What if jumbo sets to false in sharding

What is the meaning of jumbo = false in sharding?

I have followed all steps, like import both product.json and products.part2.json and still chunks showing 1 & 2 Mbs for two shards
MongoDB Enterprise mongos> db.products.find().count()
775174
MongoDB Enterprise mongos>

m103.products
shard key: { “name” : 1 }
unique: false
balancing: true
chunks:
aws-repl01 1
** aws-repl02 2**
{ “name” : { “$minKey” : 1 } } -->> { “name” : “In Meiner Mitte - CD” } on : aws-repl01 Timestamp(2, 0)
{ “name” : “In Meiner Mitte - CD” } -->> { “name” : “Tha Shiznit: Episode 1 - CD” } on : aws-repl02 Timestamp(2, 1)
{ “name” : “Tha Shiznit: Episode 1 - CD” } -->> { “name” : { “$maxKey” : 1 } } on : aws-repl02 Timestamp(1, 2)

{
id" : "m103.products-name"In Meiner Mitte - CD”",
“lastmod” : Timestamp(2, 1),
“lastmodEpoch” : ObjectId(“5d9a391d6b252936ec88f744”),
“ns” : “m103.products”,
“min” : {
“name” : “In Meiner Mitte - CD”
},
“max” : {
“name” : “Tha Shiznit: Episode 1 - CD”
},
“shard” : “aws-repl02”,
"jumbo" : false,
“history” : [
{
“validAfter” : Timestamp(1570388252, 3),
“shard” : “aws-repl02”
}
]
}

will this cause any problem like splitting or balancing, if jumbo is false and what if jumbo is true.

Thanks
Krishna

In Chapter 3, there’s a lecture that explains Jumbo Chunks. You’ll soon get to that lecture! In short, jumbo=false is a good thing.

Did you run the db.settings.save(<parameters>) part? Maybe it just needs some time to balance out. And this 1 & 2 are the number of chunks, not the size of the chunk. Look at the settings collection in the config db to check the size.

By the way, running count() on a sharded cluster doesn’t return an accurate result. You’ll need to add criteria to get an accurate count.

Hi @kmk050193,

In additions to @007_jb,

It means, the chunk that you are looking at is not a jumbo chunk. If MongoDB cannot split a chunk that exceeds the specified chunk size, MongoDB labels the chunk as jumbo by setting the jumbo flag to true.

It is not a good thing to have a jumbo chunk in your sharded environment.

{"jumbo" : false} is what you should always have in your sharded environment.

If {"jumbo":true} then the the balancer won’t move it which could potentially result in performance degradation in the sharded environment.

Please refer this post on how to clear the jumbo flag and fix the issue if you have one of these in your sharded cluster.

The count() function didn’t return the accurate count of the document because it does not filter the orphans/migrants.

I think what you are looking for here is cursor.itcount() function.

Hope it helps!

If you have any other issue, please feel free to get back to us.

Happy Learning :slight_smile:

Thanks,
Shubham Ranjan
Curriculum Support Engineer