1 |
"J. Roeleveld" <joost@××××××××.org> writes: |
2 |
|
3 |
> On Friday, April 24, 2015 10:23:01 PM lee wrote: |
4 |
>> "J. Roeleveld" <joost@××××××××.org> writes: |
5 |
>> > On Thursday, April 23, 2015 11:03:53 PM lee wrote: |
6 |
>> > Do you have anything that you find insufficiently documented or is too |
7 |
>> > difficult? |
8 |
>> sure, lots |
9 |
> |
10 |
> Have you contacted the Xen project with this? |
11 |
|
12 |
I've been asking questions on mailing lists. What do you expect? I |
13 |
could tell them "your documentation sucks" and they might say "go ahead |
14 |
and improve it then". I tried to improve it the little bit I could; |
15 |
it's on the wiki, if it's still there. |
16 |
|
17 |
>> > Containers. |
18 |
>> > Chroots don't have much when it comes to isolation. |
19 |
>> |
20 |
>> What exactly are the issues with containers? Ppl seem to work on them |
21 |
>> and to manage to make them more secure over time. |
22 |
> |
23 |
> Lack of clear documentation on how to use them. All the examples online refer |
24 |
> to systemd-only commands. |
25 |
|
26 |
True, there isn't much, if any, clear documentation. I followed the |
27 |
Gentoo wiki and it's working fine, though. |
28 |
|
29 |
>> >> > Virtualbox is nice for a quick test. I wouldn't use it for production. |
30 |
>> >> |
31 |
>> >> Why not? |
32 |
>> > |
33 |
>> > Several reasons: |
34 |
>> > |
35 |
>> > 1) I wouldn't trust a desktop application for a server |
36 |
>> |
37 |
>> So that's a gut feeling? |
38 |
> |
39 |
> No, a combination of experience and common sense. |
40 |
> A desktop application dies when the desktop dies. |
41 |
|
42 |
You cannot run it from the command line? It only runs in an X session? |
43 |
If that is so, I'm going to need something else. |
44 |
|
45 |
>> > 2) The overhead from Virtualbox is quite high (still better then VMWare's |
46 |
>> > desktop versions though) |
47 |
>> |
48 |
>> Overhead in which way? I haven't done much with virtualbox yet and |
49 |
>> merely found it rather easy to use, very useful and to just work fine. |
50 |
> |
51 |
> Virtualbox is easy when all you want is to quickly run a VM for a quick test. |
52 |
> It isn't designed to run multiple VMs with maximum performance. |
53 |
> In my experience I get on average 80% of the performance inside a Virtualbox |
54 |
> VM when compared to running them on the machine directly. With Xen, I can get |
55 |
> 95%. |
56 |
> (This is using normal work-loads, lets not talk about 3D inside a VM) |
57 |
|
58 |
Someone told me that you may find xen reducing the performance by up to |
59 |
40%. |
60 |
|
61 |
>> Compared to containers, the overhead xen requires is enormous, |
62 |
> |
63 |
> Hardly comparable. Containers run inside the same kernel. With Xen, or any |
64 |
> other virtualisation technology, you run a full OS. |
65 |
|
66 |
How is that not comparable? |
67 |
|
68 |
You don't need to run a full OS, and you're not stuck with fixed memory |
69 |
assignments without even the ability to overcommit when you use |
70 |
containers. With xen, you're stuck with what you initially assigned, |
71 |
may your VM currently use it or not. |
72 |
|
73 |
If my mail server was a xen VM, I'd have assigned 2GB to it; as a |
74 |
container, it uses less than one. If the machine I'm working on was a |
75 |
xen VM, I'd have assigned at least 16GB to it; as the host of the |
76 |
container, it costs me nothing. |
77 |
|
78 |
So obviously, the overhead required by xen is enormous. |
79 |
|
80 |
> |
81 |
>> and it |
82 |
>> doesn't give you a stable system to run VMs on because dom0 is already |
83 |
>> virtualized itself. |
84 |
> |
85 |
> Why doesn't it provide a stable system? |
86 |
> The dom0 has 1 task and 1 task only: Manage the VMs and resources provided to |
87 |
> the VMs. That part can be made extremely stable. |
88 |
|
89 |
It's already virtualized itself. With containers, I have a |
90 |
non-virtualized system as usual, as stable as they are. A container is |
91 |
like just another service I can start or stop, and I can access it |
92 |
easily because it simply resides under /etc/lxc while I can use the host |
93 |
for whatever else I'm doing. |
94 |
|
95 |
With xen, I have a virtualized system to begin with, which is wasted |
96 |
because it's only purpose is to provide a way to maintain other VMs. I |
97 |
can't fully use any of these VMs because for what I'm doing, I'd have to |
98 |
pass through my NVDIA card to one of them. IIUC, I wouldn't even be |
99 |
able to log in to the host because it won't have a graphics card, |
100 |
provided that I actually could pass the graphics card through, which |
101 |
appears to be pretty much impossible. |
102 |
|
103 |
