The switch to the new P2Pool protocol, triggered about 3 hours ago,
resulted in all nodes mining invalid shares due to a programming
error. Mining has not completely stopped; payouts are just no longer
This problem was triggered by P2Pool's version count going into the
>95% upgraded area, everyone switching to the new protocol, and then
it briefly going below the threshold, revealing a bug. The root cause
of the bug being created was me overlooking the fact that Python's
isinstance function also returns True for instances of subclasses.
Upgrading to the current version in git as soon as possible is
essential to getting the pool functioning properly again. Windows
binary is here:
After enough people upgrade and some time passes, all nodes, upgraded
or not, will resume working.