Cisco Prime 3.x fills up /var


Symptom : Prime 3.7 won’t start. Reason : /var is full, as /var/log/wtmp fills up quickly :

sudo -i
ls -l /var/log/wtmp
ade # utmpdump /var/log/wtmp | more
    [5] [06834] [tyS0] [     ] [ttyS0] [ ] [0.0.0.0] [Tue Feb 04 21:30:45 2020 CET]
    [5] [06835] [tyS1] [     ] [ttyS1] [ ] [0.0.0.0] [Tue Feb 04 21:30:45 2020 CET]
    [6] [06834] [tyS0] [LOGIN] [ttyS0] [ ] [0.0.0.0] [Tue Feb 04 21:30:45 2020 CET]
    [6] [06835] [tyS1] [LOGIN] [ttyS1] [ ] [0.0.0.0] [Tue Feb 04 21:30:45 2020 CET]
    [8] [06834] [tyS0] [     ] [ttyS0] [ ] [0.0.0.0] [Tue Feb 04 21:30:55 2020 CET]
    [8] [06835] [tyS1] [     ] [ttyS1] [ ] [0.0.0.0] [Tue Feb 04 21:30:55 2020 CET]

But what fills up so quickly ? The ttyS0 above gives a hint, let’s look at the service : oops, it starts every 10 seconds, fails and retry…

journalctl -u serial-getty@ttyS0.service
    Apr 21 09:46:55 hostname systemd[1]: serial-getty@ttyS0.service holdoff time over, scheduling restart.
    Apr 21 09:46:55 hostname systemd[1]: Stopped Serial Getty on ttyS0.
    Apr 21 09:46:55 hostname systemd[1]: Started Serial Getty on ttyS0.
    Apr 21 09:46:55 hostname agetty[28879]: /dev/ttyS0: not a tty

So, why is my Cisco Prime Appliance failing to start Getty ? Because my ESX VM’s do not have serial ports… I only wonder why Cisco has this service enabled on their VM appliance.

The fix :

sudo systemctl stop serial-getty@ttyS0.service
sudo systemctl disable serial-getty@ttyS0.service
sudo systemctl stop serial-getty@ttyS1.service
sudo systemctl disable serial-getty@ttyS1.service 
sudo -i
cat /dev/null > /var/log/wtmp 

Problem fixed. I will announce it to Cisco, hopefully it will end up in an update. Prime 3.8 runs on top of a newer RedHat version, maybe it’s fixed there.