Hi @Maysam,
That message is a Dell PCIe bus fatal error (PCI13xx family in the iDRAC Lifecycle Log), not a Check Point-specific failure.
The reason it works on R80.30 and crashes on R81.20 is almost certainly the Gaia kernel change. R80.30 ships with kernel 2.6.18 by default (there is a separate 3.10-kernel build for specific hardware), while R81.20 is based on RHEL 7.9 with kernel 3.10.0-1160. That jump brought updated NIC drivers (igb/ ixgbe/ i40e / mlx_core) and different PCIe AER, MSI-X and ASPM behaviour, which can expose a bus fault the older driver never triggered.
First question: which quad-port NIC is it exactly (Intel I350-T4, X520, X710, Broadcom 5719, 57414, QLogic)? The chipset determines the driver and whether it is certified. Please cross-check it against the Open Servers tab of the Hardware Compatibility List and sk168335 (Known Limitations for Open Servers and Virtual Machines). If the card is not listed for R81.20, that is likely your answer.
Steps I would take:
1. Update the NIC firmware itself
BIOS 2.19 and iDRAC are not enough; the NIC has its own firmware payload. Flash it via Lifecycle Controller or the Dell bootable repository ISO for the R730.
2. Identify the device behind Bus 9
Export the iDRAC Lifecycle Log to confirm which physical slot maps to bus 9, and if you can reach Expert mode, run:
lspci -nn | grep -i ethernet
dmesg | grep -i -e aer -e mps -e "bus 09"
The vendor device ID from lspci -nn is what really matters here.
3. BIOS power settings
F2 → System BIOS → System Profile Settings → System Profile = Performance. This disables C-States, C1E and aggressive PCIe power management.
4. Disable the Slot Option ROM
System BIOS → Integrated Devices (or Slot Disablement) → set Option ROM = Disabled for the slot holding the quad card, assuming you are not booting via PXE/iSCSI from it.
5. Move the card to a different slot/riser
A different riser means a different root port and different lane negotiation. On the R730, the riser 2/3 slots require CPU2 to be populated.
6. Kernel boot parameters
Edit the kernel line in GRUB and test:
pci=pcie_bus_safe
pcie_aspm=off pci=noaer
pci=pcie_bus_safe is worth trying first a Max Payload Size mismatch across the PCIe fabric is a well-known cause of exactly this "Bus Fatal Error / power cycle system" symptom on PowerEdge hardware. If any of these let it boot, you have confirmed it is a driver / PCIe negotiation issue rather than faulty hardware.
You mentioned seeing ports 0-7 during the R81.20 install but not being able to configure the new ones that fits a driver that loads and then faults.
If the card is on the HCL and none of the above helps, I would open a TAC case with the iDRAC Lifecycle Log export, the lspci -nn output and a cpinfo.
Hope this helps!