Add 'rebootable' label to ceph-pull-requests

This is to signal node selection that the job will reboot the node
after it finishes.  Without an explicit label on the node and a request
for it, any 'trusty' node would satisfy, and be rebooted; this is
a problem for nodes that 'dial in' to the Jenkins master (such as
physical slaves on the sepia network).

This corresponds with the addition of 'rebootable' tags to mita's
configuration, so that mita nodes continue to suffice for this job.

Signed-off-by: Dan Mick <dan.mick@redhat.com>

diff --git a/ceph-pull-requests/config/definitions/ceph-pull-requests.yml b/ceph-pull-requests/config/definitions/ceph-pull-requests.yml
index ea39bd95fbd75b20842c3d068bb58b2d1ed86cff..b1675a4e39b8d8de824bc757464320c37232a17b 100644 (file)
--- a/ceph-pull-requests/config/definitions/ceph-pull-requests.yml
+++ b/ceph-pull-requests/config/definitions/ceph-pull-requests.yml
@@ -3,7 +3,7 @@
     project-type: freestyle
     defaults: global
     concurrent: true
-    node: huge && (centos7 || trusty)
+    node: huge && (centos7 || trusty) && rebootable
     display-name: 'ceph: Pull Requests'
     quiet-period: 5
     block-downstream: false