fix: copy and prune data from database with move_to_static_files, before a pipeline run/unwind

[joshieDo]

Before we were only copying data from database to static files with the expectation that during live sync the pruner would gradually delete it.

However, if there is an unwind that fails/crashes/panics for whatever reason, we end up in a situation where the database has supposedly unwounded data (potentially invalid), even though static file data has been unwound.

This PR makes sure that we delete whatever data we copied pre-pipeline run/unwind. So, in case of a failed unwind, there's no leftover database data

[mattsse]

that makes sense to me,

the clone is a bit meh, not a big deal, but what if we Arc the DB in the producer instead?

[shekhirin]

The change LGTM, only need to set the pruner timeout because IIUC otherwise we'll use the default 100ms one

[joshieDo]

tested with reth stage unwind and passed, but needed this added to the pruner 8b5444e , for when the tip is the highest static file block. Otherwise Pruner wouldn't be called

[joshieDo]

@shekhirin 8b5444e approve?