Hi, new here I had a question concerning the hardware requirements part, to be more precise on the storage size. If I understood it well, in theory we should have at least two separate drives, one for managing (with etcd and other stuff) and one for data (VM and stuff), all the storage recommendations in the hardware requirements documentation is for the management node storage right (so 0.5/1TB) and the 5000+ random iops part applies to it as well and no specific recommendations for the other drive. Regarding the 5000+ random iops, is there a benchmark list somewhere with drives? Is it a number obtained with something like fio and a 4k randwrite test? Currently setting up my 3 node cluster homelab, everything works from single 1TB skhynix p41 and I was considering adding a second drive for each nodes 😁

are you running 10gb networking? the reason I ask is that your drive speed is MUCH faster than the network. Longhorn needs as much network as possible. for “prod” they recommend 10gb networking and all SSDs. for “prod” I also recommend a 3rd party storage device for greater HA and flexibility. For your homelab I will not see any real difference with the second drives. Unless you are running out of space. lol

Well I have a 2.5G nic and a 10G SFP+ nic, I was considering swapping for a dual SFP+ 10G nic in the future. I have a nas over 10G (but it's HDD storage) as well and another node that could provide storage over QSFP+ 40G.

That would work nicely. 10gb all the things. Especially for harvester/longhorn