Hi,
Sometimes I really wish there would be some user-customizable buttons in the webUI. I do a lot of scripting to automate my PVE server but it is quite annoying to run all that scripts though the console. It would be so much more comfortable, if I could run these scripts by pressing a button in the webUI.
We already got hook scripts that can be placed in the snippets storage, so PVE tasks can make use of hooks for automation. My idea would be that we could place scripts in the snippets storage and call them something like "userscript_SomeName.bash" or "userscript_SomeOtherName.sh". PVE could then scan the snippet storage for files starting with "userscript_" to see if there are userscripts present and if they are, PVE could show some buttons in the webUI to make those scripts executable.
The webUI header got a lot of empty space. Would be really nice if they could show up there like this:
I know that this wouldn't be that easy to implement, as you probably also would need to check permissions, add new privileges so it could be limited which users are allowed to run which scripts and so on. Maybe parsing of scripts for metadata so the script could specify a name that should be shown in the webUI, needed privilege and similar. Maybe different script states or scripts that allow arguments so running script for a node would pass the node name to the script and so on...
But I would really like to see that. I got so much tasks I need to run scripts for:
Does such a feature request already exist in the bugtracker or should I create a new one?
Someone else finds this useful or am I the only person here that heavily cusatomizes its PVE nodes?
Sometimes I really wish there would be some user-customizable buttons in the webUI. I do a lot of scripting to automate my PVE server but it is quite annoying to run all that scripts though the console. It would be so much more comfortable, if I could run these scripts by pressing a button in the webUI.
We already got hook scripts that can be placed in the snippets storage, so PVE tasks can make use of hooks for automation. My idea would be that we could place scripts in the snippets storage and call them something like "userscript_SomeName.bash" or "userscript_SomeOtherName.sh". PVE could then scan the snippet storage for files starting with "userscript_" to see if there are userscripts present and if they are, PVE could show some buttons in the webUI to make those scripts executable.
The webUI header got a lot of empty space. Would be really nice if they could show up there like this:
I know that this wouldn't be that easy to implement, as you probably also would need to check permissions, add new privileges so it could be limited which users are allowed to run which scripts and so on. Maybe parsing of scripts for metadata so the script could specify a name that should be shown in the webUI, needed privilege and similar. Maybe different script states or scripts that allow arguments so running script for a node would pass the node name to the script and so on...
But I would really like to see that. I got so much tasks I need to run scripts for:
- mounting filesystems
- unlocking encrypted pools/partitions
- bulk starting/stopping guests in an specific order
- disabling/enabling multiple storages at once
- booting/stopping other servers though IPMI
- unlocking encryption of remote server
- bulk manual backups
- switching between performance modes (underclocking CPU and changing CPU governor to save electricity when no running VMs actually needs that much power)
- ...
Does such a feature request already exist in the bugtracker or should I create a new one?
Someone else finds this useful or am I the only person here that heavily cusatomizes its PVE nodes?
Last edited: