Fleming Testnet v1 Release - *NOW OFFLINE IN PREPARATION FOR TESTNET V2*

[Luis_Gonzalez_Alvare]

Congratulations to the entire team and the entire community !!! I will never forget the launch date of the tesnet, since it celebrated my best friend’s birthday !!! But I would like to convey the following concern to the community. I got to know the project at the end of 2017, and at the beginning of 2018, I got on the boat … I have always had a “global” idea of ​​the project, but I do not understand and I get lost in many of its developments … In short, I am an end user, not a person who is professionally dedicated to the world of computer science / programming / developers … and others … I have just checked that the tesnet has come out, but I feel the helplessness of not being able to get on that boat … it is too complicated for me. I’m dying to have a node and become a farmer … but I can see that I still have time for it … and I will also have to put myself “in the queue” and be admitted by the protocol. I have spoken with people from the project, and they are waiting to learn more, but they have the same knowledge limitations as me. I have been saving some money to get the necessary equipment (a good team, or two, the best that can be …) instead of buying maids … and I see that it is not yet known what type of equipment It will be the most “profitable” to be a farmer … I would like to know something more about when, a person “without knowledge” but eager to enter, lobcan finally get … Is it possible to light a light of hope?

[19eddyjohn75]

@bochaco
Sorry to bother, but it’s essential to our SNapp, that users can send money to nrs. safe://cryptokid works, but cryptokid as seen below doesn’t, it use to work in the previous testnet.

eddy@hal9000:~$ safe nrs create cryptokid --link “safe://hyryyyyyr3cqeqy4fibcpkwqxu4asjphm86o9ezfkfrquhg77zafpsj3hac?v=0”
New NRS Map for “safe://cryptokid” created at: “safe://hyryygy9r6anbaqtztteohcxer396swiym1fp9wfjweyixtsduspzwy9ujhn7a”

• cryptokid safe://hyryyyyyr3cqeqy4fibcpkwqxu4asjphm86o9ezfkfrquhg77zafpsj3hac?v=0

eddy@hal9000:~$ safe keys transfer 0.342 --from ac945eb3e10cd9180c0dfcc8206a95a1449e116149e2af3bf4b16eeff87a582f --to cryptokid
Error: InvalidInput: Invalid Ed25519 public key bytes: cryptokid

eddy@hal9000:~$ safe keys transfer 0.342 --from ac945eb3e10cd9180c0dfcc8206a95a1449e116149e2af3bf4b16eeff87a582f --to safe://cryptokid
Success. TX_ID: 0

A googolplex thx for all the hard work, super exciting to have a SN up and running again.

[Traktion]

@Antifragile

Can someone from Maidsafe give us some information about how fast the network was when they run it internally? Is this slow speed problem of this life network only? I would like have some overview of whether the slowness is baked in the code and was there in testing environment, or it is because of some networking issues on public test net. Could you please post here some estimate times for upload/download/coin transfer you experienced in your local environment? Thank you.

I can only comment on baby fleming, but that was pretty quick. Obviously, all local, but actions were near instant.

I suspect there are many things that could impact on speed though. Slow nodes being allowed in, not being punished much for being slow, etc, for starters. I don’t know how much of that is in place, but I can’t remember seeing much on that side in the source code currently.

Then there is the known CLI issue that is causing various slow down and making it hard to upload whole directories. The team have an idea about that cause and it could well be causing all sorts of performance issues.

Imo, just proving it works, splits happen, things resolve, etc, are huge steps. The rest will come. I’m sure the team have learned a boat load of stuff already and the next few iterations will start to address them.

[davidpbrown]

@19eddyjohn75

$ safe keys transfer 0.342 --from pot’o’gold

27.947999994 up for grabs there?..

[Scorch]

@Nigel

got me up and running quick

Good to hear! I’m glad the Wiki could be of some use already

[Josh]

Has anyone received rewards yet?

[intrz]

Self encryption and decryption seems like it currently is using AES on CPU. Doing this on the GPU should give at least 10x performance improvement for the encryption and decryption.

At the moment though, I think there are other larger performance issues that needs fixing first.

edit: It might already be optimized enough with custom instructions on modern CPUs to not be a bottleneck

[Toivo]

Installed on Ubuntu 18.04.4 LTS. Followed the instructions, no problem at all.

• Downloaded the safe-the-planet.png
• Created coins.
• Uploaded a file: safe://hyryyry944czd9zjww7bhhnoceqwqbrzekyfdqezqrbchm4maai8yhbphurnra Toivo_through_bottle.jpg
• Downloaded the previous file

Everything fine so far. Next up: joining as a node.

[Traktion]

@Antifragile

@Traktion

I can only comment on baby fleming, but that was pretty quick. Obviously, all local, but actions were near instant.

This is important, if the code on localnet works fast, then the bottlenecks can’t be just long distances. For 60 second download latency we would need to have our vaults on the Moon:)

Anyway, for working web, we need super fast download. Upload can be slower, coin transfer can last seconds, no problem with that. But download is super important. Websites have hundreds of files and those are not loaded in parallel. Many webs need first to download some file, then start to download another, etc… So with latency in seconds, this still can work like advanced bittorent, but has very small chance to server actual websites. I of course don’t know where is the problem, whether the latency is caused by initial communication and than it is fast. I don’t know, how fast it is if I want to download 100 files in parallel, etc. But those are very important questions. And I would really appriciate if someone from the team could tell us, what speed do they expect once everything is fixed and ready to go live.

I think we can assume this isn’t even close to the expected performance of a fully operational network. It wouldn’t be feasible to wrap a browser around it, let alone these simple test cases.

