Rackspace's cloud keeps backup of all the could, it can't be used to recover data of a single client. That backup is kept in the event of emergency if anything drastic happens to Rackspace's infrastructure, that's why they offer optional offsite backup at extra if a customer requires it on individual level
Zhou has said before that they offered Rackspace $10k to help recover deleted serves; Rackspace simply couldn't help. RS would have to nuke all their cloud client's servers and restore them back (which also can be assumed isn't instantaneous process)
It's hard for me to picture a scenario where restoring a backup would "require" them to completely disrupt their clients' servers. Surely they can restore it to other equipment that's not being used? I mean, when I say I can "flip over backwards" and restore my backups for somebody, it doesn't mean I"m going to shut down all my services because I'm pretending that the backup can only be restored to the same hardware it was taken from - it means I'm going to grab some unused machine and restore to that instead.
Even the whole idea that their backups are only useful for restoring the "entire cloud", I just don't buy that. The news recently reported they hit 100k customers - let's say each customer has an average of 30 gigs of data. Do they back up their entire cloud to a single 3-petabyte file and that's why it can only be restored on an all-or-nothing basis? What kind of media do they use to store a file this big?
At least if they have made the claim that they offered $10k and it wasn't accepted, that's better than what I had thought before, which was that no significant efforts were made (I don't follow every post on this topic).
I call bullshit, no half arse company(and mind you rackspace is fairly competent) would reject $10k for such a recovery job which they should have procedures in place to do.
Yeah, none of that adds up at all.. They could restore the VM's if they really wanted too, there is no technical barrier for them to do so that money can't fix, that's for certain. I mean even if they were doing LUN level backup and had to recover 10, 50, 100 customers or something in order to pull the VM's out of the LUN, they could have restored it to some scratch space took what they needed and scrapped the rest.
We use ZFS for all our VM's, if something like that happened on our systems, we mount the earlier snapshot and boot the VM right back up again in a sandbox, take whatever we need, and that's it.
Rackspace probably has similar technologies, worst case scenario they have to restore 100 customers to a sandbox, 10k, 20k, 50k, 100k, any of these amounts would be trivial in this case. I'm sure if you got escalated to the right level of management it would have happened, had they acted QUICKLY and DECISIVELY, which unfortunately, we know is NOT what they do..
Which is why this unfortunately does all look like the long con, none of this shit adds up, its just one amateur hour shit show that just keeps going and going and going..
