Cluster/hardware sizing

[rbo]

We have to clarify the cluster/hardware sizing.

Initially, for development & testing, I started with:

Amount	Type	CPU	RAM	Disks
3	SB48	Intel Core i7-7700	64GB	2x SSD SATA 256 GB
1	SB49	Intel Core i7-6700	64GB	2x SSD SATA 256 GB

[tumido]

How many clusters we plan on deploying in Hetzner? Can we get more than one?

Also, I think Operate First would appreciate clusters dedicated to:

• storage (minimal HW requirements as per OCS docs here with raw capacity in tens of TiBs)
• computing/GPUs focused (at least similar to MOC Zero cluster - 9 nodes, in total: 260CPU cores, 400 GiB RAM, no storage needed + GPUs)

/cc @durandom @goern

[rbo]

How many clusters we plan on deploying in Hetzner? Can we get more than one?

Also, I think Operate First would appreciate clusters dedicated to:

• storage (minimal HW requirements as per OCS docs here with raw capacity in tens of TiBs)

Maybe we can try to use a compact deployment.
Don't want to waste too many resources on the control plane.

We can choose one of the AX* Servers. Minimum AX51-NVME or higher.

• computing/GPUs focused (at least similar to MOC Zero cluster - 9 nodes, in total: 260CPU cores, 400 GiB RAM, no storage needed + GPUs)

Sadly GPU is not available at Hetzner.

Controle Plane I recommend AX41-NVME or AX51-NVME
Infra Nodes: depend on what we want to deploy: OpenShift Logging?
Compute: AX101 or AX161

Maybe it makes sense to set up only one cluster to save resources ~ budget.

Try to get some cheap servers from the Serverbörse

/cc @durandom @goern

[durandom]

Let's think about our user base for this deployment.
The audience would be solution architects and others, get a namespace and you can do whatever you want.
A little bit more relaxed than the zero cluster.
And we should make sure we're running always at 80% capacity, since we pay for it. Therefore I'd start with the smallest supported setup and then grow by demand.

[goern]

BTW it look like we need some GPU for https://github.com/AICoE/Varangian
If no GPU is agailable at hetzner, @durandom do we get them from AWS?

[rbo]

Because of #8, we decided to start with:

Role	Amount	Type	CPU	RAM	Disks
Master	3	SB48	Intel Core i7-7700	64GB	2x SSD SATA 256 GB
Worker	3	SB94	Intel Xeon E5-1650V3	256GB	2x SSD SATA 480 GB

One of the workers will be used as a bootstrap during installation and added to the cluster after installation.

[sesheta]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

/lifecycle stale

[sesheta]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

/lifecycle rotten

[rbo]

initial sizing is done. We can close the issue.