@cloudant/couchbackup 2.9.12-SNAPSHOT.136 → 2.9.12

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

@@ -11,44 +11,215 @@
 // 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 describe it */
 '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 + '/';
-    }
-    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})` : ''}`;
+const assert = require('assert');
+const backup = require('../app.js').backup;
+const fs = require('fs');
+const nock = require('nock');
+
+const goodUrl = 'http://localhost:5984/db';
+// The real validateArgs function of app.js isn't
+// exported - so we call the exported backup method
+// instead. We don't get as far as a real backup when
+// testing error cases. For success cases we nock the
+// goodUrl and
+const validateArgs = function(url, opts, callback) {
+  const nullStream = fs.createWriteStream('/dev/null');
+  let cb = callback;
+  if (url === goodUrl) {
+    // Nock the goodUrl
+    nock(goodUrl).head('').reply(404, { error: 'not_found', reason: 'missing' });
+    // replace the callback to handle the nock response
+    // to avoid attempting a real backup
+    cb = function(err) {
+      nullStream.end();
+      if (err.name === 'DatabaseNotFound') {
+        // This is what we expect if we reached the backup
+        // This is success for valid args cases.
+        err = null;
+      }
+      callback(err);
+    };
   }
+  backup(url, nullStream, opts, cb);
+  return true;
 };
+
+const stderrWriteFun = process.stderr.write;
+let capturedStderr;
+
+function captureStderr() {
+  process.stderr.write = function(string, encoding, fd) {
+    capturedStderr += string;
+  };
+}
+
+function releaseStderr() {
+  process.stderr.write = stderrWriteFun;
+  capturedStderr = null;
+}
+
+function assertErrorMessage(msg, done) {
+  return function(err, data) {
+    try {
+      assert(err.message, 'There should be an error message');
+      assert(err.message.indexOf(msg) >= 0);
+      assert(data === null || data === undefined, 'There should only be an error.');
+      done();
+    } catch (e) {
+      done(e);
+    }
+  };
+}
+
+function assertNoError(done) {
+  return function(err, data) {
+    try {
+      assert(err === null, 'There should be no error message.');
+      done();
+    } catch (e) {
+      done(e);
+    }
+  };
+}
+
+describe('#unit Validate arguments', function() {
+  it('returns error for invalid URL type', function(done) {
+    validateArgs(true, {}, assertErrorMessage('Invalid URL, must be type string', done));
+  });
+  it('returns no error for valid URL type', function(done) {
+    validateArgs(goodUrl, {}, assertNoError(done));
+  });
+  it('returns error for invalid (no host) URL', function(done) {
+    validateArgs('http://', {}, assertErrorMessage('Invalid URL', done));
+  });
+  it('returns error for invalid (no protocol) URL', function(done) {
+    validateArgs('invalid', {}, assertErrorMessage('Invalid URL', done));
+  });
+  it('returns error for invalid (wrong protocol) URL', function(done) {
+    validateArgs('ftp://invalid.example.com', {}, assertErrorMessage('Invalid URL protocol.', done));
+  });
+  it('returns error for invalid (no path) URL', function(done) {
+    validateArgs('https://invalid.example.com', {}, assertErrorMessage('Invalid URL, missing path element (no database).', done));
+  });
+  it('returns error for invalid (no protocol, no host) URL', function(done) {
+    validateArgs('invalid', {}, assertErrorMessage('Invalid URL', done));
+  });
+  it('returns error for invalid buffer size type', function(done) {
+    validateArgs(goodUrl, { bufferSize: '123' }, assertErrorMessage('Invalid buffer size option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns error for zero buffer size', function(done) {
+    validateArgs(goodUrl, { bufferSize: 0 }, assertErrorMessage('Invalid buffer size option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns error for float buffer size', function(done) {
+    validateArgs(goodUrl, { bufferSize: 1.23 }, assertErrorMessage('Invalid buffer size option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns no error for valid buffer size type', function(done) {
+    validateArgs(goodUrl, { bufferSize: 123 }, assertNoError(done));
+  });
+  it('returns error for invalid log type', function(done) {
+    validateArgs(goodUrl, { log: true }, assertErrorMessage('Invalid log option, must be type string', done));
+  });
+  it('returns no error for valid log type', function(done) {
+    validateArgs(goodUrl, { log: 'log.txt' }, assertNoError(done));
+  });
+  it('returns error for invalid mode type', function(done) {
