SCP can be deployed in cluster mode. You can switch SCP from the standalone mode to cluster mode to support horizontal expansion and high availability. After you switch to the cluster mode, tenants can use the load balancing function of ADC to access the nodes in the SCP cluster. Administrators can use the IP address of the SCP cluster to log in to the active node and then manage the SCP cluster.
Precautions
Currently, SCP can be deployed only in three-node cluster mode. The cluster can tolerate only one faulty node.
The nodes in the SCP cluster must run the same SCP version. The SCP nodes added to the cluster must not belong to any resource pool or user.
The configuration of the nodes added to the cluster must be higher than or equal to the configuration of the active node in the cluster.
When the active node is disconnected, if the databases of the active and standby nodes are not synced in real time, the cluster does not automatically switch to the standby node. This means that the roles of the active and standby nodes remain unchanged. You need to manually recover the active node before the cluster can trigger a switchover.
When a node is down, if you need the node to automatically recover, keep the SCP version and SP version of the faulty node consistent with those of the active node.
Only disconnected nodes can be removed from the cluster.
You cannot switch SCP from the cluster mode to the standalone mode.
When you deploy SCP in cluster mode, you cannot clone SCP VMs. You must deploy SCP by importing the SCP template multiple times.
Recommended Configuration
A three-node cluster significantly enhances the transaction processing capability of SCP. It ensures the high concurrency and high availability of SCP in large-scale delivery scenarios. In cluster mode, it is recommended to deploy SCP in a three-node cluster. The table below describes the configuration.
{{ $t('index.defaultHeader.chromeBrowserTip') }}