Skip to main content

OpenStack

Reboot a compute node

  1. Live migrate all instances running on the compute node with the help of the OpenStack Resource Manager

  2. Ensure that no more instances are running on the compute node

    ps ax | grep qemu
  3. Reboot the compute node

    osism apply reboot -l NODE -e ireallymeanit=yes
  4. Re-enable the compute service

    openstack --os-cloud admin compute service set --enable --disable-reason "" NODE nova-compute
  5. Check compute service

    openstack --os-cloud admin compute service list --host NODE --service nova-compute

Add a new compute node

osism apply operator -u osism -l NODE
osism apply bootstrap -l NODE

If FRR is used:

osism apply frr -l NODE
osism apply common -l NODE
osism apply openvswitch -l NODE
osism apply ovn -l NODE
osism apply prometheus -l NODE
osism apply prometheus -l control
osism apply ceilometer -l NODE
osism apply neutron -l NODE
osism apply nova -l NODE

If Scaphandre is used:

osism apply scaphandre -l NODE

If Netdata is used:

osism apply netdata -l NODE

Refresh the /etc/hosts file:

osism apply hosts

Refresh the SSH client configuration file:

osism apply sshconfig

Add compute node to the known hosts file:

osism apply known-hosts

Containers that run on a compute node. Versions may differ.

$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
559e5176695c quay.io/osism/nova-compute:27.1.1.20230919 "dumb-init --single-…" 5 minutes ago Up 5 minutes (healthy) nova_compute
31248d71ab7d quay.io/osism/nova-libvirt:8.0.0.20230919 "dumb-init --single-…" 6 minutes ago Up 6 minutes (healthy) nova_libvirt
9292030d706c quay.io/osism/nova-ssh:27.1.1.20230919 "dumb-init --single-…" 6 minutes ago Up 6 minutes (healthy) nova_ssh
fda4b6fb30c8 quay.io/osism/neutron-metadata-agent:22.0.3.20230919 "dumb-init --single-…" 2 hours ago Up 2 hours (healthy) neutron_ovn_metadata_agent
0e3ec450b668 quay.io/osism/ceilometer-compute:20.0.1.20230919 "dumb-init --single-…" 6 hours ago Up 6 hours (healthy) ceilometer_compute
25ff9702e0e5 quay.io/osism/prometheus-libvirt-exporter:6.0.0.20230919 "dumb-init --single-…" 6 hours ago Up 6 hours prometheus_libvirt_exporter
1bff2e29923b quay.io/osism/prometheus-cadvisor:0.45.0.20230919 "dumb-init --single-…" 6 hours ago Up 6 hours prometheus_cadvisor
602832daf237 quay.io/osism/prometheus-node-exporter:1.4.0.20230919 "dumb-init --single-…" 6 hours ago Up 6 hours prometheus_node_exporter
d4de2f32cdf8 quay.io/osism/ovn-controller:23.6.1.20230919 "dumb-init --single-…" 6 hours ago Up 6 hours ovn_controller
3bf43ae5a94f quay.io/osism/openvswitch-vswitchd:3.1.2.20230919 "dumb-init --single-…" 7 hours ago Up 7 hours (healthy) openvswitch_vswitchd
ebc048b02ab2 quay.io/osism/openvswitch-db-server:3.1.2.20230919 "dumb-init --single-…" 7 hours ago Up 7 hours (healthy) openvswitch_db
4f33dfa66c14 hubblo/scaphandre:0.5.0 "scaphandre promethe…" 7 hours ago Up 7 hours 10.10.129.64:9155->8080/tcp scaphandre
9b1f6342dc60 quay.io/osism/cron:3.0.20230919 "dumb-init --single-…" 7 hours ago Up 7 hours cron
718aecaddde1 quay.io/osism/kolla-toolbox:16.1.1.20230919 "dumb-init --single-…" 7 hours ago Up 7 hours kolla_toolbox
f6f9422c1853 quay.io/osism/fluentd:4.5.1.20230919 "dumb-init --single-…" 7 hours ago Up 7 hours fluentd

Remove a compute node

  1. In the configuration repository remove the compute node everywhere. Then update the configuration repository on the manager with osism apply configuration

  2. Live migrate all instances running on the compute node with the help of the OpenStack Resource Manager

  3. Evacuate all instances on the compute node with the help of the OpenStack Resource Manager

  4. Ensure that no more instances are running on the compute node

    ps ax | grep qemu
  5. Stop all OpenStack Nova services on the compute node

    systemctl stop kolla-nova_ssh-container.service
    systemctl stop kolla-nova_libvirt-container.service
    systemctl stop kolla-nova_compute-container.service

  6. Delete the compute service

    $ openstack --os-cloud admin compute service list
    +--------------------------------------+----------------+---------+----------+----------+-------+----------------------------+
    | ID | Binary | Host | Zone | Status | State | Updated At |
    +--------------------------------------+----------------+---------+----------+----------+-------+----------------------------+
    | f161d739-21de-4cb0-a5d3-d21cff652697 | nova-scheduler | manager | internal | enabled | up | 2023-12-21T11:52:59.000000 |
    | 646d16db-acd9-486c-bd16-8fe2c13bf198 | nova-conductor | manager | internal | enabled | up | 2023-12-21T11:53:04.000000 |
    | 90345eb5-cf2f-47ef-becc-758ee36fb132 | nova-compute | manager | nova | disabled | down | 2023-12-21T11:53:00.000000 |
    +--------------------------------------+----------------+---------+----------+----------+-------+----------------------------+
    $ openstack --os-cloud admin compute service delete 90345eb5-cf2f-47ef-becc-758ee36fb132
  7. Refresh the facts

    osism apply facts
  8. Refresh the /etc/hosts file

    osism apply hosts
  9. Refresh the SSH client configuration file

    osism apply sshconfig
  10. Remove compute node from Prometheus monitoring

    osism apply prometheus -l monitoring