@scottalanmiller said in Windows Failover Clustering... what are your views and why?:
@Dashrender said in Windows Failover Clustering... what are your views and why?:
@Obsolesce said in Windows Failover Clustering... what are your views and why?:
The thing is, they are VMs, you can move shit around whenever, to wherever, should the need arise.... and without any downtime if needed. I think each server/service/system should have it's own "SLA" (it's almost midnight can't think of the word now) and should be placed appropriately. Only you can answer whether or not it needs HA. You can do the math to figure out exactly what the cost of each GB of SSD capacity, vCPU, Memory, etc. costs for HA placement versus non HA and decide appropriately where to put the VM. I really don't think Hard Drive life is a concern here, you'll pull through the lifespan of the drive easily, or you won't because it's defective which in that case doesn't matter anyways on your decision. So I don't think that's a factor here. It all comes down to math regarding costs vs what is being considered for placement.
wait a second - this whole box was likely built to be 100% HA - so anything running on it, or planned to be running on it was likely scoped with the expectation of being on HA, wither or not HA was needed - at least that's my expectation.
The VM's being put on it now that the OP is talking about, likely were never originally on the board for this cluster/hardware - where they? I mean - where/why are these VMs a thing now and what was the plan for their placement? Did whomever wanted these VMs get the sign off from the ones that paid for the cluster? (devil's advocate)
I think he said that the originally engineering plan was NOT this. It was designed to be non-HA for some or most of the workloads. Super high HA just for select workloads.
Correct. We had to get HA for the subset of our workload. So, this had to be built. We needed less budget to extend the storage, RAM and CPU on these three machines for the non HA compared to having to build entirely separate machines for that workload.
Purely, that was the plan. My mind, it still is. Just some folk are pushing to make all HA. Even, for example, PDQ. We do not need PDQ to be HA. It's a small VM, sure. But it 100% does not need HA. Even if 50GB, we don't need that replicated three times! If it is on a host that does die, if need be, I can start the backup on my veeam box. Once the host is fixed, I can migrate it to the fixed hardware.
Even our webservers don't need HA. Stick one on each host and use HAProxy so when one is down, the we server is taken out of the pool. Sure, make the HAProxy HA, or roll out a HAProxy cluster (I'm sure that would have built in application level HA of some form) but, sure, make that HA if need be.
We have proprietary software on top of windows that isn't made with application level HA. That's why we need the failover clustering. Those VM for sure need to migrate should a physical die. Outside of that, we just don't need it. But for that special case, we do. That's why we have it.
Even domain controllers don't need to be on the CSV. Have one local to each server with the DC in the cluster, but not on shared storage. If a host does die, you still have two DCs online and cam migrate fmso roles (if the holder died). No need to be in the CSV.
But, some folk want all in CSV, and want to waste lots of CSV storage for things that don't need it.
The point of the CSV was that as we increase our proprietary tool over the next few years, there is room to do so. If that space is full of data that doesn't need HA, we have lost the opportunity to use it.