+    validateArgs(goodUrl, { mode: true }, assertErrorMessage('Invalid mode option, must be either "full" or "shallow"', done));
+  });
+  it('returns error for invalid mode string', function(done) {
+    validateArgs(goodUrl, { mode: 'foobar' }, assertErrorMessage('Invalid mode option, must be either "full" or "shallow"', done));
+  });
+  it('returns no error for valid mode type', function(done) {
+    validateArgs(goodUrl, { mode: 'full' }, assertNoError(done));
+  });
+  it('returns error for invalid output type', function(done) {
+    validateArgs(goodUrl, { output: true }, assertErrorMessage('Invalid output option, must be type string', done));
+  });
+  it('returns no error for valid output type', function(done) {
+    validateArgs(goodUrl, { output: 'output.txt' }, assertNoError(done));
+  });
+  it('returns error for invalid parallelism type', function(done) {
+    validateArgs(goodUrl, { parallelism: '123' }, assertErrorMessage('Invalid parallelism option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns error for zero parallelism', function(done) {
+    validateArgs(goodUrl, { parallelism: 0 }, assertErrorMessage('Invalid parallelism option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns error for float parallelism', function(done) {
+    validateArgs(goodUrl, { parallelism: 1.23 }, assertErrorMessage('Invalid parallelism option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns no error for valid parallelism type', function(done) {
+    validateArgs(goodUrl, { parallelism: 123 }, assertNoError(done));
+  });
+  it('returns error for invalid request timeout type', function(done) {
+    validateArgs(goodUrl, { requestTimeout: '123' }, assertErrorMessage('Invalid request timeout option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns error for zero request timeout', function(done) {
+    validateArgs(goodUrl, { requestTimeout: 0 }, assertErrorMessage('Invalid request timeout option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns error for float request timout', function(done) {
+    validateArgs(goodUrl, { requestTimeout: 1.23 }, assertErrorMessage('Invalid request timeout option, must be a positive integer in the range (0, MAX_SAFE_INTEGER]', done));
+  });
+  it('returns no error for valid request timeout type', function(done) {
+    validateArgs(goodUrl, { requestTimeout: 123 }, assertNoError(done));
+  });
+  it('returns error for invalid resume type', function(done) {
+    validateArgs(goodUrl, { resume: 'true' }, assertErrorMessage('Invalid resume option, must be type boolean', done));
+  });
+  it('returns no error for valid resume type', function(done) {
+    validateArgs(goodUrl, { resume: false }, assertNoError(done));
+  });
+  it('returns error for invalid key type', function(done) {
+    validateArgs(goodUrl, { iamApiKey: true }, assertErrorMessage('Invalid iamApiKey option, must be type string', done));
+  });
+  it('returns error for key and URL credentials supplied', function(done) {
+    validateArgs('https://a:b@example.com/db', { iamApiKey: 'abc123' }, assertErrorMessage('URL user information must not be supplied when using IAM API key.', done));
+  });
+  it('warns for log arg in shallow mode', function(done) {
+    captureStderr();
+    try {
+      validateArgs(goodUrl, { mode: 'shallow', log: 'test' }, function(err, data) {
+        assert.ok(err);
+        assert.ok(!data);
+        assert(capturedStderr.indexOf('The options "log" and "resume" are invalid when using shallow mode.') > -1, 'Log warning message was not present');
+      });
+      done();
+    } catch (e) {
+      done(e);
+    } finally {
+      releaseStderr();
+    }
+  });
+  it('warns for resume arg in shallow mode', function(done) {
+    captureStderr();
+    try {
+      validateArgs(goodUrl, { mode: 'shallow', log: 'test', resume: true }, function(err, data) {
+        assert.ok(err);
+        assert.ok(!data);
+        assert(capturedStderr.indexOf('The options "log" and "resume" are invalid when using shallow mode.') > -1, 'Log warning message was not present');
+      });
+      done();
+    } catch (e) {
+      done(e);
+    } finally {
+      releaseStderr();
+    }
+  });
+  it('warns for parallism arg in shallow mode', function(done) {
+    captureStderr();
+    try {
+      validateArgs(goodUrl, { mode: 'shallow', parallelsim: 10 }, function(err, data) {
+        assert.ok(err);
+        assert.ok(!data);
+        assert(capturedStderr.indexOf('The option "parallelism" has no effect when using shallow mode.') > -1, 'Log warning message was not present');
+      });
+      done();
+    } catch (e) {
+      done(e);
+    } finally {
+      releaseStderr();
+    }
+  });
+});

