Uploaded image for project: 'Openshift sandboxed containers'
  1. Openshift sandboxed containers
  2. KATA-2133

Errors logged in journal on pod deletion

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Medium Medium
    • None
    • OCP 4.13
    • kata-containers
    • False
    • None
    • False
    • Hide
      QEMU was detecting the disappearance of virtiofsd and was abusively logging errors in the system journal each time a Kata workload was being deleted.

      With this release, the Kata runtime now kills QEMU before tearing down virtiofsd.

      This fix is available for OCP 4.13 and OCP 4.14 only.
      Show
      QEMU was detecting the disappearance of virtiofsd and was abusively logging errors in the system journal each time a Kata workload was being deleted. With this release, the Kata runtime now kills QEMU before tearing down virtiofsd. This fix is available for OCP 4.13 and OCP 4.14 only.
    • Bug Fix
    • Done
    • Kata Sprint #235, Kata Sprint #236
    • 0
    • 0

      Description

      Kata always log some errors on pod deletion.

      Steps to reproduce

      1. Deploy kata
      2. Start a pod using the kata runtime class
      3. Stop the pod

      Expected result

      No errors.

      Actual result

      The following lines land in the system journal :

      level=error msg="Failed to read guest console logs" console-protocol=unix console-url=/run/vc/vm/058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514/console.sock error="read unix @->/run/vc/vm/058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514/console.sock: use of closed network connection" name=containerd-shim-v2 pid=181775 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers subsystem=sandbox
level=error msg="qemu-kvm: Failed to write msg. Wrote -1 instead of 20." name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
level=error msg="qemu-kvm: Failed to set msg fds." name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
level=error msg="qemu-kvm: vhost VQ 0 ring restore failed: -22: Invalid argument (22)" name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
level=error msg="qemu-kvm: Failed to set msg fds." name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
level=error msg="qemu-kvm: vhost VQ 1 ring restore failed: -22: Invalid argument (22)" name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
evel=error msg="qemu-kvm: Failed to set msg fds." name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
level=error msg="qemu-kvm: vhost_set_vring_call failed 22" name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
level=error msg="qemu-kvm: Failed to set msg fds." name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu
level=error msg="qemu-kvm: vhost_set_vring_call failed 22" name=containerd-shim-v2 pid=181775 qemuPid=181784 sandbox=058f54c49d389618b0db45a838d9729f19b0681b055eb25fef160736c8928514 source=virtcontainers/hypervisor subsystem=qemu

      Impact

      Apart from saturating the system journal when running a lot of short lived pods, I don't expect much troubles. These errors should be investigated though and if they are not relevant, they should not go to the journal as this gives a bad feeling to the customer.

      Env

      OCP 4.13.0-0.nightly-2023-04-06-060829
      kata-containers-3.0.2-4.el9.x86_64

      Additional helpful info

      <logs, screenshot, doc links, etc.>

            beraldoleal Beraldo Leal
            rhgkurz Greg Kurz
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated:
              Resolved: