Skip to main content

MongoDB Rebuild Index

  1. Enter the storage component container

    docker exec -it $(docker ps | grep mingdaoyun-sc | awk '{print $1}') bash

  1. Create a reIndex.js script file to rebuild user-defined indexes under the mdwsrows database

    Content of reIndex.js Script File
    var targetDbName = "mdwsrows";  // Target database

    // Collection whitelist; collections in the whitelist will not have their indexes rebuilt
    var collectionWhitelist = [
        "discussion",
        "rowrelations",
        "workSheetRowTopic",
        "workSheetTopic",
        "wslogs"
    ];

    // Index whitelist; indexes in the whitelist will not be rebuilt
    var indexWhitelist = [
        "_id_",
        "idx_ctime",
        "idx_utime",
        "uk_rowid",
        "idx_tp_status",
        "idx_thirdprimary"
    ];

    // DateTime formatting function
    function formatDateTime() {
        var now = new Date();
        var utc8Time = new Date(now.getTime() + (8 * 60 * 60 * 1000));
        return utc8Time.toISOString().replace('Z', '+08:00');
    }

    // Formatting output function
    function printHeader(text) {
        print("\n" + "=".repeat(100));
        print(text);
        print("=".repeat(100));
    }

    function printSection(text) {
        print("\n" + "-".repeat(80));
        print(text);
        print("-".repeat(80));
    }

    function printTimedAction(time, action) {
        print(`\n[${time}] ${action}`);
    }

    // Format JSON with consistent indentation
    function formatJSON(obj, indent = 5) {
        return JSON.stringify(obj, null, 2)
            .split('\n')
            .map((line, i) => i === 0 ? line : ' '.repeat(indent) + line)
            .join('\n');
    }

    // Format createIndex command
    function formatCreateIndexCommand(collName, key, options) {
        return `db.${collName}.createIndex(${formatJSON(key)},\n     ${formatJSON(options)})`;
    }

    function printCommand(command) {
        print("  └─ Execute:");
        print("     " + command);
    }

    function printCompletion(seconds) {
        print(`  └─ ✓ Completed in ${seconds.toFixed(3)} seconds\n`);
    }

    // Connect to the target database
    var targetDb = db.getSiblingDB(targetDbName);
    var startTime = formatDateTime();

    printHeader("MongoDB Index Rebuild Process");
    print(`\n• Start Time: ${startTime}`);
    print(`• Target Database: ${targetDb.getName()}`);

    // Get all collections to rebuild by filtering the collection whitelist
    var collections = targetDb.getCollectionNames().filter(function(collName) {
        return !collectionWhitelist.includes(collName) &&
               !collName.startsWith('system.');
    });

    print(`• Total Collections: ${collections.length}`);

    collections.forEach(function(collName, index) {
        var coll = targetDb.getCollection(collName);
        var stats = coll.stats();
        
        // Output collection info and progress
        printSection(`Processing Collection [${index + 1}/${collections.length}]: ${collName}`);
        print(`\n• Document Count: ${stats.count}`);
        print(`• Storage Size: ${stats.storageSize} bytes`);
        
        // Get indexes to rebuild by filtering the index whitelist
        var indexes = coll.getIndexes();
        var rebuildIndexes = indexes.filter(function(idx) {
            return !indexWhitelist.includes(idx.name);
        });
        
        if (rebuildIndexes.length === 0) {
            print("\n✓ No indexes need to be rebuilt.");
            return;
        }
        
        // Output index rebuild plan
        print(`\n• Indexes to Rebuild (${rebuildIndexes.length}):`);
        rebuildIndexes.forEach(function(idx) {
            print(`  ├─ Name: ${idx.name.padEnd(20)}`);
            print(`  │  Key: ${JSON.stringify(idx.key)}`);
        });
        
        // Rebuild each index
        rebuildIndexes.forEach(function(idx) {
            var key = idx.key;
            var options = {};
            
            // Copy index configuration, excluding system attributes
            for (var prop in idx) {
                if (!["v", "ns", "background"].includes(prop)) {
                    options[prop] = idx[prop];
                }
            }
            options.background = true;  // Build index in background
            
            try {
                // Drop the old index
                var dropTime = formatDateTime();
                printTimedAction(dropTime, `Dropping Index: ${idx.name}`);
                printCommand(`db.${collName}.dropIndex("${idx.name}")`);
                
                var dropStart = new Date();
                coll.dropIndex(idx.name);
                var dropEnd = new Date();
                printCompletion((dropEnd - dropStart)/1000);
                
                // Create a new index
                var createTime = formatDateTime();
                printTimedAction(createTime, `Creating Index: ${idx.name}`);
                printCommand(formatCreateIndexCommand(collName, key, options));
                
                var createStart = new Date();
                coll.createIndex(key, options);
                var createEnd = new Date();
                printCompletion((createEnd - createStart)/1000);
                
            } catch (e) {
                print(`  └─ ✗ Error: ${e.message}`);
                print("     Skipping this index...\n");
            }
        });
    });

    var endTime = formatDateTime();
    printHeader("Process Completed");
    print(`\n• End Time: ${endTime}`);

  1. Execute the script

    nohup mongo mongodb://127.0.0.1:27017/admin --quiet reIndex.js >> reIndex_output.log 2>&1 &
    • The execution log will be output to the reIndex_output.log file

  1. Monitor the log and wait for the script to complete its execution (the log will end with Process Complete and End Time)

  2. Create a reIndexWithCmd.js script file to rebuild some indexes that HAP system relies on

    Content of reIndexWithCmd.js Script File
    // ====================================================================
    //                          CONFIGURATION
    // ====================================================================
    // Configure the databases and their whitelist collections for index rebuilding here
    // Format: "Database Name": ["Collection to skip 1", "Collection to skip 2..."]
    // Use an empty array [] if all collections in a database need index rebuilding
    var targetDatabases = {
        "mdpost": [],
        "MDHistory": []
    };
    // ====================================================================


