def _get_host_k8s_cgroup_config(self, host):
config = {}
# determine set of all logical cpus and nodes
host_cpus = self._get_host_cpu_list(host, threads=True)
host_cpuset = set([c.cpu for c in host_cpus])
host_nodeset = set([c.numa_node for c in host_cpus])
# determine set of platform logical cpus and nodes
platform_cpus = self._get_host_cpu_list(
host, function=constants.PLATFORM_FUNCTION, threads=True)
platform_cpuset = set([c.cpu for c in platform_cpus])
platform_nodeset = set([c.numa_node for c in platform_cpus])
vswitch_cpus = self._get_host_cpu_list(
host, function=constants.VSWITCH_FUNCTION, threads=True)
vswitch_cpuset = set([c.cpu for c in vswitch_cpus])
# determine set of isolcpus logical cpus and nodes
isol_cpus = self._get_host_cpu_list(
host, function=constants.ISOLATED_FUNCTION, threads=True)
isol_cpuset = set([c.cpu for c in isol_cpus])
# determine reserved sets of logical cpus in a string range set format
# to pass as options to kubelet
k8s_platform_cpuset = utils.format_range_set(platform_cpuset)
k8s_all_reserved_cpuset = utils.format_range_set(platform_cpuset
| vswitch_cpuset
| isol_cpuset)
# determine platform reserved memory
k8s_reserved_mem = 0
host_memory = self.dbapi.imemory_get_by_ihost(host.id)
numa_memory = utils.get_numa_index_list(host_memory)
for node, memory in numa_memory.items():
reserved_mib = memory[0].platform_reserved_mib
if reserved_mib is not None:
k8s_reserved_mem += reserved_mib
# determine set of nonplatform logical cpus
# TODO(jgauld): Commented out for now, using host_cpuset instead.
# nonplatform_cpuset = host_cpuset - platform_cpuset
if constants.WORKER in utils.get_personalities(host):
if self.is_openstack_compute(host):
k8s_cpuset = utils.format_range_set(platform_cpuset)
k8s_nodeset = utils.format_range_set(platform_nodeset)
else:
# kubelet cpumanager is configured with static policy.
# The resulting DefaultCPUSet excludes reserved cpus
# based on topology, and that also happens to correspond
# to the platform_cpuset. kubepods are allowed to
# span all host numa nodes.
# TODO(jgauld): Temporary workaround until we have a version
# of kubelet that excludes reserved cpus from DefaultCPUSet.
# The intent is to base k8s_cpuset on nonplatform_cpuset.
# Commented out for now, using host_cpuset instead.
# k8s_cpuset = utils.format_range_set(nonplatform_cpuset)
k8s_cpuset = utils.format_range_set(host_cpuset)
k8s_nodeset = utils.format_range_set(host_nodeset)
else:
k8s_cpuset = utils.format_range_set(host_cpuset)
k8s_nodeset = utils.format_range_set(host_nodeset)
LOG.debug('host:%s, k8s_cpuset:%s, k8s_nodeset:%s', host.hostname,
k8s_cpuset, k8s_nodeset)
# determine cpu/topology mgr policies
labels = self.dbapi.label_get_by_host(host.uuid)
for label in labels:
if label.label_key == constants.KUBE_TOPOLOGY_MANAGER_LABEL:
config.update({
'platform::kubernetes::params::k8s_topology_mgr_policy':
label.label_value
})
elif label.label_key == constants.KUBE_CPU_MANAGER_LABEL:
config.update({
'platform::kubernetes::params::k8s_cpu_mgr_policy':
label.label_value
})
config.update({
'platform::kubernetes::params::k8s_cpuset':
"\"%s\"" % k8s_cpuset,
'platform::kubernetes::params::k8s_nodeset':
"\"%s\"" % k8s_nodeset,
'platform::kubernetes::params::k8s_platform_cpuset':
"\"%s\"" % k8s_platform_cpuset,
'platform::kubernetes::params::k8s_all_reserved_cpuset':
"\"%s\"" % k8s_all_reserved_cpuset,
'platform::kubernetes::params::k8s_reserved_mem':
k8s_reserved_mem,
})
return config
def _get_host_memory_config(self, host):
config = {}
if constants.WORKER in utils.get_personalities(host):
host_memory = self.dbapi.imemory_get_by_ihost(host.id)
memory_numa_list = utils.get_numa_index_list(host_memory)
platform_cpus_no_threads = self._get_platform_cpu_list(host)
