Describe the bug
On oVirt Node 4.5.4 el8 installed on a Dell R710 on a separate drive from the PERC controller, disabled, and on the PERC controller with elrepo driver, after hosted engine is deployed. Upon starting a VM with network attached to ovirtmgmt bridge or restarting a VM the network completely locks up to the point iDRAC IP isn't pingable. Error produced below over and over until it restarts.
dracut-initqueue[xxxx]: cat: /sys/class/net/bonding_masters/ifindex: Not a directory
To reproduce
Install on Dell R710.
Deploy Hosted-Engine
Try to install Ubuntu or Windows VM and intermittently during startup/reboot/just running the network will lockup.
Changing the profile to empty allows it to work. I added it back to the Ubuntu VM while running and it caused it to immediately crash at reboot.
Expected behavior
The expected behaviour should be that the networking should not cause the entire R710 to be completely inaccessible with iDRAC even unable to be pinged or accessed.
Screenshots
Screenshot of the error from the display.
Additional context
Tried with builtin network card "Broadcom Inc. and subsidiaries NetXtreme II BCM5709 Gigabit Ethernet (PowerEdge R710 BCM5709 Gigabit Ethernet) bnx2" as the bridged ovirtmgmt and tried a whole new install with the Broadcom disabled in oVirt Node cockpit and used the PCI network card "Intel Corporation 82576 Gigabit Network Connection (Gigabit ET Dual Port Server Adapter) igb', which has the same issue.
I also tried creating a separate network bridge tied to the other Intel port on the Dual Port card and that causes the same error.
The only way it boots without some of the above issues is by setting the network to empty or being just lucky at that boot time as it doesn't happen every time, but happens often enough I cannot rely on the VMs to work.
Describe the bug On oVirt Node 4.5.4 el8 installed on a Dell R710 on a separate drive from the PERC controller, disabled, and on the PERC controller with elrepo driver, after hosted engine is deployed. Upon starting a VM with network attached to ovirtmgmt bridge or restarting a VM the network completely locks up to the point iDRAC IP isn't pingable. Error produced below over and over until it restarts.
dracut-initqueue[xxxx]: cat: /sys/class/net/bonding_masters/ifindex: Not a directory
To reproduce Install on Dell R710. Deploy Hosted-Engine Try to install Ubuntu or Windows VM and intermittently during startup/reboot/just running the network will lockup. Changing the profile to empty allows it to work. I added it back to the Ubuntu VM while running and it caused it to immediately crash at reboot.
Expected behavior The expected behaviour should be that the networking should not cause the entire R710 to be completely inaccessible with iDRAC even unable to be pinged or accessed.
Screenshots Screenshot of the error from the display.![oVirt Crash](https://user-images.githubusercontent.com/64096865/224414315-72d1ead7-1e7c-4dcc-b573-dab800c70abd.jpeg)
Additional context Tried with builtin network card "Broadcom Inc. and subsidiaries NetXtreme II BCM5709 Gigabit Ethernet (PowerEdge R710 BCM5709 Gigabit Ethernet) bnx2" as the bridged ovirtmgmt and tried a whole new install with the Broadcom disabled in oVirt Node cockpit and used the PCI network card "Intel Corporation 82576 Gigabit Network Connection (Gigabit ET Dual Port Server Adapter) igb', which has the same issue.
I also tried creating a separate network bridge tied to the other Intel port on the Dual Port card and that causes the same error.
The only way it boots without some of the above issues is by setting the network to empty or being just lucky at that boot time as it doesn't happen every time, but happens often enough I cannot rely on the VMs to work.