1 |
On Monday, January 18, 2016 06:07:33 AM Rich Freeman wrote: |
2 |
> On Mon, Jan 18, 2016 at 1:44 AM, J. Roeleveld <joost@××××××××.org> wrote: |
3 |
> > On Monday, January 18, 2016 02:02:27 AM lee wrote: |
4 |
> >> You would have a full VM for each user? |
5 |
> > |
6 |
> > Yes |
7 |
> > |
8 |
> >> That would be a huge waste of resources, |
9 |
> > |
10 |
> > Diskspace and CPU can easily be overcommitted. |
11 |
> > |
12 |
> >... |
13 |
> > |
14 |
> > The biggest reason why I don't use KVM is the lack of full snapshot |
15 |
> > functionality. Snapshotting disks is nice, but you end up with an unclean- |
16 |
> > shutdown situation and anything that's not yet committed to disk is gone. |
17 |
> |
18 |
> Seems like on linux a straightforward design would be spinning up |
19 |
> containers on demand, with snapshots underneath. Granted, somebody |
20 |
> still needs to build it, but spinning up a container per user isn't |
21 |
> much more resource-intensive than just running x2go with multiple |
22 |
> users in a single namespace which is how it works today. It certainly |
23 |
> would be less wasteful than a full VM. They also launch and shutdown |
24 |
> super-fast. |
25 |
> |
26 |
> Of course, this is a linux-only solution (or BSD I believe). You're |
27 |
> not going to be able to do this with OSX/Windows guests. |
28 |
|
29 |
A similar solution is generally done with VDI implementations as well. |
30 |
Replace "container" with VM and you have the same. |
31 |
|
32 |
-- |
33 |
Joost |