The command above downloads and unpacks the snapshot in one step: it only needs as much storage as the [unpacked] snapshot itself, as it doesn't store the archive. Now we can go back to the previous directory:
If you're already running a validator node and would like to switch to using RocksDB, you can use the following steps to keep the downtime during the upgrade to the minimum:
Before you start: ensure that your node has enough space to accommodate an additional 60GB of storage for the new snapshot
Without stopping the node: download the snapshot and unpack it in the same directory where you unpacked the legacy, ParityDB snapshot.
You should end up with two directories: db and paritydb (the latter we will later be able to delete, the former is the new RocksDB database)
Ensure your validator is not in the active committee (the best time for an upgrade is right after your validator has stopped being in the active committee).
Stop your node.
Start your node with pruning enabled (see the Running the binary section).
Now you can remove the paritydb folder to save space.
Congratulations, you're now running a RocksDB validator!