This topic describes the frequently asked questions (FAQ) about OCP clusters in the multi-cluster mode.
FAQ
Q1: Does the ConfigUrl service of an OCP cluster, a leader or follower OCP cluster, record only the Root Service information of OceanBase clusters managed by the OCP cluster or the Root Service information of all OceanBase clusters?
A: The ConfigUrl service of an OCP cluster only returns the Root Service list of OceanBase clusters managed by the OCP cluster.
Q2: Assume that my OCP cluster fails, the OceanBase cluster managed by it works properly, and the ConfigUrl service is no longer available. Do I need another OCP cluster to take over the OceanBase cluster and the OBProxy cluster for disaster recovery?
A: Yes, you do.
Q3: What will happen when my leader OCP cluster recovers after a disaster recovery switchover?
A: If your leader OCP cluster recovers after a disaster recovery switchover, it has no relationship with the new leader OCP cluster. If the original primary OceanBase cluster was switched over to a standby OceanBase cluster during disaster recovery, the original primary OceanBase cluster is no longer available.
Q4: My application system is bound to the endpoint of the primary OceanBase cluster. After a switchover of the primary and standby OceanBase clusters during disaster recovery, the endpoint of the standby OceanBase cluster involved in the application system was changed. Do I need to change the endpoint of the standby OceanBase cluster in my application system?
A: Yes. You must specify the cluster ID for the endpoint, for example, obclient -hxxx.xxx.xxx.xxx -P2881 -uroot@ocp_meta31:5 -p****** -A.