Noob here... Probably stupid noob questions

Please point me to the info if available:

  1. I have 2 machines with storage, is there a way to combine them instead of having full nodes on both machines with the same wallet?
  2. I see that you can “copy” a plot from one machine to another, but how do you get the second machine to start farming with that copied plot?
  3. Can I put plots on a slow NAS drive or does it need to be local?
  4. When you are farming, does the plot have to be on R/W media or can it be R/O?
1 Like

Dont have answers to all your questions but here goes.

  1. not sure how you would “combine” machines. You can set up a machine to be just a harvester that is farming w/o having a full node running but you would have to check the GitHub or wiki notes on how.

  2. As soon as any node sees plots in the active directories it will start farming them. There is no free lunch though. Having the same plot file in different farmers does not give you more chances of getting a reward. When under the Plots tab on the left just go to the 3 dots in the upper right (to the right of the “Add a Plot” button) and that is how you “Add Plot Directory”. These are where you tell Chia where the plots are that you want to farm from.

  3. A slow NAS is fine. People are farming on Raspberry Pi 4s.

  4. The Node needs to be able to search the plots but I doubt if there is any more writing that is done after the plot is done. Others will correct me on this I am sure. I think R/o would be ok.

2 Likes
  1. I have not figured out how to setup a node for farming which didn’t include a full download of the blockchain. So, both nodes currently have identical wallets running on them with full block chains. I heard this is not ideal, but I don’t know how to setup otherwise.
  2. I did just copy the files from my laptop over to my desktop but the GUI doesn’t see them. Is there some directory structure which you need to use on both nodes so that it will “find” them on the second machine? Im not trying to double-up here, just want to move from the laptop with small drive to desktop with large drive. This is more of a “move” than a “copy”.
  3. Ill have to look into that, thought the bandwidth requirements were such that you had to do significant I/O during farming so that it could validate your “lotto tickets” in the file being farmed.
  4. Just wondering how much I/O is required here, but from #3, it sounds like not much.
  1. It may not be ideal from a security perspective since your private key is on any full node but if you have secure control of both machines I would not worry about it.

  2. See the attachment. As long as you have the directory added, you should start farming them. Are you running the latest 1.1.0 version? Just try uninstalling and reinstalling Chia or restarting your computer. That has always fixed any issues like that of mine.Screen Shot 2021-04-23 at 7.32.04 AM

1 Like

What is your plan or purpose for having 2 machines? Do you have different plot files on each of them? If yes and if they are on the same network, consider doing a network drive on one machine to the other and add those as another plot directory, similar to what chiafarmer explained.

AFAIK to farm you need the full blockchain as that is the purpose of the farming. If you want to create plots, then you don’t need the blockchain.

The instructions that chiafarmer provided should address this, you can have multiple plot directories for the farmer to look into.

I haven’t seen a concrete confirmation on this, but on this forum and reddit I believe I saw that there is not constant read on the plots only if a certain condition is met then it goes and reads from the plot files.

Same for the writes, I saw that writes are not performed on the plots once they are complete, but don’t know for sure as I haven’t tested that myself.

1 Like

Thanks, What happened is that I copied the plot from the laptop over to the farming machine and put it into the same directory as the existing plots which were made on the farmer itself. Turns out that the farming machine said the plot was failed and wouldn’t accept it. I tried again and found out that I have to stop farming during the time I made the copy, then restart the farmer after the copy is complete, then it worked. So, im assuming the farmer tried to start using the plot before it completed copying.

I read that changing the extension to say plot-mv while moving will avoid the confusion that you mentioned. Then just rename the extension to .plot once copied.

1 Like

Sorry, but I’m positive you should not have to stop and restart the node. Something else must be going wrong with the plot or the copy. The farmer is smart enough to wait until the file copy is finished before attempting to use it. I do it all the time this way.

2 Likes

Thanks, ill try that next time.

Yea, I don’t know for sure what the problem was, I only tried this once. The first time failed while I had the farmer running, but the second time worked after I shut the farmer down. So, only 2 data points, small sample. Could be a network problem or something else.

Thanks for sharing a bit more info usao. Are all your questions answered now or still anything pending? Good luck!

1 Like

All good. I have a method which is working for me.
My next goal is to figure out how to do the command line plotting so I don’t have to run a second full node on the plotter.

That should be easy - which OS are you running? hoping you say Linux lol

If running Linux - once you install and activate chia do this:
chia plots create -t /temp/dir -d /final/dir -f farmer_key -p public_pool_key

This will run 1 plot with default values (~4600 MB RAM, k32 etc.), you can add -n X where X is how many plots you want to be queued up.

To get the farmer and public pool keys, on your main node run: chia keys show and it will display those values.

If not Linux then, nevermind haha :slight_smile:

1 Like

I could run CentOS7, but currently running Windows. Just will take some time to get all setup.

No need to change OS if you are comfortable with Windows.

Here is a guide on Reddit that might help, they are talking about parallel runs here, but you could just run that one portion with start-process etc until the sleep part.

1 Like

This helps a lot. Thanks.
Im going to take my HP servers and install ESXi, then I create a Linux VM to run the plotting. That way I can have both Windows and Linux at the same time.

@usao I don’t see why you would want to do that. You’ll just add overhead for no gain. Run a full node on bare metal and call it a day, your choice on Windows or Linux based on comfort. I hear Linux plots faster though if you want to shoot for max perf. I admin a vSphere server at work and it’s great when you need tons of VM’s, but whenever possible for performance stick to bare metal.

Someone on Reddit has this same issue. It’s because you need to STOP farming the plot you are going to copy over first. From your description of running 2 full nodes this is highly likely the case since you are farming the node youre trying to transfer.

And you don’t have to stop the farming node like your method. Just remove the plot in question from the directory so it’s not being farmed before you transfer it.

1 Like

Yea, I decided to split the farm into separate volumes on the NAS, with each node pointing to it’s own volume. Both full nodes have their own blockchain, but use the same key. I think it’s working now.

Ive updated my plan to just use local directories for each node rather than the NAS.