Mixing up consoles happens to all of us - but not always so far away.
Then I learned to install
molly-guard on all of my systems. This does not help for the Gui though...
Disable server shutdown/restart from the web interface
- Thread starter arnolem
- Start date
Then I learned to install
As for the web interface, if the Proxmox Team doesn't want to introduce another confirmation step, what about locating the confirmation window differently when the confirmation is regarding a node as when the confirmation is regarding an LXC or a VM? E.g. for the node in the top right while the rest remains in the center or vice versa? This would introduce a visual differentiation for the users while remaining very little change for the proxmox team...?
It just happened to me.
The real problem is that the design of shutdown button is the same for "node" or "vm".
Idea : If PVE team can change the picto fa-power-off to fa-plug for halt one node, it would be less confusing
The real problem is that the design of shutdown button is the same for "node" or "vm".
Idea : If PVE team can change the picto fa-power-off to fa-plug for halt one node, it would be less confusing
Last edited:
Good idea - visually different is the key here.
You may open a "Feature Request / Bug"-report over at https://bugzilla.proxmox.com to persist this idea
that sounds like a sensible thing to do and isn't hard to change. Therefore I send a patch for this: https://lore.proxmox.com/pve-devel/20250507124434.1730046-1-a.lauterer@proxmox.com/T/#uIdea : If PVE team can change the picto fa-power-off to fa-plug for halt one node, it would be less confusing
We just shut down our node after clicking on the VM line in Search. Of course we didn't read the popup. Our fault - but those node-level controls really should look more ominous. E.g. have them on separate page, like some kind of IPMI. Or at least have "Shutdown Node", or even name in the button directly. That would help already.
Came here report the exact same thing.
I used webUI to turn off a series of VMs, I was tired, misclicked and didn't read the pop-up.
We were testing, but in prod, it could have stalled the prod and endangered equipment and possibly on site employee health. I have to find ways to mitigate (yes, rbac helps, but I agree with the opinion higher in the thread, that it is a work-around).
Thank you @aaron for the patch. If I understand correctly, it hasn't yet been merged.
I simply want to be able to disable the button and will be fully content with turning the node off in terminal only.
For now, will implement rbac on all our admin accounts as there is simply no other way to mitigate and will keep a special account for node shutdown.
I used webUI to turn off a series of VMs, I was tired, misclicked and didn't read the pop-up.
We were testing, but in prod, it could have stalled the prod and endangered equipment and possibly on site employee health. I have to find ways to mitigate (yes, rbac helps, but I agree with the opinion higher in the thread, that it is a work-around).
Thank you @aaron for the patch. If I understand correctly, it hasn't yet been merged.
I simply want to be able to disable the button and will be fully content with turning the node off in terminal only.
For now, will implement rbac on all our admin accounts as there is simply no other way to mitigate and will keep a special account for node shutdown.
I mean not gonna lie this is exactly the kind of thing that shows that they don't take UX seriously.
The button should be red, it should have a different icon and the resulting popup should be red/yellow or anything indicating that this is a really big deal what you're about to do.
It should not under any circumstances look the same as a container or VM shutdown.
Absolutely insane isn't even close to describing the situation.
The sad part is that you have here 2 pages of comments, and they will all get disregarded as the opinions of "a small number of customers". This arrogance is what kills products. I spent 20 years running a UX firm (before it was even called that), and I have seen first-hand what happens to products when tech leadership ignores direct user feedback. I have a shelf full of failed hardware/software products that we were called in to fix, and many times some of the largest pain points were well-known by the teams, but they couldn't be implemented until an outsider like us told management to do it. Sure, this isn't a huge issue, but it could easily have catastrophic results at the wrong time. Remember the message send to all residents of Hawaii a couple years back? "Incoming nuclear attack -- this is not a test!" (It was a test, in fact, and the whole thing was down to bad UX -- little things matter.)