Merge pull request #65 from overleaf/spd-migration-script

Add script to migrate data from S3

services/docstore/scripts/rearchive-all-docs.js

@@ -0,0 +1,123 @@
+const DocArchiveManager = require('../app/js/DocArchiveManager').promises
+const MongoManager = require('../app/js/MongoManager').promises
+const { getCollection, ObjectId } = require('../app/js/mongodb')
+const minimist = require('minimist')
+
+async function worker(projectId) {
+  try {
+    // see if the project needs to be unarchived, and unarchive it
+    const archivedDocs = await MongoManager.getArchivedProjectDocs(projectId)
+    if (archivedDocs.length) {
+      await DocArchiveManager.unArchiveAllDocs(projectId)
+    }
+
+    // get the doc content so we can validate it
+    const docs = await MongoManager.getProjectsDocs(
+      projectId,
+      { include_deleted: false },
+      { lines: 1 }
+    )
+
+    // start archiving in the background while we check the content, if it was archived to begin with
+    let archivePromise
+    if (archivedDocs.length) {
+      archivePromise = DocArchiveManager.archiveAllDocs(projectId)
+    }
+
+    let warning = false
+
+    // validate the doc contents and log any warnings to investigate later
+    for (const doc of docs) {
+      if (!doc.lines) {
+        warning = true
+        console.error('WARN:', projectId, doc._id, 'has no content')
+      }
+      // eslint-disable-next-line no-control-regex
+      if (doc.lines && doc.lines.some((line) => line.match(/(\r|\u0000)/))) {
+        warning = true
+        console.error('WARN:', projectId, doc._id, 'has invalid characters')
+      }
+    }
+
+    // ensure the archive process has finished
+    if (archivePromise) {
+      await archivePromise
+    }
+
+    if (!warning) {
+      // log to stderr along with the other output
+      console.error('OK:', projectId)
+    }
+  } catch (err) {
+    console.error('ERROR:', projectId, err)
+  }
+}
+
+async function rearchiveAllDocs() {
+  const params = minimist(process.argv.slice(2))
+  const maxWorkers = params.w || 1
+  console.log(`Starting with ${maxWorkers} workers`)
+
+  // start from an objectId and run in ascending order, so we can resume later
+  const query = {}
+  const startId = params._[0]
+  if (startId) {
+    const validator = new RegExp('^[0-9a-fA-F]{24}$')
+    if (!validator.test(startId)) {
+      console.error('Invalid object id')
+      return
+    }
+    query._id = {
+      $gt: ObjectId(startId)
+    }
+    console.log(`Starting from object ID ${startId}`)
+  } else {
+    console.log('No object id specified. Starting from the beginning.')
+  }
+
+  const results = (await getCollection('projects'))
+    .find(query, { _id: 1 })
+    .sort({ _id: 1 })
+  let jobCount = 0
+
+  // keep going until we run out of projects
+  while (true) {
+    // get a new project to run a job with
+    const project = await results.next()
+    // if there are no more projects, wait until all the jobs have finished and exit
+    if (!project) {
+      // eslint-disable-next-line no-unmodified-loop-condition
+      while (jobCount) {
+        await new Promise((resolve) => setTimeout(resolve, 50))
+      }
+      return
+    }
+
+    // wait until there are fewer than maxWorkers jobs running
+    // eslint-disable-next-line no-unmodified-loop-condition
+    while (jobCount >= maxWorkers) {
+      await new Promise((resolve) => setTimeout(resolve, 50))
+    }
+
+    // start a new job in the background and then continue the loop
+    ++jobCount
+    worker(project._id)
+      .then(() => --jobCount)
+      .catch(() => {
+        console.error('ERROR:', project._id)
+        --jobCount
+      })
+  }
+}
+
+if (!module.parent) {
+  rearchiveAllDocs()
+    .then(() => {
+      console.log('Finished!')
+      process.exit(0)
+    })
+    .catch((err) => {
+      console.error('Something went wrong:', err)
+      process.exit(1)
+    })
+}