@cloudant/couchbackup 2.9.10-SNAPSHOT.112 → 2.9.10

package/.scannerwork/scanner-report/source-59.txt

@@ -1,5 +1,4 @@
-#!/usr/bin/env node
-// Copyright © 2017, 2021 IBM Corp. All rights reserved.
+// Copyright © 2017, 2018 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.
@@ -12,66 +11,180 @@
 // 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.
+
+// Small script which backs up a Cloudant or CouchDB database to an S3
+// bucket, using an intermediary file on disk.
+//
+// The script generates the backup object name by combining together the path
+// part of the database URL and the current time.
+
 'use strict';
 
-const error = require('../includes/error.js');
+const stream = require('stream');
 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
-};
-
-// 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));
-
-backupBatchDebug.enabled = !program.quiet;
-
-let ws = process.stdout;
-
-// open output file
-if (program.output) {
-  let flags = 'w';
-  if (program.log && program.resume) {
-    flags = 'a';
+const url = require('url');
+
+const AWS = require('aws-sdk');
+const couchbackup = require('@cloudant/couchbackup');
+const debug = require('debug')('s3-backup');
+const tmp = require('tmp');
+const VError = require('verror').VError;
+
+/*
+  Main function, run from base of file.
+*/
+function main() {
+  const argv = require('yargs')
+    .usage('Usage: $0 [options]')
+    .example('$0 -s https://user:pass@host/db -b <bucket>', 'Backup db to bucket')
+    .options({
+      source: { alias: 's', nargs: 1, demandOption: true, describe: 'Source database URL' },
+      bucket: { alias: 'b', nargs: 1, demandOption: true, describe: 'Destination bucket' },
+      prefix: { alias: 'p', nargs: 1, describe: 'Prefix for backup object key', default: 'couchbackup' },
+      s3url: { nargs: 1, describe: 'S3 endpoint URL' },
+      awsprofile: { nargs: 1, describe: 'The profile section to use in the ~/.aws/credentials file', default: 'default' }
+    })
+    .help('h').alias('h', 'help')
+    .epilog('Copyright (C) IBM 2017')
+    .argv;
+
+  const sourceUrl = argv.source;
+  const backupBucket = argv.bucket;
+  const backupName = new url.URL(sourceUrl).pathname.split('/').filter(function(x) { return x; }).join('-');
+  const backupKeyPrefix = `${argv.prefix}-${backupName}`;
+
+  const backupKey = `${backupKeyPrefix}-${new Date().toISOString()}`;
+  const backupTmpFile = tmp.fileSync();
+
+  const s3Endpoint = argv.s3url;
+  const awsProfile = argv.awsprofile;
+
+  // Creds are from ~/.aws/credentials, environment etc. (see S3 docs).
+  const awsOpts = {
+    signatureVersion: 'v4',
+    credentials: new AWS.SharedIniFileCredentials({ profile: awsProfile })
+  };
+  if (typeof s3Endpoint !== 'undefined') {
+    awsOpts.endpoint = new AWS.Endpoint(s3Endpoint);
   }
-  const fd = fs.openSync(program.output, flags);
-  ws = fs.createWriteStream(null, { fd: fd });
+  const s3 = new AWS.S3(awsOpts);
+
+  debug(`Creating a new backup of ${s(sourceUrl)} at ${backupBucket}/${backupKey}...`);
+  bucketAccessible(s3, backupBucket)
