I like to upgrade 2 nodes vcs cluster.
Let`s say that i failover the service that runs in node 1 to node2,and i hastop -local ,then stop gab and llt,then unload gab and llt.
I upgrade vcs on node1.
Now if i like to move back the service to node 1,what are those steps?
If i start vcs on node1 then vcs on node1 and 2 have separate version and had can not work.right?
i can not fail over the service from node 2 to 1.
i can off line the resource group on node 2 ,then stop vcs on node 2 .
here i can start vcs in single node on node 1->this bring the service on node 1,then upgrade vcs on node 2 and on gabconfing -c n2.
or i upgrade vcs on node 2,then start vcs on both nodes.
how is correct?
the term "failover" means to move service groups(resources) between the nodes in the same running VCS nodes.
That is the nodes running different versions of VCS cannot form a cluster. ( slave nodes cannot join the running HAD).
however, if for whatever reason, you want to online the applications on a node, you can do it manually (outside VCS).
Please be aware of possible "concurrency violation (possible data corruption) when bring online applications manually outside VCS).