Browse Source

Re-use existing ComputeNode on ironic rebalance

When a nova-compute service dies that is one of several ironic based
nova-compute services running, a node rebalance occurs to ensure there
is still an active nova-compute service dealing with requests for the
given instance that is running.

Today, when this occurs, we create a new ComputeNode entry. This change
alters that logic to detect the case of the ironic node rebalance and in
that case we re-use the existing ComputeNode entry, simply updating the
host field to match the new host it has been rebalanced onto.

Previously we hit problems with placement when we get a new
ComputeNode.uuid for the same ironic_node.uuid. This reusing of the
existing entry keeps the ComputeNode.uuid the same when the rebalance of
the ComputeNode occurs.

Without keeping the same ComputeNode.uuid placement errors out with a 409
because we attempt to create a ResourceProvider that has the same name
as an existing ResourceProvdier. Had that worked, we would have noticed
the race that occurs after we create the ResourceProvider but before we
add back the existing allocations for existing instances. Keeping the
ComputeNode.uuid the same means we simply look up the existing
ResourceProvider in placement, avoiding all this pain and tears.

Closes-Bug: #1714248
Co-Authored-By: Dmitry Tantsur <dtantsur@redhat.com>
Change-Id: I4253cffca3dbf558c875eed7e77711a31e9e3406
changes/55/508555/8
John Garbutt 4 years ago
committed by Matt Riedemann
parent
commit
e3c5e22d1f
  1. 47
      nova/compute/resource_tracker.py
  2. 86
      nova/tests/unit/compute/test_resource_tracker.py
  3. 5
      nova/virt/driver.py
  4. 5
      nova/virt/fake.py
  5. 3
      nova/virt/ironic/driver.py

47
nova/compute/resource_tracker.py

@ -492,6 +492,50 @@ class ResourceTracker(object):
return (nodename not in self.compute_nodes or
not self.driver.node_is_available(nodename))
def _check_for_nodes_rebalance(self, context, resources, nodename):
"""Check if nodes rebalance has happened.
The ironic driver maintains a hash ring mapping bare metal nodes
to compute nodes. If a compute dies, the hash ring is rebuilt, and
some of its bare metal nodes (more precisely, those not in ACTIVE
state) are assigned to other computes.
This method checks for this condition and adjusts the database
accordingly.
:param context: security context
:param resources: initial values
:param nodename: node name
:returns: True if a suitable compute node record was found, else False
"""
if not self.driver.rebalances_nodes:
return False
# Its possible ironic just did a node re-balance, so let's
# check if there is a compute node that already has the correct
# hypervisor_hostname. We can re-use that rather than create a
# new one and have to move existing placement allocations
cn_candidates = objects.ComputeNodeList.get_by_hypervisor(
context, nodename)
if len(cn_candidates) == 1:
cn = cn_candidates[0]
LOG.info("ComputeNode %(name)s moving from %(old)s to %(new)s",
{"name": nodename, "old": cn.host, "new": self.host})
cn.host = self.host
self.compute_nodes[nodename] = cn
self._copy_resources(cn, resources)
self._setup_pci_tracker(context, cn, resources)
self._update(context, cn)
return True
elif len(cn_candidates) > 1:
LOG.error(
"Found more than one ComputeNode for nodename %s. "
"Please clean up the orphaned ComputeNode records in your DB.",
nodename)
return False
def _init_compute_node(self, context, resources):
"""Initialize the compute node if it does not already exist.
@ -527,6 +571,9 @@ class ResourceTracker(object):
self._update(context, cn)
return
if self._check_for_nodes_rebalance(context, resources, nodename):
return
# there was no local copy and none in the database
# so we need to create a new compute node. This needs
# to be initialized with resource values.

