How does the node reconcile missing receipts? Say, on claim, the current frame/increment is at 10 and my backed up copy of the store file only accounts upto 5 proofs completed, would that corrupt the entire group of receipts or are the work done checked individually therefore if I can only present 5 receipts, I’m only rewarded for those 5 and the rest are ignored?
The work is checked in a way that runs from the latest available to the initial proof. If you lost the last five proofs (6-10) and had the first five (1-5), your node would check in starting from the fifth, working backwards.
If somehow (this should not be possible) you had the 6th-10th proofs, you would be stuck unless you replicated the conditions (e.g. if you switched core counts on the 6th proof from 2 to 4 cores, you would need to run the first five proofs again under 2 cores, but it would require some degree of manual intervention to make this happen)
but it would require some degree of manual intervention to make this happen
Is it possible to automate this part?
Myself tried different core counts to maximum the rewards, also some nodes have been migrated to another server (with more cores, as previous server is not economy any more due to too less cores).
I even cannot remember what core count each node has been run (and the history of the changes).
I believe this is quite common, it would be great if the claim command can just deal with “core-count” changing scenarios.
There may be a misunderstanding: if the proofs are monotonically intact (e.g. perhaps a missed backup that lost the last five proofs, but 1 to n-5 are present), there’s no problem. It’s if you had 6 through n, but somehow 1-5 was lost (which again should not be possible without altering the store or catastrophic data loss with no backups) that manual intervention would be required.
If you have your proofs, it is aware (self-evident) of the core count changes
3 Likes