The problem is that I know that the OpenVZ to LXC migration will drop some required features for my use case in terms of isolation of untrusted containers, so whilst I'm sure 3.x to 4.x migration will be fine for many users, for some it won't be until LXC (and then Proxmox 4.x) gets those features. I fully expect LXC (and so Proxmox 4.x) to get there, but it's not there yet for my use case.
Understanding Proxmox 3.4 EOL and 4.0
- Thread starter adamb
- Start date
The problem is that I know that the OpenVZ to LXC migration will drop some required features for my use case in terms of isolation of untrusted containers, so whilst I'm sure 3.x to 4.x migration will be fine for many users, for some it won't be until LXC (and then Proxmox 4.x) gets those features. I fully expect LXC (and so Proxmox 4.x) to get there, but it's not there yet for my use case.
Agreed. Use case has a LOT do with things. It's unfortunate that so many have come to depend on OpenVZ ... I've been tracking them for years and have always noticed quite a lot of instability in how that teams works. They haven't been and still are not responsive to the needs of their users unless those users have large wallets. KVM on the other hand, although is not as big a performer from being full virtualization vs containerization, has been solid since the beginning and got even more so once RedHat brought them into the fold.
I totally understand what you are saying though ... if you've come to depend on something you want to know it will be well supported going forward...
Nobody is this thread has uttered that Proxmox does not contain a smooth and clean upgrade path, on the contrary. What this thread is all about, IMHO, is that moving from 3.x to 4.x means replacing OpenVZ with LXC which all in itself is not difficult. The problem is that some features available in OpenVZ is scheduled for later implementation in LXC and that some users considers these features a show stopper for their current business case. Everybody is convinced that these missing features (production ready) will eventually come to LXC but they are not there at the moment these features are, not excluding others, live migration, live backup, and isolation from host. Live migration and live backup is closed connected and are scheduled for sometime this year - I expected medio 2016, while host isolation seems further ahead since this requires kernel support - I don't expect a production feature complete support until late in the kernel 4.5 or early 2.6.
Just my two cents;-)
