I have a 4 core node (an old intel i5) that chugs along at 100% CPU syncing about 1 block every 20s.
At 6s block time it’s clear this won’t fly. Maybe the hardware specs should be revised? Why wasn’t this an issue during the stress test?
Thanks.
… and follow up question - is there something I can do about it? Would plotting on another machine be an option somehow, or would the node still lag behind after the plotting?
Seems more people are having this issue.
We have just released some updates that should help! I appreciate your patience, It is mostly just minor issues that are causing problems, as we get it more tuned in you will see less and less hardware usage.
You can technically plot on another machine, just make sure you dont have two farmers plotting the same plot at the same time (that is how you burn a farmer and plot)
but yes you could plot on another system, then transfer the files and fire it back up that is a pretty good idea.
Thanks, good to know. Is it the case then that the minimum requirements for initial plotting/syncing are different from requirements for normal operation? It seems that salt updates can also be demanding.
It would be great to be able to plot on a more resourced machine and then run the node with a more energy efficient spec.
Couple parts to this,
A. yes, the minimum requirements to actually farm are less than what’s required to plot/sync.
B. Salt updates are a bit demanding, but mostly this is a bug currently. Farmer uses a lot of RAM · Issue #569 · subspace/subspace · GitHub (should be available in next release)
C. Ideally we will get plotting/syncing to a point where it wont be too bad either.