platform_core_count = len(platform_cpus_no_threads)
platform_nodes = []
vswitch_nodes = []
hugepages_2Ms = []
hugepages_1Gs = []
vswitch_2M_pages = []
vswitch_1G_pages = []
vm_4K_pages = []
vm_2M_pages = []
vm_1G_pages = []
vs_pages_updated = False
for node, memory_list in memory_numa_list.items():
memory = memory_list[0]
vswitch_2M_page = 0
vswitch_1G_page = 0
platform_size = memory.platform_reserved_mib
platform_node = "\"node%d:%dMB:%d\"" % (
node, platform_size, platform_core_count)
platform_nodes.append(platform_node)
vswitch_size = memory.vswitch_hugepages_size_mib
vswitch_pages = memory.vswitch_hugepages_reqd \
if memory.vswitch_hugepages_reqd is not None \
else memory.vswitch_hugepages_nr
if vswitch_pages == 0:
vswitch_pages = memory.vswitch_hugepages_nr
vswitch_node = "\"node%d:%dkB:%d\"" % (
node, vswitch_size * 1024, vswitch_pages)
vswitch_nodes.append(vswitch_node)
vm_hugepages_nr_2M = memory.vm_hugepages_nr_2M_pending \
if memory.vm_hugepages_nr_2M_pending is not None \
else memory.vm_hugepages_nr_2M
vm_hugepages_nr_1G = memory.vm_hugepages_nr_1G_pending \
if memory.vm_hugepages_nr_1G_pending is not None \
else memory.vm_hugepages_nr_1G
vm_hugepages_nr_4K = memory.vm_hugepages_nr_4K \
if memory.vm_hugepages_nr_4K is not None else 0
total_hugepages_2M = vm_hugepages_nr_2M
total_hugepages_1G = vm_hugepages_nr_1G
if memory.vswitch_hugepages_size_mib == constants.MIB_2M:
total_hugepages_2M += vswitch_pages
vswitch_2M_page += vswitch_pages
elif memory.vswitch_hugepages_size_mib == constants.MIB_1G:
total_hugepages_1G += vswitch_pages
vswitch_1G_page += vswitch_pages
vswitch_2M_pages.append(vswitch_2M_page)
vswitch_1G_pages.append(vswitch_1G_page)
hugepages_2M = "\"node%d:%dkB:%d\"" % (
node, constants.MIB_2M * 1024, total_hugepages_2M)
hugepages_1G = "\"node%d:%dkB:%d\"" % (
node, constants.MIB_1G * 1024, total_hugepages_1G)
hugepages_2Ms.append(hugepages_2M)
hugepages_1Gs.append(hugepages_1G)
vm_4K_pages.append(vm_hugepages_nr_4K)
vm_2M_pages.append(vm_hugepages_nr_2M)
vm_1G_pages.append(vm_hugepages_nr_1G)
if (memory.vswitch_hugepages_reqd and
vswitch_pages != memory.vswitch_hugepages_nr):
vs_pages_updated = True
platform_reserved_memory = "(%s)" % ' '.join(platform_nodes)
vswitch_reserved_memory = "(%s)" % ' '.join(vswitch_nodes)
nr_hugepages_2Ms = "(%s)" % ' '.join(hugepages_2Ms)
nr_hugepages_1Gs = "(%s)" % ' '.join(hugepages_1Gs)
vswitch_2M = "\"%s\"" % ','.join([str(i) for i in vswitch_2M_pages])
vswitch_1G = "\"%s\"" % ','.join([str(i) for i in vswitch_1G_pages])
vm_4K = "\"%s\"" % ','.join([str(i) for i in vm_4K_pages])
vm_2M = "\"%s\"" % ','.join([str(i) for i in vm_2M_pages])
vm_1G = "\"%s\"" % ','.join([str(i) for i in vm_1G_pages])
config.update({
'platform::compute::params::worker_base_reserved':
platform_reserved_memory,
'platform::compute::params::compute_vswitch_reserved':
vswitch_reserved_memory,
'platform::compute::hugepage::params::nr_hugepages_2M':
nr_hugepages_2Ms,
'platform::compute::hugepage::params::nr_hugepages_1G':
nr_hugepages_1Gs,
'platform::compute::hugepage::params::vswitch_2M_pages':
vswitch_2M,
'platform::compute::hugepage::params::vswitch_1G_pages':
vswitch_1G,
'platform::compute::hugepage::params::vm_4K_pages':
vm_4K,
'platform::compute::hugepage::params::vm_2M_pages':
vm_2M,
'platform::compute::hugepage::params::vm_1G_pages':
vm_1G,
})
if vs_pages_updated:
grub_hugepages_1G = "hugepagesz=1G hugepages=%d" % (
sum(vswitch_1G_pages) + sum(vm_1G_pages))
config.update({
'platform::compute::grub::params::g_hugepages':
grub_hugepages_1G,
})
if sum(vswitch_2M_pages) > 0:
config.update({
'platform::vswitch::params::hugepage_dir': '/mnt/huge-2048kB'
})
return config
def _get_host_cpu_config(self, host):
config = {}
