Make the changes for x11 to the sshd file

nano /etc/ssh/sshd_config

Change the following lines:

X11Forwarding yes

X11DisplayOffset 10

X11UseLocalhost no

 

Then restart the service

service sshd restart 

One may have to exit the session and restart the session for this to take effect.

On your local machine download an xwindows client.

You can test this is set up correctly at your own discretion with a non-IBM tool such as:

xclock

If this is not correctly set up. Please work with your system administrator to make sure your xwindows is working.

Step 2: Install Docker (CENT OS – CE)

More information can be found on the docker website listed below. I have included the relevant Cognos steps below.

(**Please note, mixing CENT OS Docker CE and RHEL is unsupported. This is given as an example of steps. You can try similar steps with your licensed Docker for RHEL EE) 

 Get Docker Engine - Community for CentOS

Remove old versions of docker:

yum remove docker \

                  docker-client \

                  docker-client-latest \

                  docker-common \

                  docker-latest \

                  docker-latest-logrotate \

                  docker-logrotate \

                  docker-engine

Install required packages. yum-utils provides the yum-config-manager utility, and device-mapper-persistent-data and lvm2 are required by the devicemapper storage driver. (these should be installed by default so you might already have them)

yum install -y yum-utils   device-mapper-persistent-data   lvm2

Remove any old docker repos:

rm /etc/yum.repos.d/docker*.repo

Set up the stable repository

yum-config-manager \

    --add-repo \

    https://download.docker.com/linux/centos/docker-ce.repo

Install the docker engine

yum install docker-ce docker-ce-cli containerd.io

List and sort the versions available on the repo

yum list docker-ce --showduplicates | sort -r

Run the following command to install a specific version:

yum install docker-ce-<VERSION_STRING> docker-ce-cli-<VERSION_STRING> containerd.io

I used the most recent at the time

sudo yum install docker-ce-18.06.3.ce-3.el7 docker-ce-cli-18.06.3.ce-3.el7 containerd.io

 start docker

systemctl start docker

Check that you have the most recent/supported version of docker:

docker version

Now you can start docker:

systemctl start docker

Potential error:

Permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock

Go to the docker.sock file and change the permissions (according to your company’s standards). In house this was resolved by:

cd /var/run/

chmod 666 docker.sock

Step 2(Windows 10) : Install Docker 

Download and install your favorite supported version of docker. I got mine from the Docker website.

Click on “Download From Docker Hub”

Click on “Get Docker”

Run the executable.

Let it download.

Be sure to leave “USE WINDOWS CONTAINERS INSTEAD OF LINUX CONTAINERS” unchecked.

The Cognos Jupyter containers are ubuntu.

You can verify this yourself once this is installed by bashing into the container.  

To bash in: docker exec -it -u 0 ca_jupyter_hub bash
To check OS: cat /etc/os-release

NAME="Ubuntu"
VERSION="18.04.2 LTS (Bionic Beaver)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 18.04.2 LTS"
VERSION_ID="18.04"
HOME_URL="https://www.ubuntu.com/"
SUPPORT_URL="https://help.ubuntu.com/"
BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/"
PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy"
VERSION_CODENAME=bionic
UBUNTU_CODENAME=bionic

 Let it finish installing.

Launch Docker and make sure it is installed correctly. You should see a little white Docker whale on the bottom corner. If it is red, engage your system admin to help deduce what is wrong.

Note: Some pitfalls here include the fact that nested virtualization may be needed.

You need to make sure that under Task Manager > Performance tab that it indicates Virtualization is Enabled.

In order to assure this, before I started and connected to the VM on the machine hosting the VM, I ran the following commands in powershell:

1. Enable-WindowsOptionalFeature -Online -FeatureName Microsoft-Hyper-V -All -Verbose
2. Enable-WindowsOptionalFeature -Online -FeatureName Containers -All -Verbose
3. bcdedit /set hypervisorlaunchtype Auto
4. Set-VMProcessor -VMName <Enter-VM-Name> -ExposeVirtualizationExtensions $true -Verbose
5. net stop vmms
6. net start vmms

Step 3: Install Jupyter

Navigate to the installers

cd /tmp/installers

Run the installer

./ ca_instl_lnxi38664_2.0.191205.bin

Select the /zip repository in the installers directory recently created:

Please note, if you receive the following error, you likely do not have enough disk space on the server:

Error processing tar file(exit status 1): write /opt/*****: no space left on device

Around 35 GB of available space is needed for the images once completed, this does not include the intermediate/temporary space that docker requires for various layers during the install. Verify and allocate a minimum of 60 GB of space to the /var/lib/docker/ directory to run the install with no issues.

