Shulou(Shulou.com)05/31 Report--

In this issue, the editor will bring you about how to create instance for openstack e version. The article is rich in content and analyzes and narrates it from a professional point of view. I hope you can get something after reading this article.

1. Horizon module:

The lanch of the tables.py is mapped to the LaunchForm of the forms.py through url, and then the parameters are passed in the page. Then call the server_create of nova.py in the api module,

2 、 python-novaclient

Then call create () in the ServerManager class of the python-novaclient module V1.1/servers.py, and _ boot () in the calling base.py to encapsulate the url request.

3. Nova module:

The release module for wsgi is in the nova/api/openstack/compute/__init__.py file. The corresponding relationship of url address after release is as follows:

Self._list () corresponds to the underlying index () function

Self._create () corresponds to the underlying create () function

Self._delete () corresponds to the underlying delete () function

Self._get () corresponds to the underlying show () function

The url of create corresponds to the create () in nova/api/openstack/compute/servers.py

In this function, get parameters such as image, network, flavor, etc., then call self.compute_api.create and then call self._schedule_run_instance, send the run_instance method of rpc_method in message queue, and call the run_instance of manager.py in scheduler module according to FLAGS.scheduler_topic configuration. The configuration file scheduler_driver= FLAGS.scheduler_driver, where the production environment uses-- scheduler_driver=nova.scheduler.simple.SimpleScheduler, so: the schedule_run_instance method in simple.py.

Driver.cast_to_compute_host sends messages to message queues.

The run_instance method of compute/manage.py, _ run_instance, _ spawn and then configure the default='nova.virt.connection.get_connection' of the file compute_driver = FLAGS.compute_driver

In get_connection, t = 'libvirt': conn = libvirt_conn.get_connection (read_only) followed by from nova.virt.libvirt import connection as libvirt_conn

So call the spawn function in libvirt and generate the xml file.

1. Horizon module call process

2. Novaclient module call process

Note: the sentence resp, body= self.api.client.post (url, body=body) is to send a post request. The wsgi in the nova,nova can automatically execute the corresponding function according to this url match.

3. Nova module call process

Note: nova's wsgi releases are all under the / nova/api/openstack/compute folder.

The available_zone parameter is not passed on.

The following statement indicates that scheduler is called

-- scheduler_driver=nova.scheduler.simple.SimpleScheduler is used in the production environment here, so:

Here is the simple scheduling algorithm:

Def _ schedule_instance (self, context, instance_opts, * _ args, * * _ kwargs):

"Picks a host that is up and has the fewest running instances."

Elevated = context.elevated ()

Availability_zone = instance_opts.get ('availability_zone')

Zone, host = FLAGS.default_schedule_zone, None

If availability_zone:

Zone, _ x, host = availability_zone.partition (':')

If host and context.is_admin:

Service = db.service_get_by_args (elevated, host, 'nova-compute')

If not utils.service_is_up (service):

Raise exception.WillNotSchedule (host=host)

Return host

Results = db.service_get_all_compute_sorted (elevated) / / get the host with the least resource utilization

In_isolation = instance_opts ['image_ref'] in FLAGS.isolated_images

Check_cores = not in_isolation or not FLAGS.skip_isolated_core_check

If zone:

Results = [(service, cores) for (service, cores) in results

If service ['availability_zone'] = = zone] / / the availability_zone here is passed to the page

Requested_networks = _ kwargs ['requested_networks']

If requested_networks:

Results = [(service, cores) for (service, cores) in results

If service ['network_ref'] in [network_uuid for (network_uuid, fixed_ip) in requested_networks] / / the requested_networks here is passed to the page

For result in results:

(service, instance_cores) = result

If in_isolation and service ['host'] not in FLAGS.isolated_hosts:

# isloated images run on isolated hosts

Continue

If service ['host'] in FLAGS.isolated_hosts and not in_isolation:

# images that aren't isolated only run on general hosts

Continue

If (check_cores and

Instance_cores + instance_opts ['vcpus'] > FLAGS.max_cores):

Msg = _ ("Not enough allocatable CPU cores remaining")

Raise exception.NoValidHost (reason=msg)

If not self._isenough_subnet (elevated, service):

Continue

If utils.service_is_up (service) and not service ['disabled']:

Return service ['host']

Msg = _ ("Is the appropriate service running?")

Raise exception.NoValidHost (reason=msg)

This is how the openstack e version shared by the editor created instance. If you happen to have similar doubts, you might as well refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.