Clear hash id database on package resync.

[simpoir]

This should avoid loop when inconsistencies are present in the database
through either corruption or server-restore. (LP: #1616116)

Testing instructions:

• launch a landscape server and register (using ppa packages is somewhat simpler as hash-db is pre-populated).
• register client against server. wait for package info
• pg_dumpall
• add a repo and wait for new package to show on client.
• restore postgres backup.
• trigger a package install from the new repo to have new package show up
• if you haven't regenerated hash-ids on the server, clients will be stuck in a loop. this is to be expected as the server is inconsistent with itself.
• client should resync once then will re-fetch hash on the next run. Some hash-id requests may go through in the meantime.

[simpoir]

A nice side-effect of dropping the hash database is a delay is induced in the package info refresh, thus throttling the refresh and avoiding the server being overloaded.

[sparkiegeek]

LGTM 👍