Doc Text:
|
Previously, a race condition occurred during the initialization of the neutron database when neutron-server was first run. This error was seen when two controllers happened to start neutron-server simultaneously. Subsequently, the startup of neutron-server and agents failed on the controller node that lost the race, and as a consequence, Neutron services failed to start on the affected controller nodes. Errors in the logs look like the following:
DBDuplicateEntry: (IntegrityError) (1062, "Duplicate entry 'datacentre-1' for key 'PRIMARY'") 'INSERT INTO ml2_vlan_allocations (physical_network, vlan_id, allocated) VALUES (%s, %s, %s)' (('datacentre', 1, 0),
With this release, the Neutron server is momentarily started and then stopped on one node, the pacemaker master, allowing this initial database setup to happen, before allowing the rest of the puppet or pacemaker configuration to happen. As a result, Neutron services are brought up on all controllers nodes without error.
|