Re: cloud backup services

[Jim Oberholtzer <midrangel@xxxxxxxxxxxxxxxxx>]

I have a question.  Have you ever tried a restore from that site?  
Unless you have the *IBM portion and the *SAVSYS at that location, what 
you have is a bunch of files that have old data in them.  You might be 
able to recover a single object, but what about a more critical 
situation?  If so, the bandwidth to get the data there in a timely 
fashion must be really impressive.

Where the storage is for a backup is really not important as long as you 
can recover from it.  Unless you've tested the recovery, you do not have 
a back up.  Period.

Complacency in regular backup routines and then lack of testing is the 
single biggest reason systems become unrecoverable.  I have rebuilt four 
systems already this year with old cobbled together data from multiple 
sources because the customer did not test the recovery to find out the 
backups were not what they thought.  Unfortunately I'm getting pretty 
good at it since I've done it so often.

I can concieve of a way to build a recoverable system using cloud as the 
back up medium,  but that would be using techniques that are not widely 
used, although have been around for some time.

Jim Oberholtzer
Chief Technical Architect
Agile Technology Architects


On 7/25/2013 10:53 PM, Nathan Andelin wrote:
> I don't want to have to figure out how much data we're FTP'ing to our hosting provider, but there have been months where we exceeded our allotted bandwidth, which triggered their automated procedures to refuse further connections, which of course impacted our nightly backups. I still think cloud backup is a good idea.
>
> -Nathan