+    .then(() => {
+      return createBackupFile(sourceUrl, backupTmpFile.name);
+    })
+    .then(() => {
+      return uploadNewBackup(s3, backupTmpFile.name, backupBucket, backupKey);
+    })
+    .then(() => {
+      debug('Backup successful!');
+      backupTmpFile.removeCallback();
+      debug('done.');
+    })
+    .catch((reason) => {
+      debug(`Error: ${reason}`);
+    });
+}
+
+/**
+ * Return a promise that resolves if the bucket is available and
+ * rejects if not.
+ *
+ * @param {any} s3 S3 client object
+ * @param {any} bucketName Bucket name
+ * @returns Promise
+ */
+function bucketAccessible(s3, bucketName) {
+  return new Promise(function(resolve, reject) {
+    const params = {
+      Bucket: bucketName
+    };
+    s3.headBucket(params, function(err, data) {
+      if (err) {
+        reject(new VError(err, 'S3 bucket not accessible'));
+      } else {
+        resolve();
+      }
+    });
+  });
+}
+
+/**
+ * Use couchbackup to create a backup of the specified database to a file path.
+ *
+ * @param {any} sourceUrl Database URL
+ * @param {any} backupTmpFilePath Path to write file
+ * @returns Promise
+ */
+function createBackupFile(sourceUrl, backupTmpFilePath) {
+  return new Promise((resolve, reject) => {
+    couchbackup.backup(
+      sourceUrl,
+      fs.createWriteStream(backupTmpFilePath),
+      (err) => {
+        if (err) {
+          return reject(new VError(err, 'CouchBackup process failed'));
+        }
+        debug('couchbackup to file done; uploading to S3');
+        resolve('creating backup file complete');
+      }
+    );
+  });
+}
+
+/**
+ * Upload a backup file to an S3 bucket.
+ *
+ * @param {any} s3 Object store client
+ * @param {any} backupTmpFilePath Path of backup file to write.
+ * @param {any} bucket Object store bucket name
+ * @param {any} key Object store key name
+ * @returns Promise
+ */
+function uploadNewBackup(s3, backupTmpFilePath, bucket, key) {
+  return new Promise((resolve, reject) => {
+    debug(`Uploading from ${backupTmpFilePath} to ${bucket}/${key}`);
+
+    function uploadFromStream(s3, bucket, key) {
+      const pass = new stream.PassThrough();
+
+      const params = {
+        Bucket: bucket,
+        Key: key,
+        Body: pass
+      };
+      s3.upload(params, function(err, data) {
+        debug('S3 upload done');
+        if (err) {
+          debug(err);
+          reject(new VError(err, 'Upload failed'));
+          return;
+        }
+        debug('Upload succeeded');
+        debug(data);
+        resolve();
+      }).httpUploadProgress = (progress) => {
+        debug(`S3 upload progress: ${progress}`);
+      };
+
+      return pass;
+    }
+
+    const inputStream = fs.createReadStream(backupTmpFilePath);
+    const s3Stream = uploadFromStream(s3, bucket, key);
+    inputStream.pipe(s3Stream);
+  });
+}
+
+/**
+ * Remove creds from a URL, e.g., before logging
+ *
+ * @param {string} url URL to safen
+ */
+function s(originalUrl) {
+  const parts = new url.URL(originalUrl);
+  return url.format(parts, { auth: false });
 }
 
