Hardware Profile for Management Nodes

[#1]

I'm in need of some advice with regard to a hardware configuration of Management nodes in a soon to be production cluster.

Hardware was purchased for an HDP Production cluster prior my involvement.

The hardware specs for 6 Management Nodes is as follows:

HP DL380

- 2x 10 core Xeon

- 128GB of RAM

- 2x 300GB SAS for OS

- 4x 400GB SAS SSD's for application related Services in a RAID 10 configuration.

We're trying to determine if the disk space will be adequate to handle a 150 Node Cluster and expansion up to 250 in the future.

[#2 · Accepted solution]

I would say go with 256 gb ram. but 128 is def do-able and often seen in the field. Go with Dual 10Gbps NIC. I would go with at least 12-2TB drives broken down into

2 OS drives with raid-1 mirror.

6 drives in raid-10 (metadata)

4 spare drives