@cloudant/couchbackup 2.9.10-SNAPSHOT.113 → 2.9.10

package/.scannerwork/scanner-report/source-10.txt

@@ -1,4 +1,3 @@
-#!/usr/bin/env node
 // Copyright © 2017, 2021 IBM Corp. All rights reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
@@ -14,64 +13,269 @@
 // limitations under the License.
 'use strict';
 
-const error = require('../includes/error.js');
+const async = require('async');
+const events = require('events');
 const fs = require('fs');
-const cliutils = require('../includes/cliutils.js');
-const couchbackup = require('../app.js');
-const parser = require('../includes/parser.js');
-const debug = require('debug');
-const backupDebug = debug('couchbackup:backup');
-const backupBatchDebug = debug('couchbackup:backup:batch');
-
-backupDebug.enabled = true;
-
-const program = parser.parseBackupArgs();
-
-const databaseUrl = cliutils.databaseUrl(program.url, program.db);
-const opts = {
-  bufferSize: program.bufferSize,
-  log: program.log,
-  mode: program.mode,
-  parallelism: program.parallelism,
-  requestTimeout: program.requestTimeout,
-  resume: program.resume,
-  iamApiKey: program.iamApiKey,
-  iamTokenUrl: program.iamTokenUrl
+const error = require('./error.js');
+const spoolchanges = require('./spoolchanges.js');
+const logfilesummary = require('./logfilesummary.js');
+const logfilegetbatches = require('./logfilegetbatches.js');
+
+/**
+ * Read documents from a database to be backed up.
+ *
+ * @param {string} db - `@cloudant/cloudant` DB object for source database.
+ * @param {number} blocksize - number of documents to download in single request
+ * @param {number} parallelism - number of concurrent downloads
+ * @param {string} log - path to log file to use
+ * @param {boolean} resume - whether to resume from an existing log file
+ * @returns EventEmitter with following events:
+ *  - `received` - called with a block of documents to write to backup
+ *  - `error` - on error
+ *  - `finished` - when backup process is finished (either complete or errored)
+ */
+module.exports = function(db, options) {
+  const ee = new events.EventEmitter();
+  const start = new Date().getTime(); // backup start time
+  const batchesPerDownloadSession = 50; // max batches to read from log file for download at a time (prevent OOM)
+
+  function proceedWithBackup() {
+    if (options.resume) {
+      // pick up from existing log file from previous run
+      downloadRemainingBatches(options.log, db, ee, start, batchesPerDownloadSession, options.parallelism);
+    } else {
+      // create new log file and process
+      spoolchanges(db, options.log, options.bufferSize, ee, function(err) {
+        if (err) {
+          ee.emit('error', err);
+        } else {
+          downloadRemainingBatches(options.log, db, ee, start, batchesPerDownloadSession, options.parallelism);
+        }
+      });
+    }
+  }
+
+  validateBulkGetSupport(db, function(err) {
+    if (err) {
+      return ee.emit('error', err);
+    } else {
+      proceedWithBackup();
+    }
+  });
+
+  return ee;
 };
 
-// log configuration to console
-console.error('='.repeat(80));
-console.error('Performing backup on ' + databaseUrl.replace(/\/\/.+@/g, '//****:****@') + ' using configuration:');
-console.error(JSON.stringify(opts, null, 2).replace(/"iamApiKey": "[^"]+"/, '"iamApiKey": "****"'));
-console.error('='.repeat(80));
+/**
+ * Validate /_bulk_get support for a specified database.
+ *
+ * @param {string} db - nodejs-cloudant db
+ * @param {function} callback - called on completion with signature (err)
+ */
+function validateBulkGetSupport(db, callback) {
+  db.service.postBulkGet({ db: db.db, docs: [] }).then(() => { callback(); }).catch(err => {
+    err = error.convertResponseError(err, function(err) {
+      switch (err.status) {
+        case undefined:
+          // There was no status code on the error
+          return err;
+        case 404:
+          return new error.BackupError('BulkGetError', 'Database does not support /_bulk_get endpoint');
+        default:
+          return new error.HTTPError(err);
+      }
+    });
+    callback(err);
+  });
+}
+
+/**
+ * Download remaining batches in a log file, splitting batches into sets
+ * to avoid enqueueing too many in one go.
+ *
+ * @param {string} log - log file name to maintain download state
+ * @param {string} db - nodejs-cloudant db
+ * @param {events.EventEmitter} ee - event emitter to emit received events on
+ * @param {time} startTime - start time for backup process
+ * @param {number} batchesPerDownloadSession - max batches to enqueue for
+ *  download at a time. As batches contain many doc IDs, this helps avoid
+ *  exhausting memory.
+ * @param {number} parallelism - number of concurrent downloads
+ * @returns function to call do download remaining batches with signature
+ *  (err, {batches: batch, docs: doccount}) {@see spoolchanges}.
+ */
+function downloadRemainingBatches(log, db, ee, startTime, batchesPerDownloadSession, parallelism) {
+  let total = 0; // running total of documents downloaded so far
+  let noRemainingBatches = false;
+
+  // Generate a set of batches (up to batchesPerDownloadSession) to download from the
+  // log file and download them. Set noRemainingBatches to `true` for last batch.
+  function downloadSingleBatchSet(done) {
+    // Fetch the doc IDs for the batches in the current set to
+    // download them.
+    function batchSetComplete(err, data) {
+      if (!err) {
+        total = data.total;
+      }
+      done(err);
+    }
+    function processRetrievedBatches(err, batches) {
+      if (!err) {
+        // process them in parallelised queue
+        processBatchSet(db, parallelism, log, batches, ee, startTime, total, batchSetComplete);
+      } else {
+        batchSetComplete(err);
+      }
+    }
 
