Proxmox VE

We used Proxmox to implement private cloud services, for clusters
of a small number of servers, from 3 to 11 with and without high
availability. Allways with ZFS file systems, and we used to install the
root pool in SSDs mirrored and use other pools with RAID 10 in groups of
four, for the virtual machines and containers, for the backups and snapshots, we used magnetic disks with RAID 10, in groups of four. Do
not use an even number of servers because does not facilitates the
implementation of High Availability, because the corosync service must
have an odd number of servers to detect a failed server for the quorum
system. We used a variety of servers, from clone PCs with AMD Ryzen with 6 cores and 12 threads with 64 GB of RAM no ECC, to high end servers with 64 cores and 128 threads per cpu and 2 cpus per server, with AMD EPYC Rome or Milan, 2 terabytes of RAM ECC.

Incentivized

• BackUp System, provides extensive propiertes and verifications. An exclusive server for BackUp Administration (PBS)
• Hardware Customization. You can select several properties to adapt the hardware to your needs
• Easy administration. You will be able to manage the server with easily accessible tools such as the web console and usage statistics.
• Certificates Administration.

• Can't manage ZFS replication or snapshots as easily as I'd like
• Encrypted disks aren't easily deployed in VMs
• Moving disks between VMs is not automated

Incentivized

Proxmox VE provides the most capable, yet stable virtualization platform in the market today. Licensing options are also competitive and cost-effective for support, and support is extremely fast and knowledgable of getting issues resolved as quickly and soundly as possible.

Incentivized

Out of every product I have used for this, Proxmox VE is the most concise, clear, and functional I have ever seen. I continue to use Proxmox VE even after occasionally comparing alternatives available because of it's usability, design concept, and great support of features. It's very unlikely I will find a product that can even compete with Proxmox VE in every angle of what Proxmox VE provides.

Incentivized

Proxmox VE's ha-cluster functionality is very much improved, though does have a not-very-often occurrence of failure. In a 2-node cluster of Proxmox VE, HA can fail causing an instance that is supposed to migrate between the two nodes stop and fail until manually recovered through the command-line tools provided. Other than this, the HA clustering capability of Proxmox VE has proven to be reliable in 3 or more clustered environments with much less chance of these failures to occur.

Incentivized

Proxmox VE's interfacing is always fast to load, both the Web interface and the command-line tool interfaces. Reporting is practically real time almost all the time, and you can see everything in mere seconds, easily able to identify if something is wrong or it everything is in tip-top shape as always desired

Incentivized

They are fast, understanding, very intelligent, know their product very well, fast, responsive, and concise. Need I say more?

Incentivized

It worked, was easy and super fast to deploy, and provided everything we needed in a matter of minutes

Incentivized

Proxmox VE is cheaper than VMware, especially upscaling an HA architecture. Compared with other free or less expensive solutions, Proxmox VE is high compatible with more types of hardware solutions and more VM types. From my point of view, Proxmox VE has no competitor at the same price level, it offers the most complete and production-ready HA solution.

Incentivized

Proxmox VE provides everything you need to quickly add new storage mediums, network and local, as well as networking interfaces, such as using Linux standard bridges and now Open-vSwitch bridges which can be even more scalable than before. Proxmox VE 4.0 dropped support for OpenVZ in favor of the more well supported and native LXC and made an upgrade path to it very simple.

Incentivized

• Proxmox has allowed to us to do more with less. We can invest in a single (or multiple if clustering) host with a decent specification, and run most of our infrastructure on it.
• Open source technologies allowed us to re-use previous skills and knowledge. There was very little onboarding required because we already knew Debian, KVM, ZFS, etc.
• Virtualisation has vastly reduced the amount of time required to maintain all our systems. Everything is so much more organised and lends itself to automation (with Ansible, in our case).

Incentivized