {
+    log.silly(`New message on channel ${channel}`)
-  res.header('Access-Control-Allow-Origin', '*')
-  res.header('Access-Control-Allow-Headers', 'Authorization, Accept, Cache-Control')
-  res.header('Access-Control-Allow-Methods', 'GET, OPTIONS')
 
-  next()
+    if (!callbacks) {
-}
-
-const setRequestId = (req, res, next) => {
-  req.requestId = uuid.v4()
-  res.header('X-Request-Id', req.requestId)
-
-  next()
-}
-
-const accountFromToken = (token, req, next) => {
-  pgPool.connect((err, client, done) => {
-    if (err) {
-      next(err)
       return
     }
 
-    client.query('SELECT oauth_access_tokens.resource_owner_id, users.account_id FROM oauth_access_tokens INNER JOIN users ON oauth_access_tokens.resource_owner_id = users.id WHERE oauth_access_tokens.token = $1 LIMIT 1', [token], (err, result) => {
+    callbacks.forEach(callback => callback(message))
-      done()
+  })
 
+  redisClient.psubscribe('timeline:*')
+
+  const subscribe = (channel, callback) => {
+    log.silly(`Adding listener for ${channel}`)
+    subs[channel] = subs[channel] || []
+    subs[channel].push(callback)
+  }
+
+  const unsubscribe = (channel, callback) => {
+    log.silly(`Removing listener for ${channel}`)
+    subs[channel] = subs[channel].filter(item => item !== callback)
+  }
+
+  const allowCrossDomain = (req, res, next) => {
+    res.header('Access-Control-Allow-Origin', '*')
+    res.header('Access-Control-Allow-Headers', 'Authorization, Accept, Cache-Control')
+    res.header('Access-Control-Allow-Methods', 'GET, OPTIONS')
+
+    next()
+  }
+
+  const setRequestId = (req, res, next) => {
+    req.requestId = uuid.v4()
+    res.header('X-Request-Id', req.requestId)
+
+    next()
+  }
+
+  const accountFromToken = (token, req, next) => {
+    pgPool.connect((err, client, done) => {
       if (err) {
         next(err)
         return
       }
 
-      if (result.rows.length === 0) {
+      client.query('SELECT oauth_access_tokens.resource_owner_id, users.account_id FROM oauth_access_tokens INNER JOIN users ON oauth_access_tokens.resource_owner_id = users.id WHERE oauth_access_tokens.token = $1 LIMIT 1', [token], (err, result) => {
-        err = new Error('Invalid access token')
+        done()
-        err.statusCode = 401
 
-        next(err)
-        return
-      }
-
-      req.accountId = result.rows[0].account_id
-
-      next()
-    })
-  })
-}
-
-const authenticationMiddleware = (req, res, next) => {
-  if (req.method === 'OPTIONS') {
-    next()
-    return
-  }
-
-  const authorization = req.get('Authorization')
-
-  if (!authorization) {
-    const err = new Error('Missing access token')
-    err.statusCode = 401
-
-    next(err)
-    return
-  }
-
-  const token = authorization.replace(/^Bearer /, '')
-
-  accountFromToken(token, req, next)
-}
-
-const errorMiddleware = (err, req, res, next) => {
-  log.error(req.requestId, err)
-  res.writeHead(err.statusCode || 500, { 'Content-Type': 'application/json' })
-  res.end(JSON.stringify({ error: err.statusCode ? `${err}` : 'An unexpected error occurred' }))
-}
-
-const placeholders = (arr, shift = 0) => arr.map((_, i) => `$${i + 1 + shift}`).join(', ');
-
-const streamFrom = (id, req, output, attachCloseHandler, needsFiltering = false) => {
-  log.verbose(req.requestId, `Starting stream from ${id} for ${req.accountId}`)
-
-  const listener = message => {
-    const { event, payload, queued_at } = JSON.parse(message)
-
-    const transmit = () => {
-      const now   = new Date().getTime()
-      const delta = now - queued_at;
-
-      log.silly(req.requestId, `Transmitting for ${req.accountId}: ${event} ${payload} Delay: ${delta}ms`)
-      output(event, payload)
-    }
-
-    // Only messages that may require filtering are statuses, since notifications
-    // are already personalized and deletes do not matter
-    if (needsFiltering && event === 'update') {
-      pgPool.connect((err, client, done) => {
         if (err) {
-          log.error(err)
+          next(err)
           return
         }
 
-        const unpackedPayload  = JSON.parse(payload)
+        if (result.rows.length === 0) {
-        const targetAccountIds = [unpackedPayload.account.id].concat(unpackedPayload.mentions.map(item => item.id)).concat(unpackedPayload.reblog ? [unpackedPayload.reblog.account.id] : [])
+          err = new Error('Invalid access token')
+          err.statusCode = 401
 
-        client.query(`SELECT target_account_id FROM blocks WHERE account_id = $1 AND target_account_id IN (${placeholders(targetAccountIds, 1)}) UNION SELECT target_account_id FROM mutes WHERE account_id = $1 AND target_account_id IN (${placeholders(targetAccountIds, 1)})`, [req.accountId].concat(targetAccountIds), (err, result) => {
+          next(err)
-          done()
+          return
+        }
 
+        req.accountId = result.rows[0].account_id
+
+        next()
+      })
+    })
+  }
+
+  const authenticationMiddleware = (req, res, next) => {
+    if (req.method === 'OPTIONS') {
+      next()
+      return
+    }
+
+    const authorization = req.get('Authorization')
+
+    if (!authorization) {
+      const err = new Error('Missing access token')
+      err.statusCode = 401
+
+      next(err)
+      return
+    }
+
+    const token = authorization.replace(/^Bearer /, '')
+
+    accountFromToken(token, req, next)
+  }
+
+  const errorMiddleware = (err, req, res, next) => {
+    log.error(req.requestId, err)
+    res.writeHead(err.statusCode || 500, { 'Content-Type': 'application/json' })
+    res.end(JSON.stringify({ error: err.statusCode ? `${err}` : 'An unexpected error occurred' }))
+  }
+
+  const placeholders = (arr, shift = 0) => arr.map((_, i) => `$${i + 1 + shift}`).join(', ');
+
+  const streamFrom = (id, req, output, attachCloseHandler, needsFiltering = false) => {
+    log.verbose(req.requestId, `Starting stream from ${id} for ${req.accountId}`)
+
+    const listener = message => {
+      const { event, payload, queued_at } = JSON.parse(message)
+
+      const transmit = () => {
+        const now   = new Date().getTime()
+        const delta = now - queued_at;
+
+        log.silly(req.requestId, `Transmitting for ${req.accountId}: ${event} ${payload} Delay: ${delta}ms`)
+        output(event, payload)
+      }
+
+      // Only messages that may require filtering are statuses, since notifications
+      // are already personalized and deletes do not matter
+      if (needsFiltering && event === 'update') {
+        pgPool.connect((err, client, done) => {
           if (err) {
             log.error(err)
             return
           }
 
-          if (result.rows.length > 0) {
+          const unpackedPayload  = JSON.parse(payload)
-            return
+          const targetAccountIds = [unpackedPayload.account.id].concat(unpackedPayload.mentions.map(item => item.id)).concat(unpackedPayload.reblog ? [unpackedPayload.reblog.account.id] : [])
-          }
 
-          transmit()
+          client.query(`SELECT target_account_id FROM blocks WHERE account_id = $1 AND target_account_id IN (${placeholders(targetAccountIds, 1)}) UNION SELECT target_account_id FROM mutes WHERE account_id = $1 AND target_account_id IN (${placeholders(targetAccountIds, 1)})`, [req.accountId].concat(targetAccountIds), (err, result) => {
+            done()
+
+            if (err) {
+              log.error(err)
+              return
+            }
+
+            if (result.rows.length > 0) {
+              return
+            }
+
+            transmit()
+          })
         })
-      })
+      } else {
-    } else {
+        transmit()
-      transmit()
+      }
+    }
+
+    subscribe(id, listener)
+    attachCloseHandler(id, listener)
+  }
+
+  // Setup stream output to HTTP
+  const streamToHttp = (req, res) => {
+    res.setHeader('Content-Type', 'text/event-stream')
+    res.setHeader('Transfer-Encoding', 'chunked')
+
+    const heartbeat = setInterval(() => res.write(':thump\n'), 15000)
+
+    req.on('close', () => {
+      log.verbose(req.requestId, `Ending stream for ${req.accountId}`)
+      clearInterval(heartbeat)
+    })
+
+    return (event, payload) => {
+      res.write(`event: ${event}\n`)
+      res.write(`data: ${payload}\n\n`)
     }
   }
 
