presignature inconsistency will persist when deploying next release with redis

[ppca]

To deploy the next release to partner nodes, which will contain redis storage, they need to run tf apply again on their side, which will definitely happen at different times per partner.
And since next deploy will have the redis storage, the presignature inconsistency caused by nodes starting at different times will not be resolved by a simple redeploy any more, it will be persisted.
One way to start fresh, is we build an option of clearing all the triples and presignatures in redis at node restart into our code, which can be turned off. And when we do the next release, we switch that option on, so when every partner has finished tf apply, we hit the deploy button on our side and everyone will start with a clean plate.
After nodes have stocked up some presigs, we push a change to the image to turn off that option. Now redis storage will keep the triples and presignatures at node restart.