Bug 1513199

_pacemaker_ rebased to version 1.1.18 The _pacemaker_ packages have been upgraded to upstream version 1.1.18, which includes the following behavioral difference: Pacemaker now probes virtual machines launched with a Pacemaker Remote connection ("guest nodes"), just as it probes any node that joins the cluster. This allows Pacemaker to catch services that were mistakenly started at boot or by hand, and to refresh its knowledge after a resource clean-up. As such, it is important in order to be able to avoid having a service running in conflicting locations. However, these probes must be executed and the results processed before any resources can be started on the guest node. This can result in a noticeable increase in start-up time. Also, if users were previously relying on the probes not being done, the probes may fail (for example, if the relevant software isn't installed on the guest). These effects can be avoided in cases where it is not possible for certain resources to run on the guest nodes. Usually there will already be -INFINITY location constraints enforcing that. Users can add `resource-discovery=never` to the location constraint options to tell pacemaker not to probe that resource on the guest nodes. (This should not be done for any resource that can run on the guest.)