if constants.WORKER in utils.get_personalities(host):
host_cpus = self._get_host_cpu_list(host, threads=True)
if not host_cpus:
return config
platform_cpus_no_threads = self._get_platform_cpu_list(host)
vswitch_cpus_no_threads = self._get_vswitch_cpu_list(host)
platform_numa_cpus = utils.get_numa_index_list(platform_cpus_no_threads)
vswitch_numa_cpus = utils.get_numa_index_list(vswitch_cpus_no_threads)
# build a list of platform reserved cpus per numa node
platform_cores = []
for node, cpus in platform_numa_cpus.items():
cpu_list = ','.join([str(c.cpu) for c in cpus])
platform_node = "\"node%d:%s\"" % (node, cpu_list)
platform_cores.append(platform_node)
# build a list of vswitch reserved cpu counts per numa node
vswitch_cores = []
for node, cpus in vswitch_numa_cpus.items():
cpu_count = len(cpus)
vswitch_node = "\"node%d:%d\"" % (node, cpu_count)
vswitch_cores.append(vswitch_node)
reserved_platform_cores = "(%s)" % ' '.join(platform_cores)
reserved_vswitch_cores = "(%s)" % ' '.join(vswitch_cores)
# all logical cpus
host_cpus = self._get_host_cpu_list(host, threads=True)
host_cpuset = set([c.cpu for c in host_cpus])
host_ranges = utils.format_range_set(host_cpuset)
n_cpus = len(host_cpuset)
# platform logical cpus
platform_cpus = self._get_host_cpu_list(
host, function=constants.PLATFORM_FUNCTION, threads=True)
platform_cpuset = set([c.cpu for c in platform_cpus])
platform_ranges = utils.format_range_set(platform_cpuset)
# vswitch logical cpus
vswitch_cpus = self._get_host_cpu_list(
host, constants.VSWITCH_FUNCTION, threads=True)
vswitch_cpuset = set([c.cpu for c in vswitch_cpus])
vswitch_ranges = utils.format_range_set(vswitch_cpuset)
# non-platform logical cpus
rcu_nocbs_cpuset = host_cpuset - platform_cpuset
rcu_nocbs_ranges = utils.format_range_set(rcu_nocbs_cpuset)
# non-vswitch logical cpus
non_vswitch_cpuset = host_cpuset - vswitch_cpuset
non_vswitch_ranges = utils.format_range_set(non_vswitch_cpuset)
cpu_options = ""
if constants.LOWLATENCY in host.subfunctions:
config.update({
'platform::compute::pmqos::low_wakeup_cpus':
"\"%s\"" % vswitch_ranges,
'platform::compute::pmqos::hight_wakeup_cpus':
"\"%s\"" % non_vswitch_ranges,
})
vswitch_ranges = rcu_nocbs_ranges
cpu_options += "nohz_full=%s " % vswitch_ranges
cpu_options += "isolcpus=%s rcu_nocbs=%s kthread_cpus=%s " \
"irqaffinity=%s" % (vswitch_ranges,
rcu_nocbs_ranges,
platform_ranges,
platform_ranges)
config.update({
'platform::compute::params::worker_cpu_list':
"\"%s\"" % host_ranges,
'platform::compute::params::platform_cpu_list':
"\"%s\"" % platform_ranges,
'platform::compute::params::reserved_vswitch_cores':
reserved_vswitch_cores,
'platform::compute::params::reserved_platform_cores':
reserved_platform_cores,
'platform::compute::grub::params::n_cpus': n_cpus,
'platform::compute::grub::params::cpu_options': cpu_options,
})
return config
def _get_host_k8s_cgroup_config(self, host):
config = {}
# determine set of all logical cpus and nodes
host_cpus = self._get_host_cpu_list(host, threads=True)
host_cpuset = set([c.cpu for c in host_cpus])
host_nodeset = set([c.numa_node for c in host_cpus])
# determine set of platform logical cpus and nodes
platform_cpus = self._get_host_cpu_list(
host, function=constants.PLATFORM_FUNCTION, threads=True)
platform_cpuset = set([c.cpu for c in platform_cpus])
platform_nodeset = set([c.numa_node for c in platform_cpus])
# determine platform reserved number of logical cpus
k8s_reserved_cpus = len(platform_cpuset)
# determine platform reserved memory
k8s_reserved_mem = 0
host_memory = self.dbapi.imemory_get_by_ihost(host.id)
numa_memory = utils.get_numa_index_list(host_memory)
for node, memory in numa_memory.items():
reserved_mib = memory[0].platform_reserved_mib
if reserved_mib is not None:
