Best multiple machines scenario?

Now please tel me what is the best scenario for having multiple PC s , plotting for the same fpk and pool contract on madmax. They are not on the same LAN, far remote. So is it:

  1. Running 1 Full node/harvester + separate farmers , using this super complicated ca certificate and TLS and other stuff from chia wiki, the proper chia way

  2. Running full node on all machines, same keys, plotting with same fpk and pool contract, storing plots on each machine

  3. Something else?

Please advise

1 Like

I have something similar to option 2. I only run 2 systems. I have a plotter and a farmer.
The plotter you can use your Mad Max if you want and plot away. Have it on the internet and on the network.
The farmer needs to be on the internet and with ALL the storage (except for what is temp HDD on the plotter)
Use the farmer as a NAS and transfer the plots over the network to the farmer.

I leave 2 full GUI nodes running 24x7 with the same key and same pool NFT (I did have to copy the synced DB and wallet DB from my farmer to my plotter to transfer a copy of the NFT as the Chia app isn’t updating properly).

I do not know if this is best. In fact, this is considered not a recommendation by lots of farmers. I do it because it is the easiest way to get up and going with Chia and I never had any issues. My farm is only now 21TB. At 18TB before pooling, I didn’t win anything ever because 18TB isn’t enough plots to be win.

Are you mining solo?

Yes and no. I still have about 15TB of original plots left to replot. For now, I am leaving them farming as they were and I only wipe 1 drive on my farmer server/NAS at a time. I am planning to be 100% pool plots this month with 20+ TB.

@kreaninw would say, use a power efficient raspberry pi, and non-powered USB hubs because of back-feed.

2 Likes

@Chia.Switzerland would say, nooo raspberry pi’s are a nono they’re too slow, no one wins on raspberry pi.

1 Like

Are all devices in the same LAN or not?

@khominhvi :sweat_smile: :sweat_smile: :sweat_smile:

Not yet, really :sweat_smile:

However, at first, I was going to answer the OP that whether the 2 PCs are on the same network shouldn’t matter as long as he can make them see each other. The complicated part would be only that since he will have to be able to port forwarding and also pass the firewall. Those 2 would depend entirely on his ISP and his OS.

I personally have my plotter as a harvester and my Pi as a farmer running a full node. But they’re on the same network though. I only control both of them over the internet using the VPN. Therefore I don’t have my experience setting a far away harvester and full node machine :sweat_smile:

If devices are on different LANs he could just run them alll as invidual nodes and leave it be. When both nodes use pool only plots he should be to see his total amout of blocks while not using network drives. Corrrect me please if I am wrong.

Apparently, they’re not according to the OP :sweat_smile:

Why does the OP not just use network shares? Save yourself some trouble, right click the folders the the device which is not running the node and share it in your network (would be your plotter). Then add the network drives on the device which is running the node / harvester / farmer / wallet in windows and add them to the config. No trouble with setting up another chia harvester / farmer / whatever on the second device.

As far as I understand the OP is using Window and not Linux. On Windows it’s just a right click on the folder to share it over the network.

This even works over WAN…

On linux it is a bit more of a pain in the ass but works as well. You need to install a samba server first, and edit the config which paths to share to which users. There are tutorials on the net for that.

I still do not get why one does have 15 TB across a WAN… IT Administrator abusing free storage on company’s pc? :smiley:

BTW. if there are multiple PCs in one LAN, you could just run one farmer / full node in the LAN and a harvester on the remaining devices.

The complicated CA certificate stuff (guide) was originally written because people were running multiple full nodes at the same location. This was unnecessary as you could have the other machines connect to one full node as remote harvesters. This helped people save bandwidth. Not really an issue if you’re only running 1 or 2 full nodes.

There’s nothing stopping you from doing this, it is probably the easiest setup. If they’re on different networks at different locations it’ll work best. If they’re on the same network it’ll still work, just consumes more bandwidth than needed, and also means you’ll have to manage each machine separately, because I don’t think your other machines/plots will appear in your Chia-Gui/Interface.

Technically, if you have too many full nodes at one location, to the point it slows down your internet, it could affect farmer response times. But probably unlikely to happen unless you’re running multiple machines on a low bandwidth internet connection.

I would recommend: for each physical location run one full node, and have the other machines connect to it as remote harvesters. I think this would be the optimal solution.

But know that whatever machine your store the plots on and want to farm on you’ll need to have on 24/7. So ideally, the machine that consumes the least power (because Chia is supposed to be greener and more environmentally friendly).

BTW can you please post the log entries for
Loaded a total of x plots of size y TiB, in z seconds for a Pie? I am really curious to see the number for multiple farm sizes.

I believe this is subject matter would be best suited in a separate topic. Maybe you would like to start a separate topic exploring the performance of pi based farms?

Also, for pi related stats, @kreaninw would be in a better position to answer your question as he has the largest pi farm I know of. :innocent:

1 Like

NO they are on separate nets at the mo. However in aboyt I month time I am moving it to solar powered location, so then they will be on the same LAN.

Right now, I am doing exactly that. However, I am not able to see no.2’s plots on the space pool’s dash.

Linux, not the same LAN

Thanks man. You really are an expert on Chia, its a pleasure to learn from you :slight_smile:

Hahaha, thanks, but I’m no expert. Just someone who is a little further down the Chia road, and passing along what I’ve learned along the way. :+1:

Hi Guys,

I’ve read this thread a few times and thanks for the info you have shared.

If I may ask for your help, because I didn’t find a suitable answer for a little bigger farm.

My complete plot count - 5210 (1612 - NFT the rest OG) but it’s not relevant.

Currently I’m having one Full Node PC with around 110 external 5TB hard drives all connected through hard-disk controllers.

Full Node PC runs only Chia GUI and Farm.

(I’m plotting on three different machines that are absolutely not connected to the Internet using MadMax)

My problem is I have more HDDs and I don’t want to connect them all to one PC. It’s not even possible and trust me - I know it’s a super ugly solution.

Is there any suggestion you can give?

What is the best in my case?

But a few RasberyPis and set them up according to the super complicated documentation in Chia Github or just connect them through LAN ? Or?

I appreciate all the answers and thanks in front.