testnode: restart nfs-kernel-server and don't enable it.

We found an issue where ansible would fail to enable nfs-kernel-server
on vivid. It's already enabled by default anyway, so let's remove that
to avoid the error.

Also, ceph-qa-chef used to only restart nfs-kernel-server for ubuntu and
didn't enable so let's go back to that functionality.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

diff --git a/roles/testnode/handlers/main.yml b/roles/testnode/handlers/main.yml
index 18d4412916d5707ff897e058ade185d0e6137c39..b33c1e3c5c8ccd85b317848cbce02afefec1f395 100644 (file)
--- a/roles/testnode/handlers/main.yml
+++ b/roles/testnode/handlers/main.yml
@@ -15,11 +15,10 @@
     state: started
     enabled: yes
 
-- name: start nfs-server
+- name: restart nfs-server
   service:
     name: "{{ nfs_service }}"
-    state: started 
-    enabled: yes
+    state: restarted
 
 - name: restart cron
   service:

diff --git a/roles/testnode/tasks/nfs.yml b/roles/testnode/tasks/nfs.yml
index e06d606874ea657cae91ecebf54fd737cfe8469d..9ba641f94bea7fbfe63f7a5ef3f08bd5f40a0b8c 100644 (file)
--- a/roles/testnode/tasks/nfs.yml
+++ b/roles/testnode/tasks/nfs.yml
@@ -8,4 +8,4 @@
     mode: 0644
   notify:
     - start rpcbind
-    - start nfs-server
+    - restart nfs-server