Hide Forgot
Description of problem: After deployment of RHOS with sriov, the number of active VFs are configured in “/sys/class/net/em2/device/sriov_numvfs”. I dont find persistent entry in any configuration file on the node. If we reboot the node, changes will go away. Also I don’t see iommu enabled on the compute node. So required parameters are: intel_iommu=on <== for pci passthrough igb.max_vfs=7 <== no of vfs to be enabled on intel nic Workaround: `grubby --update-kernel=ALL --args="intel_iommu=on igb.max_vfs=7` Version-Release number of selected component (if applicable): RHOS10 How reproducible: 100% Steps to Reproduce: 1.deploy RHOS10 with sriov templates 2. 3. Actual results: no persistent entry for required boot parameters to configure SRIOV using director Expected results: There should be persistent entry for required parameters on compute nodes. Additional info:
Steps to Reproduce: 1.deploy RHOS10 with sriov templates 2.confirm the required parameters are there: "/sys/class/net/em2/device/sriov_numvfs". 3. "intel_iommu=on" will not be there in kernel boot parameter list(/proc/cmdline) <== as "intel_iommu=on" is absent, you will not be able to spawn instance with sriov port. 4.reboot compute node. 5. "/sys/class/net/em2/device/sriov_numvfs" will be set to 0
(In reply to Pratik Pravin Bandarkar from comment #1) > Steps to Reproduce: > 1.deploy RHOS10 with sriov templates > 2.confirm the required parameters are there: > "/sys/class/net/em2/device/sriov_numvfs". > 3. "intel_iommu=on" will not be there in kernel boot parameter > list(/proc/cmdline) <== as "intel_iommu=on" is absent, you will not be able > to spawn instance with sriov port. > 4.reboot compute node. > 5. "/sys/class/net/em2/device/sriov_numvfs" will be set to 0 for iommu: ---------- Can you provide the list of templates and the parameters that you have used? You should be using first-boot [1] script along with "ComputeKernelArgs" parameter to provide the iommu configuration. for sriov_numvfs: ---------------- can you provide below files from the sriov compute node? /etc/sysconfig/allocate_vfs /sbin/ifup-local /etc/udev/rules.d/70-tripleo-reset-sriov.rules [1] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/10/html-single/network_functions_virtualization_configuration_guide/#ap-sriov-first-boot
Closing this bug as there has been no response to the request for more information. From reviewing the first boot templates, I believe it should persist the given ComputeKernelArgs in /etc/default/grub. If there is indeed an issue with this then please re-open and include the requested info.