Last week mysql-related stuff made me to reconfigure gandi-vps server and delete old instance after that. UI dialog asked me if I want to keep system partition but silently deleted custom created one with my Stanford-related stuff.

It was restored aster request to the support for free (it seems that standard rate is euro 70) so the story is was over. The one thing that has been slightly bothering me that statement "you have to detach the disks before deleting a VPS" is true but wrong.

Worst of all the answer is known for 25 years at least. So I wrote a feedback to gandi customer service but a thought about how many things we do wrong even if we already knew the answer is really depressing.

My feedback: This comment is related to case support-en #3584703 there you can find all background. Case is resolved, but I hope that our collaboration is not over and as I'm thinking to place some severs for our current projects on your cloud hosting, I feel like I should try to give a more detailed feedback.

It's now 25 years since Ivar Jacobson firstly developed use cases. Ok, count 5 years to development and discussion, 20 years.

The way of thinking behind use cases is about achieving user goals step-by-step. It's the easy part. Difficult one is that on EVERY STEP of every use case scenario which is user-system dialogue, SYSTEM is responsible for protecting interests of all stackeholders on every step.

Bearing that in mind you wouldn't have the situation "ok, pleb, it's your job to remember about detaching a partition to keep it safe". Protect stackholder interests on every step. A guarantee or assurance - user doesn't lose his data until he *explicitely* asks to do so. If you need to recursively delete all server-attached resources - ask about every one, or at least about important ones.

I don't know why I'm trying to tell all this stuff, but I feel it's important. It's not about make two/ten checkboxes or not, it's about right and wrong.