I am opening this debate here, because dwarfpool is ignorant to all our requests, so there is no other way to inform them about something but here.
After a customer complained to us regarding rejects on dawrfpool, we analysed the thing and noticed following:
After some time of inactivity (if no shares are being sent from us), dwarfpool would stop sending new jobs. This issue is resolved after we again submit share (which would be stale of course, because dwarfpool was not sending new jobs). But every next submitted share would also be rejected (for unknown reasons), even though, it is NOT stale (analyse packets from 60 to 65 in wireshark: https://github.com/nicehash/Specifications/raw/master/issues/dwarf_s.pcapng
So, what that means is following; if you have slow miner and your share submission falls below certain threshold set by dwarfpool, all your following shares will be rejected, even thought these are perfectly valid
. In simple english for those who don't understand stuff about shares; if you have slow miner, dwarfpool may simply suddenly ignore all your work and never credit you for it, until you restart your miner