466 lines
15 KiB
Python
466 lines
15 KiB
Python
# Copyright (c) 2013 Mirantis Inc.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
# implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import datetime
|
|
|
|
from novaclient import exceptions as nova_exceptions
|
|
from oslo.config import cfg
|
|
import six
|
|
|
|
from savanna import conductor as c
|
|
from savanna import context
|
|
from savanna.openstack.common import excutils
|
|
from savanna.openstack.common import log as logging
|
|
from savanna.service import networks
|
|
from savanna.service import volumes
|
|
from savanna.utils import general as g
|
|
from savanna.utils.openstack import nova
|
|
|
|
|
|
conductor = c.API
|
|
CONF = cfg.CONF
|
|
LOG = logging.getLogger(__name__)
|
|
|
|
|
|
def create_cluster(cluster):
|
|
ctx = context.ctx()
|
|
try:
|
|
# create all instances
|
|
conductor.cluster_update(ctx, cluster, {"status": "Spawning"})
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
_create_instances(cluster)
|
|
|
|
# wait for all instances are up and networks ready
|
|
cluster = conductor.cluster_update(ctx, cluster, {"status": "Waiting"})
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
|
|
instances = get_instances(cluster)
|
|
|
|
_await_active(instances)
|
|
|
|
_assign_floating_ips(instances)
|
|
|
|
_await_networks(instances)
|
|
|
|
cluster = conductor.cluster_get(ctx, cluster)
|
|
|
|
# attach volumes
|
|
volumes.attach(cluster)
|
|
|
|
# prepare all instances
|
|
cluster = conductor.cluster_update(ctx, cluster,
|
|
{"status": "Preparing"})
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
|
|
_configure_instances(cluster)
|
|
except Exception as ex:
|
|
LOG.warn("Can't start cluster '%s' (reason: %s)", cluster.name, ex)
|
|
with excutils.save_and_reraise_exception():
|
|
cluster = conductor.cluster_update(ctx, cluster,
|
|
{"status": "Error",
|
|
"status_description": str(ex)})
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
_rollback_cluster_creation(cluster, ex)
|
|
|
|
|
|
def get_instances(cluster, instances_ids=None):
|
|
inst_map = {}
|
|
for node_group in cluster.node_groups:
|
|
for instance in node_group.instances:
|
|
inst_map[instance.id] = instance
|
|
|
|
if instances_ids is not None:
|
|
return [inst_map[id] for id in instances_ids]
|
|
else:
|
|
return [v for v in six.itervalues(inst_map)]
|
|
|
|
|
|
def scale_cluster(cluster, node_group_id_map, plugin):
|
|
ctx = context.ctx()
|
|
|
|
instance_ids = []
|
|
try:
|
|
instance_ids = _scale_cluster_instances(
|
|
cluster, node_group_id_map, plugin)
|
|
|
|
cluster = conductor.cluster_get(ctx, cluster)
|
|
cluster = clean_cluster_from_empty_ng(cluster)
|
|
|
|
instances = get_instances(cluster, instance_ids)
|
|
|
|
_await_active(instances)
|
|
|
|
_assign_floating_ips(instances)
|
|
|
|
_await_networks(instances)
|
|
|
|
cluster = conductor.cluster_get(context, cluster)
|
|
|
|
volumes.attach_to_instances(get_instances(cluster, instance_ids))
|
|
|
|
except Exception as ex:
|
|
LOG.warn("Can't scale cluster '%s' (reason: %s)", cluster.name, ex)
|
|
with excutils.save_and_reraise_exception():
|
|
cluster = conductor.cluster_get(ctx, cluster)
|
|
_rollback_cluster_scaling(cluster,
|
|
get_instances(cluster, instance_ids),
|
|
ex)
|
|
instance_ids = []
|
|
|
