Hi. I would like to build box of nodes. I wonder if scale it. I’m planning to build box with 8x raspberry pi + 8x 1TB disks and switch with poe. Is it ok?
Better will be 8x raspberry pi and 8x 1tb disk or 4x raspberry pi and 4x 2tb disks?
What I should know about building my “miner”?
1x rpi and 8tb hdd… After its full buy next…
But if you have more public ip buy more rpi… 3x public ip in diferent /24 segment 3x rpi
That’s my plan I’m getting 4Public IP addresses and a single machine with 24Tb and it will just contain 4 VMs that will allow for scalability in future
Those public ip adresses have to be from different /24 subnets too though.
Be careful with regards to costs if planning to get multiple IPs and multiple HDD & machines.
Nodes receive data pretty slowly, so you should start small, and ramp up little by little when needed, as suggested by @Krystof.
Otherwise you would end up with a monster setup whose ROI would take tens of years…
Have a look at this to estimate your potential gains:
(Copy the sheet to your own Google account for modifying it)
I have a question about the IP addresses
So what is meant by the /24 subnet
Please note I am talking about the Public IP Addresses
will IP address
72.52.58.1
And address
72.52.58.2
Looked as 2 different nodes
Or still as a single node
That is the same subnet.
They would be treated as a single node. They are in the same /24 subnet.
Here is an example of 2 IPs that would be seen as different nodes:
Thanks a lot for explaining it
Thank you. I filled this sheet, but on YouTube channels I have seen different results between 2 vm’s - 1 vm with 1 tb and 1 vm with 6 tb. In option with 6 tb author earned only 2x more than with 1tb disk.
About ip adresses - this is problem, because probably I have only 1 public IP and options for port forwarding. Can Open VPN on my server solve this problem?
This can vary depending on many factors… if a node had free space while ot a lot of test data was being heavily sent out for future tests, it could earn more during the test phase, for instance.
Also, the type of disk, network quality, machine power… Many things can impact nodes’ performances.
The estimators is just an estimator… But in average, it proved to be quite accurate.
Besides its author is regularly updating it to reflect past months observed performances.
In theory I guess, if the vpn provider allows for port forwarding… But that would add another point of possible failure to your setup, and additional latency.
This said, I’m really not used to VPNs, someone with knowledge on that should hop in here to give you better insights on that matter ^^
Also note that the vpn provider would cost some money too, so one should carefully estimate if it’s worth it or not. Especially as one vpn per node would be needed…
Recently I’ve seen a node with <2TB earn more money than my 7TB node… But those are exceptions.
As for VPNs, they work but might not work well or might also be used by others and therefore share the subnet.
even my nodes in same pc ern different money on same size and same age, may be there is some more relevent data in node that used more often.