Podman and var tmp

#linux #podman

Issues during building

At the end of your image build process you will see the following lines

[2/2] STEP 10/10: CMD ["bash","/root/nginx_run.sh"]
--> Using cache d539d506d5b8887f48df658583778af3e5be9f45aaee8dc75e3eba8137641217
[2/2] COMMIT nginx:latest
--> d539d506d5b8
Successfully tagged localhost/nginx:latest
d539d506d5b8887f48df658583778af3e5be9f45aaee8dc75e3eba8137641217

During the COMMIT stage the files are being written to /var/tmp/buildah##########. If /var/tmp is too small or currently has too many files your build will fail with the lines "no space left on device"

You can safely delete any buildah########## that exist assuming they are older than "now" and are not in use.

For more information on when /var/tmp is automatically cleaned on the host see How is TMP cleared

Issues during pulling

You may see the above failures when pulling images as well.

/var/tmp is too small

Possible workarounds for /var/tmp being too small for your image

Redefine TMPDIR

Just define TMPDIR like the example below:

$ export TMPDIR="${HOME}/tmp/buildah"
$ mkdir -p "${TMPDIR}"

RHEL Docs

There are two solutions.