XCP-ng
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Groups
    • Register
    • Login

    Could not change "-machine" by 'xe vm-param-set'

    Scheduled Pinned Locked Moved Compute
    6 Posts 3 Posters 549 Views 3 Watching
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • L Offline
      lyan
      last edited by

      With default xcp-ng 8.2, I saw the hvm is created with "qemu-dm-3 -machine pc-0.10,accel=xen,"

      I checked xe vm-param-setthat there is no such option change "-machine", I tried "other-config" got no such key error.

      I am new to xcp-ng, wondering if there is a way to change the machine model for qemu, such as "pc-i440fx-42".

      Thanks.

      1 Reply Last reply Reply Quote 0
      • olivierlambertO Offline
        olivierlambert Vates 🪐 Co-Founder CEO
        last edited by

        Hi,

        You can't change that in XCP-ng. Can you tell us what do you want to achieve?

        L 1 Reply Last reply Reply Quote 0
        • L Offline
          lyan @olivierlambert
          last edited by

          Thanks for the reply, @olivierlambert. I recently hit a bug that Xen could not create msi interrupt for HVM, I noticed that we qemu-dm is using pc-0.1 as the machine model by default, as the starter, I kind of want to change it to i440fx-4.2 which is aligned to the qemu version. I can not change it from the cmd line atm, maybe it is hard-wired in the code?

          T 1 Reply Last reply Reply Quote 0
          • olivierlambertO olivierlambert moved this topic from Xen Orchestra on
          • olivierlambertO Offline
            olivierlambert Vates 🪐 Co-Founder CEO
            last edited by

            I think it's slightly more complicated than that. Adding @ThierryEscande in the loop

            1 Reply Last reply Reply Quote 0
            • T Offline
              ThierryEscande Vates 🪐 XCP-ng Team @lyan
              last edited by

              Hi @lyan,

              This might be related to a known issue about PCI passthrough with nvme devices: the kernel tries to allocate more MSI-X vectors than the guest can handle. You can try to increase the number of guest IRQs with the Xen boot parameter extra_guest_irqs. The default is 64 and you can increase it to 128 with:

              /opt/xensource/libexec/xen-cmdline --set-xen "extra_guest_irqs=128"
              

              A reboot of the host is needed.

              L 1 Reply Last reply Reply Quote 1
              • L Offline
                lyan @ThierryEscande
                last edited by

                Hi, @ThierryEscande, thanks for the input here。

                It is pretty interesting that I kind also located the same place by following the interrupter path.
                https://github.com/liayan/xen-xcp-ng/blob/main/xen/common/domain.c#L250

                It quite fits our case atm that issues happened after hitting the limit of 64. Unfortunately, I can not test it yet because our host is stuck in a weird POST status. Will update here later.

                1 Reply Last reply Reply Quote 1
                • First post
                  Last post