SERVER-31920 refactor validate_collections.js to accept a list of nodes in addition to a DB

2 files changed, 128 insertions, 120 deletions

diff --git a/jstests/hooks/run_validate_collections.js b/jstests/hooks/run_validate_collections.js
index d95c9eb5c9f..772c2366113 100644
--- a/jstests/hooks/run_validate_collections.js
+++ b/jstests/hooks/run_validate_collections.js
@@ -4,22 +4,15 @@
 
 (function() {
     assert.eq(typeof db, 'object', 'Invalid `db` object, is the shell connected to a mongod?');
-    load('jstests/libs/parallelTester.js');
 
-    function getDirectConnections(conn) {
+    function getConnectionStrings(conn) {
         // If conn does not point to a repl set, then this function returns [conn].
         const res = conn.adminCommand({isMaster: 1});
-        const connections = [];
-
         if (res.hasOwnProperty('hosts')) {
-            for (let hostString of res.hosts) {
-                connections.push(new Mongo(hostString));
-            }
+            return res.hosts;
         } else {
-            connections.push(conn);
+            return [conn.host];
         }
-
-        return connections;
     }
 
     function getConfigConnStr() {
@@ -42,8 +35,8 @@
         return db.isMaster().msg === 'isdbgrid';
     }
 
-    function getServerList() {
-        const serverList = [];
+    function getHostList() {
+        let hostList = [];
 
         if (isMongos()) {
             // We're connected to a sharded cluster through a mongos.
@@ -51,7 +44,7 @@
             // 1) Add all the config servers to the server list.
             const configConnStr = getConfigConnStr();
             const configServerReplSetConn = new Mongo(configConnStr);
-            serverList.push(...getDirectConnections(configServerReplSetConn));
+            hostList = getConnectionStrings(configServerReplSetConn);
 
             // 2) Add shard members to the server list.
             const configDB = db.getSiblingDB('config');
@@ -60,61 +53,17 @@
             while (cursor.hasNext()) {
                 const shard = cursor.next();
                 const shardReplSetConn = new Mongo(shard.host);
-                serverList.push(...getDirectConnections(shardReplSetConn));
+                hostList.push(...getConnectionStrings(shardReplSetConn));
             }
         } else {
             // We're connected to a mongod.
-            serverList.push(...getDirectConnections(db.getMongo()));
+            hostList = getConnectionStrings(db.getMongo());
         }
 
-        return serverList;
+        return hostList;
     }
 
-    // Run a separate thread to validate collections on each server in parallel.
-    var validateCollectionsThread = function(host, testData) {
-        load('jstests/hooks/validate_collections.js');  // For validateCollections.
-        TestData = testData;  // Pass the TestData object from main thread.
-
-        try {
-            print('Running validate() on ' + host);
-            const conn = new Mongo(host);
-            conn.setSlaveOk();
-            jsTest.authenticate(conn);
-
-            const dbNames = conn.getDBNames();
-            for (let dbName of dbNames) {
-                if (!validateCollections(conn.getDB(dbName), {full: true})) {
-                    return {ok: 0};
-                }
-            }
-            return {ok: 1};
-        } catch (e) {
-            print('Exception caught in scoped thread running validationCollections on server: ' +
-                  host);
-            return {ok: 0, error: e.toString(), stack: e.stack};
-        }
-    };
-
-    // We run the scoped threads in a try/finally block in case any thread throws an exception, in
-    // which case we want to still join all the threads.
-    let threads = [];
-    const serverList = getServerList();
+    load('jstests/hooks/validate_collections.js');  // For CollectionValidator.
+    new CollectionValidator().validateNodes(getHostList());
 
