Allow the pipeline to fail

Adding the `path` attribute, to the 'terraform/cf-certs/upload.tf' is
causing the task to fail. This is expected behaviour, as it redefines
the certificate.

Terraform would be trying to remove the old certificates, causing amazon
to fail, as these are still in use. The idea is, that the new
certificates will get created and defined to be used. Once that happens,
the old ones, will be peacefully removed.

This problem, will occur only in our DEV and CI environments. Staging
and prod should not be affected.

This commit, could be reverted after a week after merging this PR.

concourse/pipelines/create-cloudfoundry.yml

@@ -706,6 +706,10 @@ jobs:
                   tar xzf cf-certs/cf-certs.tar.gz -C generated-certificates
                   tar xzf bosh-CA/bosh-CA.tar.gz
 
+                  # FIXME We're expecting this task to fail.
+                  # That's ok, whilst we're updating the certificates.
+                  set +e
+
                   terraform apply -var-file=paas-cf/terraform/{{aws_account}}.tfvars \
                     -var system_domain_crt="$(cat generated-certificates/system_domain.crt)" \
                     -var system_domain_key="$(cat generated-certificates/system_domain.key)" \
@@ -717,6 +721,9 @@ jobs:
                     -state-out=updated-tfstate/cf-certs.tfstate \
                     paas-cf/terraform/cf-certs
                 fi
+
+                # FIXME
+                echo "We're expecting this task to fail. That's fine. - Rafal"
         ensure:
           put: cf-certs-tfstate
           params: