...
IOS-XE router might crash when we receive two(or more) of same(duplicated) fragments on different interfaces. %CPPDRV-3-LOCKDOWN: R0/0: cpp_cp: QFP0.0 CPP Driver LOCKDOWN encountered due to previous fatal error (HW: QFP interrupt).@ %CPPDRV-3-LOCKDOWN: R0/0: fman_fp_image: QFP0.0 CPP Driver LOCKDOWN encountered due to previous fatal error (HW: QFP interrupt).@ %CPPDRV-3-LOCKDOWN: R0/0: cpp_ha: QFP0.0 CPP Driver LOCKDOWN encountered due to previous fatal error (HW: QFP interrupt).@ %IOSXE-1-PLATFORM: R0/0: kernel: QFP0.0: Fatal Fault: HW reported: QFP interrupt@ These syslog messages might be seen before the crash %FRAG-3-REASSEMBLY_ERR: Reassembly/VFR encountered an error: the first fragment's copy of packet state does not exist %FRAG-3-REASSEMBLY_DBG: Reassembly/VFR encountered an error: IPv4 vFR complete with wrong refcnt, actual refcnt 4, expected refcnt 1 %FRAG-3-REASSEMBLY_ERR: Reassembly/VFR encountered an error: the first fragment's copy of packet state does not exist
Virtual fragmentation reassembly (VFR) has been configured. This is a VFR day-1 issue and very corner case. We may hit this issue only when same fragments are received on different interfaces at the same time. And these interfaces must have VFR configured via "ip virtual-reassembly" explicitly or enabled automatically by some features (such as NAT, Cisco IOS XE Firewall, IPSec) on them.
Remove Virtual fragmentation reassembly (VFR) configuration from the interfaces. Starting from 16.9.x and later releases, "Virtual fragmentation reassembly (VFR) is automatically enabled by some features (such as NAT, Cisco IOS XE Firewall, IPSec) ...". It is not required to manually configure the "ip virtual-reassembly" for these features after 16.9.x. Please refer to ?IP Addressing: Fragmentation and Reassembly Configuration Guide? for additional information. https://www.cisco.com/c/en/us/td/docs/ios-xml/ios/ipaddr_frre/configuration/xe-16/frre-xe-16-book/virt-frag-reassembly.html
Kernel crashes might be seen after the CCP ucode crash [463768.148140] QFP0.0: Fatal Fault: HW reported: QFP interrupt [464158.128801] systemd-journald[109]: Failed to send WATCHDOG=1 notification message: Connection refused [464175.912943] Uhhuh. NMI received for unknown reason 3d on CPU 0. [464175.912944] Do you have a strange power saving mode enabled? [464175.912944] Kernel panic - not syncing: NMI: Not continuing