Community Test 13 November - offline

So Im guessing that test network cost you ~$70 or so for 3 days? In that case we owe you lots of beer.

EDIT: even at $720/month then for what the project will gain from a community testnet up 100% of the time or close then I still think that is incredible marketing/learming value if there is $2m in the BGF. Yes its more than we individually can afford but in the bigger scheme of things its chickenfeed.

3 Likes

No only half that, beacuse I use the cheaper 2vcpu-4gb droplets :slight_smile:

3 Likes

So we could say $360/month then or thereabouts? <$500 to be safe?

2 Likes

*sarcasm filter disengaged

It’s almost as if you all need some kind of accounting method or token built into the testnet code to keep track of the resources different community members are contributing… then sell those to the bgf to cover bandwidth and hardware costs… :rofl:

*sarcasm filter reengaged

5 Likes

Thats the project for the week after next.
Its very high spec, very sturdy
Were you planning on dancing on it?

Looks like Pi has joined the out of stock brigade.
At least over here, checked most of the usual suspects and seeing back in stock Feb 2022.
Obs can get scalped on Amazon as usual though.

2 Likes

I blame these damn chinese stealing all the chips
You are quite right the prices on ebay are horrendous

2 Likes

Ram usage used to be muuuch higher.

We’ve erred on the side of caution since then as optimising for mem hasn’t been that important vs stability in test runs.

That said, I’ve also spun up some smaller droplet networks and they seemed fine. We just haven’t rigourously hammered them yet. So I’d say keep going w/ smaller ones. And see how you go. Unless you see nodes stopped logging (aka dead), then they shoulnd’t likely be hitting the ceiling (there are some resource usage stats printed to logs, so you may be able to glean more there).

At the moment the most intense node usage we see is largely CPU bound during section startup/DKG (which is what we’re largely focussed on stabilising just now). But if there’s no splits going on that shou;dnt be the issue.

Ram usage shouldn’t creep up too much (although apparently does still looking at @Vort 's post above; though nothing like before we’ll have to take a look at that).

I do local testnets a fair bit and have 45 nodes running on one machine frequently. So they should not be wildly resource intensive. (caveat: i’m not trying to chuck up massive files yet though)


Agreed. I think the community testing here is most definitely useful.

While we haven’t had change to dive into logs too much here yet. We are defeinitely following along! @chriso is mad keen to get back to debugging the CLI and offering more support over here. So hopefully he’ll be back on that soon enough!

But aye just getting folk used to doing this, testing out what works (size of droplets is itself great info) and what doesn’t , developing, testing and setting up tooling is all great stuff :+1: :muscle: :bowing_man:

12 Likes

Stupid question. I haven’t read much about this thread. Is this a real operating safe test network? If I join this net, can I upload files which you guys can download?

1 Like

Yes that is the plan. To test it out and see what bugs can be found.

You can’t join it, because it is offline.

When someone will launch similar test network, then yes, you will be able to use it like you said. Until it crashes because of bugs.

A huge amount of dev work has been done since this was live. As I understand the CLI has not yet caught up but once it does we could try again.
(It may well have been out of whack at the time of this test).

6 Likes