Bug 1343714 - rhel-osp-director: 7.3->8.0 upgrade of deployment with external ceph. The /etc/ceph/ceph.conf on compute lists wrong monitors.
Summary: rhel-osp-director: 7.3->8.0 upgrade of deployment with external ceph. The /e...
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: rhosp-director
Version: 8.0 (Liberty)
Hardware: Unspecified
OS: Unspecified
Target Milestone: async
: 10.0 (Newton)
Assignee: Giulio Fidente
QA Contact: Alexander Chuzhoy
Depends On:
TreeView+ depends on / blocked
Reported: 2016-06-07 18:21 UTC by Alexander Chuzhoy
Modified: 2016-07-19 20:59 UTC (History)
8 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Last Closed: 2016-07-19 20:59:55 UTC
Target Upstream Version:

Attachments (Terms of Use)

Description Alexander Chuzhoy 2016-06-07 18:21:01 UTC
rhel-osp-director:  7.3->8.0 upgrade of deployment with external ceph. The /etc/ceph/ceph.conf on compute lists wrong monitors.


Steps to reproduce:
1. Deploy 7.3 with external ceph.
2. Populate the overcloud (successfully launch an instance residing on external ceph)
3. Upgrade the setup to 8.0
4. Attempt to launch another instance

The instance goes into error state.

Checking the /etc/ceph/ceph.conf file on compute:
osd_pool_default_pgp_num = 32
osd_pool_default_min_size = 1
auth_service_required = cephx
fsid = 1c50eddf-56fb-470f-a72d-47a960f520a0
auth_supported = cephx
auth_cluster_required = cephx
mon_host =,,
auth_client_required = cephx
osd_pool_default_size = 3
osd_pool_default_pg_num = 32
ms_bind_ipv6 = False

osd_journal_size = 1024

The mon_host lines actually contains IP addresses of controllers.

On computes replace the IP addresses in the mon_host line with the IP address of external ceph monitors and restart systemctl restart openstack-nova-compute.service

Comment 5 Alexander Chuzhoy 2016-07-11 17:58:58 UTC
Deployment command:
openstack overcloud deploy --templates --control-scale 3 --compute-scale 1 --swift-storage-scale 0 --block-storage-scale 0 --neutron-tunnel-types vxlan,gre --neutron-network-type vxlan,gre --neutron-network-vlan-ranges datacentre:118:143 --neutron-bridge-mappings datacentre:br-ex  --ntp-server clock.redhat.com --timeout 90 -e /usr/share/openstack-tripleo-heat-templates/environments/storage-environment.yaml -e /usr/share/openstack-tripleo-heat-templates/environments/network-isolation.yaml -e network-environment.yaml -e ~/ssl-heat-templates/environments/enable-tls.yaml -e ~/ssl-heat-templates/environments/inject-trust-anchor.yaml --ceph-storage-scale 0 -e puppet-ceph-external.yaml

Comment 6 Giulio Fidente 2016-07-11 18:02:45 UTC
Assuming we were using the same environment file for both the initial deployment and the upgrade, could you attach the environment file with the external-ceph parameters and paste the upgrade cmdline?

Comment 9 Alexander Chuzhoy 2016-07-19 20:59:55 UTC
Indeed the upgrade command was missing some of the templates used to deploy.
Once added, was able to successfully upgrade.
Closing as not a bug.
Thanks Giulio!

Note You need to log in before you can comment on or make changes to this bug.