After installing OpenShift Origin, your system might need extra configuration to ensure your hosts consistently run smoothly.

While these are classified as run-once tasks, you can perform any of these at any time if any circumstances change.

NTP synchronization

NTP (Network Time Protocol) is for keeping hosts in sync with the world clock. Time synchronization is important for time sensitive operations, such as log keeping and time stamps, and is highly recommended for Kuberbetes, which OpenShift Origin is built on. OpenShift Origin operations include etcd leader election, health checks for pods and some other issues, and helps prevent time skew problems.

Depending on your instance, NTP might not be enabled by default. To verify that a host is configured to use NTP:

$ timedatectl
      Local time: Thu 2017-12-21 14:58:34 UTC
  Universal time: Thu 2017-12-21 14:58:34 UTC
        RTC time: Thu 2017-12-21 14:58:34
       Time zone: Etc/UTC (UTC, +0000)
     NTP enabled: yes
NTP synchronized: yes
 RTC in local TZ: no
      DST active: n/a

If both NTP enabled and NTP synchronized are yes, then NTP synchronization is active.

If no, install and enabling the ntp or chrony rpm package.

For NTP:

# timedatectl set-ntp true

For chrony:

# yum install chrony
# systemctl enable chronyd --now

Time synchronization should be enabled on all hosts in the cluster, whether using NTP or any other method.

For more information about the timedatectl command, timezones, and clock configuration, see Configuring the date and time and UTC, Timezones, and DST.

Entropy

OpenShift Origin uses entropy to generate random numbers for objects such as IDs or SSL traffic. These operations wait until there is enough entropy to complete the task. Without enough entropy, the kernel is not able to generate these random numbers with sufficient speed, which can lead to timeouts and the refusal of secure connections.

To check available entropy:

$ cat /proc/sys/kernel/random/entropy_avail
2683

The available entropy should be verified on all node hosts in the cluster. Ideally, this value should be above 1000.

Red Hat recommends monitoring this value and issuing an alert if the value is under 800.

The most reliable way to determine the available entropy is with the rngtest command from the rng-tools:

$ cat /dev/random | rngtest -c 100

If the above takes too long to complete, then there is not enough entropy available.

Depending on your environment, entropy can be increased in multiple ways. For more information, see the following blog post: https://developers.redhat.com/blog/2017/10/05/entropy-rhel-based-cloud-instances/

Generally, you can increase entropy by installing the rng-tools package and enabling the rngd service:

# yum install rng-tools
# systemctl enable --now rngd

Once the rngd service has started, entropy should increase to a sufficient level.

Checking the default storage class

For proper functionality of dynamically provisioned persistent storage, the default storage class needs to be defined. During the installation, this default storage class is defined for common cloud providers, such as Amazon Web Services (AWS), Google Cloud Platform (GCP), and more. To verify that the default storage class is defined:

$ oc get storageclass
NAME                 TYPE
ssd                  kubernetes.io/gce-pd
standard (default)   kubernetes.io/gce-pd

The above output is taken from an OpenShift Origin instance running on GCP, where two kinds of persistent storage are available: standard (HDD) and SSD. Notice the standard storage class is configured as the default. If there is no storage class defined, or none is set as a default, see the Dynamic Provisioning and Creating Storage Classes section.