Chapter 36. Creating and running containers

Procedure

1. Create the directory you want to use for the container. For example:

   # mkdir cyclictest

2. Change into that directory:

   # cd cyclictest

3. Log in to a host that provides a container registry service:

   # podman login registry.redhat.io
   Username: my_customer_portal_login
   Password: ***
   Login Succeeded!

4. Create a Containerfile:

   # vim Containerfile

5. If you are building from a custom Containerfile modify your Containerfile and build it. Here is an example with cyclictest. If not creating your own image, you can also pull the realtime-tests-container image to run cyclictest:

   # podman build -t cyclictest .

Procedure

1. Run a container using the podman run command:

   # podman run --device=/dev/cpu_dma_latency --cap-add ipc_lock --cap-add sys_nice --cap-add sys_rawio --rm -ti cyclictest
   
   /dev/cpu_dma_latency set to 0us
   policy: fifo: loadavg: 0.08 0.10 0.09 2/947 15
   
   T: 0 ( 8) P:95 I:1000 C: 3209 Min: 1 Act: 1 Avg: 1 Max:  14
   
   T: 1 ( 9) P:95 I:1500 C: 2137 Min: 1 Act: 2 Avg: 1 Max:  23
   
   T: 2 (10) P:95 I:2000 C: 1601 Min: 1 Act: 2 Avg: 2 Max:   7
   
   T: 3 (11) P:95 I:2500 C: 1280 Min: 1 Act: 2 Avg: 2 Max:  72
   
   T: 4 (12) P:95 I:3000 C: 1066 Min: 1 Act: 1 Avg: 1 Max:   7
   
   T: 5 (13) P:95 I:3500 C:  913 Min: 1 Act: 2 Avg: 2 Max:  87
   
   T: 6 (14) P:95 I:4000 C:  798 Min: 1 Act: 1 Avg: 2 Max:   7
   
   T: 7 (15) P:95 I:4500 C:  709 Min: 1 Act: 2 Avg: 2 Max:  29

   This example shows the podman run command with the required, real time-specific options. For example:

   • The first in first out (FIFO) scheduler policy is made available for workloads running inside the container through the --cap-add=sys_nice option. This option also allows setting the CPU affinity of threads, another important configuration dimension when tuning a real time workload.

   • The --device=/dev/cpu_dma_latency option makes the host device available inside the container (subsequently used by the cyclictest workload to configure the CPU idle time management). If the specified device is not made available, an error similar to the message below is displayed:

     WARN: stat /dev/cpu_dma_latency failed: No such file or directory

     When confronted with error messages such as these, refer to the podman-run(1) manual page. To get a specific workload running inside a container, other Podman options might be helpful.

     In some cases, you also need to add the --device=/dev/cpu option to add that directory hierarchy, mapping per-CPU device files such as /dev/cpu/*/msr.