Yes, that makes more sense indeed. Maybe a simple "not-a-perfect-translation"-bug.In the CLI [1], the "add" command still confuses me, I believe it should be called "join".
Yes, that makes more sense indeed. Maybe a simple "not-a-perfect-translation"-bug.In the CLI [1], the "add" command still confuses me, I believe it should be called "join".
Is there a BZ requesting that addition to the GUI?With a round-robin served DNS A record, you can see e.g. "pve.internal" and it could be any of the dozen nodes there.
Is there a BZ requesting that addition to the GUI?
or provide patchesBut I absolutely encourage everyone who wants to improve the state of things to contribute to bugreports
I just saw this now ... round robin will not work that well, the cookies will not match and you'll probably have to relogin and ask yourself, why this is the case. Best to use a reverse proxy with sticky sessions, so that the reverse proxy will now where you've been routed before. There are a couple of threads discussing this.With a round-robin served DNS A record, you can see e.g. "pve.internal" and it could be any of the dozen nodes there. Similar with reverse proxy. Similar with every post on this forum with a "screenshot" that does not include address line.
or provide patches
I just saw this now ... round robin will not work that well, the cookies will not match and you'll probably have to relogin and ask yourself, why this is the case. Best to use a reverse proxy with sticky sessions, so that the reverse proxy will now where you've been routed before. There are a couple of threads discussing this.
Um, wasn't adding something to show which node the browser is connecting to your idea @esi_y?I think the OP was talking of GUI alone (adding nodename).
Um, wasn't adding something to show which node the browser is connecting to your idea @esi_y?
GUI does not provide any warning on lost quorum situation, but severely limits e.g. access to shell, also proxying connections to other nodes without clear error message (those are generic failures of termproxy).
The Summary interface change isn't removing any information. Instead it's condensing part of the layout (the text heavy bit), adding a "System Health Status" piece, and moving the single column of vertical charts to be two columns instead. The two columns of charts fits into the default sized view a bit easier, to show everything all at once:This might be good to reduce "function overload" for newcomers and non-technical staff that audits/uses only a small set of Proxmox VE's functionality, but for all power users it would be a big disruption of their workflow experience.
One IMO more realistic idea to solve this without interfering with power/long-term users is adding a new "simple view" which reduces what's visible by default on the user interface to the minimum, like basic load and health overviews of nodes, guests and tasks (like backup) and guest management (start/stop, migrate).
The update part is easy and already there ... everything else is not determinable from the hypervisor. You need integration into the hardware (this is done on other hypervisors with software from the hardware vendor, I don't see that happending for PVE), VM and container health ... how is that even defined??? As said numerous times before, get a decend monitoring solution, PVE will not and cannot provide this."System Health Status"
This is a typical homelab request and PVE is not about homelabbing.
Not sure how accurate that is. The guy making the Youtube video with the screenshots and stuff is coming at this from a decent VMware vSphere/ESX background.This is a typical homelab request
Heh, my PG efforts in recent days are mostly here as the PostgreSQL project doesn't yet really have a handle on Docker stuff:former contributor to PostgreSQL
Somewhat related to this something I wanted for a while, copying over a vimwiki entry from over a year ago:
- System Health status pane in the Node "Summary" view (and prob similar in the VM/container "Summary" view) → TBD
The submenu won't change anytime soon, when users are not happy about that it often correlates with they're having used other hypervisors for a while and now checking out PVE; but we're not a copy and do not want to be one. Our reasoning for moving to this layout is, among other things, better space usage overall, especially with the more navigation entries we gain and also less mouse movement required after selecting the datacenter, a node or a virtual guest and then choosing a sub-panel.Move the node vertical submenu to be horizontal, then reorganise the Summary view to be more compact → TBD (maybe)
- Not sure if I'll make a BZ of this one. @esi_y already pointed out that this is how the Proxmox webUI used to be back in 2015.