You question about if is possible to run in 4.3 the old manual upgrade instead of utility upgrade as shown in this image
The only way to upgrade cluster nodes individually is to take them out of the cluster and upgrade them as standalone nodes. This would require to take care of any scheduled jobs, making sure none of them runs on a node that is out of the cluster.
In order to upgrade the nodes individually, the following procedure can be used to minimize downtime.
1. Start at a time when no scheduled job is about to run.
2. Remove one node from each cluster site, one at a time. Make sure the first node is completely out of the cluster and reports itself as standalone node, before removing the second node, etc.
3. Reset the cluster configuration on the nodes that were taken out. Unlock them and remove any scheduled jobs that might otherwise be running in the future on those nodes. Upgrade them to 4.3.1 + 4.3.1.01 individually.
4. During the main upgrade with outage stop the remaining cluster. This starts the downtime.
5. Reset the cluster configuration on the master node and upgrade it to 4.3.1 + 4.3.1.01.
6. Configure a new cluster with the upgraded master and the nodes that had been upgraded upfront already, then start the cluster. This ends the downtime.
7. Gradually upgrade the remaining nodes similar to step 3, and add them to the cluster one by one. Again, make sure not to start adding a second node before the first node is fully integrated into the cluster, shows itself in sync and shows as in sync on the Configuration > Clustering page on the master node.