86
nova/tests/unit/compute/test_resource_tracker.py

@ -435,6 +435,7 @@ def setup_rt(hostname, virt_resources=_VIRT_DRIVER_AVAIL_RESOURCES,
vd.get_inventory.side_effect = NotImplementedError
vd.get_host_ip_addr.return_value = _NODENAME
vd.estimate_instance_overhead.side_effect = estimate_overhead
vd.rebalances_nodes = False
with test.nested(
mock.patch('nova.scheduler.client.SchedulerClient',
@ -1012,6 +1013,39 @@ class TestInitComputeNode(BaseTestCase):
self.assertFalse(create_mock.called)
self.assertTrue(update_mock.called)
@mock.patch('nova.objects.ComputeNodeList.get_by_hypervisor')
@mock.patch('nova.objects.PciDeviceList.get_by_compute_node',
return_value=objects.PciDeviceList())
@mock.patch('nova.objects.ComputeNode.create')
@mock.patch('nova.objects.ComputeNode.get_by_host_and_nodename')
@mock.patch('nova.compute.resource_tracker.ResourceTracker.'
'_update')
def test_compute_node_rebalanced(self, update_mock, get_mock, create_mock,
pci_mock, get_by_hypervisor_mock):
self._setup_rt()
self.driver_mock.rebalances_nodes = True
cn = copy.deepcopy(_COMPUTE_NODE_FIXTURES[0])
cn.host = "old-host"
def fake_get_all(_ctx, nodename):
return [cn]
get_mock.side_effect = exc.NotFound
get_by_hypervisor_mock.side_effect = fake_get_all
resources = copy.deepcopy(_VIRT_DRIVER_AVAIL_RESOURCES)
self.rt._init_compute_node(mock.sentinel.ctx, resources)
get_mock.assert_called_once_with(mock.sentinel.ctx, _HOSTNAME,
_NODENAME)
get_by_hypervisor_mock.assert_called_once_with(mock.sentinel.ctx,
_NODENAME)
create_mock.assert_not_called()
update_mock.assert_called_once_with(mock.sentinel.ctx, cn)
self.assertEqual(_HOSTNAME, self.rt.compute_nodes[_NODENAME].host)
@mock.patch('nova.objects.ComputeNodeList.get_by_hypervisor')
@mock.patch('nova.objects.PciDeviceList.get_by_compute_node',
return_value=objects.PciDeviceList(objects=[]))
@mock.patch('nova.objects.ComputeNode.create')
@ -1019,10 +1053,55 @@ class TestInitComputeNode(BaseTestCase):
@mock.patch('nova.compute.resource_tracker.ResourceTracker.'
'_update')
def test_compute_node_created_on_empty(self, update_mock, get_mock,
create_mock, pci_tracker_mock):
create_mock, pci_tracker_mock,
get_by_hypervisor_mock):
get_by_hypervisor_mock.return_value = []
self._test_compute_node_created(update_mock, get_mock, create_mock,
pci_tracker_mock,
get_by_hypervisor_mock)
@mock.patch('nova.objects.ComputeNodeList.get_by_hypervisor')
@mock.patch('nova.objects.PciDeviceList.get_by_compute_node',
return_value=objects.PciDeviceList(objects=[]))
@mock.patch('nova.objects.ComputeNode.create')
@mock.patch('nova.objects.ComputeNode.get_by_host_and_nodename')
@mock.patch('nova.compute.resource_tracker.ResourceTracker.'
'_update')
def test_compute_node_created_on_empty_rebalance(self, update_mock,
get_mock,
create_mock,
pci_tracker_mock,
get_by_hypervisor_mock):
get_by_hypervisor_mock.return_value = []
self._test_compute_node_created(update_mock, get_mock, create_mock,
pci_tracker_mock,
get_by_hypervisor_mock,
rebalances_nodes=True)
@mock.patch('nova.objects.ComputeNodeList.get_by_hypervisor')
@mock.patch('nova.objects.PciDeviceList.get_by_compute_node',
return_value=objects.PciDeviceList(objects=[]))
@mock.patch('nova.objects.ComputeNode.create')
@mock.patch('nova.objects.ComputeNode.get_by_host_and_nodename')
@mock.patch('nova.compute.resource_tracker.ResourceTracker.'
'_update')
def test_compute_node_created_too_many(self, update_mock, get_mock,
create_mock, pci_tracker_mock,
get_by_hypervisor_mock):
get_by_hypervisor_mock.return_value = ["fake_node_1", "fake_node_2"]
self._test_compute_node_created(update_mock, get_mock, create_mock,
pci_tracker_mock,
get_by_hypervisor_mock,
rebalances_nodes=True)
def _test_compute_node_created(self, update_mock, get_mock,
create_mock, pci_tracker_mock,
get_by_hypervisor_mock,
rebalances_nodes=False):
self.flags(cpu_allocation_ratio=1.0, ram_allocation_ratio=1.0,
disk_allocation_ratio=1.0)
self._setup_rt()
self.driver_mock.rebalances_nodes = rebalances_nodes
get_mock.side_effect = exc.NotFound
@ -1088,6 +1167,11 @@ class TestInitComputeNode(BaseTestCase):
cn = self.rt.compute_nodes[_NODENAME]
get_mock.assert_called_once_with(mock.sentinel.ctx, _HOSTNAME,
_NODENAME)
if rebalances_nodes:
get_by_hypervisor_mock.assert_called_once_with(
mock.sentinel.ctx, _NODENAME)
else:
get_by_hypervisor_mock.assert_not_called()
create_mock.assert_called_once_with()
self.assertTrue(obj_base.obj_equal_prims(expected_compute, cn))
pci_tracker_mock.assert_called_once_with(mock.sentinel.ctx,

5
nova/virt/driver.py

@ -134,6 +134,11 @@ class ComputeDriver(object):
requires_allocation_refresh = False
# Indicates if this driver will rebalance nodes among compute service
# hosts. This is really here for ironic and should not be used by any
# other driver.
rebalances_nodes = False
def __init__(self, virtapi):
self.virtapi = virtapi
self._compute_event_callback = None

5
nova/virt/fake.py

@ -161,9 +161,12 @@ class FakeDriver(driver.ComputeDriver):
self._mounts = {}
self._interfaces = {}
self.active_migrations = {}
self._nodes = self._init_nodes()
def _init_nodes(self):
if not _FAKE_NODES:
set_nodes([CONF.host])
self._nodes = copy.copy(_FAKE_NODES)
return copy.copy(_FAKE_NODES)
def init_host(self, host):
return

3
nova/virt/ironic/driver.py

@ -141,6 +141,9 @@ class IronicDriver(virt_driver.ComputeDriver):
# migration has been completed.
requires_allocation_refresh = True
# This driver is capable of rebalancing nodes between computes.
rebalances_nodes = True
def __init__(self, virtapi, read_only=False):
super(IronicDriver, self).__init__(virtapi)
global ironic

Loading…
Cancel
Save