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

@@ -1,4 +1,4 @@
-// Copyright © 2017 IBM Corp. All rights reserved.
+// Copyright © 2018, 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.
@@ -11,36 +11,82 @@
 // 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 describe it */
 'use strict';
 
-// stolen from http://strongloop.com/strongblog/practical-examples-of-the-new-node-js-streams-api/
-const stream = require('stream');
-
-module.exports = function() {
-  const liner = new stream.Transform({ objectMode: true });
-
-  liner._transform = function(chunk, encoding, done) {
-    let data = chunk.toString();
-    if (this._lastLineData) {
-      data = this._lastLineData + data;
-    }
-
-    const lines = data.split('\n');
-    this._lastLineData = lines.splice(lines.length - 1, 1)[0];
-
-    for (const i in lines) {
-      this.push(lines[i]);
-    }
-    done();
-  };
-
-  liner._flush = function(done) {
-    if (this._lastLineData) {
-      this.push(this._lastLineData);
-    }
-    this._lastLineData = null;
-    done();
-  };
-
-  return liner;
-};
+const fs = require('fs');
+const readline = require('readline');
+const u = require('./citestutils.js');
+const uuid = require('uuid').v4;
+
+const params = { useApi: true };
+
+describe(u.scenario('Concurrent database backups', params), function() {
+  it('should run concurrent API database backups correctly #slower', function(done) {
+    // Allow up to 900 s to backup and compare (it should be much faster)!
+    u.setTimeout(this, 900);
+
+    let doneCount = 0;
+    let doneErr;
+    const finished = function(err) {
+      doneCount++;
+      if (doneCount === 2) {
+        done(doneErr || err);
+      }
+      doneErr = err;
+    };
+
+    const checkForEmptyBatches = function(fileName, cb) {
+      let foundEmptyBatch = false;
+
+      const rd = readline.createInterface({
+        input: fs.createReadStream(fileName),
+        output: fs.createWriteStream('/dev/null'),
+        terminal: false
+      });
+
+      rd.on('line', function(line) {
+        if (JSON.parse(line).length === 0) {
+          // Note: Empty batch arrays indicate that the running backup is
+          // incorrectly sharing a log file with another ongoing backup job.
+          foundEmptyBatch = true;
+        }
+      });
+
+      rd.on('close', function() {
+        if (foundEmptyBatch) {
+          cb(new Error(`Log file '${fileName}' contains empty batches`));
+        } else {
+          cb();
+        }
+      });
+    };
+
+    // [1] Run 'largedb2g' database backup
+    const actualBackup1 = `./${uuid()}`;
+    const output1 = fs.createWriteStream(actualBackup1);
+    output1.on('open', function() {
+      u.testBackup(params, 'largedb2g', output1, function(err) {
+        if (err) {
+          finished(err);
+        } else {
+          checkForEmptyBatches(actualBackup1, finished);
+        }
+      });
+    });
+
+    // [2] Run 'largedb1g' database backup
+    const actualBackup2 = `./${uuid()}`;
+    const output2 = fs.createWriteStream(actualBackup2);
+    output2.on('open', function() {
+      u.testBackup(params, 'largedb1g', output2, function(err) {
+        if (err) {
+          finished(err);
+        } else {
+          checkForEmptyBatches(actualBackup2, finished);
+        }
+      });
+    });
+  });
+});

package/.scannerwork/scanner-report/source-58.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017, 2022 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.
@@ -11,101 +11,104 @@
 // 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 fs = require('fs');
-const liner = require('./liner.js');
-const change = require('./change.js');
-const error = require('./error.js');
-const debug = require('debug')('couchbackup:spoolchanges');
+const url = require('url');
+const toxy = require('toxy');
+// Import the common hooks
+require('../test/hooks.js');
 
-/**
- * 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;
+const tpoisons = toxy.poisons;
+const trules = toxy.rules;
 
-  // 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++;
-    }
-  };
+function setupProxy(poison) {
+  const backendUrl = new url.URL(process.env.COUCH_BACKEND_URL);
+  const proxy = toxy({
+    auth: `${backendUrl.username}:${backendUrl.password}`,
+    changeOrigin: true
+  });
 
-  // 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;
-      }
-    }
-  };
+  // Forward traffic to DB
+  proxy.forward(process.env.COUCH_BACKEND_URL);
 
-  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}`));
-        }
-      });
+  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);
   }
 
-  getChanges();
-};
+  // Catch remaining traffic
+  proxy.all('/*');
+  return proxy;
+}
+
+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');
+  });
+});