Reason : No extra properties allowed is killing me

Aurum · December 7, 2018, 4:54am

Please delete thread…syntax error found.

Kryztoval · December 7, 2018, 6:52am

“keys” is mandatory.

Is that your whole configuration?
can you paste it inside “code” tags?

hanssv.chain · December 7, 2018, 8:00am

No extra properties allowed means you have something in your configuration that is not expected by the node. Try removing the instances: 3 and see if it works. If it does it means you failed to build the multi_gp-branch and thus the node does not expect that parameter.

hanssv.chain · December 7, 2018, 1:40pm

If it works without instances: but not with it, it means you have not followed the guide on how to build the multi_gpu miner branch. I suggest you try that then…

Aurum · December 7, 2018, 1:43pm

You’re fast, I edited above post about Types at the same time.
The last line of my first post above shows exactly what I did to install Multi-GPU cuda29. I got confused maybe with directories /node in one guide and /epoch in another.

hanssv.chain · December 7, 2018, 1:47pm

The miner (i.e. the cuda29 executable) is the same for single and multi GPU - it is the ~~epoch~~ aeternity node that is different.

Aurum · December 7, 2018, 1:49pm

Is that cuda29 file type right???

hanssv.chain · December 7, 2018, 1:55pm

Ok, great then you don’t have a problem with adding instances: 3 to you epoch.yaml.

hanssv.chain · December 7, 2018, 2:14pm

Deploy AE Mainnet CUDA MultiGPU Miner. | by Chris | Medium looks correct AFAIK.

Aurum · December 7, 2018, 2:20pm

So, if I do an install that will mine with mean29-avx2 then do I have a node that will mine with a single GPU using cuda29???
I’ve got several rigs running mean29-avx2 fine with their own open ports.
How can I remove everything and start over for a fresh try???
Sorry, I’m so ignorant, I’m not a coder.

Aurum · December 7, 2018, 2:32pm

I deployed my node as described in your guide:

github.com

aeternity/aeternity/blob/v1.0.0/docs/installation.md

# Install an epoch node using a release binary

This document describes how to install an epoch node using a release binary.

In order to install an epoch node using a release binary, you need to:
* Retrieve the release binary corresponding to your platform;
* Install the required dependencies;
* Deploy the node.

## Retrieve release binary

The release binaries are published on [GitHub](https://github.com/aeternity/epoch/releases) and are tested on the following platforms:
* Ubuntu 16.04.3 LTS (x86-64);
* Ubuntu 18.04 LTS (x86-64);
* macOS High Sierra 10.13 (x86-64);
* macOS Mojave 10.14 (x86-64).

## Install dependencies

### Ubuntu package

This file has been truncated. show original

My lines are:
mkdir ~/epoch
cd ~/epoch
tar xf ~/Downloads/epoch-1.0.1-ubuntu-x86_64.tar.gz
Maybe I should’ve put it in ~/node

hanssv.chain · December 7, 2018, 2:38pm

The directory does not matter.
Note that this node is a Single instance node, and in order to use the GPU miner you need to build and install it.

Aurum · December 7, 2018, 2:40pm

But you said above:
The miner (i.e. the cuda29 executable) is the same for single and multi GPU

hanssv.chain · December 7, 2018, 2:42pm

Yes, that is still true.

However, it is not shipped in the release you still have to build it and place it in the right directory?! The release only contains the CPU miner.

hanssv.chain · December 7, 2018, 2:51pm

You have said multiple times that you followed Chris guide for the Multi GPU miner setup, if so, you have already built the GPU miner (hence my answer that it is the same as the single GPU miner)…

Otherwise the documentation for building the cuda miner is in aeternity/cuda-miner.md at master · aeternity/aeternity · GitHub

Kryztoval · December 8, 2018, 8:08am

I compiled and tried gpu mining and it works, I saw several people installed and ran this on ubuntu 18.04.1lts which, byw is not an old version of linux, is the latest LTS version there is.

if you did follow the github that hanssv kinddly provided you should have no issues.

Aurum · December 8, 2018, 2:08pm

What does this ambiguous phrase “non-root user with ALL sudo privileges” actually mean to a lay end user??? The man sudoers file is 2421 lines long. The word “privileges” appears once on Line 7. The word “privilege” does not appear.
I can use the id command to see if I’m root:
aurum@Rig-05:~$ id
uid=1000(aurum) gid=1000(aurum) groups=1000(aurum),4(adm),24(cdrom),27(sudo),30(dip),46(plugdev),115(lpadmin),117(nopasswdlogin),127(sambashare)
Ok I’m not root, I’m aurum and a member of group sudo. I can open sudoers.tmp using sudo visudo:

privilege specification

root ALL=(ALL:ALL) ALL

Members of the admin group may gain root privileges

%admin ALL=(ALL) ALL

Allow members of group sudo to execute any command

%sudo ALL=(ALL:ALL) ALL

See sudoers(5) for more information on “#include” directives:

#includedir /etc/sudoers.d

aurum ALL=(ALL) NOPASSWD: ALL

Hmm, do I comply with being a “non-root user with ALL sudo privileges” or not??? It looks to me that the %sudo group is equivalent to root. Rules are applied in order and when multiple rules apply the last entry is used. So aurum ALL=(ALL) NOPASSWD: ALL takes precedence over %sudo ALL=(ALL:ALL) ALL. But aurum only has two "ALL"s while %sudo has four "ALL"s. Hmm, maybe I need three ALLs. Maybe four.
It would be nice if the instructions said something like:
Assure that you are a non-root user with ALL sudo privileges by having this line placed at the end of your sudoers.tmp file delimited by spaces (or is it tabs):

ALL=(ALL:ALL) ALL
or
aurum ALL=(ALL:ALL) ALL NOPASSWD: ALL

It would be nice if one could be certain that they’ve complied with a requirement so they can have less to worry about when alpha code does not work.

Kryztoval · December 9, 2018, 8:11am

it means youi have to create a user and make it a part of the “sudo” group in ubuntu.