Consider disabling dnf-makecache.timer

[jlebon]

See openshift/os#1476 (comment).

$ podman run -ti --rm quay.io/centos-bootc/centos-bootc:stream9 systemctl is-enabled dnf-makecache.timer
enabled

But cc @dcantrell for his thoughts on the suggestion there.

[travier]

Should we make this unit condition on !/run/ostree-booted?

[jlebon]

Should we make this unit condition on !/run/ostree-booted?

It already is. The problem is more that it pulls in network-online.target, which is not recommended. We have a test in FCOS/RHCOS to make sure it doesn't inadvertently get activated with the default units.

[cgwalters]

I think this is something to fix upstream in dnf, honestly. It should gracefully wait for network availability and not error out.

(Incidentally we have a related problem in rpm-ostree-countme.timer I've noticed the service failing, but the service shouldn't fail if networking isn't working)