Add more Storage to ISAR ODF

[rbo]

Let's add more unused NVMe's to the ISAR ODF.

@dmoessne already removed on NVME from a UCS Box.

Node	Disks	CPU	RAM
ucs55	still have 1x1.8 TB NVME	60 cores	1TB
ucs56	still have 1x1.8 TB NVME	60 cores	1TB
ucs57	is the one @dmoessne already removed	60 cores	1TB
storm2	2x1.8 TB NVME,4x 1.5 TB SATA SSD's	72 Threads	756GB
storm3	2x1.8 TB NVME,4x 1.5 TB SATA SSD's	88 Threads	756GB
storm5	2x1.8 TB NVME,4x 1.5 TB SATA SSD's	88 Threads	756GB
storm6	3x1.8 TB NVME,4x 1.5 TB SATA SSD's	88 Threads	756GB

[rbo]

🤔 Maybe it makes more sense to get rid of ceph nodes and move the ceph to storm3,storm5 and storm6 because they have almost the same hardware configuration.

Because our ceph nodes are not RHEL 10 ready:

[rbo]

@DanielFroehlich what do you think?

[DanielFroehlich]

I had the same idea, but more from a capacity perspective - and make use of nice hardware.
I think the RHEL9/10 topic will still take a while to actually manifest. I doubt that ODF will switch to an UBI10 base image soon.
But we could prep ourself for that by e.g. adding already storm5+6 to the ODF cluster with one or two NVMEs. I would use storm5+6 because they have more cores, keep in mind each disks eats aways cores reserved for ODF.
Let keep this in mind and discuss next year, once we have the fusion in place. Maybe we drop ODF totally in favour of connecting ISAR to the fusion?

[DanielFroehlich]

As wee need to make more room for the fusion (#315), I think it would make sense to actually migrate storage away from the ceph nodes to the storm nodes. That would allow us to shutdown and remove the (small, old) ceph nodes, and keep the old but powerful UCS nodes (even more useful if the CPU upgrade idea #339 actually works)

[rbo]

Looks good: https://access.redhat.com/solutions/7134647