MCD-pull: run after network-online.target in Azure

This was originally reordered such that ovs-configuration was able to
run when the reboot was skipped. See: openshift#3858

This broke ARO since they require the network to be ready for the pull
to happen (and generally, probably best for the network to be ready
before attempting to pull the new OS image).

This fix aims to apply just to Azure, which is not ideal since it drifts
the service definition, but if don't-reboot-on-metal and ARO have
different dependency chains, this is the easiest middle ground.

templates/common/azure/units/machine-config-daemon-pull.service.yaml

@@ -0,0 +1,24 @@
+name: "machine-config-daemon-pull.service"
+enabled: true
+contents: |
+  [Unit]
+  Description=Machine Config Daemon Pull
+  # Make sure it runs only on OSTree booted system
+  ConditionPathExists=/run/ostree-booted
+  # This "stamp file" is unlinked when we complete
+  # machine-config-daemon-firstboot.service
+  ConditionPathExists=/etc/ignition-machine-config-encapsulated.json
+  # Run after crio-wipe so the pulled MCD image is protected against a corrupted storage from a forced shutdown
+  Wants=crio-wipe.service network-online.target
+  After=crio-wipe.service network-online.target
+
+  [Service]
+  Type=oneshot
+  RemainAfterExit=yes
+  ExecStart=/bin/sh -c "while ! /usr/bin/podman pull --authfile=/var/lib/kubelet/config.json '{{ .Images.machineConfigOperator }}'; do sleep 1; done"
+  {{if .Proxy -}}
+  EnvironmentFile=/etc/mco/proxy.env
+  {{end -}}
+
+  [Install]
+  RequiredBy=machine-config-daemon-firstboot.service