L1(FullSync): syncer stuck on side chain during full sync

[fmoletta]

As we download batches of block headers from peers we might receive a batch from a peer that is not following the canonical chain. As we have no way of knowing whether the blocks are canonic or not, we will proceed to request the block bodies for said batch and fail to do so as the blocks are not canonic. This may lead us to fail the sync and lose our peers as we asume they don't have the blocks we need.

This was encountered of fusaka-devnet-3 testnet with the following logs:

Upon checking the block explorer we can see that the hashes for blocks 462 & 483 do not match the ones on the canonical chain.
The logs following the ones on the screenshot consist of an endless stream of "Looking for Peers"