What is our primary use case?
It is implemented in several Private Cloud Data Centers, from 5 to 16 nodes, from 256 GB RAM to 1 TB RAM, all using Supermicro Systems, running SQL servers, Application Servers ( SAP, Prothevs and some tailored Apps developed in house ) VDI and Files Repositories . all smooth and trouble free .
How has it helped my organization?
With live migration, there is no downtime when any node needs service or upgrades. Plus, when I need more memory, I just allocate it; more disk space? same. It is possible to administer almost anything remotely, now it takes months till a customer needs a weekend visit to service something, before Proxmox, it used to happens once a month!
What is most valuable?
Live migration :
Is simplifies the daily work, and reduces the downtime to nearly 0 .
Clonning and Snapshots :
Verry useful for development testing and backup .
Histogram about i/o, memory, CPU and network usage :
Helps to find the issue when some VM are running in some kind of bottleneck .
What needs improvement?
Mostly simplify web interface, simple things like the use of right mouse button to change VM attributes, make it more intuitive, but it is just cosmetic improvements, the core functions are all there.
Version 6.x improved the support for Containers to the PRO level .
For how long have I used the solution?
re than 10 years, since version 1.6 all the way to version 6.2
What do I think about the stability of the solution?
Never had any issue with stability, even in the very earliest stages of development...
What do I think about the scalability of the solution?
Not for now, the 16 nodes per cluster limit have been removed in the 5.x release !
How are customer service and technical support?
Customer Service: OK, in fact the little we've used it, the site documentation was really good.Technical Support: Same as above.
Which solution did I use previously and why did I switch?
Yes, Citrix XEN Server and EMC VMWare Server. I left Citrix for security, there are too many "black box" in the binaries that cannot be audited. In the VMWare case, for licenses politics, it uses a "per socket" schema that punishes high density servers with 2 or 4 sockets .
Is is the tool of choice used in the entire Facebook infra-structure, so it must be "at least" good enough .
How was the initial setup?
The first test was a little complex, in the time of the 1.4 version, but since the 2.0 version it is extremely simple, takes less then an hour to deploy a node and connect it to an existing cluster.
The 6.x release made things even easier, most maintenance tasks are now in the web interface
What about the implementation team?
I am an IT consultant, and I do everything myself (with my team, we are 4 people!)
What was our ROI?
Outstanding, worth every single cent spent on it!
What's my experience with pricing, setup cost, and licensing?
The deployment itself cost nothing, it is Opensource software! The monthly subscription costs 17 euros per node, and it covers anything we needed so far...There are more options, but it is more then our customers needed until now .
Which other solutions did I evaluate?
Several, From Red Hat to Ubuntu, and everything I can find with a Google search, I spent 6 months with these research and tests!
The Facebook contract came well after this tests .
What other advice do I have?
Do it! The simplicity will surprise you!
Which deployment model are you using for this solution?
Which version of this solution are you currently using?