How to query slow logs and create indexes by mongo

This article is about how mongo queries slow logs and creates indexes. The editor thinks it is very practical, so share it with you as a reference and follow the editor to have a look.

1. Check whether the mongodb slow log is on.

Use LogDB

Db.getProfilingStatus ()

two。 Open the slow log and set the operation for more than 100 milliseconds as a slow operation

Db.setProfilingLevel (1100)

3. View slow log content

Db.system.profile.find () .sort ({$natural:-1})

Journal,

Shard3:PRIMARY > db.system.profile.find () .sort ({$natural:-1})

{"op": "query", "ns": "LogDB.Flashxxx", "command": {"find": "Flashxxx", "filter": {"UserId": {"$in": ["1111111"]}}, "shardVersion": [Timestamp (0,0), ObjectId ("0000000000000000000000000"), "lsid": {"id": UUID ("d08179ba-ecbd-4f9d-92e7-cd491e5a1d8a") "uid": BinData (0, "iwsImy7Wfk2mAp4o/uuD+I9mpETQTxb5PXo26pIQkS4=")}, "$clusterTime": {"clusterTime": Timestamp (1558696199, 3), "signature": {"hash": BinData (0, "Ht7EWhYpVLiy+DbH8Cu1Ijs6PNk="), "keyId": NumberLong ("6647500774429425691")}}, "$client": {"driver": {"name": "mongo-csharp-driver", "version": "2.7.0.0"} "os": {"type": "Windows", "name": "Microsoft Windows 6.3.9600", "architecture": "x8600064", "version": "6.3.9600"}, "platform": ".NET Framework 4.0.0.0", "mongos": {"host": "new-mongo01:20000", "client": "10.205.33.32 name 59518" "version": "3.6.9"}, "$configServerState": {"opTime": {"ts": Timestamp (1558696196, 12), "t": NumberLong (3)}}, "$db": "LogDB"}, "keysExamined": 0, "docsExamined": 37428, "cursorExhausted": true, "numYield": 37428 "locks": {"Global": {"acquireCount": {"r": NumberLong}}, "Database": {"acquireCount": {"r": NumberLong (293)}}, "Collection": {"acquireCount": {"r": NumberLong (293)}, "nreturned": 1, "responseLength": 673, "protocol": "op_msg" "millis": 104, "planSummary": "COLLSCAN", "execStats": {"stage": "COLLSCAN", "filter": {"UserId": {"$eq": "11111111115"}}, "nReturned": 1, "executionTimeMillisEstimate": 100, "works": 37430, "advanced": 1, "needTime": 37428, "needYield": 0, "saveState": 37428 "restoreState": 292, "isEOF": 1, "invalidates": 0, "direction": "forward", "docsExamined": 37428}, "ts": ISODate ("2019-05-24T11:09:59.346Z"), "client": "10.205.34.91", "allUsers": [{"user": "_ system", "db": "local"}] "user": "_ _ system@local"}

Note that COLLSCAN is a full table scan.

View the execution plan:

Db.FlashClientData.find ({UserId: 1000493111}) .explain ()

Shard3:PRIMARY > db.FlashClientData.find ({UserId: 100049111}). Explain ()

{

"queryPlanner": {

"plannerVersion": 1

"namespace": "LogDB.Flashxxx"

"indexFilterSet": false

"parsedQuery": {

"UserId": {

"$eq": 10x0493111

}

}

"winningPlan": {

"stage": "COLLSCAN"