1 |
On 02/02/2017 04:05 PM, Rich Freeman wrote: |
2 |
> On Thu, Feb 2, 2017 at 3:35 PM, james <garftd@×××××××.net> wrote: |
3 |
> |
4 |
>> |
5 |
>> I think that unikernels are something everyone should be aware of |
6 |
>> as they purport to be the latest trend in securing all sorts of systems. |
7 |
>> (a brief read). |
8 |
>> |
9 |
> |
10 |
> Not really for all sorts, more for servers. Otherwise I get it, and |
11 |
> at this point now that I run almost everything in containers I tend to |
12 |
> be more inclined to run different distros in those containers. |
13 |
> |
14 |
>> This is only the case because profiles are in general in a mess and there |
15 |
>> are little in the way of conventions. What is so sacrosanct about upstream |
16 |
>> for a truly embedded gentoo system or a gentoo based IoT device? |
17 |
> |
18 |
> Nothing, in that space. |
19 |
> |
20 |
> The problem is the new user experience. When somebody is new to |
21 |
> Gentoo and not super-knowledgeable the first thing they're going to do |
22 |
> is set up a desktop. Now, they might not call it a desktop. They |
23 |
> might not even run X11 on it. But, they're basically falling into |
24 |
> that desktop user experience where whatever they do install "just |
25 |
> works" and is feature-complete. |
26 |
|
27 |
Wow, I'm shocked. Perhaps you have forgotten that is was "I" that |
28 |
bemoaned and protested ad nauseam for a simple, basic desktop |
29 |
for the noob moving to or just testing gentoo? Needless to say the |
30 |
majority of long term folks on gentoo-user, prefer to run those noobs |
31 |
off. (Remember I was on gentoo-user for 12 years, by the hour.) |
32 |
Now our gentoo distro is catering and concerned about these folks? |
33 |
(excuse me but I haft to roll around on the floor a bit to get some |
34 |
belly laughter out....) ok, I'm back now. Nobody give a damn about the |
35 |
gentoo noob; that's why it is gentoo policy not to have an installer. |
36 |
|
37 |
|
38 |
|
39 |
> It is true that we also attract advanced users who are looking for |
40 |
> something different. They have no issues getting any distro to dance |
41 |
> for them, and they're picking Gentoo because it is best suited for |
42 |
> their specific need. These users are much more likely to be |
43 |
> interested in minimal configurations, embedded systems, the hardened |
44 |
> profiles, and so on. |
45 |
|
46 |
This pool is growing and many are contributing.... hardened on top |
47 |
of minimal.... very very cool. |
48 |
|
49 |
> However, the problem is that if we optimize mainly for the second |
50 |
> group we basically lose the first group entirely, and I suspect that |
51 |
> is overall going to be the bigger group. |
52 |
|
53 |
Nope, sorry, I have to disagree. Please explain why we cannot, in the |
54 |
profiles, support both groups. The (minimal) embedded profile pathway |
55 |
need not be mentioned in the handbook. But, perhaps in the |
56 |
gentoo-embedded-handbook it could be introduced ? If the embedded devs are |
57 |
offended, then it could be unser it's own profile:: spartan, monopod, |
58 |
minimized or any self identifying moniker. |
59 |
|
60 |
|
61 |
> If what you want is a "unikernel profile" for Gentoo then you're going |
62 |
> to be changing a LOT of assumptions. |
63 |
|
64 |
Oops, hit the brakes! Unikernels the way that unikerel.org describes |
65 |
them is more of a enhanced state machine boot_code or an executive |
66 |
or a linux kernel plus one lib. Sorry but that's not the only vision |
67 |
and mine is to build highly targeted 'minimal' gentoo systems that can |
68 |
dynamically shed and load new frameworks (groups of packages and codes |
69 |
and such) on top of the 'default set' or embedded set of packages. So it |
70 |
can become a full blown mail server or a singular monotonic device, just |
71 |
sniffing ethernet, without a reboot. I'm betting the farm that my vision |
72 |
of minimal/embedded gentoo will be far more successful than those folks |
73 |
pushing proprietary Unikernel products. Still the generic moniker |
74 |
'unikernel' is the closest commonly used moniker to where I'm driving |
75 |
too, I so 'lifted' it from those folks. Goals are similar but mine is a |
76 |
minimized gentoo, at it's core and dynamically flexible without reboot. |
77 |
|
78 |
|
79 |
|
80 |