-backupBatchDebug.enabled = !program.quiet;
+    readBatchSetIdsFromLogFile(log, batchesPerDownloadSession, function(err, batchSetIds) {
+      if (err) {
+        ee.emit('error', err);
+        // Stop processing changes file for fatal errors
+        noRemainingBatches = true;
+        done();
+      } else {
+        if (batchSetIds.length === 0) {
+          noRemainingBatches = true;
+          return done();
+        }
+        logfilegetbatches(log, batchSetIds, processRetrievedBatches);
+      }
+    });
+  }
 
-let ws = process.stdout;
+  // Return true if all batches in log file have been downloaded
+  function isFinished(callback) { callback(null, noRemainingBatches); }
 
-// open output file
-if (program.output) {
-  let flags = 'w';
-  if (program.log && program.resume) {
-    flags = 'a';
+  function onComplete() {
+    ee.emit('finished', { total: total });
   }
-  const fd = fs.openSync(program.output, flags);
-  ws = fs.createWriteStream(null, { fd: fd });
+
+  async.doUntil(downloadSingleBatchSet, isFinished, onComplete);
 }
 
-backupDebug('Fetching all database changes...');
-
-return couchbackup.backup(
-  databaseUrl,
-  ws,
-  opts,
-  error.terminationCallback
-).on('changes', function(batch) {
-  backupBatchDebug('Total batches received:', batch + 1);
-}).on('written', function(obj) {
-  backupBatchDebug('Written batch ID:', obj.batch, 'Total document revisions written:', obj.total, 'Time:', obj.time);
-}).on('error', function(e) {
-  backupDebug('ERROR', e);
-}).on('finished', function(obj) {
-  backupDebug('Finished - Total document revisions written:', obj.total);
-});
+/**
+ * Return a set of uncompleted download batch IDs from the log file.
+ *
+ * @param {string} log - log file path
+ * @param {number} batchesPerDownloadSession - maximum IDs to return
+ * @param {function} callback - sign (err, batchSetIds array)
+ */
+function readBatchSetIdsFromLogFile(log, batchesPerDownloadSession, callback) {
+  logfilesummary(log, function processSummary(err, summary) {
+    if (!err) {
+      if (!summary.changesComplete) {
+        callback(new error.BackupError('IncompleteChangesInLogFile',
+          'WARNING: Changes did not finish spooling'));
+        return;
+      }
+      if (Object.keys(summary.batches).length === 0) {
+        return callback(null, []);
+      }
+
+      // batch IDs are the property names of summary.batches
+      const batchSetIds = getPropertyNames(summary.batches, batchesPerDownloadSession);
+      callback(null, batchSetIds);
+    } else {
+      callback(err);
+    }
+  });
+}
+
+/**
+ * Download a set of batches retrieved from a log file. When a download is
+ * complete, add a line to the logfile indicating such.
+ *
+ * @param {any} db - nodejs-cloudant database
+ * @param {any} parallelism - number of concurrent requests to make
+ * @param {any} log - log file to drive downloads from
+ * @param {any} batches - batches to download
+ * @param {any} ee - event emitter for progress. This funciton emits
+ *  received and error events.
+ * @param {any} start - time backup started, to report deltas
+ * @param {any} grandtotal - count of documents downloaded prior to this set
+ *  of batches
+ * @param {any} callback - completion callback, (err, {total: number}).
+ */
+function processBatchSet(db, parallelism, log, batches, ee, start, grandtotal, callback) {
+  let hasErrored = false;
+  let total = grandtotal;
+
+  // queue to process the fetch requests in an orderly fashion using _bulk_get
+  const q = async.queue(function(payload, done) {
+    const output = [];
+    const thisBatch = payload.batch;
+    delete payload.batch;
+    delete payload.command;
+
+    function logCompletedBatch(batch) {
+      if (log) {
+        fs.appendFile(log, ':d batch' + thisBatch + '\n', done);
+      } else {
+        done();
+      }
+    }
+
+    // do the /db/_bulk_get request
+    db.service.postBulkGet({
+      db: db.db,
+      revs: true,
+      docs: payload.docs
+    }).then(response => {
+      // create an output array with the docs returned
+      response.result.results.forEach(function(d) {
+        if (d.docs) {
+          d.docs.forEach(function(doc) {
+            if (doc.ok) {
+              output.push(doc.ok);
+            }
+          });
+        }
+      });
+      total += output.length;
+      const t = (new Date().getTime() - start) / 1000;
+      ee.emit('received', {
+        batch: thisBatch,
+        data: output,
+        length: output.length,
+        time: t,
+        total: total
+      }, q, logCompletedBatch);
+    }).catch(err => {
+      if (!hasErrored) {
+        hasErrored = true;
+        err = error.convertResponseError(err);
+        // Kill the queue for fatal errors
+        q.kill();
+        ee.emit('error', err);
+      }
+      done();
+    });
+  }, parallelism);
+
+  for (const i in batches) {
+    q.push(batches[i]);
+  }
+
+  q.drain(function() {
+    callback(null, { total: total });
+  });
+}
+
+/**
+ * Returns first N properties on an object.
+ *
+ * @param {object} obj - object with properties
+ * @param {number} count - number of properties to return
+ */
+function getPropertyNames(obj, count) {
+  // decide which batch numbers to deal with
+  const batchestofetch = [];
+  let j = 0;
+  for (const i in obj) {
+    batchestofetch.push(parseInt(i));
+    j++;
+    if (j >= count) break;
+  }
+  return batchestofetch;
+}