-    try {
-        serverList.forEach(server => {
-            const thread = new ScopedThread(validateCollectionsThread, server.host, TestData);
-            threads.push(thread);
-            thread.start();
-        });
-    } finally {
-        // Wait for each thread to finish. Throw an error if any thread fails.
-        const returnData = threads.map(thread => {
-            thread.join();
-            return thread.returnData();
-        });
-
-        returnData.forEach(res => {
-            assert.commandWorked(res, 'Collection validation failed');
-        });
-    }
 })();
diff --git a/jstests/hooks/validate_collections.js b/jstests/hooks/validate_collections.js
index 31e5d26affe..71136a3e4cc 100644
--- a/jstests/hooks/validate_collections.js
+++ b/jstests/hooks/validate_collections.js
@@ -1,75 +1,134 @@
 // Wrapper around the validate command that can be used to validate index key counts.
 'use strict';
 
-function validateCollections(db, obj) {
-    function dumpCollection(coll, limit) {
-        print('Printing indexes in: ' + coll.getFullName());
-        printjson(coll.getIndexes());
-
-        print('Printing the first ' + limit + ' documents in: ' + coll.getFullName());
-        var res = coll.find().limit(limit);
-        while (res.hasNext()) {
-            printjson(res.next());
-        }
+function CollectionValidator() {
+    load('jstests/libs/parallelTester.js');
+
+    if (!(this instanceof CollectionValidator)) {
+        throw new Error('Please use "new CollectionValidator()"');
     }
 
-    assert.eq(typeof db, 'object', 'Invalid `db` object, is the shell connected to a mongod?');
-    assert.eq(typeof obj, 'object', 'The `obj` argument must be an object');
-    assert(obj.hasOwnProperty('full'), 'Please specify whether to use full validation');
+    this.validateCollections = function(db, obj) {
+        function dumpCollection(coll, limit) {
+            print('Printing indexes in: ' + coll.getFullName());
+            printjson(coll.getIndexes());
 
-    var full = obj.full;
+            print('Printing the first ' + limit + ' documents in: ' + coll.getFullName());
+            const res = coll.find().limit(limit);
+            while (res.hasNext()) {
+                printjson(res.next());
+            }
+        }
 
-    var success = true;
+        assert.eq(typeof db, 'object', 'Invalid `db` object, is the shell connected to a mongod?');
+        assert.eq(typeof obj, 'object', 'The `obj` argument must be an object');
+        assert(obj.hasOwnProperty('full'), 'Please specify whether to use full validation');
 
-    var adminDB = db.getSiblingDB("admin");
+        const full = obj.full;
 
-    // Don't run validate on view namespaces.
-    let filter = {type: "collection"};
-    if (jsTest.options().skipValidationOnInvalidViewDefinitions) {
-        // If skipValidationOnInvalidViewDefinitions=true, then we avoid resolving the view catalog
-        // on the admin database.
-        //
-        // TODO SERVER-25493: Remove the $exists clause once performing an initial sync from
-        // versions of MongoDB <= 3.2 is no longer supported.
-        filter = {$or: [filter, {type: {$exists: false}}]};
-    }
+        let success = true;
 
-    // Optionally skip collections.
-    if (Array.isArray(jsTest.options().skipValidationNamespaces) &&
-        jsTest.options().skipValidationNamespaces.length > 0) {
-        let skippedCollections = [];
-        for (let ns of jsTest.options().skipValidationNamespaces) {
-            // Strip off the database name from 'ns' to extract the collName.
-            const collName = ns.replace(new RegExp('^' + db.getName() + '\.'), '');
-            // Skip the collection 'collName' if the db name was removed from 'ns'.
-            if (collName !== ns) {
-                skippedCollections.push({name: {$ne: collName}});
+        // Don't run validate on view namespaces.
+        let filter = {type: "collection"};
+        if (jsTest.options().skipValidationOnInvalidViewDefinitions) {
+            // If skipValidationOnInvalidViewDefinitions=true, then we avoid resolving the view
+            // catalog on the admin database.
+            //
+            // TODO SERVER-25493: Remove the $exists clause once performing an initial sync from
+            // versions of MongoDB <= 3.2 is no longer supported.
+            filter = {$or: [filter, {type: {$exists: false}}]};
+        }
+
+        // Optionally skip collections.
+        if (Array.isArray(jsTest.options().skipValidationNamespaces) &&
+            jsTest.options().skipValidationNamespaces.length > 0) {
+            let skippedCollections = [];
+            for (let ns of jsTest.options().skipValidationNamespaces) {
+                // Strip off the database name from 'ns' to extract the collName.
+                const collName = ns.replace(new RegExp('^' + db.getName() + '\.'), '');
+                // Skip the collection 'collName' if the db name was removed from 'ns'.
+                if (collName !== ns) {
+                    skippedCollections.push({name: {$ne: collName}});
+                }
             }
+            filter = {$and: [filter, ...skippedCollections]};
         }
-        filter = {$and: [filter, ...skippedCollections]};
-    }
 
