Restore log needs more clarity when object already exists and isn't different

[ncdc]

While debugging #1138, what I discovered is the following sequence of events:

1. Ark tries to restore a PV
2. It gets an already-exists error
3. There are no differences between fromBackup and fromCluster
4. Therefore it's a no-op

The only relevant/related log line for the above is

time="2019-01-03T14:36:29Z" level=info msg="Restoring PersistentVolume: pvc-18ceea3e-0a8d-11e9-aef1-02f9e6d3ca22" backup=pont logSource="pkg/restore/restore.go:796" restore=heptio-ark/pont-20190103153555

The restore has no warnings or errors.

I think we should consider changing the log line from Restoring <type>: <name> ... to Attempting to restore <type>: ...

I also think we should consider adding a log line to the effect Did not restore <type>: <name> because it already exists and is identical to the backed up copy

[ncdc]

Relevant source code lines:

(attempting to restore)
https://github.com/heptio/ark/blob/fd8bf14bb67508dbabacfbe1662ed08f9ecc0d10/pkg/restore/restore.go#L796

(did not restore X b/c no-op)
https://github.com/heptio/ark/blob/fd8bf14bb67508dbabacfbe1662ed08f9ecc0d10/pkg/restore/restore.go#L851