I’m testing ReductSotre with single server/instance, single drive.
Can I scale it up to something like one instance with multiple drives? and multiple instances across multiple servers?
Can ReductStore work like a cluster?
1 Like
ReductStore supports the one writer + many read-only replica approach: Disaster Recovery | ReductStore
This makes the data highly available, but ingestion always goes through one writer.
The replicas can automatically run and scale on a cluster if you provide a network folder or S3-compatible storage.
To scale ingestion, run multiple ReductStore instances behind a load balancer or reverse-proxy and redirect traffic per bucket.
This method is a bit tricky and cannot be scaled automatically, so you need to create routes and rules for segmentation.
ReductStore currently doesn’t have multi-disk support.
However, since buckets are mapped to folders, you can mount disks to that folder at the operating system level. I would use a RAID array instead, though.
I hope it helps.