Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
Too much stake too fast #5443
The root cause of the TdS stage 0 dry run 2 cluster failure was due to too much stake coming online from epoch 6 to epoch 7.
Epoch 6 Stake
Epoch 7 Stake
@mvines the stake changing from 0-66% is an artifact of epoch stakes, which are a snapshot. this is basically the expected behavior.
Say we take a snapshot halfway through epoch 5 (for epoch 7), if a stake activates in the very next slot, the next time it ends up in epoch stakes will be halfway through 6 (for epoch 8), in which case it's had 2 epochs of ramp (5 and 6)
I don't think we need to hack together a super quick fix for this if it moves us away from a more complete solution. Is that #5426 or do we need more in addition? I think it's reasonable to continue to expect that in a TdS setup, we'll go from 0 delegated stake to a large N very quickly as participants jump on at the start.