• 0 Posts
  • 54 Comments
Joined 1 year ago
cake
Cake day: June 11th, 2023

help-circle
  • dai@lemmy.worldtolinuxmemes@lemmy.world32GB Ram and Linux
    link
    fedilink
    arrow-up
    2
    ·
    edit-2
    5 hours ago

    Is Orca that resource intensive? I’m running it in a container with KasmVNC and have never really checked out the resource usage. Admittedly it’s on one of my local servers in another room. I guess it’s how large your projects are too.

    Edit: maybe it’s just my small projects









  • The volumes of cash that Microsoft throw at retailers (custom builders / big box) is astronomical. Worked for a relatively small retailer with some international buying power. EOFY “MDF” from Microsoft was an absurd figure.

    Our builders would belt out 3 - 6 machines per day, depending on complexity of the custom build, the pre-built machines were in the 6+ per day range.

    Considering the vast majority of those machines were running windows (some sold without an os), from a quick estimate after too many beers we were out of pocket 10% at most of the bulk buy price for licence keys after our “market development funds” came through.

    It’s fucking crook.



  • I’ve not tested the method linked but yeah it would seem like it’s possible via this method.

    My lone VM doesn’t need a connection to those drives so I’ve not had a point to.

    You could probably run OMV in an LXC and skip the overheads of a VM entirely. LXC are containers, you can just edit the config files for the containers on the host Proxmox and pass drives right through.

    Your containers will need to be privileged, you can also clone a container and make it privileged if you have something setup already as unprivileged!


  • Yeah there is a workaround for using bind-mounts in Proxmox VMs: https://gist.github.com/Drallas/7e4a6f6f36610eeb0bbb5d011c8ca0be

    If you wanted, and your drives are mounted to the Proxmox host (and not to a VM), try an LXC for the services you are running, if you require a VM then the above workaround would be recommended after backing up your data.

    I’ve got my drives mounted in a container as shown here:

    Basicboi config, but it’s quick and gets the job done.

    I’d originally gone down the same route as you had with VMs and shares, but it’s was all too much after a while.

    I’m almost rid of all my VMs, home assistant is currently the last package I’ve yet to migrate. Migrated my frigate to a docker container under nixos, tailscale exit node under nixos too while the vast majority of other packages are already in LXC.


  • Ahh the shouting from the rooftops wasn’t aimed at you, but the general group of people in similar threads. Lots of people shill tailscale as it’s a great service for nothing but there needs to be a level of caution with it too.

    I’m quite new to the self hosting game myself, but services like tailscale which have so much insight / reach into our networks are something that in the end, should be self hosted.

    If your using SMB locally between VMs maybe try proxmox, https//clan.lol/ is something I’m looking into to replace Proxmox down the line. I share bind-mounts currently between multiple LXC from the host Proxmox OS, configuration is pretty easy, and there are lots of tutorials online for getting started.



  • Tailscale is great, but it’s not something that should be shouted from the rooftops.

    I use tailscale with nginx / pihole for my home services BUT there will be a point where the “free” tier of their service will be gutted / monetized and your once so free, private service won’t be so free.

    Tailscale are SAAS (software as a service), once their venture capital funds look like their running dry, the funds will be coming from your data, limiting the service with a push to subscription models or a combination.

    Nebula is one such alternative, headscale is another. Wire guard (which tailscale is based on) again is another.




  • Ahh man gotta love NVIDIA. Most of my machines have an NVIDIA GPU, but I’ve had only a few minor issues along the way. Mostly from me not reading things correctly.

    Saw issues with flickering electron applications, sleep somehow running the GPU until my battery was drained, hyprland just saying not today and random crashes here and there.

    Systems are pretty stable now, laptop runs fine in hybrid mode (AMD / NVIDIA) and I removed almost all electron applications.

    I’ve found if I can’t figure something out I’ll start a new module for another package. But I guess if it’s something “mission critical” in your case a GPU then it’s pretty hard to do much else.



  • Had my boss trying to grab a pdf (crosswords, colouring pages, printed for kids in a pub) while using Chrome without any adblock extensions.

    The volume of ads, trick links, and shite on that one website in particular was outstanding. She asked me if a link was OK to click. Promptly pointed out she should use Firefox (which has unlock and other extensions added) instead of chrome as the link she had clicked was for some sketchy software and not a crossword.

    I can’t imagine the internet without ad blockers. Ublock is a great addition, removing elements from pages is a huge advantage. So many sites sling rubbish wherever they can.