Mainnode + smb connected plot folders?

Support
#1

Hey follow chia enthusiasts.

So I’m considering alternatives to the main node/harvester approach. Seems a bit buggy and not well documented tbh. Well at least I’m going to try make it work, though also need to consider a backup plan as follows:

What if we mount the harvester drives as SMB shares and just have the mainnode do everything directly adding folders as network shares to the main node?

These network shares would then represents actual drives on harvesters. Thus the harvester would basically not even run chia and just be an NFS server. Not sure if performance would be good. IE can it seek into a random file location and respond to challenges efficiently with this approach vs using a proper harvester server? (chia start harvester)

#2

I do not recommend this. There have been issues with farming plots over the network on a NAS as opposed to locally attached storage or a Harvester running locally on the NAS.

Updates have been made however that if a harvester is taking longer than 5 seconds to respond to a challenge that it shows up as an error in the log.

2 Likes
#3

So, mahybe a related question if we had the same plot on two machines, each full node with diff keys… will chia freakout? What if same plot is shared with same config on two diff fullnodes?