GaussDB Notes and Constraints

To ensure the stability and security of GaussDB, certain constraints are put in place for access or permissions control. Table 1 describes such constraints.

Table 1 Function constraints

Function	Constraints
Database access	Security group rules must be added to allow the ECSs to access the GaussDB instances. By default, a GaussDB instance cannot be accessed by an ECS in a different security group. To allow it, you must add an inbound rule to the GaussDB security group. The default port number of GaussDB instances is 8000.
Deployment	ECSs where DB instances are deployed are not directly visible to you. You can only access the DB instances through IP addresses and database ports.
Database root permissions	Only the root user permissions are available on the instance creation page.
GaussDB instance reboot	GaussDB instances cannot be rebooted through commands. They must be rebooted on the management console.
GaussDB backup files	GaussDB backup files are stored in OBS buckets and are not visible to you.
Configuration modification	By default, the specifications cannot be reduced. If you need to reduce the specifications, contact customer service. Before you change the instance specifications, ensure that the instance is available. The specifications of an instance cannot be changed if the instance or its node is abnormal, or the storage is full. During the specification change for HA (1 primary + 2 standby) instances, a primary/standby switchover is triggered. During the switchover, services are interrupted for about 1 minute. After you change the CPU/memory specifications of an instance, the instance will be rebooted and services will be interrupted. Therefore, select off-peak hours to perform this operation. Rebooting a DB instance will clear the cached memory in it. You are advised to reboot it during off-peak hours.
Failover	For primary/standby instances, services are unavailable for about 10 seconds when the primary node is being switched to the standby node.
Data restoration	To prevent data loss, you are advised to back up critical data before data restoration.
Storage	If the storage of an instance is full, no more data can be written to databases. You are advised to periodically check the storage and scale it up as needed.
Performance tuning	Performance tuning may need to reboot the instance and services may be interrupted.

last updated: 2025-10-17 09:42 UTC - commit: f74f6a5b880367e02ae05babd56ab31579b53964