FlexNet Manager Blog (Archived) — 0_MattR (Flexera Software)
What are Containers in FNMS 2020 R1?
Containers are a new type of application virtualisation that provides much greater flexibility and scalability then other technologies. The most popular of these is Docker and in FNMS 2020 R1 the upgraded agent will allow you to discover and inventory docker containers giving you complete visibility into docker instances in your environment and the software running on them.
Some key points about this functionality:
- An agent upgrade IS required (Cloud and On-Premises)
- Server components also need to be updated (On-Premises customers only)
- Docker capability is only on our Linux agents in this release, Windows will be updated later.
- This release is focused on Discovery and Inventory, software running in containers will not be included in license reconcile or consumption calculations
- The new agent monitors the Docker Engine service on the device to collect inventory which means you can only discover containers on devices where you have the agent installed and so cloud based Kubernetes or other Container Orchestration services is not currently included.
Getting started with Docker discovery and inventoryHow to enable Docker Container inventory?
Enabling docker inventory will modify installation counts and so by default it is disabled and will need to be enabled.
You can do this by:
- Open Discovery & Inventory – Settings
- Find “Container Scanning” section
- Enable checkbox “Enable detection of Docker and running inventory agent inside Docker containers”
- Click Save in bottom right of screen
Once enabled, this will update agent policy and will be deployed to beacons and then agents need to update policy, run inventory, upload this inventory and then full inventory import and reconcile needs to complete.
This process usually takes 24-48 hours before it’s all processed.
Viewing list of Containers in your environment
To validate that inventory has succeeded, go back to Discovery & Inventory menu and there is now a new Containers section with a view called “All Containers”, this view shows all containers in your environment.
Licensing Managers
One of the use cases we were trying to address for SAM Managers was “How many of my applications are running in containers”? We know from discussions that you want to see what impact containerised applications are having so we’ve added a new “Containers” column to the following views:
- Application views e.g. Install Applications, All Applications etc
- Unlicensed Installations
These Container columns show the subset of installations from containers so from the Unlicensed Installations above, you have 4 installations of Ubuntu and 3 come from Containers which means that as Container applications aren’t included in license reconcile, you have 1 that can be addressed by the reconcile and 3 from containers which you need to review and manually allocate where necessary.
Hardware Asset Managers
If you are managing specific devices and want to see whether they are running containers there is a “Container” tab which lists the Container and Docker Images.
If you then want to see what applications these devices are running within containers you can go to the Applications tab and you will see a new sub-tab called Containers which will show you all the Containers associated with this device.
These are the main views we’ve implemented, I will be adding a Q&A section at a later date once I’ve finished collating all of the recent questions.
Hi Matt,
A few questions:
Best regards,
Markward
Hi @mfranz ,
I'll ask one of the engineers to add further input if needed but in summary we've added an extra daemon service to the usage engine of the agent that monitors the docker engine service and once it identifies a docker image has been created it uses the zero footprint inventory (i.e. ndtrack.sh on Linux) to scan the contents of the image to produce ndi content.
As each docker image is immutable i.e. it cannot be changed, we know that any container deployed from that image will have the same contents, this allows for on-going updates instead of just a one-time snapshot.
So to track both the software in containers and the state of containers / images we use a combination of the inventory agent and the usage agent respectively; the latter will track when the containers were started, stopped and destroyed and these are used to populate the fields you see in the All Containers and properties views.
Hi Matt, I have just been asked if Flexera supports Kubernetes Clusters (in my situation they are talking about Redhats Openshift) ... only knowing a little bit about Containers and the different type of environments that house them , for example Docker and RedHats Openshift , I am a little bit confused after reading what Flexera has delivered in this particular Release.
Perhaps you could answer the question that has been asked of me .. I understand from reading the information that I could find doing searches - that the "updated" agent will be required for sure and that it will only identify the type of container (Docker only) but not the applications running within it .. or atleast that is what I understood.
I apologize in advance for my lack of knowledge between containers, kubernetes etc. and if the information is located somewhere else but I too am prohibited in viewing YouTube videos so I could not see what was discussed in a previous webinar where the different phases was explained. Perhaps you could copy the key areas of discussion into a response.
Thanks Again
Bruce Giles
Hi @bruce_giles ,
The 2020 R1 version on container functionality will carry out discovery and inventory of docker containers so will identify the Host, Container Image and Containers and will also run ndtrack to identify software running in the containers.
We don't currently support Kubernetes / OpenShift however this is something I've been asked for a lot recently so it is going to either by 1st or 2nd on the container roadmap which means that current expectations is they will be delivered in the first half of next year however we're still finalising the roadmap for Q1 and Q2 so I can't give exact release details at this point.
Does this help?
Hi Matt, that is Perfect !! ... thanks again, really appreciate it !!
Bruce
Hi,
Will the agent report correctly on Oracle deployments within containers?
Thanks
Yes it will. We run the same agent code inside containers, that we do outside.
Hi,
thanks for the article. Very interesting.
In my Env. I've checked docker on a RHEL8 system as well as on a SLES 15 System. I've set up a simple Apache-server in a docker on both and scanned the systems too, and I know for sure, that both OS have different kind of dockerization-technology implemented.
So for them (the OS-Vendors), it is not really necessary to have the "initial docker/container-binaries or installation packs" in place. They do it somehow different.
RHEL8 see this article:
https://access.redhat.com/solutions/3696691
SLES15:https://documentation.suse.com/de-de/sles/15-GA/html/SLES-all/book-sles-docker.html
It seems like the OS-Vendors have established different approaches, to achieve the same goal, and it seems that they don't need the initial docker-installation as provided by docker itself "https://docs.docker.com/engine/install/ "My question is:
How does Flexera keep up with this little "own-revolution" regarding dockerization and different techniques, to deliver a correct recognition by the Agent for those different OSes, if the initial docker installation is "missing" or is "hidden" to the Agent?regards,
Matthias
PS: used Agents and FNMS-Server for my checks are on 2020 R2 - on-prem
Hi @mschwach ,
Building additional inventory capabilities around containers is time consuming as there are several scenarios to consider for each one; the most common being that containers are created / destroyed in very short time periods and so monitoring needs to continuously running.
As such, while we are aware that there are multiple different container technologies in the market, the plan is to focus on the most common / most in demand technologies. The next one we'll be focusing on is Kubernetes.
After that, we will gauge demand for other technologies and consider each on a case by case basis, the best way to submit these is through the Ideas portal.
Hope this helps.
Hi @mrichardson ,
thanks a lot for the answer. It is sufficient for me.
Will you also pick up sub-capacity calculations in regard to IBM Kubernetes?
https://www.ibm.com/software/passportadvantage/containerlicenses.html
https://www.ibm.com/software/passportadvantage/containerfaqov.html