[Passed] Proposal to fund development of NuBox (Peerbox port)

Great, I understand this can be used for both holding NSR and USNBT just as the official PC client, right?
Do you think you can make the Nu client work again on all Raspberry models? Or does that need stable Nu 2.1 source?

That would be a nice bonus. Getting the network safer.

I like it!
I just wonder, is it possible for an android tablet/phone to be routed and run linux and then run nu and ppc wallets and mint?
Or keep the android and run some kind of linux emulation and then run the wallets?
Has anyone tried this out?

Debian is Debian. It could work on Ubuntu and similar derivatives of Debian, but due to subtle differences in system design might mess up something.
But yes, just like with Peerbox nowdays - one would be able to install it on server or something.

Yes, I have seen that Nu depends on older and unsupported libdb. This can be solved with “hack” but it is not a long term solution and Nu devs should do something about it.

About updates (indefinite), well that is not so much work. I did say - as long as they are compatible with existing design it should be very simple and not so time consuming. It is basic support, I consider it as an obligation.

It should typically be withing 24h, but if there are some problems with build (like with 2.1 now) it could go up to 72h.

Yes, this is possible. I can maintain both versions in the same repository.

Only me. There is no other way of doing it due to GPG keys. However the binaries themselves can be hosted on official NuBits.com if shareholders agree with that. This would enable webmaster of Nu to deal with them too.

Yes, you are right.

Your sentence makes me think that there is some known problem with some specific version of Nu client and Raspberry? I am not aware of such problems, can you give me more details?
Other than that, it should work for all Raspberry Pi version out of the box.

If you have access to the bootloader you should be able to install/run any Linux that has support for the architecture.
At least Gentoo should do :wink:
If you don’t want to load an alternative Linux, root rights should allow you to open shells and install/compile applications.
But I’m not very familiar with doing that on mobile devices.
Sometimes I struggle with my RaspberryPis :wink:

The current v2.03 client requires a lot of memory and CPU. Way more than a peercoin client and it makes running a stable node on previous generations of Raspberries difficult. This issue is being addressed in the newer v2.1 client, but we still don’t have a stable official release for that. This thread shows some of the issues with the v2.1 client: Help us test Nubits v2.1.0. Hopefully this will be addressed soon, but you might want to wait for that before starting work on your project.

Some additional information regarding RaPi and hardware requirements - both RaPis have 1 GB RAM.

RaPi1:

nud getinfo | grep '"version"'
    "version" : "v2.0.3-dirty-beta",
[...]
PID USER      PR  NI  VIRT  RES  SHR S  %CPU %MEM    TIME+  COMMAND
974 pi        20   0 1011m 840m  19m S  26.2 86.4  14874:23 nud

RaPi2

nud getinfo | grep '"version"'
    "version" : "v2.1.1-RC1-4-g047e5fe-dirty-beta",
[...]
top -p $(pgrep -d',' nud)	
[...]
PID USER      PR  NI  VIRT  RES  SHR S  %CPU %MEM    TIME+  COMMAND
3381 pi        20   0  634m 391m 3592 S   6.5 40.2  12877:00 nud

RaPi1 runs for a long time, but as you can see both the RAM consumption as well as the CPU consumption are quite high - and this instance of nud has nothing to do, but broadcast liquidity information for some NuBots!
RaPi2 needs much less RAM and CPU and the total CPU time is only that high because of syncing the blockchain.

Ok, I see. I don’t think this would be a problem though. Some optimization to gain more useful RAM can be made, like I have done already with old release of Peerbox.
Maybe Nu should move to newer Bitcoin codebase as they have introduced many performance optimizations. However the biggest gain would be moving to leveldb instead of old Berkeleydb.

Anyway, I do not think this issues would tamper with development of the project.

Nu 2.1 code is already using leveldb and that’s why the RAM consumption went down quite far.
Berkeleydb is only continued for wallet support, but not for the blockchain.

Well then why do you worry :smiley:

Nu 2.1 … is unfortunately not ready for production yet.

Well I am saying that that should not interfere with development of NuBox.
Nu binary would simply be updated to 2.1 branch when the branch is ready.

I like the idea.
Being able to decide the monetary policies of a decentralized bank and transfer synthetic fiat anonymously on a raspberrypi would be awesome.
2 questions:

  • can u reuse the work done on nu raspberry pi binaries?
  • what is your hourly rate?

No, that is not possible. Proper Debian packages are needed.

  • what is your hourly rate?

I did not calculate sum needed for this project based on hourly rate. It was more like a round figure that should probably cover the hours of coding and testing.

2 Likes

I am preparing to finalize the motion.
One last question for Nu Shareholders and users.
How would you prefer installation of NuBox?

In the form of adding repository sudo apt-get install nubox but installing Raspbian on your own?

or

Should I periodically issue full, pre-configured Raspbian image along with the repository?

As a nushareholder, I’d give this project a green light. I am especially interested in the web GUI for configuring the wallets / motion votes and what not, so you wouldn’t even have to SSH into the Pi. I would buy myself 3 new Pis and mint nushares, blockshares and peercoins separately on each device. The fourth pi would share the internet. It’s because the 1 GiB RAM limit because having 3 wallets minting on the same pi is probably impossible.

1 Like

Peercoin is very lightweight by the way. It’s easy to pair with Nud or Bcexchanged.

I like this better. It not only faster to download but also preserve my configuration files, script, logs, and things I have installed.

Probably. But keep in mind that resource footprint is strongly correlated to how many coins/shares there are in the wallet.

edited

Why isnt this a grant proposal?