Chapter 3. Known Issues

DNS addressing provided by the external xip.io service is unavailable without a working Internet connection.

After signing up for a Red Hat Developers account, you must wait approximately fifteen minutes to allow the new account information to propagate to the Red Hat Subscription Management system before the Container Development Environment can be registered.

The nodejs-4-rhel7 Docker-formatted container image, which is available from the Red Hat Atomic Registry, does not work when the Container Development Environment is connected to the Internet through a proxy server.

To enable automatic registration of the Container Development Environment Vagrant box, credentials for to Red Hat Subscription must be defined in the Vagrantfiles that are supplied in the Red Hat Container Tools ZIP file. Credentials defined in the user’s global Vagrantfile in /home/<username>/.vagrant.d/ are ignored by the vagrant-registration plugin.

An insecure certificate causes an error when using oc login. Use the --insecure-skip-tls-verify=true option with the oc login command to prevent the warning about insecure connection.

When using the Container Development Environment behind a proxy server, the vagrant reload command does not work properly, which results in OpenShift applications not loading. To avoid this issue, halt and restart your Container Development Environment Vagrant box using the vagrant halt and vagrant start commands.

Using the --path option with the install-cli command of the vagrant-service-manager plugin does not work as expected. The following error is shown:

Binary now available at /path/to/CDK/2.3/components/rhel/rhel-ose/cli
run binary as:
oc <command>
run following command to configure your shell:
eval "$(VAGRANT_NO_COLOR=1 vagrant service-manager install-cli openshift --path cli | tr -d '\r')"

# Binary now available at /path/to/CDK/2.3/components/rhel/rhel-ose/cli
# run binary as:
# oc <command>
export PATH=/path/to/CDK/2.3/components/rhel/rhel-ose:$PATH

# run following command to configure your shell:
# eval "$(VAGRANT_NO_COLOR=1 vagrant service-manager install-cli openshift --path cli | tr -d '\r')"

If either of the environment variables, PROXY_USER or PROXY_PASSWORD, is unset, the vagrant up command does not work as expected. If they are unset, Vagrant may also log many events more rapidly than expected. There are two workarounds:

In some circumstances, the filesystem may become full more quickly than expected. This impact can be minimized by running oadm-prune periodically. Filesystems that fill up can have an additional volume added to the affected volume group, to extend the runtime of the system. As the filesystem becomes full, growing log files may also fill the remaining space.

The install-cli command provided by the vagrant-service-manager plugin does not install the 1.3.1 version recommended for OpenShift Container Platform version 3.3.1.3, which is supplied by the Container Development Environment by default. The correct version can be installed using the following command:

vagrant service-manager install-cli openshift --cli-version 1.3.1

$ vagrant service-manager install-cli openshift --cli-version 1.3.1

In some circumstances, the filesystem may become full more quickly than expected. This impact can be minimized by running oadm-prune periodically. Filesystems that fill up can have an additional volume added to the affected volume group, to extend the runtime of the system. As the filesystem becomes full, growing log files may also the remaining space.

When the vagrant-registration plugin is used in interactive mode in the Cygwin Bash terminal, the entered password is visible in the standard input. The following warning will be printed:

Error! Your console doesn't support hiding input. We'll ask for
input again below, but we WILL NOT be able to hide input. If this
is a problem for you, ctrl-C to exit and fix your stdin.

Error! Your console doesn't support hiding input. We'll ask for
input again below, but we WILL NOT be able to hide input. If this
is a problem for you, ctrl-C to exit and fix your stdin.

The /usr/bin/docker wrapper on Red Hat Enterprise Linux 7 overwrites the DOCKER_CERT_PATH environment variable, which means that the Docker client does not work properly with the Docker service provided by the Container Development Environment. To work around this issue, use directly the docker-current command.

On Red Hat Enterprise Linux 7.2, the vagrant destroy command does not properly delete the virtual machine storage image. To work around this libvirt bug, you need to change the ownership of the image file and perform several steps using the virsh command.

Installing the vagrant-service-manager plugin may fail on Red Hat Enterprise Linux 7 due to missing dependencies. You can install the missing packages with:

sudo yum -y install @'Development Tools'
sudo yum -y install libvirt-devel rh-ruby22-ruby-devel

# sudo yum -y install @'Development Tools'
# sudo yum -y install libvirt-devel rh-ruby22-ruby-devel

The Vagrant plugins should then be installed:

vagrant plugin install vagrant-service-manager
vagrant plugin install vagrant-registration vagrant-sshfs
vagrant plugin list

# vagrant plugin install vagrant-service-manager
# vagrant plugin install vagrant-registration vagrant-sshfs
# vagrant plugin list

Kubernetes startup will cause the openshift service in a VM to stop. Starting openshift after Kubernetes may also fail with an error. This problem can be verified with the command vagrant service-manager status openshift.