MongoDB CursorNotFound Error on collection.find() for a few hundred small records

Question

I'm running on Mongo 3.6.6 (on a small Mongo Atlas cluster, not sharded) using the native Node JS driver (v. 3.0.10)

My code looks like this:

const records = await collection.find({
  userId: ObjectId(userId),
  status: 'completed',
  lastUpdated: {
    $exists: true,
    $gte: '2018-06-10T21:24:12.000Z'
  }
}).toArray();

I'm seeing this error occasionally:

{
  "name": "MongoError",
  "message": "cursor id 16621292331349 not found",
  "ok": 0,
  "errmsg": "cursor id 16621292331349 not found",
  "code": 43,
  "codeName": "CursorNotFound",
  "operationTime": "6581469650867978275",
  "$clusterTime": {
    "clusterTime": "6581469650867978275",
    "signature": {
      "hash": "aWuGeAxOib4XWr1AOoowQL8yBmQ=",
      "keyId": "6547661618229018626"
    }
  }
}

This is happening for queries that return a few hundred records at most. The records are a few hundred bytes each.

I looked online for what the issue might be but most of what I found is talking about cursor timeouts for very large operations that take longer than 10 minutes to complete. I can't tell exactly how long the failed queries took from my logs, but it's at most two seconds (probably much, much shorter than that).

I tested running the query with the same values as one that errored out and the execution time from explain was just a few milliseconds:

"executionStats" : {
    "executionSuccess" : true, 
    "nReturned" : NumberInt(248), 
    "executionTimeMillis" : NumberInt(3), 
    "totalKeysExamined" : NumberInt(741), 
    "totalDocsExamined" : NumberInt(741), 
    "executionStages" : {...}
    }, 
    "allPlansExecution" : []
    ]
} 

Any ideas? Could intermittent network latency cause this error? How would I mitigate that? Thanks

Answer 1

You can try these 3 things:

---

a) Set the cursor to false

db.collection.find().noCursorTimeout();

You must close the cursor at some point with cursor.close();

---

b) Or reduce the batch size

db.inventory.find().batchSize(10);

---

c) Retry when the cursor expires:

let processed = 0;
let updated = 0;

while(true) {
    const cursor = db.snapshots.find().sort({ _id: 1 }).skip(processed);

    try {
        while (cursor.hasNext()) {
            const doc = cursor.next();

            ++processed;

            if (doc.stream && doc.roundedDate && !doc.sid) {
                db.snapshots.update({
                    _id: doc._id
                }, { $set: {
                    sid: `${ doc.stream.valueOf() }-${ doc.roundedDate }`
                }});

                ++updated;
            } 
        }

        break; // Done processing all, exit outer loop
    } catch (err) {
        if (err.code !== 43) {
            // Something else than a timeout went wrong. Abort loop.

            throw err;
        }
    }
}

Answer 2

First of all, if your data is too big it's not a good idea to use toArray() method, instead it's better to use forEach() and loop throw the data. Just like this :

const records = await collection.find({
  userId: ObjectId(userId),
  status: 'completed',
  lastUpdated: {
    $exists: true,
    $gte: '2018-06-10T21:24:12.000Z'
  }
});

records.forEach((record) => {
    //do somthing ...
});

Second, you can use {allowDiskUse: true} option for getting large data.

const records = await collection.find({
  userId: ObjectId(userId),
  status: 'completed',
  lastUpdated: {
    $exists: true,
    $gte: '2018-06-10T21:24:12.000Z'
  }
},
{allowDiskUse: true});