Does anybody has done the servicenow integration with Red hat satellite server? Any help will be appreciated.

[Ajay Sharma2]

Does anybody has done the servicenow integration with Red hat satellite server? Any help will be appreciated.

[Priyanka Vasant]

Hi,

I have attached few link that might help you.

Click here.

Link 1

Link 3

--

Regards,

Priyanka Asawale

[Severin Launiau]

Hi Ajay,

First off, which version of Red Hat Satellite are you using?

Then, is that about orchestration or CMDB inventory?

About the inventory, I did a custom integration a while back with Sat5 using their XMLRPC interface. Their API provided very limited information, so I only picked up the list of active systems and their score, nothing more.

On Sat6, things are way better. I actually intend to develop a custom pattern for that. One of the nice things on Sat6 is that you can access facts of the systems registered in satellite. And apparently facts can be retrieved via their API.

A couple of facts you may want to consider:

dmi::system::uuid

dmi::system::serial_number

virt::is_guest (to know if it's a VM)

virt::uuid

uname::nodename (the FQDN)

With that, you should be covered for your identification rules. Note that in some environments, the serial # is not enough for uniqueness, for example on some old RHEV 3.x VMs. Use the UUID + serial # if you can set up a hybrid rule, or just rely on the FQDN. And if you have a rule on serial # only, add a check for "virtual" set to FALSE. You likely want to update the OOTB Linux Server pattern to submit that virtual attribute.

About your implementation of Sat6, make sure you have a support SME around to watch for performances. Sat6 requires quite a bit of tuning if you have a massive implementation (multi org, content views, thousands of systems). The underlying infrastructure is very important, especially your storage backend and the apache config. Engage your Red Hat support and your TAM if you or your Sat6 SMEs notice a service degradation.

A few tips about mixing up data sources for the same classes of items:

- class downgrade / upgrade: have a look at CI Reclassify. Maybe you have some data sources that contain just secondary, metadata information. They may not have all the information you need to correctly determine the class. You could pull your inventory from the midserver agent, but that's quite heavy. Instead, you can consider writing a pre sensor script to upgrade the class when the record is already registered. Have a look at this.

- discovery sources and last discovered date: make a decision about these attributes. Do you want your pattern to set the source to "ServiceNow"? Do you want to track each individual source in its own boolean or timestamp attribute?

- schedule: how often do you need that data refreshed? If you have a large set of items and many relationships, it may take quite a bit of time to process all of that. I am actually considering using a cluster of dedicated agents for running patterns like Kubernetes / OpenShift. These agents likely need some tuning, since they should have the exact opposite config of regular discovery agents that run many queries in parallel (WMI / SSH / SNMP).

- records life cycle: what's your plan when a record just disappear from that source? What if the subscription agent on the guest hosts just doesn't connect anymore, let's say because the server certificate expired? Have you set up some auto clean up capabilities at the source or in ServiceNow? Do you have autoscaling features enabled on your virtualization platform and if so, what would be the impact on your CMDB?

Have fun with patterns!

Séverin