Have we ever had more than 1 split?
I feel like 3 4 sections is a worthy goal.
5 Likes
That’s what I was wondering.
Three historical split events gets you 4 sections with a section prefix length of 2. That’s a difficult way to think about it though. Section prefix length (L) is the better/easier metric that describes the network tree structure. Number of sections equals 2^L.
4 Likes
Hate to be even more difficult. That only holds true is there is evenly spread randomness in chunk size-xor address.
Lets say that there is unevenness and statistically this is quite likely as more splits occur.
1
/ \
10 11 This is always the case
/ \ Either 10 or 11 must split first (exact time is highly unlikely)
100 101
/ \
1000 1001 Say 100 has a lot less space available than 11 then this is likely
And if 100 or 101 has less space than 11 did and fills before 11 splits then the tree structure is quite unbalanced and your L is 4 but only 7 sections. 2^L is only the maximum sections existing and in a reasonably random spread os size-XOR uploading then I would expect between 2^(L-1) and 2^L
But if unevenness while network grows then that would not hold true and could be less than 2^(L-1)
5 Likes
There’s certainly pleasure in seeing this one to just endure, too. Even without my own node in. If you restart, I would try join of course, but I could wait a week, too.
Do you mean folk join, get out, and restart their own nodes, or that you’d need to restart the whole network?
If the former, I can tell you I wouldn’t get out of the Network immediately after join. But if joining was really easy I might close my node after few hours, after seeing it works. Then comes evening and I have to find a place for my machines to run through the night without bothering anyone in the house, and that’s when I may close my node.
If the latter, I don’t see a reason to restart the network more than once or twice max.
Anyway, it might be nice and motivating to get a response from the network after joining, detailing a bit what you have done and how it matters for the network:
Your node had joined. Network now has 4 sections and 132 nodes. Thank you for participating!
But that is of course for the future networks.
It seems my big put is still chunking along!
3 Likes
Good morning!
I’m seeing this.
...
The network is not accepting nodes right now. Retrying after 3 minutes
The network is not accepting nodes right now. Retrying after 3 minutes
Encountered a timeout while trying to join the network. Retrying after 3 minutes.
The network is not accepting nodes right now. Retrying after 3 minutes
The network is not accepting nodes right now. Retrying after 3 minutes
The network is not accepting nodes right now. Retrying after 3 minutes
The network is not accepting nodes right now. Retrying after 3 minutes
The network is not accepting nodes right now. Retrying after 3 minutes
Encountered a timeout while trying to join the network. Retrying after 3 minutes.
The network is not accepting nodes right now. Retrying after 3 minutes
The network is not accepting nodes right now. Retrying after 3 minutes
...
If “Encountered a timeout” means you should restart your node, why not just have the process kill itself? Rewording would certainly be good, but why even keep trying “after 3 minutes”, if something is terminally ill?
I’m thinking along the lines of @Toivo’s cows: If a cow tries going through a gate, and there is no nice milkmaid calling it, but rather some drunken cowboy accidentally shoots and wounds the cow, why tell the cow to keep trying every three minutes? If there is no hope of recovery, I say, just end its suffering.
2 Likes
Glad I caught this test. I was able to get the safe client updated and connected without any issues. I only tested ‘safe cat’ for a few files.
8 Likes
Ah no, if you see this it’s still trying to connect then.
There is no queue so to speak. It’s just whatever node first returns an aggregated sig from enough elders.
Indeed. This is actually what we need to trigger if the upload verification fails.
So this morning i can get safe://hygyy1yybp9edeorgb1wyu9yyoofd9nuz3czc9339kx7gdcumq9fx8r6sj9ey > BegBlagandSteal.mp3 super fast. Which means a chunk was not returning yesterday morning. (But since folk had it, it was there), and was reuploaded during the day yesteryday.
I have some time this morn before the small thing wakes up, will see if I can see anything enlightening.
@Sascha the daily releases are back in action. So yeh, Friday had a fresh one. It should (i think) be compatible. There’s been no changs to the messaging API. But if you want v 0.44, you can run the script itself (after downloading it) w/ install.sh -v=0.44.0
4 Likes
I can see most adults have only ~6-8.5gb stored. So there will be no nodes joining anytime soon I’m afraid.
I’m going to bring this down just now.
11 Likes
Wow, this thing lasted more than a day and you guys didn’t manage to break it. Not bad.
13 Likes
Was just thinking same thing. Think everyone chucked everything they had at this test 
11 Likes
Positive that the network was stable. 2022 could be the year.
8 Likes
That’s 36 hours after start, roughly speaking.
Maybe something like 3GB nodes would be a good value for quicker splits without overwhelming the network?
EDIT: By the way, its seems that the 8,3GB recursive put managed to put about 95% of the files! Just a few Failed to obtain any response OR verification failed. Encouraging indeed!
9 Likes
The network is not accepting nodes right now. Retrying after 3 minutes
Got that all last night.
Indeed, my thought process there was not… very well-thought-out! 
Yes, that’s the whole premise of self encryption and xor addressing. It should be true on average. There may be short periods of “time” when one side in a local branch gets ahead of another in the process of growing from 2^L to 2^(L+1) but your graphic is extreme.
1 Like
I’ve only just been able to play with the install script today, I know the network is offline but I’m guessing my error isn’t related to that - can anyone advise why I might be getting:
felixmc@az-safe:~/.safe/cli$ safe networks switch $COMNET
Switching to ‘playground’ network…
Creating ‘/home/felixmc/.safe/node’ folder for network connection info
Fetching ‘playground’ network connection information from ‘https://safe-testnet-tool.s3.eu-west-2.amazonaws.com/public-node_connection_info.config’ …
Error:
0: expected value at line 1 column 1
Location:
/rustc/db9d1b20bba1968c1ec1fc49616d4742c1725b4b/library/core/src/result.rs:1911
.safe/cli/config.json appears to only contain one line:
{“networks”:{“playground”:{“ConnInfoLocation”:“https://safe-testnet-tool.s3.eu-west-2.amazonaws.com/public-node_connection_info.config”}}}
2 Likes
I suspect it’s just cos it’s offline. Clicking on that link gives
AccessDeniedAccess DeniedJWWJ558V41W5B0SVpVuEniISofZvXOcJblhI2pZBgvcMveRfwzdsSuwNkHPQO2AjE5OnnW/ZoxxJci6NZMJPRwaPflU=
4 Likes
Ok, thanks. I thought I’d see:
The network is not accepting nodes right now.
As per Sacha’s note this morning.
2 Likes
That message is when it’s online but doesn’t need any new nodes to form the network. You can upload and download data, but you only store chunks (and eventually earn SNT) when you’re part of the network, so there’s a queue to join.
2 Likes
First timer here: is there a network that we can join and test with each other? I tried with the announced one but never got the network config. What’s the custom here to spin up new networks?
4 Likes