I'm working on a project consisting of an integrated software solution and a private cloud. I'm trying to evaluate Proxmox compared to other solutions and I've run in to some performance issues with the storage design that I had in mind.
I was hoping someone familiar with the product could give me some ideas. No need to write a long response if you don't like, I'm pretty good at research & testing once I have an idea what to look for.
I know my way around Linux but I'm intermediate at best, at the moment all I need to do is get a proof of concept working; when the time comes I'll hire someone to do the implementation and knowledge transfer.
So here is the scenario that I am trying to solve:
I need to design an virtualization/cloud architecture for small low impact environments. Ideally it should be able to scale from 1 machine to around 50 machines or more. I can probably get away with a minimum of two nodes but I'm trying to keep the initial buy in as low as possible. Budgets are tight these days and I'm trying to be sensitive to that.
I'm used to using VMware and SANs to accomplish this but I have small budget for this project and I can't afford to go that route this time. So I'm not sure how to replicate the features that I'm used to getting with a SAN with Linux based solutions. I've seen people suggest using LVM & DRBD but I'm not sure how I would scale that beyond two machines without creating a management nightmare. I’m also not sure LVM will let me take advantage of the hardware that I’ve purchased like a SAN solution would.
There is going to be a variety of use cases for this solution so I'm trying to get the best performance I can, that way I don't have to worry about it later. My original design included a 16gb RAM cache (for caching reads), 5x120gb SSDs for frequently accessed data, 3x3tb Large HDDs for archival (snapshots, infrequently accessed data, large files).
The requirements I have to meet
The design has to be able to scale from 1 or 2 servers to many
Data redundancy without being wasteful of drive space
live migration
Snapshotting of virtual machines & data
Offsite disaster recovery
Won't require expert knowledge to maintain & upgrade
Hard drive parity with reasonable balance between performance, parity & usable drive space
Thank you for taking the time to read my post and I greatly appreciate any thoughts you might have. =)
I was hoping someone familiar with the product could give me some ideas. No need to write a long response if you don't like, I'm pretty good at research & testing once I have an idea what to look for.
I know my way around Linux but I'm intermediate at best, at the moment all I need to do is get a proof of concept working; when the time comes I'll hire someone to do the implementation and knowledge transfer.
So here is the scenario that I am trying to solve:
I need to design an virtualization/cloud architecture for small low impact environments. Ideally it should be able to scale from 1 machine to around 50 machines or more. I can probably get away with a minimum of two nodes but I'm trying to keep the initial buy in as low as possible. Budgets are tight these days and I'm trying to be sensitive to that.
I'm used to using VMware and SANs to accomplish this but I have small budget for this project and I can't afford to go that route this time. So I'm not sure how to replicate the features that I'm used to getting with a SAN with Linux based solutions. I've seen people suggest using LVM & DRBD but I'm not sure how I would scale that beyond two machines without creating a management nightmare. I’m also not sure LVM will let me take advantage of the hardware that I’ve purchased like a SAN solution would.
There is going to be a variety of use cases for this solution so I'm trying to get the best performance I can, that way I don't have to worry about it later. My original design included a 16gb RAM cache (for caching reads), 5x120gb SSDs for frequently accessed data, 3x3tb Large HDDs for archival (snapshots, infrequently accessed data, large files).
The requirements I have to meet
The design has to be able to scale from 1 or 2 servers to many
Data redundancy without being wasteful of drive space
live migration
Snapshotting of virtual machines & data
Offsite disaster recovery
Won't require expert knowledge to maintain & upgrade
Hard drive parity with reasonable balance between performance, parity & usable drive space
Thank you for taking the time to read my post and I greatly appreciate any thoughts you might have. =)
Last edited: