Mastering MongoDB Connection, Authentication, and Query Optimization

Common Connection Issues

Each connection creates a thread (≈1 MB memory) and frequent creation/destruction incurs high overhead.

Connection Resource Control

Limit max connections via maxConns at startup and configure client pool size with maxPoolSize , e.g.:

mongoClient = new MongoClient("mongodb://root:****@host1:port1,host2:port2/admin?replicaSet=repl00&maxPoolSize=100");

Authentication Failures

Enable authentication, specify the auth database (usually admin ) in the connection string, and use the appropriate mechanism. Example command‑line login:

mongodb30/bin/mongo -uxyz -pxyz 10.142.1.1:7003/abc --authenticationDatabase admin

Different MongoDB versions may require matching client versions.

Cursor Timeout Settings

Control socket and cursor timeouts (e.g., socketTimeoutMS , cursorTimeoutMS ) in drivers such as PHP.

SCRAM‑SHA‑1 vs MONGODB‑CR

SCRAM‑SHA‑1 offers stronger hashing but generates a server‑nonce via /dev/urandom , which can be a performance bottleneck. The article recommends switching to MONGODB‑CR:

mongodb30/bin/mongo -uxyz -pxyz 10.142.1.1:7003/abc --authenticationDatabase admin --authenticationMechanism MONGODB-CR

Read Preference Tuning

Choose appropriate read preference (primary, primaryPreferred, secondary, secondaryPreferred, nearest) based on latency and write load; generally secondaryPreferred or primaryPreferred for high‑write workloads.

Index Optimization

Common index types: unique, hashed, sparse, geo, TTL, text. Use {background:true} when building indexes to avoid write locks.

db.test.ensureIndex({"column1":1},{"unique":true,"sparse":true,background:true});

Follow the “left‑most prefix” rule; an index {a:1,b:1,c:1} can serve queries on a, a‑b, or a‑b‑c.

Understanding Query Plans

Use explain() to view execution details. Look for IXSCAN (index scan), COLLSCAN (collection scan), and SORT stages.

Performance Tips

Avoid Ctrl+C; use db.currentOp() and db.killOp() to cancel operations.

Set cursor.maxTimeMS(N) to bound execution time.

Project only needed fields and limit result size.

Batch inserts with bulkWrite() , use upsert:true for idempotent writes.

Prefer index‑driven updates and batch them to reduce load.

Pagination Strategies

Simple skip/limit works but degrades with large offsets. Use “range pagination” by remembering the last document’s sort key:

db.test.find({name:"li",num:{$gt:305}}).limit(3)

Array Handling

Appending to array tails logs only the new element; inserting elsewhere logs the whole array. Indexing large arrays is costly; limit element count and prefer $push with $slice.

Overall, the guide addresses frequent MongoDB pitfalls and provides practical configuration and optimization recommendations for developers.