|
cluster = conductor.cluster_get(ctx, cluster)
|
|
clean_cluster_from_empty_ng(cluster)
|
|
if cluster.status == 'Decommissioning':
|
|
cluster = conductor.cluster_update(ctx, cluster,
|
|
{"status": "Error"})
|
|
else:
|
|
cluster = conductor.cluster_update(ctx, cluster,
|
|
{"status": "Active"})
|
|
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
|
|
# we should be here with valid cluster: if instances creation
|
|
# was not successful all extra-instances will be removed above
|
|
if instance_ids:
|
|
_configure_instances(cluster)
|
|
return instance_ids
|
|
|
|
|
|
def _generate_anti_affinity_groups(cluster):
|
|
aa_groups = {}
|
|
|
|
for node_group in cluster.node_groups:
|
|
for instance in node_group.instances:
|
|
if instance.instance_id:
|
|
for process in node_group.node_processes:
|
|
if process in cluster.anti_affinity:
|
|
aa_group = aa_groups.get(process, [])
|
|
aa_group.append(instance.instance_id)
|
|
aa_groups[process] = aa_group
|
|
|
|
return aa_groups
|
|
|
|
|
|
def _create_instances(cluster):
|
|
ctx = context.ctx()
|
|
|
|
#aa_groups = _generate_anti_affinity_groups(cluster)
|
|
aa_groups = {}
|
|
for node_group in cluster.node_groups:
|
|
count = node_group.count
|
|
conductor.node_group_update(ctx, node_group, {'count': 0})
|
|
userdata = _generate_user_data_script(node_group)
|
|
for idx in xrange(1, count + 1):
|
|
_run_instance(cluster, node_group, idx, aa_groups, userdata)
|
|
|
|
|
|
def _scale_cluster_instances(cluster, node_group_id_map, plugin):
|
|
ctx = context.ctx()
|
|
aa_groups = _generate_anti_affinity_groups(cluster)
|
|
instances_to_delete = []
|
|
node_groups_to_enlarge = []
|
|
|
|
for node_group in cluster.node_groups:
|
|
if node_group.id not in node_group_id_map:
|
|
continue
|
|
|
|
new_count = node_group_id_map[node_group.id]
|
|
if new_count < node_group.count:
|
|
instances_to_delete += node_group.instances[new_count:
|
|
node_group.count]
|
|
else:
|
|
node_groups_to_enlarge.append(node_group)
|
|
|
|
if instances_to_delete:
|
|
conductor.cluster_update(ctx, cluster, {"status": "Decommissioning"})
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
plugin.decommission_nodes(cluster, instances_to_delete)
|
|
cluster = conductor.cluster_update(ctx, cluster,
|
|
{"status": "Deleting Instances"})
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
for instance in instances_to_delete:
|
|
_shutdown_instance(instance)
|
|
|
|
cluster = conductor.cluster_get(ctx, cluster)
|
|
|
|
instances_to_add = []
|
|
if node_groups_to_enlarge:
|
|
cluster = conductor.cluster_update(ctx, cluster,
|
|
{"status": "Adding Instances"})
|
|
LOG.info(g.format_cluster_status(cluster))
|
|
for node_group in node_groups_to_enlarge:
|
|
count = node_group_id_map[node_group.id]
|
|
userdata = _generate_user_data_script(node_group)
|
|
for idx in xrange(node_group.count + 1, count + 1):
|
|
instance_id = _run_instance(cluster, node_group, idx,
|
|
aa_groups, userdata)
|
|
instances_to_add.append(instance_id)
|
|
|
|
return instances_to_add
|
|
|
|
|
|
def _find_by_id(lst, id):
|
|
for obj in lst:
|
|
if obj.id == id:
|
|
return obj
|
|
|
|
return None
|
|
|
|
|
|
def _run_instance(cluster, node_group, idx, aa_groups, userdata):
|
|
"""Create instance using nova client and persist them into DB."""