Step 4(RHEL): Configure Jupyter 

CD to the scripts directory of the install. Because we are on Linux, we will go to the UNIX directory.

cd /opt/ibm/cognos/jupyter/dist/scripts/unix

If you execute an LS here you will see a few scrips and configuration files

Before anything, open the config.conf file and confirm the accuracy of the information presented. For our purposes we can imagine this having 2 important parts. The first part related to the SSL component and the second part related to the Configuration component. While other sections are important, they are note covered in this demonstration.

The SSL component will be discussed in greater detail in the following technote:

How to secure Jupyter for Cognos Analytics 11.1.2 + with 3rd party certificates

The configuration section on RHEL will be comprised of the hostname and port.

1. The hostname should refer to the Jupyter server and the port should be open and accessible.

By default RHEL instances will define the hostname as $(hostname) and this will cause Jupyter to start up using the hostname specified by the machine hosting the install. You can define this yourself or on RHEL you can check what this will resolve to by running:

echo $HOSTNAME

2. The port number is the port with which the Jupyter server will start on and be reachable externally. By default this is 8000.

You will need to be sure that the Cognos Analytics Server and the browser can reach the hostname:port. A simple test would be to go to the Cognos Analytics server (where the dispatcher/application tier is installed) and telnet to the Jupyter server $HOSTNAME and port number you are defining in the config.conf file.

Example: Go to the cognos server and type in the following

Step 4 (Windows) : Configure Jupyter 

Navigate to the install location. You can do this in file explorer or by cd’ing there in a command prompt.  

 

By Default: C:\Program Files\ibm\cognos\jupyter\dist\scripts\windows

Edit the config.conf file making the HOST_NAME variable set as either the fully qualified hostname or the IP address of the Jupyter server.  This should be reachable by the cognos server.

Step 5(RHEL) : Run the installer script 

Navigate to the unix directory and run the ./installer.

cd /opt/ibm/cognos/jupyter/dist/scripts/unix

./install.sh

When this has started up correctly you should see the following entry within the install.sh start up:

Check the containers are running

docker ps

If you paste this in the browser, you should be able to see the following message, which indicates your browser can successfully communicate with the Jupyter server. This is not an error code, this is expected.

http://servername:8000

Common installer issues (RHEL)

RHEL (11.1.4 and lower)

cd to the build script

cd /opt/ibm/cognos/jupyter/dist/scripts/unix/

edit the build file

nano build.sh

Resolve this by changing the docker group to dockerroot in the group id section like this:

Old versions of Docker

The chown flag used by Jupyter, is only supported as of Docker version 17.09 and up. Upgrade your version of Docker using the steps outlined above. One can consult the Docker documentation .

Offline Installs

Change directory to the tmp folder:

cd /tmp

Now Download Pixiedust using a wget

wget https://files.pythonhosted.org/packages/16/ba/7488f06b48238205562f9d63aaae2303c060c5dfd63b1ddd3bd9d4656eb1/pixiedust-1.1.18.tar.gz 

Then extract it

tar -xvf pixiedust-1.1.18.tar.gz

Navigate to the following directory

/ibm/cognos/jupyter/dist/scripts/

Edit the file

nano additional_packages.txt

Comment out all the lines with # 

Add the following line:

--no-index --find-links=/tmp/pixiedust-1.1.18  pixiedust.

Step 5 (Windows) : Run the installer script 

Execute the install script as an administrator:

Alternatively (if you would like to see error codes that pop up if it does not start the first time), open a command prompt as an administrator and cd to the location. Following that run the install.bat file 

cd C:\Program Files\ibm\cognos\jupyter\dist\scripts\windows

install.bat

If the script finished indicating it started successfully, you can run a docker ps to see if the containers are up.

docker ps

Step 6: Verify the Cognos Configuration

Verify the dispatcher URI for external application readys with the same syntax as follows:

http://servername.domain.com:9300/bi/v1/disp

Step 7: Configure the Cognos administration portal

Please access the dispatcher URI in your browser:

http://servername.domain.com:9300/bi/v1/disp

Click on the Manage Button

Navigate to Configuration > System > Environment (tab)

For the Jupyter service location place the URL that was specified on start-up during step 7.

http://jupyter1.fyre.ibm.com:8000

Click apply and refresh the browser (you may need to clear cache and start a new session)

Click on New > Notebook

 

And test your notebook!