It is the first iteration of a fully distributed data network, warts and all. I’m sure the team will outline expectations, but I wouldn’t assume this is close to representative of the final network.

Cut the team some slack though. They have worked their backsides off to get this test net out. Let them pick over the bones. The answers will come.

[urrtag]

@StephenC

@davidpbrown

Does TryJoinLater suggest the node is stopped, until manual retry or will it naturally try to join later?

TryJoinLater means you’ll need to manually retry - making this automatic would defeat the purpose of this anti Sybil attack measure (I know it’s frustrating though)!

But doesn’t it punish just the “law abiding users”, an attacker would spin as fast as they can and push the normal users away?

[StephenC]

Ok, so sounds like whatever different version of the Visual C++ Redistributable for VS 2015 you had just doesn’t work. Hopefully you uninstalling that and using the linked version got you going again.

I’d like to get a screenshot laden set of instructions on here for you, but I don’t have a Windows machine. I could set up a VM but I won’t have time to do this today - maybe somebody on the forum can help?

[dirvine]

@Vort

I was using custom launch command, which included --skip-igd (since I do not need it), --public-addr and --local-addr since it was required.

For clients it’s best to set nothing. If you do then they try to reuse a port it may linger (an OS thing) and that can be 30secs or more and sometimes port reuse is also an issue at the OS. So clients always random port etc.

For node only ever set local/external addresses if you have manually port forwarded your router, otherwise leave all this blank too, it will magically just work. For manual port forward there may be periods on startup it looks hung if you restart fast. This is just OS fun and games though.

[Mendrit]

And it would be great to see the progress from very slow network to something much better and user friendly. Some sites I visit takes several seconds to load, so safe sites could be one day faster.

[Traktion]

@Antifragile

I have pretty good experience with realtime performance. We developed a realtime server where we have 10k user online at one moment. And we had first performance issues with 20 users online, than practically we had to fix like 100+ various kind of deadlocks and rewrote pretty everything over time to be able to handle that load on a single machine. So I know very well, there can be very stupid shortcuts in code development that does not seem like a big deal, but that make it impossible to operate once there is some amount of users online.

How long did it take you to identify those bottlenecks? A few hours or several days and weeks of studying the logs, testing theories, etc?

Clearly, the team expect to deliver much lower latency than this test net has delivered. The proposed network use cases require it. I’m just wondering what answers you are expecting them to give so soon.

Anyway, many a layman reads this forum and does not know your background or expertise. They can easily read that 60s response times may be expected, when clearly they aren’t. That isn’t a great perception for them to walk away with.

The network has tested a pile of stuff and much like Elon’s recent rockets, it still needs work. Let’s judge on the successes and give the team time to work on the failures, that’s all I’m suggesting.

[GeekOverdose]

I’ve followed the steps:
Downloaded the safe-the-planet.png
Created coins.
Uploaded a file with container URL of: safe://hyryyry3z7ma7tkgqjieun3r1gwimhzzfzbrkf54rb3bx6o81bmdqupk8ghnra

BUT. I cannot fetch the file in the container for some reason

If I cat the container I get the container URL (safe://hyryyry3z7ma7tkgqjieun3r1gwimhzzfzbrkf54rb3bx6o81bmdqupk8ghnra) and the URL of the one file inside it (safe://hygoygymh1jkcahwp5nur5b4e9eig87uos6dxx3tqfi8fm439xsqdgsdtgw)

But if I try to cat the file itself (safe://hygoygymh1jkcahwp5nur5b4e9eig87uos6dxx3tqfi8fm439xsqdgsdtgw) then I get this:
Error: NetDataError: Failed to GET Public Blob: SelfEncryption(Storage)

[davidpbrown]

$ safe nrs create moneygoeshere --link safe://hyryyyygibj47cms9wz5587iua7zijjbjw5ie6ecxhjapiud876i5ssruja
New NRS Map for "safe://moneygoeshere" created at: "safe://hyryygyh1jdf8oextg36mgmpgqmointhhgf3wnoyjwr698g8psjf4frneywn7a"
+  moneygoeshere  safe://hyryyyygibj47cms9wz5587iua7zijjbjw5ie6ecxhjapiud876i5ssruja

but then

$ safe keys transfer 10 --from ac945eb3e10cd9180c0dfcc8206a95a1449e116149e2af3bf4b16eeff87a582f --to safe://moneygoeshere
Error: InvalidInput: Failed to parse the 'to' URL: safe://moneygoeshere
$ safe dog safe://moneygoeshere
Error: ContentNotFound: Content not found at safe://moneygoeshere

???

[tfa]

@dirvine

For node only ever set local/external addresses if you have manually port forwarded your router

Another use case is to control the port number used by sn_node. For this you currently need to pass both --local-addr and --public-addr parameters with the same socket address. It works, but this isn’t user friendly and using port forwarding syntax seems inappropriate for this case.

I have proposed a PR to simplify this use case.

[dirvine]

@tfa

Another use case is to control the port number used by sn_node.

For what reason @tfa?

[tfa]

• opening a single port number in the firewall
• dedicating the same port number for all my nodes (in the past maidsafe used to use 5483)

[neo]

@maidsafe

Cannot start node due to error: Routing(TryJoinLater)

@JimCollinson @StephenC Do not know if anyone said anything about this yet, but it is an overloaded response. I took it as being that the s/w could not get through my router but in the OP it says

@maidsafe

therefore you may be attempting to join the Network but your logs tell you that the Network is not accepting new nodes at this time:

Thoughts?

Does the node continue to retry to join or is it a try once only and need to rerun the command?