@cloudant/couchbackup 2.9.13 → 2.9.14-SNAPSHOT.147

package/.scannerwork/scanner-report/source-55.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017, 2021 IBM Corp. All rights reserved.
+// Copyright © 2017, 2022 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.
@@ -13,117 +13,99 @@
 // limitations under the License.
 'use strict';
 
-const cliutils = require('./cliutils.js');
-const config = require('./config.js');
+const fs = require('fs');
+const liner = require('./liner.js');
+const change = require('./change.js');
 const error = require('./error.js');
-const path = require('path');
-const pkg = require('../package.json');
-
-function parseBackupArgs() {
-  const program = require('commander');
-
-  // Option CLI defaults
-  const defaults = config.cliDefaults();
-
-  // Options set by environment variables
-  const envVarOptions = {};
-  config.applyEnvironmentVariables(envVarOptions);
-
-  program
-    .version(pkg.version)
-    .description('Backup a CouchDB/Cloudant database to a backup text file.')
-    .usage('[options...]')
-    .option('-b, --buffer-size <n>',
-      cliutils.getUsage('number of documents fetched at once', defaults.bufferSize),
-      Number)
-    .option('-d, --db <db>',
-      cliutils.getUsage('name of the database to backup', defaults.db))
-    .option('-k, --iam-api-key <API key>',
-      cliutils.getUsage('IAM API key to access the Cloudant server'))
-    .option('-l, --log <file>',
-      cliutils.getUsage('file to store logging information during backup; invalid in "shallow" mode', 'a temporary file'),
-      path.normalize)
-    .option('-m, --mode <mode>',
-      cliutils.getUsage('"shallow" if only a superficial backup is done (ignoring conflicts and revision tokens), else "full" for complete backup', defaults.mode),
-      (mode) => { return mode.toLowerCase(); })
-    .option('-o, --output <file>',
-      cliutils.getUsage('file name to store the backup data', 'stdout'),
-      path.normalize)
-    .option('-p, --parallelism <n>',
-      cliutils.getUsage('number of HTTP requests to perform in parallel when performing a backup; ignored in "shallow" mode', defaults.parallelism),
-      Number)
-    .option('-q, --quiet',
-      cliutils.getUsage('suppress batch messages', defaults.quiet))
-    .option('-r, --resume',
-      cliutils.getUsage('continue a previous backup from its last known position; invalid in "shallow" mode', defaults.resume))
-    .option('-t, --request-timeout <n>',
-      cliutils.getUsage('milliseconds to wait for a response to a HTTP request before retrying the request', defaults.requestTimeout),
-      Number)
-    .option('-u, --url <url>',
-      cliutils.getUsage('URL of the CouchDB/Cloudant server', defaults.url))
-    .parse(process.argv);
-
-  // Remove defaults that don't apply when using shallow mode
-  if (program.opts().mode === 'shallow' || envVarOptions.mode === 'shallow') {
-    delete defaults.parallelism;
-    delete defaults.log;
-    delete defaults.resume;
-  }
-
-  // Apply the options in order so that the CLI overrides env vars and env variables
-  // override defaults.
-  const opts = Object.assign({}, defaults, envVarOptions, program.opts());
-
-  if (opts.resume && (opts.log === defaults.log)) {
-    // If resuming and the log file arg is the newly generated tmp name from defaults then we know that --log wasn't specified.
-    // We have to do this check here for the CLI case because of the default.
-    error.terminationCallback(new error.BackupError('NoLogFileName', 'To resume a backup, a log file must be specified'));
+const debug = require('debug')('couchbackup:spoolchanges');
+
+/**
+ * Write log file for all changes from a database, ready for downloading
+ * in batches.
+ *
+ * @param {string} dbUrl - URL of database
+ * @param {string} log - path to log file to use
+ * @param {number} bufferSize - the number of changes per batch/log line
+ * @param {function(err)} callback - a callback to run on completion
+ */
+module.exports = function(db, log, bufferSize, ee, callback) {
+  // list of document ids to process
+  const buffer = [];
+  let batch = 0;
+  let lastSeq = null;
+  const logStream = fs.createWriteStream(log);
+  let pending = 0;
+  // The number of changes to fetch per request
+  const limit = 100000;
+
+  // send documents ids to the queue in batches of bufferSize + the last batch
+  const processBuffer = function(lastOne) {
+    if (buffer.length >= bufferSize || (lastOne && buffer.length > 0)) {
+      debug('writing', buffer.length, 'changes to the backup file');
+      const b = { docs: buffer.splice(0, bufferSize), batch: batch };
+      logStream.write(':t batch' + batch + ' ' + JSON.stringify(b.docs) + '\n');
+      ee.emit('changes', batch);
+      batch++;
+    }
+  };
+
+  // called once per received change
+  const onChange = function(c) {
+    if (c) {
+      if (c.error) {
+        ee.emit('error', new error.BackupError('InvalidChange', `Received invalid change: ${c}`));
+      } else if (c.changes) {
+        const obj = { id: c.id };
+        buffer.push(obj);
+        processBuffer(false);
+      } else if (c.last_seq) {
+        lastSeq = c.last_seq;
+        pending = c.pending;
+      }
+    }
+  };
+
+  function getChanges(since = 0) {
+    debug('making changes request since ' + since);
+    return db.service.postChangesAsStream({ db: db.db, since: since, limit: limit, seqInterval: limit })
+      .then(response => {
+        response.result.pipe(liner())
+          .on('error', function(err) {
+            logStream.end();
+            callback(err);
+          })
+          .pipe(change(onChange))
+          .on('error', function(err) {
+            logStream.end();
+            callback(err);
+          })
+          .on('finish', function() {
+            processBuffer(true);
+            if (!lastSeq) {
+              logStream.end();
+              debug('changes request terminated before last_seq was sent');
+              callback(new error.BackupError('SpoolChangesError', 'Changes request terminated before last_seq was sent'));
+            } else {
+              debug(`changes request completed with last_seq: ${lastSeq} and ${pending} changes pending.`);
+              if (pending > 0) {
+                // Return the next promise
+                return getChanges(lastSeq);
+              } else {
+                debug('finished streaming database changes');
+                logStream.end(':changes_complete ' + lastSeq + '\n', 'utf8', callback);
+              }
+            }
+          });
+      })
+      .catch(err => {
+        logStream.end();
+        if (err.status && err.status >= 400) {
+          callback(error.convertResponseError(err));
+        } else if (err.name !== 'SpoolChangesError') {
+          callback(new error.BackupError('SpoolChangesError', `Failed changes request - ${err.message}`));
+        }
+      });
   }
 
