Description of problem: We need to change the behavior of the VM when it encounters an IO error to report. Currently, the default and only behavior is to pause the VM. Version-Release number of selected component (if applicable): Version-Release number of selected component (if applicable): oc version Client Version: 4.14.0-rc.2 Kustomize Version: v5.0.1 Server Version: 4.14.0-rc.2 Kubernetes Version: v1.27.6+1648878 oc get csv -n openshift-cnv NAME DISPLAY VERSION REPLACES PHASE kubevirt-hyperconverged-operator.v4.14.0 OpenShift Virtualization 4.14.0 kubevirt-hyperconverged-operator.v4.13.3 Succeeded openshift-pipelines-operator-rh.v1.11.0 Red Hat OpenShift Pipelines 1.11.0 Succeeded How reproducible: 100% Steps to Reproduce: 1. Map an iscsi LUN from the Netapp Storage Provider to 3 worker nodes 2. Create a PV referencing the iscsi LUN and detailing the 3 worker nodes 3. Create a shared PVC based on the iscsi LUN with volumeMode block and accessMode ReadWriteMany 4. Create 3 Windows12 R2 vms each with their own OS disk and referencing the 2nd shared iscsi LUN based PVC 5. Install the virtio-win-guest-tools to update the drivers on each of the VMs 6. Via Disk Manager access the iscsi LUN, create a partition on all available space and install NTFS 7. Install Windows Shared Cluster software on each of the 3 Windows VMS 8. Install and configure Active Directory/DNS Domain Controller on one of the VMs 9. Run the Failover Cluster - Validate Configuration tool and select to run only the Storage - ISCSI Reservation Test 10. During the scsi3 persistent reservation REGISTER AND IGNORE EXISTING the node/vm accessed crashes with the error: "Failure issuing call to Persistent REGISTER AND IGNORE EXISTENT on Test Disk 0 from node xxxxx.." *** Directly connecting 3 stand alone(Non Openshift Virtualization) bare metal machines installed with Windows to the same Netapp iscsi LUN and running the same failover validation tool with the scsi3 reservation tests works fine *** This happens on multiple clusters using a different NetApp. And that it has also been tried with a RHEL 9 system presenting the iSCSI targets. Actual results: The scsi3 persistent reservation REGISTER AND IGNORE EXISTING test fails and crashes the target Windows node Expected results: The scsi3 persistent reservation REGISTER AND IGNORE EXISTING test should pass not crash the target Windows node Additional info:
Published backport: https://github.com/kubevirt/kubevirt/pull/10730
Tested with code: oc get csv -n openshift-cnv NAME DISPLAY VERSION REPLACES PHASE kubevirt-hyperconverged-operator.v4.14.1 OpenShift Virtualization 4.14.1 kubevirt-hyperconverged-operator.v4.14.0 Succeeded openshift-pipelines-operator-rh.v1.11.0 Red Hat OpenShift Pipelines 1.11.0 Succeeded [cloud-user@ocp-psi-executor ~]$ oc version Client Version: 4.14.0-rc.2 Kustomize Version: v5.0.1 Server Version: 4.14.0-rc.2 Kubernetes Version: v1.27.6+1648878 The PR https://github.com/kubevirt/kubevirt/pull/10730 fixed this issue and the validation test passes now Moving to VERIFIED!
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Important: OpenShift Virtualization 4.14.1 security and bug fix update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2023:7704