Hi everyone,
I’d like to continue this discussion, as I have a similar situation to what Vince described earlier.
I currently have 6 shards, each holding a single chunk of approximately 260 GB. The shard key is a hashed GUID.
Now, I want to add 3 new shards and rebalance the existing data evenly across all 9 shards.
I ran the following command:
db.adminCommand({
moveChunk: “db.coll”,
find: { HashedGUID: “-3074457345618258602” },
to: “mongodb-load-mongodb-sharded-shard-6”
})
This successfully moved the entire chunk to the new shard — but it was not split. So the whole 260 GB chunk ends up on one shard.
What is the correct way to rebalance data across new shards in MongoDB 8, given that each chunk is so large? I want the existing data to be redistributed evenly across all 9 shards (including the new ones).
Thanks a lot.
Output of getShardDistribution()
{
data: “259.54GiB”,
docs: 116906238,
chunks: 1,
“estimated data per chunk”: “259.54GiB”,
“estimated docs per chunk”: 116906238
}
Totals Across All Shards
{
data: “1557.11GiB”,
docs: 701375773,
chunks: 6,
“Shard mongodb-load-mongodb-sharded-shard-0”: [
“16.66 % data”,
“16.66 % docs in cluster”,
“2KiB avg obj size on shard”
],
“Shard mongodb-load-mongodb-sharded-shard-1”: […],
“Shard mongodb-load-mongodb-sharded-shard-2”: […],
“Shard mongodb-load-mongodb-sharded-shard-3”: […],
“Shard mongodb-load-mongodb-sharded-shard-4”: […],
“Shard mongodb-load-mongodb-sharded-shard-5”: […]
}
(All shards have identical load, ~16.66% each, with 1 chunk per shard.)