This changes the way "auto-verify" works: instead of separate "auto_verify" I/O methods, the auto verify code now just marks files for check (has_file='M'), and then the regular check mechanics can do the actual checking. This greatly simplifies things.
The check method on nearline now handles release/restore as needed and the exists() check has been moved into the async.
This changes the way "auto-verify" works: instead of separate "auto_verify" I/O methods, the auto verify code now just marks files for check (
has_file='M'
), and then the regular check mechanics can do the actual checking. This greatly simplifies things.The
check
method on nearline now handles release/restore as needed and theexists()
check has been moved into the async.This is a partial fix for #183.