-  subscribe(id, listener)
+  // Setup stream end for HTTP
-  attachCloseHandler(id, listener)
+  const streamHttpEnd = req => (id, listener) => {
-}
+    req.on('close', () => {
-
+      unsubscribe(id, listener)
-// Setup stream output to HTTP
+    })
-const streamToHttp = (req, res) => {
-  res.setHeader('Content-Type', 'text/event-stream')
-  res.setHeader('Transfer-Encoding', 'chunked')
-
-  const heartbeat = setInterval(() => res.write(':thump\n'), 15000)
-
-  req.on('close', () => {
-    log.verbose(req.requestId, `Ending stream for ${req.accountId}`)
-    clearInterval(heartbeat)
-  })
-
-  return (event, payload) => {
-    res.write(`event: ${event}\n`)
-    res.write(`data: ${payload}\n\n`)
   }
-}
 
-// Setup stream end for HTTP
+  // Setup stream output to WebSockets
-const streamHttpEnd = req => (id, listener) => {
+  const streamToWs = (req, ws) => {
-  req.on('close', () => {
+    const heartbeat = setInterval(() => ws.ping(), 15000)
-    unsubscribe(id, listener)
-  })
-}
 
-// Setup stream output to WebSockets
+    ws.on('close', () => {
-const streamToWs = (req, ws) => {
+      log.verbose(req.requestId, `Ending stream for ${req.accountId}`)
-  const heartbeat = setInterval(() => ws.ping(), 15000)
+      clearInterval(heartbeat)
+    })
 
-  ws.on('close', () => {
+    return (event, payload) => {
-    log.verbose(req.requestId, `Ending stream for ${req.accountId}`)
+      if (ws.readyState !== ws.OPEN) {
-    clearInterval(heartbeat)
+        log.error(req.requestId, 'Tried writing to closed socket')
-  })
+        return
+      }
 
-  return (event, payload) => {
+      ws.send(JSON.stringify({ event, payload }))
-    if (ws.readyState !== ws.OPEN) {
-      log.error(req.requestId, 'Tried writing to closed socket')
-      return
     }
-
-    ws.send(JSON.stringify({ event, payload }))
   }
-}
 
-// Setup stream end for WebSockets
+  // Setup stream end for WebSockets
-const streamWsEnd = ws => (id, listener) => {
+  const streamWsEnd = ws => (id, listener) => {
-  ws.on('close', () => {
+    ws.on('close', () => {
-    unsubscribe(id, listener)
+      unsubscribe(id, listener)
+    })
+
+    ws.on('error', e => {
+      unsubscribe(id, listener)
+    })
+  }
+
+  app.use(setRequestId)
+  app.use(allowCrossDomain)
+  app.use(authenticationMiddleware)
+  app.use(errorMiddleware)
+
+  app.get('/api/v1/streaming/user', (req, res) => {
+    streamFrom(`timeline:${req.accountId}`, req, streamToHttp(req, res), streamHttpEnd(req))
   })
 