|
|
ctx = context.ctx()
|
|
name = '%s-%s-%03d' % (cluster.name, node_group.name, idx)
|
|
|
|
# aa_groups: node process -> instance ids
|
|
aa_ids = []
|
|
for node_process in node_group.node_processes:
|
|
aa_ids += aa_groups.get(node_process) or []
|
|
|
|
# create instances only at hosts w/ no instances w/ aa-enabled processes
|
|
hints = {'different_host': list(set(aa_ids))} if aa_ids else None
|
|
|
|
if CONF.use_neutron:
|
|
net_id = cluster.neutron_management_network
|
|
nics = [{"net-id": net_id, "v4-fixed-ip": ""}]
|
|
|
|
nova_instance = nova.client().servers.create(
|
|
name, node_group.get_image_id(), node_group.flavor_id,
|
|
scheduler_hints=hints, userdata=userdata,
|
|
key_name=cluster.user_keypair_id,
|
|
nics=nics)
|
|
else:
|
|
nova_instance = nova.client().servers.create(
|
|
name, node_group.get_image_id(), node_group.flavor_id,
|
|
scheduler_hints=hints, userdata=userdata,
|
|
key_name=cluster.user_keypair_id)
|
|
|
|
instance_id = conductor.instance_add(ctx, node_group,
|
|
{"instance_id": nova_instance.id,
|
|
"instance_name": name})
|
|
# save instance id to aa_groups to support aa feature
|
|
for node_process in node_group.node_processes:
|
|
if node_process in cluster.anti_affinity:
|
|
aa_group_ids = aa_groups.get(node_process, [])
|
|
aa_group_ids.append(nova_instance.id)
|
|
aa_groups[node_process] = aa_group_ids
|
|
|
|
return instance_id
|
|
|
|
|
|
def _generate_user_data_script(node_group):
|
|
script_template = """#!/bin/bash
|
|
echo "%(public_key)s" >> %(user_home)s/.ssh/authorized_keys
|
|
echo "%(private_key)s" > %(user_home)s/.ssh/id_rsa
|
|
"""
|
|
cluster = node_group.cluster
|
|
if nova.get_node_group_image_username(node_group) == "root":
|
|
user_home = "/root/"
|
|
else:
|
|
user_home = "/home/%s/" % nova.get_node_group_image_username(
|
|
node_group)
|
|
|
|
return script_template % {
|
|
"public_key": cluster.management_public_key,
|
|
"private_key": cluster.management_private_key,
|
|
"user_home": user_home
|
|
}
|
|
|
|
|
|
def _assign_floating_ips(instances):
|
|
for instance in instances:
|
|
node_group = instance.node_group
|
|
if node_group.floating_ip_pool:
|
|
networks.assign_floating_ip(instance.instance_id,
|
|
node_group.floating_ip_pool)
|
|
|
|
|
|
def _check_cluster_exists(cluster):
|
|
ctx = context.ctx()
|
|
# check if cluster still exists (it might have been removed)
|
|
cluster = conductor.cluster_get(ctx, cluster)
|
|
return cluster is not None
|
|
|
|
|
|
def _await_networks(instances):
|
|
if not instances:
|
|
return
|
|
|
|
ips_assigned = set()
|
|
while len(ips_assigned) != len(instances):
|
|
if not _check_cluster_exists(instances[0].node_group.cluster):
|
|
return
|
|
for instance in instances:
|
|
if instance.id not in ips_assigned:
|
|
if networks.init_instances_ips(instance):
|
|
ips_assigned.add(instance.id)
|
|
|
|
context.sleep(1)
|
|
|
|
ctx = context.ctx()
|
|
cluster = conductor.cluster_get(ctx, instances[0].node_group.cluster)
|
|
instances = get_instances(cluster, ips_assigned)
|
|
|
|
accessible_instances = set()
|
|
while len(accessible_instances) != len(instances):
|
|
if not _check_cluster_exists(instances[0].node_group.cluster):
|
|
return
|
|
for instance in instances:
|
|
if instance.id not in accessible_instances:
|
|
if _check_if_accessible(instance):
|
|
accessible_instances.add(instance.id)
|
|
|
|
context.sleep(1)
|
|
|
|
|
|
def _await_active(instances):
|
|
"""Await all instances are in Active status and available."""