package/.scannerwork/scanner-report/source-11.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017 IBM Corp. All rights reserved.
+// Copyright © 2017, 2021 IBM Corp. All rights reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -11,104 +11,108 @@
 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 // See the License for the specific language governing permissions and
 // limitations under the License.
-
-/* global after before describe */
 'use strict';
 
-const url = require('url');
-const toxy = require('toxy');
-// Import the common hooks
-require('../test/hooks.js');
-
-const tpoisons = toxy.poisons;
-const trules = toxy.rules;
-
-function setupProxy(poison) {
-  const backendUrl = new url.URL(process.env.COUCH_BACKEND_URL);
-  const proxy = toxy({
-    auth: `${backendUrl.username}:${backendUrl.password}`,
-    changeOrigin: true
-  });
-
-  // Forward traffic to DB
-  proxy.forward(process.env.COUCH_BACKEND_URL);
-
-  switch (poison) {
-    case 'normal':
-      // No poisons to add
-      break;
-    case 'bandwidth-limit':
-      // https://github.com/h2non/toxy#bandwidth
-      // Note the implementation of bandwidth is simplistic and the threshold
-      // delay is applied to every write of the buffer, so use the smallest
-      // delay possible and adjust the rate using the bytes size instead.
-      proxy
-        .poison(tpoisons.bandwidth({ bytes: 512, threshold: 1 })); // 0.5 MB/s
-      break;
-    case 'latency':
-      // https://github.com/h2non/toxy#latency
-      proxy
-        .poison(tpoisons.latency({ max: 1500, min: 250 }))
-        .withRule(trules.probability(60));
-      break;
-    case 'slow-read':
-      // https://github.com/h2non/toxy#slow-read
-      // Note this only impacts read of data from requests so only for non-GET
-      // In practice this means that it impacts restore much more than backup
-      // since although backup POSTs to _bulk_get the content is much smaller
-      // than what is POSTed to _bulk_docs for a restore.
-      // Similarly to bandwidth-limit use a 1 ms threshold
-      proxy
-        .poison(tpoisons.slowRead({ chunk: 256, threshold: 1 }))
-        // Slow read for 10 % of the time e.g. 10 ms in every 100
-        .withRule(trules.timeThreshold({ duration: 10, period: 100 }));
-      break;
-    case 'rate-limit':
-      // https://github.com/h2non/toxy#rate-limit
-      // Simulate the Cloudant free plan with 20 lookups ps and 10 writes ps
-      proxy.post('/*/_bulk_get')
-        .poison(tpoisons.rateLimit({ limit: 20, threshold: 1000 }));
-      proxy.post('/*/_bulk_docs')
-        .poison(tpoisons.rateLimit({ limit: 10, threshold: 1000 }));
-      break;
-    default:
-      throw Error('Unknown toxy poison ' + poison);
+const path = require('path');
+const tmp = require('tmp');
+
+/**
+  Return API default settings.
+*/
+function apiDefaults() {
+  return {
+    parallelism: 5,
+    bufferSize: 500,
+    requestTimeout: 120000,
+    log: tmp.tmpNameSync(),
+    resume: false,
+    mode: 'full'
+  };
+}
+
+/**
+  Return CLI default settings.
+*/
+function cliDefaults() {
+  const defaults = apiDefaults();
+
+  // add additional legacy settings
+  defaults.db = 'test';
+  defaults.url = 'http://localhost:5984';
+
+  // add CLI only option
+  defaults.quiet = false;
+
+  return defaults;
+}
+
+/**
+  Override settings **in-place** with environment variables.
+*/
+function applyEnvironmentVariables(opts) {
+  // if we have a custom CouchDB url
+  if (typeof process.env.COUCH_URL !== 'undefined') {
+    opts.url = process.env.COUCH_URL;
+  }
+
+  // if we have a specified databases
+  if (typeof process.env.COUCH_DATABASE !== 'undefined') {
+    opts.db = process.env.COUCH_DATABASE;
+  }
+
+  // if we have a specified buffer size
+  if (typeof process.env.COUCH_BUFFER_SIZE !== 'undefined') {
+    opts.bufferSize = parseInt(process.env.COUCH_BUFFER_SIZE);
+  }
+
+  // if we have a specified parallelism
+  if (typeof process.env.COUCH_PARALLELISM !== 'undefined') {
+    opts.parallelism = parseInt(process.env.COUCH_PARALLELISM);
   }
 
