In one of my PC’s, when there is a fault on the disk (disk not recognized/ready), the farmer on that PC just stop/crash. So can this be fixed in next release?
I’d like to suggest the farmer will still run as usual and it knows to skip (exclude) that broken plot/disk until the next restart. At the same time, an ERROR message will display in the log about the broken plot/disk so people is made aware and take action asap.
Thanks,
There is currently some logic for which all farms need to be ready. Farmer is not prepared for things to appear/disappear at will, at least not right now.
I can see how this might be desirable, but at the same time farmer crashing altogether brings more attention and will likely remain the default even if something like this gets implemented.
The way I thought about it, Farming supervisor and sub-farmer · Issue #2402 · subspace/subspace · GitHub would support some level of this, specifically it’ll freeze some internal operations until remote farm is available again, such that it would be possible to upgrade remote farms without parent farmer crashing, but not sure about local farms yet.
I see. It seems it is related to the mechanism where the farmer stores pieces across all of its plots. And maybe more than that which I’m not aware.
It’s not a must to change anyway. Though it is supported by Chia. Anyway Subspace is very different to Chia.
I think we will support something like this eventually, but it will require architecture refactoring to be supported properly. I agree it is probably a useful capability to have.
It’s indeed more of a psychological feeling of the ‘perfect’ software than the actual value it brings. Not a mandatory one of course.
I’m just proposing by collecting some good features from other blockchains I know. Not to say, Subspace has a lot of unique features that other blockchains can only dream to have.
Dynamic reward to fight pool for instance. I’d love to see it before mainnet please. Be prepared for netspace to reach 1 EiB or 10EiB, nothing is impossible.
Thanks.
In my discussion today, I found out more people has subspace farmer crash if their disk/plot is crashed. So while many of us are farming with dozens of disk, I think the feature is needed before mainnet now (changed my mind).
Thanks.
Should be addressed by Improve error handling by nazar-pc · Pull Request #2639 · subspace/subspace · GitHub
Appreciated this. I’ll be a very much welcome feature by community.