Restore worker pool has slightly different behavior regarding temprorary failures

What's the issue?
The updated restore worker pool (using gocbcore retries) has slightly different behavior in regards to temporary failures.

Previously, these operations would be retried indefinitely, however, now they'll be retried until we reach the operation deadline.

What's the fix?
I don't believe we have a simple fix here since we can't change the deadline value supplied to gocbcore; this will probably need some more thought.