Separate both AWX jobs

If the first one fails (unreachable node for example), the second script
won't be run.
Having 2 separate stages ensure both will always run.

We might want to create a template to deploy on all nodes at the same
time.

.gitlab-ci.yml

 stages:
   - deploy
 
-deploy:
+deploy-lcr:
   tags:
     - docker
   stage: deploy
@@ -14,6 +14,15 @@ deploy:
       -J "deploy-cs-studio@lcr"
       --tags csstudio-ess-opis
       -e "csstudio_ess_opis_version=$CI_COMMIT_REF_NAME" --monitor
+  only:
+    - master@ics-software/ess-opis
+
+deploy-nx:
+  tags:
+    - docker
+  stage: deploy
+  image: registry.esss.lu.se/ics-docker/tower-cli:3.3
+  script:
     - >
       tower-cli job launch
       -h torn.tn.esss.lu.se