Hang on a sec...
Were there no system backups or data backups? Not even taped ones?
NOT a service that should be recommended, then. The idea is that cloud storage should be backed up on a parallel server, in case of outages, surely?
Workers relying on Atlassian's cloudy team-tracking software have reverted to whiteboards and spreadsheets after a service outage made key project data vanish. A Reg reader contacted us to say firms using Atlassian’s JIRA service had lost their all data, and said Atlassian had been unable to recover it despite a week of trying …
Agreed, multiple storage failures, your cloud provider exploding because of the Large Hadron Collider or anything else are not an excuse for LOSING data.
Possibly making that data temporarily unavailable but the words "data recovery specialists" really translate as "whoops, we have no backup".
Even if the whole rack exploded - where's your backup from last week, last month, last year?
High availability and redundancy are no substitute for backups. I hope all the listed companies just cancel their contract and ask for a complete refund.
Actually, it's not vapourware and it's not modern, the marketing drones just have a "modern" word for it. Bottom line is it is still outsourced storage technology. Maybe it makes sense, maybe it doesn't. That will depend on your contract (does it require disaster recovery abilities from the vendor), its precise details (does the vendor do a yearly disaster recovery test), and the how well the outsourcing contract is monitored (are you allowed onsite during the disaster recovery test). For it to work, that's all a heck of a lot of work, and the cost savings over do it yourself are likely to be marginal, not huge. If there is a huge price savings, somebody probably left out an important component.
I love the one where the cloud provider is on stage telling the story of the history of his 6 month old cloud business and about how they started the business in a garage. Oh wait, did they tell customers their apps and data were in a garage? Or, a secure, N+1 redundant data center facility?
Here come the regulators ...
That disk failure can cause such a problem. Do you not make sure that your data is replicated across multiple storage devices/sites/tapes/whatever? You might call it backup, archiving or simply 'making a copy' but don't these people think in terms of full redundancy of complete datasets?
Yes, I know the bean counters don't like it, but isn't being laughed out of the business worse than spending the money to ensure you have a stable platform on which to offer your services?
I often think in cases like this it must be much more than a failed disk or 3.
Basically - and this will burn cloud users more and more - a chain of clouds (you heard it used here first) is going to cause this to happen more and more. The SaaS is only as reliable as the weakest cloud in the chain ;) - and as the end-user you have no idea where or control over that weakspot.
Welcome to XaaS - and you yhought your internal IT was bad ...
... 'complicated server farm that cost a lot of money and we hope is pretty good'.
I know there isn't a standard definition of 'cloud', but at the very least it should mean that someone can start yanking cables without risk of significant data loss. (Ideally, this should extend to availability.)
Doesn't sound like they have a 'cloud', more like a 'crappy SAN and a load of servers'.
Biting the hand that feeds IT © 1998–2022