nova-specs/specs/newton/approved/hypervisor-notification.rst

Add notifications for hypervisor

https://blueprints.launchpad.net/nova/+spec/hypervisor-notification

Currently, no notification will be sent for compute node state change, so it is not possible for external system to get notifications when there are compute nodes created, updated or deleted.

Having such notifications help external system to get the up to date compute node status and metrics.

Problem description

Use Cases

The external system like Searchlight[1] wants to index the compute nodes which makes the query for large number of compute nodes more fast and efficient.

The maintainer wants to get the notifications when there are compute nodes added or removed.

Proposed change

Versioned notifications will be added for the following actions:

• ComputeNode.create
• ComputeNode.save
• ComputeNode.destroy

Note

Notification will be sent only if any of the specified fields( "vcpus", "memory_mb", "local_gb", "vcpus_used", "local_gb_used", "hypervisor_hostname", "disk_available_least", "running_vms", "current_workload") changed to avoid unnecessary notifications with the same content.

Alternatives

None

Data model impact

No database schema change is needed.

The following new objects will be added to compute_node for create and update:

@base.NovaObjectRegistry.register
class ComputeNodeNotification(notification.NotificationBase):
    # Version 1.0: Initial version
    VERSION = '1.0'
    fields = {
        'payload': fields.ObjectField('ComputeNodePayload')
    }

@base.NovaObjectRegistry.register
class ComputeNodePayload(notification.NotificationPayloadBase):
    # Version 1.0: Initial version
    SCHEMA = {
        'id': ('compute_node', 'id'),
        'uuid': ('compute_node', 'uuid'),
        'host': ('compute_node', 'host'),
        'vcpus': ('compute_node', 'vcpus'),
        'memory_mb': ('compute_node', 'memory_mb'),
        'local_gb': ('compute_node', 'local_gb'),
        'vcpus_used': ('compute_node', 'vcpus_used'),
        'memory_mb_used': ('compute_node', 'memory_mb_used'),
        'local_gb_used': ('compute_node', 'local_gb_used'),
        'hypervisor_type': ('compute_node', 'hypervisor_type'),
        'hypervisor_version': ('compute_node', 'hypervisor_version'),
        'hypervisor_hostname': ('compute_node', 'hypervisor_hostname'),
        'free_ram_mb': ('compute_node', 'free_ram_mb'),
        'free_disk_gb': ('compute_node', 'free_disk_gb'),
        'current_workload': ('compute_node', 'current_workload'),
        'running_vms': ('compute_node', 'running_vms'),
        'disk_available_least': ('compute_node', 'disk_available_least'),
        'host_ip': ('compute_node', 'host_ip'),
    }
    VERSION = '1.0'
    fields = {
        'id': fields.IntegerField(),
        'uuid': fields.UUIDField(),
        'host': fields.StringField(nullable=True),
        'vcpus': fields.IntegerField(),
        'memory_mb': fields.IntegerField(),
        'local_gb': fields.IntegerField(),
        'vcpus_used': fields.IntegerField(),
        'memory_mb_used': fields.IntegerField(),
        'local_gb_used': fields.IntegerField(),
        'hypervisor_type': fields.StringField(),
        'hypervisor_version': fields.IntegerField(),
        'hypervisor_hostname': fields.StringField(nullable=True),
        'free_ram_mb': fields.IntegerField(nullable=True),
        'free_disk_gb': fields.IntegerField(nullable=True),
        'current_workload': fields.IntegerField(nullable=True),
        'running_vms': fields.IntegerField(nullable=True),
        'disk_available_least': fields.IntegerField(nullable=True),
        'host_ip': fields.IPAddressField(nullable=True),
    }
    def __init__(self, compute_node):
        super(ComputeNodePayload, self).__init__()
        self.populate_schema(compute_node=compute_node)

The following new objects will be added to compute_node for delete:

@base.NovaObjectRegistry.register
class ComputeNodeDeleteNotification(notification.NotificationBase):
    # Version 1.0: Initial version
    VERSION = '1.0'
    fields = {
        'payload': fields.ObjectField('ComputeNodeDeletePayload')
    }

@base.NovaObjectRegistry.register
class ComputeNodeDeletePayload(notification.NotificationPayloadBase):
    # Version 1.0: Initial version
    SCHEMA = {
        'id': ('compute_node', 'id'),
        'uuid': ('compute_node', 'uuid'),
    }
    VERSION = '1.0'
    fields = {
        'id': fields.IntegerField(),
        'uuid': fields.UUIDField(),
    }
    def __init__(self, compute_node):
        super(ComputeNodeDeletePayload, self).__init__()
        self.populate_schema(compute_node=compute_node)

The definition of NotificationBase can be found [2].

REST API impact

None

Security impact

None

Notifications impact

New notifications compute_node.create (will be sent after a compute node created), compute_node.update (will be sent after the non static fields of a compute node updated) and compute_node.delete (will be sent after a compute node deleted) will be introduced with INFO priority and payload of the notifications will be the serialized form of the already existing ComputeNode versioned object.

Other end user impact

None

Performance Impact

None

Other deployer impact

None

Developer impact

None

Implementation

Assignee(s)

Primary assignee:

liyingjun

Work Items

• Send new notifications when a compute node created, updated or deleted.

Dependencies

None

Testing

Besides unit test new functional test cases will be added to cover the new notifications.

Documentation Impact

None

References

[1]: Searchlight: http://docs.openstack.org/developer/searchlight/index.html [2]: Versioned notification: http://docs.openstack.org/developer/nova/notifications.html#versioned-notifications

History

Revisions

Release Name	Description
Newton	Introduced