Run the ceph-disk workunit with a mon setup with ceph-deploy. It is
prefered to the install done by the ceph tasks because it relies on the
init system and udev rules which are going to be tested by the ceph-disk
workunit.
http://tracker.ceph.com/issues/11881 Refs: #11881
Signed-off-by: Loic Dachary <loic@dachary.org>
--- /dev/null
+../../../distros/supported
\ No newline at end of file
--- /dev/null
+overrides:
+ ceph-deploy:
+ separate_journal_disk:
+ only_mon: true
+ conf:
+ global:
+ mon pg warn min per osd: 2
+ osd pool default size: 2
+ osd crush chooseleaf type: 0 # failure domain == osd
+ osd pg bits: 2
+ osd pgp bits: 2
--- /dev/null
+overrides:
+ ceph-deploy:
+ conf:
+ global:
+ debug ms: 1
+ osd:
+ debug osd: 10
+ mon:
+ debug mon: 10
+roles:
+- - mon.a
+ - client.0
+openstack:
+ machine:
+ disk: 10 # GB
+ ram: 2000 # MB
+ cpus: 1
+ volumes: # attached to each instance
+ count: 3
+ size: 1 # GB
+tasks:
+- install:
+ extras: yes
+- print: "**** done install extras"
+- ssh_keys:
+- print: "**** done ssh_keys"
+- ceph-deploy:
+- print: "**** done ceph-deploy"
+#
+# Keep this around for debugging purposes. If uncommented the target
+# will pause and the workunit can be run and debug manually.
+#
+# - exec:
+# client.0:
+# - sleep 1000000000 # forever
+#
+- workunit:
+ clients:
+ all:
+ - ceph-disk/ceph-disk.sh