None never reaches mining state

Nodes & Mining
#1

Greetings everybody!
I deployed a node on AWS using the nknx fast deploy so I can learn how this whole process works. It was created, has ID and everything. However, since it was made, all it does is to go from sync started to wait for syncing to sync finished and then offline… never did it do any mining.

nStatus says this:

  • addr: “tcp://3.67.37.5:30001”,
  • currTimeStamp: 1632859533,
  • height: 3143012,
  • id: “b831ecd6467417ad7b68a8ec24b7dc30bbcdc026fe053f91e5adacb9525c31ed”,
  • jsonRpcPort: 30003,
  • proposalSubmitted: 0,
  • protocolVersion: 40,
  • publicKey: “aab94d90f26063440ef6abf367674b8672f93d4ee0cc79c664c96cf013418f27”,
  • relayMessageCount: 2162139,
  • syncState: “WAIT_FOR_SYNCING”,
  • tlsJsonRpcDomain: 3-67-37-5.ipv4.nknlabs.io,
  • tlsJsonRpcPort: 30005,
  • tlsWebsocketDomain: 3-67-37-5.ipv4.nknlabs.io,
  • tlsWebsocketPort: 30004,
  • uptime: 18778,
  • version: v2.1.6,
  • websocketPort: 30002

Is it normal? How long does it usually take to actually mine. I did reboot it once… still no improvement. I read through the similar post already, no success.

#2

Hi, what kind of VM did you use on AWS? e.g. T2 Micro, CPU/RAM/SDD? Maybe it is due to the VM too small, or is one of those on-demand types that is not running all the time? Just guessing here. Please provide more info. And you can also use AWS Console to check the CPU, memory, disk, and network utilization of your VM.

#3

Thank you for the quick reply!
It is at2.micro with 30GiB storage. The default storage was not enough so I put as much as I could for free.
CPU is at about 11%, there is still enough space on the disk. I could use ssh to login, the services are running, the two ports are open.
It does also transfer quite a lot of data as I already exceeded the free usage limit so now I will need to pay… But, I will give it more time.

#4

Well, seem to work now. It took about 10 days from start to start of mining. Seems a lot to me… Do you think it is normal?

#5

This is not normal. I have used fast deploy in the past, and it took a few hours or half a day. If you have the opportunity, maybe you can try a different cloud provider in a different region (e.g. Digital Ocean, Linode, or Vultr)?

#6

Hi again! You see, few hours I can understand but not 10 days. Also, it only mines for very short times, the rest of the time is spent with sync and waiting for sync.
Furthermore, there is a lot of traffic through that node. This leads to the AWS free tier quota to be depleted in 3 days. But no rewards whatsoever. Just few minutes ago I checked, there were 91 nodes active, I had control of about 1% of the network. What am I doing wrong? Like I said, I used this node as an experiment. I have a lot of routers idling around that could be used but not under such conditions.

By the way, at 1% control of the network I would be making about 5000 dollars per month… That would be clearly very nice.

#7

Hello!

AWS is clearly not recommended to build NKN node as, as you stated, NKN create a lot of network traffic (between 1 and 3TB per month).

Regarding your 1% control of the network this information is incorrect. A node used by the OpenAPI have left the network, it just needs some time to refresh (the node count is around 110.000 as of today). You can find more about the revenue estimation here : https://wiki.my-nkn.cloud/index.php/FAQ#Is_there_a_revenue_calculator.3F_When_will_I_get_my_rewards.3F

Best,

#8

Greetings!

You know, I was very surprised when I went to nStatus last night and saw only 91 active nodes. However. I am sure it was just a mistake somewhere…
But, what bothers me a bit is the fact that even after a lot of traffic there is no reward.
After reading that “The reward system is for most of it a lottery system where every node in “PERSIST_FINISHED” mode can win.” and that most parameters are fantasy I can only conclude the project is basically a scam. How can I even dream of participating in such a lottery if the node is only 5% of the time in Persist_Finished state?
Also, for a clear service I expect a clear reward. Imagine you are a taxi driver and the client pays randomly. How long it takes you to stop working?Is my fault, I should have read more.
Next question: how do I transfer the node to a different provider? Is there some tutorial or clear instructions?

#9

Hi!

Most of the nodes stay stable in PERSIST_FINISHED for months (you can find some stats about the nodes in the network on https://nkn.ovh). This issue might come, as Bruce stated, from a too small configuration for your node.