-  return opts;
-}
-
-function parseRestoreArgs() {
-  const program = require('commander');
-
-  // Option CLI defaults
-  const defaults = config.cliDefaults();
-
-  // Options set by environment variables
-  const envVarOptions = {};
-  config.applyEnvironmentVariables(envVarOptions);
-
-  program
-    .version(pkg.version)
-    .description('Restore a CouchDB/Cloudant database from a backup text file.')
-    .usage('[options...]')
-    .option('-b, --buffer-size <n>',
-      cliutils.getUsage('number of documents restored at once', defaults.bufferSize),
-      Number)
-    .option('-d, --db <db>',
-      cliutils.getUsage('name of the new, existing database to restore to', defaults.db))
-    .option('-k, --iam-api-key <API key>',
-      cliutils.getUsage('IAM API key to access the Cloudant server'))
-    .option('-p, --parallelism <n>',
-      cliutils.getUsage('number of HTTP requests to perform in parallel when restoring a backup', defaults.parallelism),
-      Number)
-    .option('-q, --quiet',
-      cliutils.getUsage('suppress batch messages', defaults.quiet))
-    .option('-t, --request-timeout <n>',
-      cliutils.getUsage('milliseconds to wait for a response to a HTTP request before retrying the request', defaults.requestTimeout),
-      Number)
-    .option('-u, --url <url>',
-      cliutils.getUsage('URL of the CouchDB/Cloudant server', defaults.url))
-    .parse(process.argv);
-
-  // Apply the options in order so that the CLI overrides env vars and env variables
-  // override defaults.
-  const opts = Object.assign({}, defaults, envVarOptions, program.opts());
-
-  return opts;
-}
-
-module.exports = {
-  parseBackupArgs: parseBackupArgs,
-  parseRestoreArgs: parseRestoreArgs
+  getChanges();
 };

package/.scannerwork/scanner-report/source-56.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017, 2021 IBM Corp. All rights reserved.
+// Copyright © 2017 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.
@@ -13,269 +13,63 @@
 // limitations under the License.
 'use strict';
 
-const async = require('async');
-const events = require('events');
 const fs = require('fs');
-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;
-};
-
-/**
- * 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);
+const stream = require('stream');
+const liner = require('./liner.js');
+
+const onLine = function(onCommand, batches) {
+  const change = new stream.Transform({ objectMode: true });
+  change._transform = function(line, encoding, done) {
+    if (line && line[0] === ':') {
+      const obj = {
+        command: null,
+        batch: null,
+        docs: []
+      };
+
+      let matches;
+
+      // extract command
+      matches = line.match(/^:([a-z_]+) ?/);
+      if (matches) {
+        obj.command = matches[1];
       }
-    });
-    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);
-      }
-    }
 
-    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);
+      // extract batch
+      matches = line.match(/ batch([0-9]+)/);
+      if (matches) {
+        obj.batch = parseInt(matches[1]);
       }
-    });
-  }
 
-  // Return true if all batches in log file have been downloaded
-  function isFinished(callback) { callback(null, noRemainingBatches); }
-
-  function onComplete() {
-    ee.emit('finished', { total: total });
-  }
-
-  async.doUntil(downloadSingleBatchSet, isFinished, onComplete);
-}
-
-/**
- * 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, []);
+      // if this is one we want
+      if (obj.command === 't' && batches.indexOf(obj.batch) > -1) {
+        const json = line.replace(/^.* batch[0-9]+ /, '').trim();
+        obj.docs = JSON.parse(json);
+        onCommand(obj);
       }
-
-      // 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;
+    done();
+  };
+  return change;
+};
 
-  // 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;
+module.exports = function(log, batches, callback) {
+  // our sense of state
+  const retval = { };
 
-    function logCompletedBatch(batch) {
-      if (log) {
-        fs.appendFile(log, ':d batch' + thisBatch + '\n', done);
-      } else {
-        done();
-      }
-    }
+  // called with each line from the log file
+  const onCommand = function(obj) {
+    retval[obj.batch] = obj;
+  };
 
-    // 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();
+  // stream through the previous log file
+  fs.createReadStream(log)
+    .pipe(liner())
+    .pipe(onLine(onCommand, batches))
+    .on('error', function(err) {
+      callback(err);
+    })
+    .on('finish', function() {
+      callback(null, retval);
     });
-  }, 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;
-}
+};