bc570e74f2
When e.g. ceph storage role is used and a ceph storage node is the first node to run on during deployment [1] the facts get set using this host. As a result role_networks [2] will always only contain Storage and StorageMgmt : TASK [tripleo-ssh-known-hosts : Set ssh_known_hosts fact] ****************************************************************************** Thursday 28 November 2019 14:16:00 +0000 (0:00:00.747) 0:00:13.681 ***** ok: [ceph-0] Therefore all nodes will only get additional hostnames set in the ssh_known_hosts file for the Storage and StorageMgmt network even if the they don't have this network in their role, like the compute [4] StorageMgmt: [192.168.24.42]*,[compute-0.redhat.local]*,[compute-0]*,[172.17.3.29]*,[compute-0.storage]*,[compute-0.storage.redhat.local]*,[192.168.24.8]*,[compute-0.storagemgmt]*,[compute-0.storagemgmt.redhat.local]*, ssh-rsa In such a situation live migration is broken due to host key verification failure [1] https://github.com/openstack/tripleo-ansible/blob/master/tripleo_ansible/roles/tripleo-ssh-known-hosts/tasks/main.yml#L47 [2] https://github.com/openstack/tripleo-ansible/blob/master/tripleo_ansible/roles/tripleo-ssh-known-hosts/tasks/main.yml#L54 [3] https://github.com/openstack/tripleo-heat-templates/blob/master/roles/CephStorage.yaml#L7-L11 [4] https://github.com/openstack/tripleo-heat-templates/blob/master/roles/Compute.yaml#L11-L17 Change-Id: I2491ccd9f1bbe1d9cd2848301956803c5935a6f3 Closes-Bug: #1854368 |
||
|---|---|---|
| .. | ||
| defaults | ||
| meta | ||
| molecule/default | ||
| tasks | ||