> Forget openrc vs systemd, there |
81 |
> is no reason to have any init implementation on the thing. Forget |
82 |
> linux vs bsd, there is also no reason to have a kernel in a container. |
83 |
> We don't need any editor because you're probably going to do any |
84 |
> config file editing from outside of the container. And that @system |
85 |
> set that has all that bootstrapping stuff is probably way overkill if |
86 |
> all you ultimately need is a single package to work (and maybe not all |
87 |
> of that package). Heck, your overall install approach also should be |
88 |
> questioned. Rather than build your unikernel from inside its own |
89 |
> container, you should be building from a more complete image and just |
90 |
> installing the minimum RDEPENDs in the production container (as with |
91 |
> catalyst or the chromiumos builds). And you probably wouldn't be |
92 |
> upgrading such things in place either, you'd just be creating newer |
93 |
> instances and cutting over from the old. |
94 |
|
95 |
I agree with you on this, absolutely. But you are far off from my |
96 |
branch of the profile tree and my pathway forward, so it's accurate |
97 |
but completely uncharacteristic to what I'm developing. |
98 |
|
99 |
|
100 |
> I don't question that it would be great for Gentoo to support all of |
101 |
> this stuff. I just think that we need to be careful not to destroy |
102 |
> the experience of somebody who just wants a "typical" install in order |
103 |
> to do it. Somebody who doesn't want to take the time to tweak how |
104 |
> their java implementation works probably wants the default install to |
105 |
> be something that meets the Oracle standard. Now, somebody who is |
106 |
> into tailoring can look at their application and tweak the living |
107 |
> daylights out of it, but that shouldn't be what you get when you run |
108 |
> "emerge icedtea" or whatever. |
109 |
|
110 |
|
111 |
I just do not see your 'either or scenario' as the only possibility. For |
112 |
example, in a recent gentoo-dev thread, those profile owners (devs) |
113 |
were asked to step forward and state the viability going forward of a |
114 |
profile review thread. If one of those devs wants one of those itemized |
115 |
profiles to continue to exist, is it going to be forcibly deleted? (no) |
116 |
|
117 |
So all I have to do is convince (beg?) one dev to have a place in the |
118 |
profile tree that is not subservient to upstream dictates? Perhaps |
119 |
another way forward for my work? (Glep-70?) |
120 |
|
121 |
|
122 |
> Sure, you could do all that with a profile, but the problem is: |
123 |
> 1. Maintainers aren't going to necessarily invest in that profile. |
124 |
|
125 |
Can I proxy-maintain a profile for minimized gentoo clusters? |
126 |
Can I share a profile with another compatible (need) profile ? |
127 |
|
128 |
|
129 |
> 2. New users won't necessarily use that new profile. |
130 |
|
131 |
That would be excellent. That's why there are many choices for profiles, |
132 |
right? |
133 |
|
134 |
|
135 |
> And when those things doesn't happen users look at Gentoo as the OS |
136 |
> that nothing works right on. |
137 |
|
138 |
Non-sequitur.... |
139 |
user's will be on a handbook delineated profile choice; capable users |
140 |
will be able to navigate handbook choices and other available profiles, |
141 |
or seek (neddy?) guidance. The profile I'm talking about could |
142 |
be logically under the old branch of 'experimental' if that survives |
143 |
too. But would be best as close to the root of the profile tree, as |
144 |
possible. The least amount of packages installed, is best, and located |
145 |
in a different branch than the default or whatever the new name is where |
146 |
a plethora of upstream issues exit. None of this would be a burden |
147 |
to ebuilds, if they do not work, so be it, I can and will fix what I |
148 |
need from upstream. Every week, that list grows shorter and shorter, |
149 |
particularly for the minimal builds. |
150 |
|
151 |
|
152 |
Hopefully I have justified and succinctly stated my vision, is |
153 |
accommodation not possible? All I really need is a minimized (a least or |
154 |
very low set of packages) profile that is not so concerned with most of |
155 |
the upstream projects and the noise found therein. |
156 |
|
157 |
|
158 |
hth, |
159 |
James |