Announcement

Collapse
No announcement yet.

oVirt 4.2 Brings A New Admin Panel, NVIDIA vGPU Support

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • oVirt 4.2 Brings A New Admin Panel, NVIDIA vGPU Support

    Phoronix: oVirt 4.2 Brings A New Admin Panel, NVIDIA vGPU Support

    Red Hat developers working on the oVirt virtualization management platform have announced the release of oVirt 4.2...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Am I a bad person if I want to try oVirt+Kubernetes+Cockpit in place of Proxmox one of these days?

    Now I just need to find spare hardware...

    Comment


    • #3
      Originally posted by Niarbeht View Post
      Am I a bad person if I want to try oVirt+Kubernetes+Cockpit in place of Proxmox one of these days?

      Now I just need to find spare hardware...
      Can't say if you're a bad person or not, but I can say that Proxmox is 100x easier to set up and use. oVirt is extremely complicated. We use Red Hat Virtualization at work (the Red Hat branded commercial version of oVirt) and there is no way we'd be able to run this thing without the Red Hat documentation (not available to the public) and the ability to open support cases when we run into trouble. And we've opened a lot of support cases over the past year.

      Basically, the only reason to choose RHV/oVirt is if you're already a pure Red Hat shop, so that you have a single source of support for the entire software ecosystem. If you're a home hobbyist, Proxmox is the way to go.

      That said, we are eagerly anticipating the RHV 4.2 release because of the OVS networking.

      Comment


      • #4
        Originally posted by Niarbeht View Post
        Am I a bad person if I want to try oVirt+Kubernetes+Cockpit in place of Proxmox one of these days?

        Now I just need to find spare hardware...
        I'm curious, would something like CoreOS/Fedora Atomic inside of Proxmox not also work? I have some virtual machines that does not need/want Kubernetes which is why I'm thinking of reusing existing hardware. The services that I run are not very demanding, however the way people interact/manage services are very demanding which sucks. Sorry for rambling

        Comment


        • #5
          At this point I would advice anyone looking at this not to try it. Really, I operate oVirt (without the RHEV doc) since 2 years and it is a nightmare.

          The thing never ever rebooted correctly after a cluster power cycle.

          Never ever updated itself without bringing down the whole datacenter.

          Breaks on regular CentOS system updates (so much I stopped doing them)

          If so full of race conditions that often it will magically wake up from the dead from no reasons after refusing to start for an hour.

          Their guess additions for monitoring VMs are in such bad shape the distribution package (if any) crash with obvious python errors.

          4.2 removes the ability to use HTMLv5 SPICE viewer for no reason while warned that this is absolutely critical to use RDP over firewalls.

          The Android app is useless and stops connecting after a couple hours and require to be re-installed (with cache flushed)

          They seem clueless about real world use case, especially media ones and claim "nobody do that" in bug reports while supporting imaginary use cases.

          I knowledge they might not care about < 10 hosts setup without contract support, but for things like continuous integration clusters where there is > 100 VM templates and > 500 container running at once, libvirtd isn't enough. I hoped for a while that new versions would make it more usable, but beside the hosted-engine VM setup that solved a bunch of headbanging, it's really not.

          Comment


          • #6
            how come nobody ever talks about projects like open nebula ?

            Comment


            • #7
              Originally posted by Elv13 View Post
              At this point I would advice anyone looking at this not to try it. Really, I operate oVirt (without the RHEV doc) since 2 years and it is a nightmare.

              The thing never ever rebooted correctly after a cluster power cycle.

              Never ever updated itself without bringing down the whole datacenter.

              Breaks on regular CentOS system updates (so much I stopped doing them)

              If so full of race conditions that often it will magically wake up from the dead from no reasons after refusing to start for an hour.

              Their guess additions for monitoring VMs are in such bad shape the distribution package (if any) crash with obvious python errors.

              4.2 removes the ability to use HTMLv5 SPICE viewer for no reason while warned that this is absolutely critical to use RDP over firewalls.

              The Android app is useless and stops connecting after a couple hours and require to be re-installed (with cache flushed)

              They seem clueless about real world use case, especially media ones and claim "nobody do that" in bug reports while supporting imaginary use cases.

              I knowledge they might not care about < 10 hosts setup without contract support, but for things like continuous integration clusters where there is > 100 VM templates and > 500 container running at once, libvirtd isn't enough. I hoped for a while that new versions would make it more usable, but beside the hosted-engine VM setup that solved a bunch of headbanging, it's really not.
              YMMV.
              We've been using oVirt in production over CentOS for ~2 years now with ~ 100 active VMs and self hosted engine.
              It may be my elite IT manager, but as far as I remember most of the post-install issues we had were with the 3-node GlusterFS and less with oVirt itself. (Though, the installation itself was indeed a nightmare)

              That said, we never field CentOS or oVirt updates on the production server without testing it on our test cluster.

              ... Though the lack of HTML viewer is a killer for us.

              - Gilboa

              oVirt-HV1: Intel S2600C0, 2xE5-2658V2, 128GB, 8x2TB, 4x480GB SSD, GTX1080 (to-VM), Dell U3219Q, U2415, U2412M.
              oVirt-HV2: Intel S2400GP2, 2xE5-2448L, 120GB, 8x2TB, 4x480GB SSD, GTX730 (to-VM).
              oVirt-HV3: Gigabyte B85M-HD3, E3-1245V3, 32GB, 4x1TB, 2x480GB SSD, GTX980 (to-VM).
              Devel-2: Asus H110M-K, i5-6500, 16GB, 3x1TB + 128GB-SSD, F33.

              Comment

              Working...
              X