What if any responsibility does a host have when new drives fail and all data is lost?
 

Just curious, does the host does have any responsibilities (I take it most will add to their TOS that they are not responsible) if all data and databases are lost due to a hardware failure on new hard drives on a fully managed dedicated server.

Question is purely hypothetical..

This is one of the only reasons I've been BEGRUDGINGLY moving my stuff to "cloud" hosting.

I'd say none unless it's stated that they backup your stuff or made some sort of promise.

Does not sound like a bad idea right now, I should have entertained the idea long ago.

Cloud hosting is not a magic bullet. Your hardware hosting the virtual instance can still die and all the data on it. You should consider your VM host disposable and provide contingency, make regular backups and take snapshots.

None but it's always a good idea to consider a raid array on your servers.

What you expect the host to do? pay for all your losses in profit and costs to rebuild your business because of a hard drive (that they dont even make) decides to take a shit?

Any proper cloud host has a redundant distributed SAN storage array. They don't need "magic".

That being said, I always have onsite and remote backups.

No, not at all. Just checking to see what the industry standard is.

If the server you purchased has backups then you should be able to have them restore your sites after they install your new HDD. For managed servers this is usually the standard.

Not if you host with a real cloud provider.

I do not disagree with you on the fact that you should always make backups. But a TRUE cloud infrastructure doesnt include a single point of failure. You need to make sure your data is not on the hard drives of the actual server, but on a SAN that has redundancy. In a true Virtual environment, if you server completely shits the bed, your VM should be easily transfered to the next physical host, because all the data lies on the SAN. The network should have redundant switches, redundant servers, redundant power supplies, etc and the data center should have a generator. This will technologically leave you with no single points of failure. Most people dont do this.

Find a provider that uses VMware virtualization platform. It is second to none, make sure they are using a quality name brand SAN like EMC, NetApp, HP etc. The problem is most people are not knowledgeable on these technologies and hear 'Cloud' and get all excited and are easy sales targets.

But yes, If you dont have a backup everyday, you are not taking your business seriously.

I'm not sure there would be any recourse as such, but perhaps they should have impressed upon you that a single drive failing could lose your data. Basic software RAID1 is supported by most OSs and is a relatively cheap way to keep you going in the event of a drive dying. If your server has hotswap bays, you don't even need to shut down in order to replace the failed drive.

A couple of years ago I had a server with brand new enterprise drives handed over; within less than a week one of the drives failed...