New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
OpenVz providers : what happens after Nov 2019?
OpenVz 6 goes end-of-life in November 2019, that's less than 2 years
I'm curious to know what OpenVz providers plan to do at (or before) that time:
- Migrate containers to OpenVZ/Virtuozzo 7;
- Move to another container virtualization (e.g., LXC);
- Move to full virtualization, e.g. KVM, probably with price changes;
- Something else...
Option #1 seems the logical choice but from what I've read it's problematic, at least for now.
Comments appreciated.
Thanked by 1bersy
Comments
Right now number 1 seems to be the obvious option however it's probably something that a more firm decision will be made on next year as the supporting software catches up.
Its overdue already, Kernel from 2006 which is not more supported hold together by kernel care.
KVM obviously, what else?
Nobody should be buying OpenVZ services anymore when KVM is really well priced these days.
Everyone switches to proxmox and save even more money.
@JoeMerit
Well priced !== dirt cheap
Number 1. We can't even considering number 3 because even though our KVM plans are cheaper than our OpenVZ plans, people still buy OpenVZ for some reason.
Honestly, i'll probably just really press for people to move to slices instead. I may need to add some "active ovz clients only" plans just to get it done, but I think it's the best option.
I've had unpredictable results when it comes to converting an OVZ to a KVM so I can't just 'force' the process myself either.
Francisco
@Neoon People still remember that crappy movie?
Much like they still OpenVZ
You better hope that CentOS 7 plays happy with the networking if you do migrate. Takes ages to repair
The world turns upside down (becomes flat, no longer round), nuclear war breaks out, and we reset living in caves, hiding from the shadows of the wildabeast.
I think some of the provider maybe just continue use the vz6
I can answer this with certainty:
Leave old nodes on VZ6 and sell new packages on VZ7. Whatever happens to the old nodes happens. Some may warn "can't migrate, EOL, take your chances that this continues to function without discovered vulnerability or order a new server" and some won't.
I'd bet money on it. With what I know today, it's what I'd do. Maybe even offer first X months free of new server. Sometimes you have to work toward retiring old servers and sometimes you can't reasonably migrate. The industry goes through this every few years but most just blame the providers for this or that and never talk about the real reasons.
@jarland is basically correct.
I stopped selling OpenVZ years ago and have encouraged people to move to KVM and I still have tons of people using OpenVZ.
As long as there are customers paying for that, then the old OpenVZ nodes will be kept online as they are basically forever.
Worst case scenario, OpenVZ nested in KVM as a last resort.
The breakage between OpenVZ and Virtuozzo 7 is real. There is no true upgrade path from OpenVZ.
Holy shit! Did I just step out of a time machine?
It's nice to see you again too KuJoe. I still remember when you first started in this business.
And I still remember when it was just you and a small handful of other options out there for LEBs. Glad to see you're still around.
What kind of problems? I have converted several OVZ containers to KVM VM's and visa versa. It's not too hard. Figured it out in about half a day with a little trial and error. I just rsync. It's a one liner.
Nice to see old providers coming back! and at times people getting confused between Ramhost & Ramnode.
Yeah its simple to convert your own environment with your own setup that you understand with your own passwords etc, now scale that to 8000 containers with every possible setup you can imagine including clustered nodes etc and no end-user passwords, then add that to an already significant work load
Pretty much what I've been thinking. Thanks everyone.
what is wrong with OpenVZ 7 ?
It doesn't support user level quotas, for example...
The problem is, is that if you have a OpenVz 6 container then your provider cannot seamlessly upgrade it to OpenVz/Virtuozzo 7 once OpenVz 6 goes end-of-life.
So you're left with out-of-date technology that is a security risk, or you migrate to a supported technology (KVM, LXC or Virtuozzo 7). As your provider cannot do the migration for you, you need to do it yourself.
This means creating a new VM and migrating data. If you currently have a discount OpenVz 6 box, your provider may not necessarily honour that discount price for a new VM
The good news is, this is still two years out. But it doesn't hurt to start planning.
It support: https://docs.openvz.org/openvz_users_guide.webhelp/_managing_disk_quotas.html
I apologize for not being clear. What I meant is that the only big advantage I can see in OVZ against KVM was the simfs. I don't like to run a container inside ploop or other loop-like devices. And Openvz 7 doesn't support user level quotas inside simfs-bind mounts (yet) ->
https://openvz.org/Simfs_filesystem_for_Virtuozzo_7_Containers
I have no problem migrating myself if providers put up Virtuozzo 7 nodes for old OpenVZ 6 customers.
I know KVM is the new cool thing, but the amount of overhead just blows my mind. I only use KVM when I want to run FreeBSD or when I need to do some ultra specific shit. OVZ is good for almost anything. I hope Virtuozzo 7 becomes mainstream.
Wasn't aware of this. Ive 3-4 OpenVZ VPS's (VPSDime, XVMLabs, Time4VPS & VirMach), a couple I imagine will go away at that point. The Time4VPS is a storage server, so I'll be curious to see where that goes.
For me as a customer, I've been perfectly happy with a bunch of OpenVZ machines - because they've been able to do the task just as well as a KVM alternative - yet for less money, so as much as OpenVZ can suck - it still has it's use cases - there's no reason to throw a full blown virt such as KVM after something that doesn't require it.
If KVM is so much better - someone should be able to find a KVM node with 3 cores, 125gb disk (rather fast pl0x), 2 gig ram and 4TB traffic in stockholm for 5 euro :-D can't do? Such shame.
I was going to make a joke about slabbing individual VPSs, but the more I think about it the more of a good idea it might be if OpenVZ 7 doesn't improve a lot (albeit impossible to implement for people using commercial control panels).