|
|
if not instances:
|
|
return
|
|
|
|
active_ids = set()
|
|
while len(active_ids) != len(instances):
|
|
if not _check_cluster_exists(instances[0].node_group.cluster):
|
|
return
|
|
for instance in instances:
|
|
if instance.id not in active_ids:
|
|
if _check_if_active(instance):
|
|
active_ids.add(instance.id)
|
|
|
|
context.sleep(1)
|
|
|
|
|
|
def _check_if_active(instance):
|
|
|
|
server = nova.get_instance_info(instance)
|
|
if server.status == 'ERROR':
|
|
# TODO(slukjanov): replace with specific error
|
|
raise RuntimeError("node %s has error status" % server.name)
|
|
|
|
return server.status == 'ACTIVE'
|
|
|
|
|
|
def _check_if_accessible(instance):
|
|
|
|
if not instance.management_ip:
|
|
return False
|
|
|
|
try:
|
|
# check if ssh is accessible and cloud-init
|
|
# script is finished generating id_rsa
|
|
exit_code, stdout = instance.remote.execute_command(
|
|
"ls .ssh/id_rsa", raise_when_error=False)
|
|
|
|
if exit_code:
|
|
return False
|
|
except Exception as ex:
|
|
LOG.debug("Can't login to node %s (%s), reason %s",
|
|
instance.instance_name, instance.management_ip, ex)
|
|
return False
|
|
|
|
LOG.debug('Instance %s is accessible' % instance.instance_name)
|
|
return True
|
|
|
|
|
|
def _configure_instances(cluster):
|
|
"""Configure active instances.
|
|
|
|
* generate /etc/hosts
|
|
* setup passwordless login
|
|
* etc.
|
|
"""
|
|
hosts = _generate_etc_hosts(cluster)
|
|
for node_group in cluster.node_groups:
|
|
for instance in node_group.instances:
|
|
LOG.debug('Configuring instance %s' % instance.instance_name)
|
|
with instance.remote as r:
|
|
r.write_file_to('etc-hosts', hosts)
|
|
r.execute_command('sudo mv etc-hosts /etc/hosts')
|
|
|
|
r.execute_command('sudo chown $USER:$USER .ssh/id_rsa')
|
|
r.execute_command('chmod 400 .ssh/id_rsa')
|
|
|
|
|
|
def _generate_etc_hosts(cluster):
|
|
hosts = "127.0.0.1 localhost\n"
|
|
for node_group in cluster.node_groups:
|
|
for instance in node_group.instances:
|
|
hosts += "%s %s %s\n" % (instance.internal_ip,
|
|
instance.fqdn,
|
|
instance.hostname)
|
|
|
|
return hosts
|
|
|
|
|
|
def _rollback_cluster_creation(cluster, ex):
|
|
"""Shutdown all instances and update cluster status."""
|
|
LOG.info("Cluster '%s' creation rollback (reason: %s)", cluster.name, ex)
|
|
|
|
shutdown_cluster(cluster)
|
|
|
|
|
|
def _rollback_cluster_scaling(cluster, instances, ex):
|
|
"""Attempt to rollback cluster scaling."""
|
|
LOG.info("Cluster '%s' scaling rollback (reason: %s)", cluster.name, ex)
|
|
try:
|
|
volumes.detach_from_instances(instances)
|
|
finally:
|
|
for i in instances:
|
|
_shutdown_instance(i)
|
|
|
|
|
|
def _clean_job_executions(cluster):
|
|
ctx = context.ctx()
|
|
for je in conductor.job_execution_get_all(ctx, cluster_id=cluster.id):
|
|
update = {"cluster_id": None,
|
|
"end_time": datetime.datetime.now()}
|
|
conductor.job_execution_update(ctx, je, update)
|
|
|
|
|
|
def _shutdown_instances(cluster):
|
|
for node_group in cluster.node_groups:
|
|
for instance in node_group.instances:
|
|
_shutdown_instance(instance)
|
|
|
|
|
|
def _shutdown_instance(instance):
|
|
ctx = context.ctx()
|
|
try:
|
|
if instance.node_group.floating_ip_pool:
|
|
networks.delete_floating_ip(instance.instance_id)
|
|
nova.client().servers.delete(instance.instance_id)
|
|
except nova_exceptions.NotFound:
|
|
#Just ignore non-existing instances
|
|
pass
|
|
|
|
conductor.instance_remove(ctx, instance)
|
|
|
|
|
|
def shutdown_cluster(cluster):
|
|
"""Shutdown specified cluster and all related resources."""
|
|
try:
|
|
volumes.detach(cluster)
|
|
finally:
|
|
_shutdown_instances(cluster)
|
|
_clean_job_executions(cluster)
|
|
|
|
|
|
def clean_cluster_from_empty_ng(cluster):
|
|
ctx = context.ctx()
|
|
for ng in cluster.node_groups:
|
|
if ng.count == 0:
|
|
conductor.node_group_remove(ctx, ng)
|
|
|
|
return conductor.cluster_get(ctx, cluster)
|