Storagenode Recovery Mode

I hope so. Because we want to give all components of the network to the Community.

Your node doesn’t contains files, it contains pieces. One piece from 80 pieces of the customer’s file.
The node cannot be trusted regarding data consistence, it should be audited by the authored entity, the audit service in this case.
To audit your node the satellite should download all pieces from your node for the same cost as repair, but repair nothing, only check, then download 29 pieces to reconstruct the file, then split it to 80 and upload them to nodes. The upload from any cloud provider is cost money (see Is there such a thing as an "un-disqualify"? - #32 by Alexey for numbers).
This would be significantly more expensive than just repair when the number of pieces would fall below the threshold (35 pieces by default), when the worker would need to download only once, not twice in case of full audit.
So, if you would like to pay for it, you will pay for audit 20TB * $10 = $200, then for repair (29/80 from the lost space), if you have lost 1TB, it would be 29/80 * 1 * 80 * $10 = $290 plus upload costs from the cloud provider.
In case of regular worker it will be triggered only when the number of pieces would fall below the threshold (which can be not happen if customer deletes their files for example).

1 Like

just run raid 1 or raid 10 50% less data but resistant to drive fails or risk it if i did not already have a raid 10 setup i would have gone with single drive for node then add a raid 1 parity after driver is full or x time goes passed to minimise risk the main question is if just running 2 nodes is cheaper even if 1 node gets lost than running a parity if your talking 10 year plan