|
This is the top level HA topology, which addresses even a site disaster recovery. It is the most complex environment to manage, with repetitive administration tasks for each one of the cells. On the other hand, it is also the most flexible environment and allows you to acquiesce a full cell at a time without having any availability issues.
Figure 5-1 shows this architecture.
![]()
Figure 5-1 WebSphere topology with multiple cells
For the scenario depicted in Figure 5-1, server affinity must be enabled at the Load Balancer to make sure that subsequent requests are routed to the same cell. Another option is to use two different URLs to access the different cells.
Any variation of this topology can be used, as long as the following steps can be performed:
1. | Isolate a branch (cell). |
2. | Wait for all sessions on this branch to expire. |
Important: You have to wait for the sessions to end, because there is no replication between the cells and thus no failover. |
3. | Update the offline branch. |
4. | Make the updated branch available again. This branch now serves the new application version while the other branch keeps on serving the old application version. |
5. | Repeat steps 1 to 4 for the second branch. |