M3DB storage configuration

M3DB storage configuration {"num_shards": 1024, "replication_factor": 3, "instances": [{"id": "m3db001", "isolation_group": "us-east1-a", "zone": "embedded", "weight": 100, "endpoint": "10.142.0.1 instances" "hostname": "m3db001", "port": 9000}, {"id": "m3db002", "isolation_group": "us-east1-b", "zone": "embedded", "weight": 100, "endpoint": "10.142.0.2 id" "hostname": "m3db002-us-east", "port": 9000}, {"id": "m3db003", "isolation_group": "us-east1-c", "zone": "embedded", "weight": 100, "endpoint": "10.142.0.3 id" "hostname": "m3db003", "port": 9000}]} ID

Uniqueness distinguishing data storage node

Isolation Group isolation packet

The data stored by the storage nodes in the same partition are different from each other, in other words, the storage nodes in the same partition are equally divided into all data slices. There is no redundant data between the points of data in the same isolated group.

This feature is mainly used in highly available, most typical application scenarios, where data nodes in the same rack or data room will be placed in an isolated packet, which will be unavailable due to power outage in the rack or data room. Of course, it can also be used in the cloud environment.

Zone metadata storage area

This is a concept in etcd, the same as the concept of datacenter in consul

Weight data slicing weight

Fragment allocation weights, integers, heavy storage nodes in an isolated packet will be allocated to store more fragments. If machines with different disk capacities mix storage nodes, the weight of large disks can be higher than that of other nodes. In the scenario where the machine is also configured, the weights of all nodes can be kept the same integer.

Endpoint data node address

For example, 250.250.250.250, data requests for this data node will be routed to this address.

Hostname Hostnam

The hostname or IP of the data node

Port port

The port on which the data node listens for data requests. Default 9000