feat(mempool): mempool migration protocol

[Trantorian1]

Is there an existing issue?

• I have searched the existing issues

Motivation

Currently, there is no way for a node to migrate its mempool to a new node as part of a database migration. This is an issue for #393 since there is no way for the warp update sender to communicate transactions in the mempool with the warp update receiver, so any transactions there will be invariably dropped once the migration has completed.

Request

• Implement a protocol for migrating the mempool between nodes and update feat(warp): added warp update to madara and other changes #393 to take advantage of it.
• Add rigorous e2e tests for mempool migration.

Solution

A good place for this would be a new admin rpc endpoint. The issue of ownership of transactions is tricky though: we would have to shutdown the rpc on node A, enable it on node B, start synchronizing all transactions still in the mempool of node A at the time, execute them while receiving further transactions on node B, and then execute these new transactions once all the txs from node A have been processed.

Are you willing to help with this request?

Yes!