We have some business data with a large amount of data, which are stored in sharding way. This can scatter the data to different machines.
We also have some log data, which are made into capped according to one collection per tenant. only a certain amount of data is reserved for batch processing in the background. the number of records in each collection is not very large, and it is not meaningful to use s-hardening. moreover, capped collection cannot be s-hardening either.
The problem now is that the number of such collections is still quite large. If they are all concentrated on one server, the load on that server will become larger, which will affect the reading and writing of the collection originally sharding listed above.
Is there any way to better distribute these capped collection to various shardservers?
Shard Svr1, Shard Svr2, Shard Svr3
log_1, log_2, log_3
log_4, log_5, log_6
In database, you can use themovePrimarySeparate unsharded collection from different database.