fixity failures needlessly repeated to DAITSS events table

[grf]

One requirement for fixity checking was to log only the most recent fixity success (doing an update on pre-existing success record.

However every fixity failure was to be logged into the events table.

Well, collect-fixities reconciles silo information from the silos nightly without keeping track of the silo's timestamp on the fixity check. So the fixity checks get needlessly repeated nightly.

This is an easy fix.

Also: there should be an easy way to re-run a fixity check on a single package (even from tape). This is pretty easy as well.