- The geographical location of the server for the ApsaraDB for MongoDB instance that
you create. You must specify a region when you create an ApsaraDB for MongoDB instance.
The region cannot be changed after the instance is created.
- When you create an ApsaraDB for MongoDB instance, you must use an Elastic Compute
Service (ECS) instance. ApsaraDB for MongoDB supports access over the internal network.
The specified region must be the same as that of the ECS instance. For more information,
see Connect to an ApsaraDB for MongoDB instance over the internal network across zones.
- A physical area with an independent power supply and network in a region.
- Resources in zones that are in the same region can communicate with each other over
the internal network. The network latency within a zone is lower that across zones.
Faults are isolated between zones.
- In the single-zone deployment mode, the three nodes of an ApsaraDB for MongoDB replica
set instance are deployed in the same zone. If an ECS instance and an ApsaraDB for
MongoDB instance are both deployed in the same zone, network latency is reduced.
- An ApsaraDB for MongoDB instance, which is the basic unit of ApsaraDB for MongoDB.
- An instance is the runtime environment for ApsaraDB for MongoDB and exists as a separate
process on a host.
- You can create, modify, and delete an instance by using the ApsaraDB for MongoDB console.
Instances are independent and their resources are isolated. They do not compete for
resources such as CPU, memory, and I/O.
- Each instance has its own unique features, such as database engine and version. ApsaraDB
for MongoDB controls instance behavior by using the parameters that correspond to
||The maximum memory that an instance can use.
- The disk size that you select when you create an ApsaraDB for MongoDB instance.
- The disk capacity occupied by the instance includes datasets and the space required
for normal instance operations, such as the system database, database rollback log,
redo log, and indexes.
- Make sure that an ApsaraDB for MongoDB instance has sufficient disk capacity to store
data. Otherwise, the instance may be locked. If the instance is locked due to insufficient
disk capacity, you can unlock the instance by expanding the disk capacity.
||The maximum number of read/write operations performed per second on block devices
at a granularity of 4 KB.
The maximum computing power of an ApsaraDB for MongoDB instance.
A single Intel Xeon series CPU core has at least 2.3 GHz of computing power with hyper-threading
The number of Transmission Control Protocol (TCP) connections between a client and
an ApsaraDB for MongoDB instance.
If the client uses a connection pool, the connections between the client and the instance
are persistent connections. Otherwise, they are short-lived connections.
|sharded cluster instance
||An ApsaraDB for MongoDB sharded cluster instance. You can purchase multiple mongos
nodes, multiple shard nodes, and a Configserver node to create an ApsaraDB for MongoDB
sharded cluster instance, which serves as a MongoDB distributed database system.
- The routing service that processes requests. All requests must be coordinated by using
mongos that serves as a request distribution center and forwards data requests to
the corresponding shard server.
- You can use multiple mongos nodes to process requests. If one mongos node fails, other
mongos nodes can continue to process the requests.
- An ApsaraDB for MongoDB instance that holds a subset of the sharded data.
- Each shard can be deployed as a three-node replica set to increase availability. You
can create multiple shards to improve read and write performance and expand storage
capacity. This way, you can implement a distributed database system based on your
application performance and storage requirements.
- A configuration server that stores all database metadata for mongos nodes and shards
in an ApsaraDB for MongoDB sharded cluster instance. Mongos nodes cache shard data
and data routing information in their memory, whereas Configservers store such data.
- When mongos nodes in a sharded cluster instance are started for the first time or
shut down and then restarted, they load configuration information from the Configserver
node. If the information of the Configserver node changes, all mongos nodes are notified
to update their status. This ensures that mongos nodes can always obtain the correct
- Configserver nodes store metadata of shards and routers and have high requirements
for service availability and data reliability. ApsaraDB for MongoDB uses three-node
replica sets to ensure the reliability of the Configserver nodes.