-  // Catch remaining traffic
-  proxy.all('/*');
-  return proxy;
+  // if we have a specified request timeout
+  if (typeof process.env.COUCH_REQUEST_TIMEOUT !== 'undefined') {
+    opts.requestTimeout = parseInt(process.env.COUCH_REQUEST_TIMEOUT);
+  }
+
+  // if we have a specified log file
+  if (typeof process.env.COUCH_LOG !== 'undefined') {
+    opts.log = path.normalize(process.env.COUCH_LOG);
+  }
+
+  // if we are instructed to resume
+  if (typeof process.env.COUCH_RESUME !== 'undefined' && process.env.COUCH_RESUME === 'true') {
+    opts.resume = true;
+  }
+
+  // if we are given an output filename
+  if (typeof process.env.COUCH_OUTPUT !== 'undefined') {
+    opts.output = path.normalize(process.env.COUCH_OUTPUT);
+  }
+
+  // if we only want a shallow copy
+  if (typeof process.env.COUCH_MODE !== 'undefined' && process.env.COUCH_MODE === 'shallow') {
+    opts.mode = 'shallow';
+  }
+
+  // if we are instructed to be quiet
+  if (typeof process.env.COUCH_QUIET !== 'undefined' && process.env.COUCH_QUIET === 'true') {
+    opts.quiet = true;
+  }
+
+  // if we have a specified API key
+  if (typeof process.env.CLOUDANT_IAM_API_KEY !== 'undefined') {
+    opts.iamApiKey = process.env.CLOUDANT_IAM_API_KEY;
+  }
+
+  // if we have a specified IAM token endpoint
+  if (typeof process.env.CLOUDANT_IAM_TOKEN_URL !== 'undefined') {
+    opts.iamTokenUrl = process.env.CLOUDANT_IAM_TOKEN_URL;
+  }
 }
 
-const poisons = [
-  'normal',
-  'bandwidth-limit',
-  'latency',
-  'slow-read',
-  'rate-limit'
-];
-
-poisons.forEach(function(poison) {
-  describe('unreliable network tests (using toxy poison ' + poison + ')', function() {
-    let proxy;
-
-    before('start toxy server', function() {
-      proxy = setupProxy(poison);
-      console.log('Using toxy poison ' + poison);
-
-      // For these tests COUCH_URL points to the toxy proxy on localhost whereas
-      // COUCH_BACKEND_URL is the real CouchDb instance.
-      const toxyUrl = new url.URL(process.env.COUCH_URL);
-      // Listen on the specified hostname only, so if using localhost we don't
-      // need external connections.
-      proxy.listen(toxyUrl.port, toxyUrl.hostname);
-    });
-
-    after('stop toxy server', function() {
-      proxy.close();
-    });
-
-    delete require.cache[require.resolve('../test/ci_e2e.js')];
-    require('../test/ci_e2e.js');
-  });
-});
+module.exports = {
+  apiDefaults: apiDefaults,
+  cliDefaults: cliDefaults,
+  applyEnvironmentVariables: applyEnvironmentVariables
+};