How to failover in log shipping

If you don't include the transaction log, then it's not really log shipping, is it?

Your full backup is restored to that point in time, and that's where it will stay until you either (a) apply a log or (b) apply a differential or another full backup. Yes, the restore does also bring over the transaction log, but only the state of the transaction log at that same point in time. I guess technically you could argue that you shipped the log once, but the point of log shipping is to keep the secondary server up to date more often than the full backups (which are typically very large in comparison to periodic log backups).

So if your intention is to do this backup / restore once a day, and you don't need to keep the copy up to date throughout the day, just do a simple backup and restore, and stop calling it log shipping.

If you want the copy to stay relatively up to date throughout the day, then yes, it will be mandatory to take periodic log backups on the primary server and restore them on the secondary server. Note that you have to weigh what's more important: keeping a copy of the data up to date, or letting users query stale data uninterrupted. If you restore a log to the secondary server every 15 minutes, you have to kick all the read-only users out every 15 minutes, because you need exclusive access to the database in order to restore the log.

You can read more information about log shipping here:

http://msdn.microsoft.com/en-us/library/ms187103(v=sql.90).aspx

If you were referring to Database Mirroring when you say mirroring (as opposed to SAN mirroring), then don't. It has been deprecated since 2012 now so it is a potential candidate for complete removal in the future releases.

Go with Always On Availability Groups. The difference in DBA overhead should be mostly the same. What is different, IMHO, is the learning curve. Log shipping is dead simple to learn, implement and operate. That's largely because it was originally created by MSIT DBAs (decades ago) as a simple, flexible DR solution which was subsequently productized. DB Mirroring applies the same concepts with a few powerful enhancements (e.g. shipping blocks instead of backups gives a sizeable perf boost).

Always On AG extends these concepts but added dependencies that users now need to learn to use and manage (e.g. WSFC). It's not rocket science but may take some effort to pickup if you're not already familiar with them. However, once you figure it out, the benefits are pretty significant (grouping resources, readable secondaries, much higher transfer perf are among the top ones).