Not having updated the Cluster OS since its install (why break something when it works?) I find that Windows Server 2016 is reaching the point at which an upgrade to 2019 is called for.
I KNOW Microsoft do not recommend in place upgrades for Server OS, however it does work (mentioned in at least two places in Microsoft official 'how-to' documents - one about server upgrades in general the other in the Cluster Upgrade FAQ)
I have a two node cluster, using Stawind vSAN Free.
I think that the two possible workflows are
- move all the roles to Node2,
- upgrade the server OS in-place for Node1
- move the roles back to Node1
- upgrade the server OS in-place for Node2,
- move all the roles to Node2
- evict Node1,
- clean install 2019 on node 1 (without deleting the SW Volume),
- reinstall Starwind
- restore the SW Config
- add Node1 back in to the cluster
- rinse and repeat for Node2.
In this second workflow the re-install Starwind bit is the one that is kind of concerning - not that it is difficult, just because I haven't done it and want to minimise the chances for me to roayally screw things up!
For both there are also probably a couple of re-syncs needed as well (each one takes a while - about 30 hours; yes I know, I should have a 10G interface but I don't ) - plus an upgrade to Starwind latest version, but that can happen before the OS upgrade happens on any node.
anyone had an experience doing this either way or got any hints/tips?
Regards
Ben