Node Manager UX and Issues

Ant-Node (was Safe-node)
543

Sorry for the inconvenience, I’ll need to take a deeper look into this. Not sure when I’ll get to it.

1 Like
544

Haha no need for that. The convenience far outways the inconvenience :grin:

1 Like
545

I will look at it at some point, because it’s an unintentional regression.

3 Likes
546

I guess Chris and the team are busy today. I’m unable to test my app because PunchBowl fails on certain files and my local testnet is borked.

I’m getting lots of side-tasks done which is good, but does anyone know how to wipe everything safenode-manager does from a machine to get a clean start?

It says the faucet is running but the process isn’t there and there’s nothing listening on 8000, so I think it is wrong and I don’t know how to clear it. I can’t use ‘stop’ or ‘reset’ because it want sudo and when I do that cargo fails, which I need to run locally. :man_shrugging:

547

You should check top after a reset quite often you will find many node processes still running. Or grep for them, I find top easier.

I make a habit of sudo killall safenode now.

And this.

1 Like
548

this works for me for a full tear down

#!/usr/bin/env bash

sudo pkill -e safe

# stop nodes

for i in {1..100}
do
 # your-unix-command-here
 sudo systemctl disable --now safenode$i
done

sudo rm /etc/systemd/system/safenode*
sudo systemctl daemon-reload
rm -rf $HOME/.local/share/safe/node

sudo rm -rf /var/safenode-manager
sudo rm -rf /var/log/safenode
2 Likes
549

beaten to the punch :slight_smile:

2 Likes
550

I’m back in business. Thanks all of you :partying_face:

5 Likes
551

@chriso So I just did a upgrade of safenode-manager and one of my scripts died, showed no data from the status and balance options

I have to now use sudo for add/start/status/balance (and maybe more). Not sure if this was mentioned before, but thought you need to know.

I know you’ll look into it so no need for a response

3 Likes
552

I think that one essential feature would be a possibility to quickly close down a selected portion of nodes, like 10%, 20%…

When people are running too many nodes and run into problems with other connections in the household, they are going to kill nodes. And if killing a subset is not made easy, they’ll kill all. Maybe a pause could also be applied.

This happened to me today when I reached the capacity of our network, and then my spouse needed bandwidth quickly.

6 Likes
553

Thanks for the useful suggestion. I’ve added it to the issue:

8 Likes
554

I am just a simple man, this sudo and no sudo is making me do stupid things.

6 Likes
555

Can you be a bit more specific here please? I would like to help, but I need to know more about what you’re finding troubling.

1 Like
556

When to know using sudo & when you don’t have to and the version currently requires sudo for things it did not before and if you forget it just gives output that seems genuine (like no nodes found) but is obviously wrong

1 Like
557

OK, I think there might have been a regression somewhere, but what would be really helpful is if you could show me the output or refer to specific commands or scenarios where there has been a problem. Otherwise I don’t have much to go on.

1 Like
558

Ahh, I thought it was all part a plan that has not yet been announced.

In short my difficulty seems to be that you can run commands with or without sudo.
Nothing really prevents you from mixing them up which ends in endless confusion.

Seems very simple but I have been getting myself into all kinds of a mess.

@aatonnomicc noted that when run without sudo everything goes to .local as it did in the past. Perhaps he can comment on that finding.

1 Like
559

AFAIK all commands in safenode-manager require sudo or it doesn’t work.

I just tried status and without sudo there is no output and it just exits normally

560

OK. I think the confusion here is due to the introduction of the ability to run user-mode services.

Here is a snippet from the recently updated README:

The reason for the introduction of user-mode services was to make the TUI usable for those less technical. Prompting for passwords is not really feasible, both from a technical and a UX point of view.

There may need to be some distinction made in the commands to indicate the use of user-mode vs. system-wide service, because clearly people are making mistakes there. I am open to suggestions as to how to handle things in terms of the usage of flags or whatever.

4 Likes
561

Ahh ok. I am just happy to know that I am not entirely losing my mind. I WILL however accept that sticking to the same pattern should not require more than 3 working brain cells. I have been falling short there.

562

It’s fine. I acknowledge that mistakes will inevitably be made here. Perhaps we could have system and user subcommands? Or use flags?

Another option could be to take user mode services completely out of the CLI and just have that supported in the TUI.

2 Likes