Today I've updated the last of our six ASR1002s to IOS XE 3.10.1S. All other are fine, but this one not so much - since the reload it's been stuck at ~80% CPU load, mainly due to "PPP IP Route" process activity. Compared to other devices with similar load (in terms of number of PPPoE sessions) on this one the "PPP IP Route" process is invoked much more often:
asr1#sh proc cpu | i five|Invok|PPP_IP_R
CPU utilization for five seconds: 88%/6%; one minute: 77%; five minutes: 78%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
188 3311079 447339 7401 30.13% 29.22% 30.05% 0 PPP IP Route
asr1#sh proc mem | i Alloc|PPP_IP_R
PID TTY Allocated Freed Holding Getbufs Retbufs Process
188 0 450586736 244132644 1779752 20038777 20038777 PPP IP Route
(all these in about three hours of uptime, with ~3400 PPPoE sessions currently active; the "Holding" memory counter is quite stable, all others are permanently increasing)
Everything else (evolution of interface traffic rates or number of PPPoE sessions, RADIUS activity) looks normal for the period of time since the reload.
I'm a bit reluctant to reloading it again at this time of day - is there anything I can use to further debug the issue?
Thank you.