    // Formatting output function (no modification needed)
    function printHeader(text) {
        print("\n" + "=".repeat(100));
        print(text);
        print("=".repeat(100));
    }

    function printSection(text) {
        print("\n" + "-".repeat(80));
        print(text);
        print("-".repeat(80));
    }

    function formatDateTime() {
        var now = new Date();
        var utc8Time = new Date(now.getTime() + (8 * 60 * 60 * 1000));
        return utc8Time.toISOString().replace('Z', '+08:00');
    }

    function formatJSON(obj, indent = 5) {
        if (obj === undefined || obj === null) {
            return "Not available";
        }
        return JSON.stringify(obj, null, 2)
            .split('\n')
            .map((line, i) => i === 0 ? line : ' '.repeat(indent) + line)
            .join('\n');
    }

    function formatFileSize(bytes) {
        if (bytes === undefined || bytes === null) return "N/A";
        return (bytes / 1024 / 1024).toFixed(2) + " MB";
    }

    function printTimedAction(time, action) {
        print(`\n[${time}] ${action}`);
    }

    function printCompletion(seconds) {
        print(`  └─ ✓ Completed in ${seconds.toFixed(3)} seconds\n`);
    }

    // ====================================================================
    //                          SCRIPT EXECUTION
    // ====================================================================

    var overallStartTime = formatDateTime();
    printHeader("MongoDB Multi-Database Index Rebuild Process Started");
    print(`\n• Overall Start Time: ${overallStartTime}`);
    print(`• Databases to Process: ${Object.keys(targetDatabases).join(', ')}`);

    // Iterate through all configured databases
    for (var dbName in targetDatabases) {
        if (targetDatabases.hasOwnProperty(dbName)) {
            
            printHeader(`Processing Database: [ ${dbName} ]`);

            // Connect to the target database
            var currentDb = db.getSiblingDB(dbName);
            
            // Get the collection blacklist (collections to skip) for the current database
            var excludedCollections = targetDatabases[dbName];

            print(`\n• Target Database: ${currentDb.getName()}`);
            print(`• Excluded Collections: ${formatJSON(excludedCollections)}`);

            // Get all collections
            var collections = currentDb.getCollectionNames();
            // Filter out collections in the blacklist and system collections
            var validCollections = collections.filter(collection => 
                !excludedCollections.includes(collection) && !collection.startsWith('system.')
            );

            print(`• Total Collections to Process in this DB: ${validCollections.length}`);
            
            if (validCollections.length === 0) {
                print("\nNo collections to process in this database. Moving to the next one.");
                continue;
            }

            validCollections.forEach(function(collection, index) {
                try {
                    // Get collection statistics
                    var stats = currentDb[collection].stats();
                    
                    // Output collection info and progress
                    printSection(`Processing Collection [${index + 1}/${validCollections.length}]: ${collection}`);
                    print(`\n• Collection Statistics:`);
                    print(`  ├─ Storage Size: ${formatFileSize(stats.storageSize)}`);
                    print(`  └─ Document Count: ${(stats.count || 0).toLocaleString()}`);
                    
                    // Get current index information
                    var indexes = currentDb[collection].getIndexes();
                    print(`\n• Current Indexes (${indexes.length}):`);
                    indexes.forEach(function(idx, i) {
                        const isLast = i === indexes.length - 1;
                        print(`  ${isLast ? '└' : '├'}─ Name: ${idx.name.padEnd(20)}`);
                        print(`  ${isLast ? ' ' : '│'}  Key: ${formatJSON(idx.key)}`);
                    });
                    
                    // Execute reIndex
                    var execTime = formatDateTime();
                    printTimedAction(execTime, "Executing reIndex()");
                    print("  └─ Execute:");
                    print(`     db.getSiblingDB('${dbName}').getCollection('${collection}').reIndex()`);
                    
                    var startExec = new Date();
                    var result = currentDb[collection].reIndex();
                    var endExec = new Date();
                    
                    if (result.ok === 1) {
                        printCompletion((endExec - startExec)/1000);
                        
                        print("• Operation Results:");
                        print(`  ├─ Previous Index Count: ${result.nIndexesWas}`);
                        print(`  ├─ Current Index Count: ${result.nIndexes}`);
                        
                        if (result.operationTime !== undefined) {
                            print(`  ├─ Operation Time: ${formatJSON(result.operationTime)}`);
                        } else {
                            print(`  ├─ Operation Time: Not available (non-replica set deployment)`);
                        }
                        
                        if (result.$clusterTime !== undefined) {
                            print(`  └─ Cluster Time: ${formatJSON(result.$clusterTime)}`);
                        } else {
                            print(`  └─ Cluster Time: Not available (non-replica set deployment)`);
                        }
                        
                    } else {
                        print(`  └─ ✗ ReIndex failed: ${formatJSON(result)}`);
                    }
                } catch (e) {
                    print(`  └─ ✗ Error: ${e.message}`);
                    print("     Skipping this collection...\n");
                }
            });
        }
    }

    var overallEndTime = formatDateTime();
    printHeader("Process Completed");
    print(`\n• Overall End Time: ${overallEndTime}`);

  1. Execute the script

    nohup mongo mongodb://127.0.0.1:27017/admin --quiet reIndexWithCmd.js >> reIndexWithCmd_output.log 2>&1 &
    • The execution log will be output to the reIndexWithCmd_output.log file
  1. Monitor the log and wait for the script to complete its execution (the log will end with Process Complete and End Time)