1 |
On 6/8/20 8:27 PM, Alexandru N. Barloiu wrote: |
2 |
> On Mon, 2020-06-08 at 20:16 -0400, james wrote: |
3 |
>> Any pointers to codes that create a cluster and run on 64Bit arm low |
4 |
>> power boards is welcome to post to this thread, or drop me a private |
5 |
>> note. |
6 |
> |
7 |
> There is no such thing as cluster for arm. It's just daemons. You equip |
8 |
> each pi with the things it's going to need. You treat them as normal |
9 |
> computers. |
10 |
|
11 |
Huh. Well, I've run across dozens of projects, some as old as 2015. |
12 |
Sure, I have not 'dug into' the details, but they seem to be quite common:: |
13 |
|
14 |
https://magpi.raspberrypi.org/articles/build-a-raspberry-pi-cluster-computer |
15 |
|
16 |
https://makezine.com/projects/build-a-compact-4-node-raspberry-pi-cluster/ |
17 |
|
18 |
https://projects.raspberrypi.org/en/projects/build-an-octapi/ |
19 |
|
20 |
https://www.hackster.io/aallan/a-4-node-raspberry-pi-cluster-e19273 |
21 |
|
22 |
Granted, the term 'cluster' in the linux world is as open as the word |
23 |
'cocktail' at a social gathering; ymmv. |
24 |
|
25 |
|
26 |
> |
27 |
> If you don't know how to start... |
28 |
> https://gentoo.osuosl.org/experimental/arm64/ |
29 |
|
30 |
|
31 |
Great link. But I typically avoid all things 'systemD' centric. (no |
32 |
discussion, just my preference). |
33 |
|
34 |
However I did find this stage 4: |
35 |
|
36 |
|
37 |
stage4-arm64-minimal-20190115.tar.bz2 |
38 |
|
39 |
at https://gentoo.osuosl.org/experimental/arm64/old/ |
40 |
|
41 |
Any idea how chip specific this stage 4 for arm64 is? |
42 |
|
43 |
> |
44 |
> there are actually some modern stage3 images. I suggest you google how |
45 |
> to emulate arm64 using qemu-static. google crossdev as well. There are |
46 |
> wonderful resources on the forums, some of which I participated in. |
47 |
|
48 |
Not applicable. there are always a myriad of nuances with this approach, |
49 |
as I often stray into unique and exotic hardware extensions. Some run |
50 |
'clusters' on a collective of ity-bity IoT devices, cause they are |
51 |
fairly close together over Rf links. Folks at the companies that build |
52 |
chipsets, are very advanced in this venue. Most of it is DoD related and |
53 |
quite hush_hush. A billionaires club, so to accurately categorize. But |
54 |
there is no issue with gentoo folks finding their own pathways forward |
55 |
with clustering arm/micro devices. It is the future and even IoT |
56 |
security semantics will be based on each (I0T) nodes performance metrics |
57 |
as opposed to traditional security (bloated) codes. These IoT comm |
58 |
links, are like a predictable wave. Monitoring the wave, in the RF |
59 |
domain, shows where and when a small portion of (for example) field IoT |
60 |
sensors are stressed (under a heavier load than normal. So you do not |
61 |
have to strictly depend on specific codes and filters to detect anomalies. |
62 |
|
63 |
Monitor and matching of various domains yields startling results. BATM |
64 |
it is more of an art from than consistent technology. Surely the good |
65 |
folks of Gentoo will validate a pathway forward. |
66 |
|
67 |
What I have discovered is there are an enormous amount of very technical |
68 |
folks that routinely use gentoo, but keep it a secret. |
69 |
|
70 |
|
71 |
> Good luck and happy hacking. |
72 |
|
73 |
|
74 |
Gentoo, hacking and exotic hardware are more of an addiction than a |
75 |
source of joy. Be at peace. and |
76 |
|
77 |
THANKS for the link, |
78 |
James |