-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);
-});
+main();

package/.scannerwork/scanner-report/source-6.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017, 2018 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.
@@ -13,42 +13,152 @@
 // limitations under the License.
 'use strict';
 
-/**
- * Utility methods for the command line interface.
- * @module cliutils
- * @see module:cliutils
- */
-
-const url = require('url');
-
-module.exports = {
-
-  /**
-   * Combine a base URL and a database name, ensuring at least single slash
-   * between root and database name. This allows users to have Couch behind
-   * proxies that mount Couch's / endpoint at some other mount point.
-   * @param {string} root - root URL
-   * @param {string} databaseName - database name
-   * @return concatenated URL.
-   *
-   * @private
-   */
-  databaseUrl: function databaseUrl(root, databaseName) {
-    if (!root.endsWith('/')) {
-      root = root + '/';
+const async = require('async');
+const stream = require('stream');
+const error = require('./error.js');
+const debug = require('debug')('couchbackup:writer');
+
+module.exports = function(db, bufferSize, parallelism, ee) {
+  const writer = new stream.Transform({ objectMode: true });
+  let buffer = [];
+  let written = 0;
+  let linenumber = 0;
+
+  // this is the queue of chunks that are written to the database
+  // the queue's payload will be an array of documents to be written,
+  // the size of the array will be bufferSize. The variable parallelism
+  // determines how many HTTP requests will occur at any one time.
+  const q = async.queue(function(payload, cb) {
+    // if we are restoring known revisions, we need to supply new_edits=false
+    if (payload.docs && payload.docs[0] && payload.docs[0]._rev) {
+      payload.new_edits = false;
+      debug('Using new_edits false mode.');
+    }
+
+    if (!didError) {
+      db.service.postBulkDocs({
+        db: db.db,
+        bulkDocs: payload
+      }).then(response => {
+        if (!response.result || (payload.new_edits === false && response.result.length > 0)) {
+          throw new Error(`Error writing batch with new_edits:${payload.new_edits !== false}` +
+            ` and ${response.result ? response.result.length : 'unavailable'} items`);
+        }
+        written += payload.docs.length;
+        writer.emit('restored', { documents: payload.docs.length, total: written });
+        cb();
+      }).catch(err => {
+        err = error.convertResponseError(err);
+        debug(`Error writing docs ${err.name} ${err.message}`);
+        cb(err, payload);
+      });
+    }
+  }, parallelism);
+
+  let didError = false;
+
+  // write the contents of the buffer to CouchDB in blocks of bufferSize
+  function processBuffer(flush, callback) {
+    function taskCallback(err, payload) {
+      if (err && !didError) {
+        debug(`Queue task failed with error ${err.name}`);
+        didError = true;
+        q.kill();
+        writer.emit('error', err);
+      }
+    }
+
+    if (flush || buffer.length >= bufferSize) {
+      // work through the buffer to break off bufferSize chunks
+      // and feed the chunks to the queue
+      do {
+        // split the buffer into bufferSize chunks
+        const toSend = buffer.splice(0, bufferSize);
+
+        // and add the chunk to the queue
+        debug(`Adding ${toSend.length} to the write queue.`);
+        q.push({ docs: toSend }, taskCallback);
+      } while (buffer.length >= bufferSize);
+
+      // send any leftover documents to the queue
+      if (flush && buffer.length > 0) {
+        debug(`Adding remaining ${buffer.length} to the write queue.`);
+        q.push({ docs: buffer }, taskCallback);
+      }
+
+      // wait until the queue size falls to a reasonable level
+      async.until(
+        // wait until the queue length drops to twice the paralellism
+        // or until empty on the last write
+        function(callback) {
+          // if we encountered an error, stop this until loop
+          if (didError) {
+            return callback(null, true);
+          }
+          if (flush) {
+            callback(null, q.idle() && q.length() === 0);
+          } else {
+            callback(null, q.length() <= parallelism * 2);
+          }
+        },
+        function(cb) {
+          setTimeout(cb, 20);
+        },
+
+        function() {
+          if (flush && !didError) {
+            writer.emit('finished', { total: written });
+          }
+          // callback when we're happy with the queue size
+          callback();
+        });
+    } else {
+      callback();
     }
-    return new url.URL(encodeURIComponent(databaseName), root).toString();
-  },
-
-  /**
-   * Generate CLI argument usage text.
-   *
-   * @param {string} description - argument description.
-   * @param {string} defaultValue - default argument value.
-   *
-   * @private
-   */
-  getUsage: function getUsage(description, defaultValue) {
-    return `${description} ${defaultValue !== undefined ? ` (default: ${defaultValue})` : ''}`;
   }
+
+  // take an object
+  writer._transform = function(obj, encoding, done) {
+    // each obj that arrives here is a line from the backup file
+    // it should contain an array of objects. The length of the array
+    // depends on the bufferSize at backup time.
+    linenumber++;
+    if (!didError && obj !== '') {
+      // see if it parses as JSON
+      try {
+        const arr = JSON.parse(obj);
+
+        // if it's an array with a length
+        if (typeof arr === 'object' && arr.length > 0) {
+          // push each document into a buffer
+          buffer = buffer.concat(arr);
+
+          // pause the stream
+          // it's likely that the speed with which data can be read from disk
+          // may exceed the rate it can be written to CouchDB. To prevent
+          // the whole file being buffered in memory, we pause the stream here.
+          // it is resumed, when processBuffer calls back and we call done()
+          this.pause();
+
+          // break the buffer in to bufferSize chunks to be written to the database
+          processBuffer(false, done);
+        } else {
+          ee.emit('error', new error.BackupError('BackupFileJsonError', `Error on line ${linenumber} of backup file - not an array`));
+          done();
+        }
+      } catch (e) {
+        ee.emit('error', new error.BackupError('BackupFileJsonError', `Error on line ${linenumber} of backup file - cannot parse as JSON`));
+        // Could be an incomplete write that was subsequently resumed
+        done();
+      }
+    } else {
+      done();
+    }
+  };
+
+  // called when we need to flush everything
+  writer._flush = function(done) {
+    processBuffer(true, done);
+  };
+  return writer;
 };