Skip to main content

MongoDB Index Rebuild

  1. Enter the storage component container

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

  1. Create the reIndex.js script file

    Script file content for reIndex.js, the script defaults to the mdwsrows database 
    var targetDbName = "mdwsrows";  // Target database

    // Collection whitelist, indexes in whitelisted collections are not rebuilt
    var collectionWhitelist = [
        "discussion",
        "rowrelations",
        "workSheetRowTopic",
        "workSheetTopic",
        "wslogs"
    ];

    // Index whitelist, whitelisted indexes are not rebuilt
    var indexWhitelist = [
        "_id_",
        "idx_ctime",
        "idx_utime",
        "uk_rowid",
        "idx_tp_status",
        "idx_thirdprimary"
    ];

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

    // Function to format output
    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}`);
    }

    // Function to format JSON, 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');
    }

    // Function to 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 specified 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, filtering out whitelist collections
    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 information 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, filtering out whitelist indexes
        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 the background
            
            try {
                // Drop 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 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 (the end of the log will output Process Complete and End Time)