The VMs would reside on LVM volumes and be hardly accessible --- though |
104 |
now I'd use ZFS subvolumes, making that as easy as with containers. |
105 |
|
106 |
Power management wouldn't work. The xen documentation sucks. Everything |
107 |
would be difficult and troublesome. It's extremely difficult to install |
108 |
a VM --- I never figured out how to actually do that. I'd be hugely |
109 |
wasting resources. |
110 |
|
111 |
I'd never have the feeling that there is a stable platform to work with, |
112 |
and it's not something I would want to have to maintain. |
113 |
|
114 |
> My Lab machine (which only runs VMs for testing and development) currently has |
115 |
> an uptime of over a year. In that time I've had VMs crashing because of bad |
116 |
> code inside the VM. Not noticing any issues there. Neither with stability nor |
117 |
> with performance. |
118 |
> My only interaction with the dom0 there is the create/destroy/start/stop/... |
119 |
> VMs. |
120 |
|
121 |
How can you use it for testing when it's so ridiculously difficult to |
122 |
install a VM? How do you do updates without rebooting when a new kernel |
123 |
version comes along? How do you adjust resource allocations depending |
124 |
on changing workloads? |
125 |
|
126 |
>> I don't know how that compares to virtualbox --- I |
127 |
>> didn't have time to look into it and it just worked, allowing me to run |
128 |
>> a VM on the fly on the same machine I'm working on without any ado. |
129 |
> |
130 |
> For that scenario, VirtualBox is quite well suited. I wouldn't run Xen on my |
131 |
> desktop or laptop. |
132 |
> |
133 |
>> That VM was simply a copy of a VM taken from a vmware server, and the |
134 |
>> copy could be used without any conversion or anything. |
135 |
> |
136 |
> Good luck doing that when you installed the VMWare client tools and drivers |
137 |
> inside a MS Windows VM. |
138 |
|
139 |
That's what it was. |
140 |
|
141 |
It gets troublesome when the VM is distributed across multiple files --- |
142 |
I started trying to convert that and never had the time to finish. |
143 |
|
144 |
>> You can't do |
145 |
>> that with xen because you'll be having lots of trouble to convert the |
146 |
>> VM, to convert the machine you're working on to xen and to get it to |
147 |
>> work, to work around all the problems xen brings about ... Some days |
148 |
>> later you might finally have it working --- which is out of the question |
149 |
>> because the VM is needed right away. And virtualbox does just that. |
150 |
> |
151 |
> Look into the pre-configured versions of Xen, like what Citrix offers. |
152 |
> I can import VMs from VMWare as well without issue. (Apart from the VMWare |
153 |
> client tools as mentioned, but Virtualbox has the same issues) |
154 |
|
155 |
Citrix is commercial, isn't it? |
156 |
|
157 |
IIRC, I tried to get xen to work with Centos (no chance), something else |
158 |
which I don't remember and finally Debian. Debian worked, but it is a |
159 |
total mess with too much ancient software and their backports, requiring |
160 |
backports kernels for xen and because of kernel bugs and problems with |
161 |
dracut/initramfs-tools (which are supposed to be fixed now). |
162 |
|
163 |
Add to that the sucking documentation of xen, things like power |
164 |
management not working and all the quirks like the clocks being offset |
165 |
despite the docs saying that they will be synced automagically, the |
166 |
impossibility to install an OS in a VM and the enormous waste of |
167 |
resources, and I really don't want to use xen --- particularly not for |
168 |
production. It's simply too troublesome. |
169 |
|
170 |
>> I was really surprised that virtualbox worked that well. Maybe xen will |
171 |
>> get there some time. |
172 |
> |
173 |
> Xen already is there. |
174 |
|
175 |
Not by far, it doesn't even have good documentation yet and is too |
176 |
troublesome and too difficult to use. |
177 |
|
178 |
> Please understand that Xen and Virtualbox have their own usecases: |
179 |
> |
180 |
> Xen is for dedicated hosts running VMs 24/7 |
181 |
|
182 |
If you can get it to work and if you have the time to get it there, it's |
183 |
a nice solution for VMs. That's very big ifs, not even to mention to |
184 |
work around the quirks. |
185 |
|
186 |
> Virtualbox is for testing stuff quickly on a laptop/desktop |
187 |
|
188 |
If it works only for that, what's the alternative? Sooner than later, |
189 |
I'll have to set up some Windoze VMs at work, and I really don't want to |
190 |
use xen. |
191 |
|
192 |
|
193 |
-- |
194 |
Again we must be afraid of speaking of daemons for fear that daemons |
195 |
might swallow us. Finally, this fear has become reasonable. |