Hi All,
I am working to bring automation with Ansible into a large corporate network with a lot of Nexus 5/6K many of which are running relatively old software (uptime of 3 years is not uncommon). I have started to use Ansible nxos_facts (v2.9) to gather information from these Nexus switches. I use "network_cli" connection plugin. It has been noticed that from time to time log messages appear in Nexus switches as below.
%SYSMGR-2-LAST_CORE_BASIC_TRACE: : PID 1753 with message python(non-sysmgr) crashed, core will be saved
These messages seem to coincide with Ansible logging in and the suspicion is that Ansible is triggering these crashes. The commands that be seen in "show accounting log all" don't give a clue.
Sat Jun 20 08:09:20 2020:type=update:id=::added user cnd_awx
Sat Jun 20 08:09:21 2020:type=start:id=<jumphost>@pts/0:user=<user>:cmd=
Sat Jun 20 08:09:22 2020:type=update:id=<jumphost>@pts/0:user=<user>:cmd=terminal length 0 (SUCCESS)
Sat Jun 20 08:09:23 2020:type=update:id=<jumphost>@pts/0:user=<user>:cmd=terminal width 511 (SUCCESS)
Sat Jun 20 08:09:33 2020:type=update:id=<jumphost>@pts/0:user=<user>:cmd=dir (SUCCESS)
Sat Jun 20 08:10:10 2020:type=stop:id=<jumphost>@pts/0:user=<user>:cmd=shell terminated because the ssh session
closed
The TACACS accounting records don't show anything more.
1) Is there a connection between the on box python crash and Ansible running?
2) What commands does Ansible run against when facts are gathered and why are these commands not appearing in accounting logs?