Hi all,
I have a couple of servers and I want to create a system for production with proxmox. I searched a lot recently, but all experiences have different scenario. I am asking for hardware perspective.
My system includes: 4 supermicro 1U chassis, all servers can have 8 SSDs. Motherboards are gigabyte H470 HD3, processors are core i5 10400 LGA1200. All servers have 64GB of ram (I can increase up to 128GB of ram).
Using this system, I want to achieve high availability. And easy to replace failed hardware. Because of this, I want to use cluster system for 3 of servers, and one server will be using network attached storage (if one server is not sufficient for this I can add 1 or 2 more servers with same configuration). All servers have 512 GB SSD WD Blue, but I will change SSDs to more premium ones, and network attached storage I am planning to buy 1TB SSDs.
Currently, these servers are brand new. They will be used for some VMs and for some dockers. My plan is to ensure high availability with clustering, dockers are already supplying this situation, but they are rely on storage server.
My wish is to easily replace SSDs, because these servers will be at datacentre, and datacentre will replace failed SSDs.
Please could you help me to make proper hardware configuration? Currently, I do not have any HBA or RAID card.
For running proxmox cluster, is it a good choice buying HBAs and using zfs, or directly buying LSI raid card for each server which will join into cluster?
For running network attached storage, , is it a good choice directly buying LSI raid card? Or how can you achieve high availability of storage.
All servers will communicate through private FS switch (S3900-24T4S, 24-Port 10/100/1000BASE-T Gigabit).
If you can help me to decide, it will be much appreciated. I am so confused. My main concern is which way will be best way to lowerage man power when disaster occurs.
If raid card dies in one of cluster servers, what do I expect in a cluster environment? Is it enough to replace died raid, installing proxmox to that server and joining as a node as brand new node?
Thank you so much for your helps
I have a couple of servers and I want to create a system for production with proxmox. I searched a lot recently, but all experiences have different scenario. I am asking for hardware perspective.
My system includes: 4 supermicro 1U chassis, all servers can have 8 SSDs. Motherboards are gigabyte H470 HD3, processors are core i5 10400 LGA1200. All servers have 64GB of ram (I can increase up to 128GB of ram).
Using this system, I want to achieve high availability. And easy to replace failed hardware. Because of this, I want to use cluster system for 3 of servers, and one server will be using network attached storage (if one server is not sufficient for this I can add 1 or 2 more servers with same configuration). All servers have 512 GB SSD WD Blue, but I will change SSDs to more premium ones, and network attached storage I am planning to buy 1TB SSDs.
Currently, these servers are brand new. They will be used for some VMs and for some dockers. My plan is to ensure high availability with clustering, dockers are already supplying this situation, but they are rely on storage server.
My wish is to easily replace SSDs, because these servers will be at datacentre, and datacentre will replace failed SSDs.
Please could you help me to make proper hardware configuration? Currently, I do not have any HBA or RAID card.
For running proxmox cluster, is it a good choice buying HBAs and using zfs, or directly buying LSI raid card for each server which will join into cluster?
For running network attached storage, , is it a good choice directly buying LSI raid card? Or how can you achieve high availability of storage.
All servers will communicate through private FS switch (S3900-24T4S, 24-Port 10/100/1000BASE-T Gigabit).
If you can help me to decide, it will be much appreciated. I am so confused. My main concern is which way will be best way to lowerage man power when disaster occurs.
If raid card dies in one of cluster servers, what do I expect in a cluster environment? Is it enough to replace died raid, installing proxmox to that server and joining as a node as brand new node?
Thank you so much for your helps
