Upgrade Path and Cluster Redundancy Deployment
Active-Standby Mode
When an active cluster becomes inaccessible for APs, external DPs (vSZ-D), and ICX switches, a standby cluster restores the latest configuration of the out-of-service (OOS) active cluster, then takes over all external devices (including APs, external DPs, and ICXswitches). The AP or ICX switch capacity is limited by the AP or ICX switch High Availability (HA) licenses on the standby cluster and the services license limits from the failed active cluster. When the active cluster returns to the in-service state, the end user can "rehome" all APs, external DPs, and ICX switches back to the active cluster.
The behavior of the standby cluster changes automatically when there is a configuration change in the following deployment types:
- One-to-one (one active
cluster to one standby cluster) deployment
The standby cluster restores the configuration from the active cluster after the configuration synchronization is completed. The standby cluster is always in backup mode and ready to receive the APs, external DPs, and ICX switches from the out-of-service active cluster.
- Many-to-one (two or three
active clusters to one standby cluster) deployment
The time taken by the standby cluster between detecting the active cluster is out-of-service and being ready to serve APs and external DPs has been enhanced.
Upgrade Path
| Base Version | Upgrade Version | Topology | Upgrade Flow |
|---|---|---|---|
| 3.6.x | 5.x | 1-to-1 | SOP C Only 1-to-1 deployment in R3.6 |
| 5.0.x | 5.1.x | 1-to-1 | SOP B |
| N-to-1 (N ≧ 2) | SOP A | ||
| 5.1.x | 5.2.x | 1-to-1 | SOP B |
| N-to-1 (N ≧ 2) | SOP A | ||
| 5.2.x | 6.0.x | 1-to-1 | SOP B |
| N-to-1 (N ≧ 2) | SOP A | ||
| 6.0.x | 6.1.x | 1-to-1 | SOP B |
| N-to-1 (N ≧ 2) | SOP A |
Active-Active Mode
When there are multiple clusters, one cluster can be the configuration source cluster, and all other active cluster restores its configuration periodically to make sure the configuration between the clusters are synchronized constantly. When the active cluster becomes inaccessible for APs and external DPs (vSZ-D), they fail over to the target active cluster with priority. Refer to SOP Z for more information.