-    let collInfo = db.getCollectionInfos(filter);
-    for (var collDocument of collInfo) {
-        var coll = db.getCollection(collDocument["name"]);
-        var res = coll.validate(full);
-
-        if (!res.ok || !res.valid) {
-            if (jsTest.options().skipValidationOnNamespaceNotFound &&
-                res.errmsg === 'ns not found') {
-                // During a 'stopStart' backup/restore on the secondary node, the actual list of
-                // collections can be out of date if ops are still being applied from the oplog. In
-                // this case we skip the collection if the ns was not found at time of validation
-                // and continue to next.
-                print('Skipping collection validation for ' + coll.getFullName() +
-                      ' since collection was not found');
-                continue;
+        let collInfo = db.getCollectionInfos(filter);
+        for (let collDocument of collInfo) {
+            const coll = db.getCollection(collDocument["name"]);
+            const res = coll.validate(full);
+
+            if (!res.ok || !res.valid) {
+                if (jsTest.options().skipValidationOnNamespaceNotFound &&
+                    res.errmsg === 'ns not found') {
+                    // During a 'stopStart' backup/restore on the secondary node, the actual list of
+                    // collections can be out of date if ops are still being applied from the oplog.
+                    // In this case we skip the collection if the ns was not found at time of
+                    // validation and continue to next.
+                    print('Skipping collection validation for ' + coll.getFullName() +
+                          ' since collection was not found');
+                    continue;
+                }
+                print('Collection validation failed with response: ' + tojson(res));
+                dumpCollection(coll, 100);
+                success = false;
             }
-            print('Collection validation failed with response: ' + tojson(res));
-            dumpCollection(coll, 100);
-            success = false;
         }
-    }
 
-    return success;
+        return success;
+    };
+
+    // Run a separate thread to validate collections on each server in parallel.
+    const validateCollectionsThread = function(validatorFunc, host, testData) {
+        TestData = testData;  // Pass the TestData object from main thread.
+
+        try {
+            print('Running validate() on ' + host);
+            const conn = new Mongo(host);
+            conn.setSlaveOk();
+            jsTest.authenticate(conn);
+
+            const dbNames = conn.getDBNames();
+            for (let dbName of dbNames) {
+                if (!validatorFunc(conn.getDB(dbName), {full: true})) {
+                    return {ok: 0};
+                }
+            }
+            return {ok: 1};
+        } catch (e) {
+            print('Exception caught in scoped thread running validationCollections on server: ' +
+                  host);
+            return {ok: 0, error: e.toString(), stack: e.stack};
+        }
+    };
+
+    this.validateNodes = function(hostList) {
+        // We run the scoped threads in a try/finally block in case any thread throws an exception,
+        // in which case we want to still join all the threads.
+        let threads = [];
+
+        try {
+            hostList.forEach(host => {
+                const thread = new ScopedThread(
+                    validateCollectionsThread, this.validateCollections, host, TestData);
+                threads.push(thread);
+                thread.start();
+            });
+        } finally {
+            // Wait for each thread to finish. Throw an error if any thread fails.
+            const returnData = threads.map(thread => {
+                thread.join();
+                return thread.returnData();
+            });
+
+            returnData.forEach(res => {
+                assert.commandWorked(res, 'Collection validation failed');
+            });
+        }
+    };
 }
+
+// Ensure compatability with existing callers. Cannot use `const` or `let` here since this file may
+// be loaded more than once.
+var validateCollections = new CollectionValidator().validateCollections;