Run ssh command using the "allow_unsupported_shells" flag

[Anders Westlund]

Hi.

We are trying to get information from a Cisco device using ssh.

According to this documentation:
https://docs.servicenow.com/bundle/madrid-it-operations-management/page/product/discovery/reference/r_Parameters.html

I understand it as when using the "allow_unsupported_shells" flag no shell validation is done, and the command specified in the "ECC queue name" is run directly after ssh login. When I examine the logs (I have enabled the ssh_debug flag) it looks like the shell validation (i.e. the "echo $0" command)  is run both with and without the "allow_unsupported_shells" flag.

Is it possible to  avoid the "echo $0" when performing as ssh command via a probe, and move directly to the actual command and get it working according to documentation linked above?

I have enabled the "use_snc_ssh" flag (also validated that this is enabled by looking in the ssh logs) and we use MID server version:
madrid-12-18-2018__patch5-06-26-2019_07-01-2019_1425

Using the logs on the mid server I have also validated that the login is successful and that we get access to the device.

[ggarcia1]

We've since been able to discover our UCS/UCCX devices using a pattern instead of Probe/Sensors.

The I don't fully understand why, but for our devices it seems to be related to response time of the CLI. The pattern seems to wait longer for the CLI to load and actually complete the query, whereas the probe just returns the start up statement.

So if you end up here, try something like this; which is pretty much lifted from the Call Manager pattern:

[Hitoshi Ozawa]

I just created a REST service in our internal network to ssh to cisco devices. Just calling the REST API to the service via MID server.

Wrote the service in Flask + Paramiko in few lines and using the standard ServiceNow REST tools to call it. I'm able to gather all the information on several devices in one REST call.