-  ws.on('error', e => {
+  app.get('/api/v1/streaming/public', (req, res) => {
-    unsubscribe(id, listener)
+    streamFrom('timeline:public', req, streamToHttp(req, res), streamHttpEnd(req), true)
+  })
+
+  app.get('/api/v1/streaming/public/local', (req, res) => {
+    streamFrom('timeline:public:local', req, streamToHttp(req, res), streamHttpEnd(req), true)
+  })
+
+  app.get('/api/v1/streaming/hashtag', (req, res) => {
+    streamFrom(`timeline:hashtag:${req.params.tag}`, req, streamToHttp(req, res), streamHttpEnd(req), true)
+  })
+
+  app.get('/api/v1/streaming/hashtag/local', (req, res) => {
+    streamFrom(`timeline:hashtag:${req.params.tag}:local`, req, streamToHttp(req, res), streamHttpEnd(req), true)
+  })
+
+  wss.on('connection', ws => {
+    const location = url.parse(ws.upgradeReq.url, true)
+    const token    = location.query.access_token
+    const req      = { requestId: uuid.v4() }
+
+    accountFromToken(token, req, err => {
+      if (err) {
+        log.error(req.requestId, err)
+        ws.close()
+        return
+      }
+
+      switch(location.query.stream) {
+      case 'user':
+        streamFrom(`timeline:${req.accountId}`, req, streamToWs(req, ws), streamWsEnd(ws))
+        break;
+      case 'public':
+        streamFrom('timeline:public', req, streamToWs(req, ws), streamWsEnd(ws), true)
+        break;
+      case 'public:local':
+        streamFrom('timeline:public:local', req, streamToWs(req, ws), streamWsEnd(ws), true)
+        break;
+      case 'hashtag':
+        streamFrom(`timeline:hashtag:${location.query.tag}`, req, streamToWs(req, ws), streamWsEnd(ws), true)
+        break;
+      case 'hashtag:local':
+        streamFrom(`timeline:hashtag:${location.query.tag}:local`, req, streamToWs(req, ws), streamWsEnd(ws), true)
+        break;
+      default:
+        ws.close()
+      }
+    })
+  })
+
+  server.listen(process.env.PORT || 4000, () => {
+    log.level = process.env.LOG_LEVEL || 'verbose'
+    log.info(`Starting streaming API server worker on port ${server.address().port}`)
   })
 }
-
-app.use(setRequestId)
-app.use(allowCrossDomain)
-app.use(authenticationMiddleware)
-app.use(errorMiddleware)
-
-app.get('/api/v1/streaming/user', (req, res) => {
-  streamFrom(`timeline:${req.accountId}`, req, streamToHttp(req, res), streamHttpEnd(req))
-})
-
-app.get('/api/v1/streaming/public', (req, res) => {
-  streamFrom('timeline:public', req, streamToHttp(req, res), streamHttpEnd(req), true)
-})
-
-app.get('/api/v1/streaming/public/local', (req, res) => {
-  streamFrom('timeline:public:local', req, streamToHttp(req, res), streamHttpEnd(req), true)
-})
-
-app.get('/api/v1/streaming/hashtag', (req, res) => {
-  streamFrom(`timeline:hashtag:${req.params.tag}`, req, streamToHttp(req, res), streamHttpEnd(req), true)
-})
-
-app.get('/api/v1/streaming/hashtag/local', (req, res) => {
-  streamFrom(`timeline:hashtag:${req.params.tag}:local`, req, streamToHttp(req, res), streamHttpEnd(req), true)
-})
-
-wss.on('connection', ws => {
-  const location = url.parse(ws.upgradeReq.url, true)
-  const token    = location.query.access_token
-  const req      = { requestId: uuid.v4() }
-
-  accountFromToken(token, req, err => {
-    if (err) {
-      log.error(req.requestId, err)
-      ws.close()
-      return
-    }
-
-    switch(location.query.stream) {
-    case 'user':
-      streamFrom(`timeline:${req.accountId}`, req, streamToWs(req, ws), streamWsEnd(ws))
-      break;
-    case 'public':
-      streamFrom('timeline:public', req, streamToWs(req, ws), streamWsEnd(ws), true)
-      break;
-    case 'public:local':
-      streamFrom('timeline:public:local', req, streamToWs(req, ws), streamWsEnd(ws), true)
-      break;
-    case 'hashtag':
-      streamFrom(`timeline:hashtag:${location.query.tag}`, req, streamToWs(req, ws), streamWsEnd(ws), true)
-      break;
-    case 'hashtag:local':
-      streamFrom(`timeline:hashtag:${location.query.tag}:local`, req, streamToWs(req, ws), streamWsEnd(ws), true)
-      break;
-    default:
-      ws.close()
-    }
-  })
-})
-
-server.listen(process.env.PORT || 4000, () => {
-  log.level = process.env.LOG_LEVEL || 'verbose'
-  log.info(`Starting streaming API server on port ${server.address().port}`)
-})