Regarding the scam well … I would suggest you to do some research before making this kind of accusation so let’s not talk about it :slight_smile: Your comparison is valid in some types of network but not the way NKN is designed. In the NKN network nodes “compete against” each node to be “the best node”, not “helping” each other. If you feel uncomfortable about the tech I would really suggest you to not go further :slight_smile:

Regarding your last question, yes, you can easily move your node from a provider to another. All the information you need are available in the wiki https://wiki.my-nkn.cloud/index.php/After_installation

Best,

1 Like
#10

Greetings!

Well, good remark: what is too small? I read the node can operate on a RaspberryPi. What I have there is a lot larger. Please, tell me, what is a good configuration? Do I need more RAM, more storage, processing power? The load on the CPU never exceeded 15%…
Now, how do the nodes compete? I see nodes receiving rewards every hour. How does that factor “luck” into it? Nothing that regular is related to luck. Also, exchanging the tokens works in mysterious ways. Sometimes you can, sometimes you can’t.
I will check how to move the node to another supplier but I am slowly losing confidence.
I am quite comfortable with the tech, try me.

#11

I read the link you sent me. Basically it will require me to re-deploy the node on another supplier, starting form zero and just replace the wallet. I was hoping to avoid all the downloading and installing and just transfer the virtual machine as a whole. Will try it and let you know.

#12

NKN mining does not need much resources, especially during stable running phase (“persist_finished”). My typical node is a $5 VPS, 1 CPU core, 1GB RAM, and 25-30GB SSD. and my CPU utilization is typically less than 30%. Faster Internet always help, especially in terms of latency to other nodes. SSD is definitely preferred over HDD, since there are quite a bit of disk I/O for transactions and block data.

The only time that there will be a lot more CPU utilization is when the node is doing blockchain database pruning. This only happens every time you reboot the node, typically due to upgrade to new software release. Depending on how long ago was the last pruning, it can take a few hours of close to 100% CPU utilization. But again, this is temporary and happens very infrequently.

But from what you mentioned, the type of VPS you have should suffice.

You second question is about mining fairness and predictability. This topic has been beaten to death, thus I suggest you search the forum. The wallet that you see getting frequent mining rewards, belong to the big miners who aggregate hundreds or even thousands of mining nodes to one beneficiary wallet address. They are not from a single NKN node.

#13

Thank you! That is also what I thought, therefore I am still surprised that it takes so long time to start mining and that the state does not persist. What puzzles me is how it can be deployed on a RaspberryPi if it needs so much storage.

#14

32GB MicroSD is cheap, and even 128GB is affordable. And I do run NKN on Raspberry Pi at home.

#15

You know, I will try it. So far is still not clear why the node behaves so poorly. Also, what bothers me is that big miners thing you said earlier…

#16

My AWS experience is similar: 2 days waiting for sync and I started with ChainDB pruned to 3138675. I also have a t2.micro that seems to be barely sweating … other than network traffic. It should finish today, compacting at 3182400 now.

#17

Thank you, at least I am not the only one.
I switched to Linode, it works a lot better. You should know the AWS has a transfer limit for the free tier which will be exhausted quickly once it works properly.
The Linode however works at 70% of the capacity, this is also not normal.
Anyway, no block yet even though the traffic was quite good. This is still not normal.

#18

I’ll check it out. Digital Ocean locked my account without minutes of deploying an NKN droplet. I’m playing with NKN as a learning exercise. So far the only node that is 100% functional is Windows at my home. At least that one is truly free and I have AT&T 1Gb Fiber with great latency up/down US East Coast. Over 8.1M relays in last 24 hours.

#19

My AWS node is finally behaving. It was super slow to relay, slow enough it would have to re-sync every 15-20 minutes. since it was pretty much useless, I decided to update/upgrade everything. After that it wasn’t any better so I just turned it off: actually took 3 attempt to “stop instance” before it finally stopped. I turned it on about 10 minutes later and it work great now: synced up within 10 minutes and has been relaying >400K/hour the last hour.

While messing with that, I created a WIN2019 instance on Vultr. Yeah, WINDOWS: in under 90 minutes I transferred the 16.4GB ChainDB from my home WIN10 node, pruned, got a new ID, and achieved PERSIST_FINISHED. I’m not going to argue OS, but I’ll take 90 minutes of 30 hours any day of the week. $100 free credit I can run that puppy 5 months and see how many block rewards it can collect.

#20

Greetings!
My purpose was the same, to try and learn something. The node is working now, not as fast as yours but still, quite OK. However no rewards yet. I was planning to use the rewards and make more nodes. I don’t plan to use my credit card anymore to pay for mainnet coins to get the ID. If you know of a way to have it truly free, let me know but as far as I am concerned, it requires about 5 dollars of investing.