k8s_reserved_mem += reserved_mib
# determine set of nonplatform logical cpus
# TODO(jgauld): Commented out for now, using host_cpuset instead.
# nonplatform_cpuset = host_cpuset - platform_cpuset
if constants.WORKER in utils.get_personalities(host):
if self.is_openstack_compute(host):
k8s_cpuset = utils.format_range_set(platform_cpuset)
k8s_nodeset = utils.format_range_set(platform_nodeset)
else:
# kubelet cpumanager is configured with static policy.
# The resulting DefaultCPUSet excludes reserved cpus
# based on topology, and that also happens to correspond
# to the platform_cpuset. kubepods are allowed to
# span all host numa nodes.
# TODO(jgauld): Temporary workaround until we have a version
# of kubelet that excludes reserved cpus from DefaultCPUSet.
# The intent is to base k8s_cpuset on nonplatform_cpuset.
# Commented out for now, using host_cpuset instead.
# k8s_cpuset = utils.format_range_set(nonplatform_cpuset)
k8s_cpuset = utils.format_range_set(host_cpuset)
k8s_nodeset = utils.format_range_set(host_nodeset)
else:
k8s_cpuset = utils.format_range_set(host_cpuset)
k8s_nodeset = utils.format_range_set(host_nodeset)
LOG.debug('host:%s, k8s_cpuset:%s, k8s_nodeset:%s', host.hostname,
k8s_cpuset, k8s_nodeset)
config.update({
'platform::kubernetes::params::k8s_cpuset':
"\"%s\"" % k8s_cpuset,
'platform::kubernetes::params::k8s_nodeset':
"\"%s\"" % k8s_nodeset,
'platform::kubernetes::params::k8s_reserved_cpus':
k8s_reserved_cpus,
'platform::kubernetes::params::k8s_reserved_mem':
k8s_reserved_mem,
})
return config
def _get_host_cpu_config(self, host):
config = {}
if constants.WORKER in utils.get_personalities(host):
host_cpus = self._get_host_cpu_list(host, threads=True)
if not host_cpus:
return config
platform_cpus_no_threads = self._get_platform_cpu_list(host)
vswitch_cpus_no_threads = self._get_vswitch_cpu_list(host)
platform_numa_cpus = utils.get_numa_index_list(
platform_cpus_no_threads)
vswitch_numa_cpus = utils.get_numa_index_list(
vswitch_cpus_no_threads)
# build a list of platform reserved cpus per numa node
platform_cores = []
for node, cpus in platform_numa_cpus.items():
cpu_list = ','.join([str(c.cpu) for c in cpus])
platform_node = "\"node%d:%s\"" % (node, cpu_list)
platform_cores.append(platform_node)
# build a list of vswitch reserved cpu counts per numa node
vswitch_cores = []
for node, cpus in vswitch_numa_cpus.items():
cpu_count = len(cpus)
vswitch_node = "\"node%d:%d\"" % (node, cpu_count)
vswitch_cores.append(vswitch_node)
reserved_platform_cores = "(%s)" % ' '.join(platform_cores)
reserved_vswitch_cores = "(%s)" % ' '.join(vswitch_cores)
# all logical cpus
host_cpus = self._get_host_cpu_list(host, threads=True)
host_cpuset = set([c.cpu for c in host_cpus])
host_ranges = utils.format_range_set(host_cpuset)
n_cpus = len(host_cpuset)
# platform logical cpus
platform_cpus = self._get_host_cpu_list(
host, function=constants.PLATFORM_FUNCTION, threads=True)
platform_cpuset = set([c.cpu for c in platform_cpus])
platform_ranges = utils.format_range_set(platform_cpuset)
# vswitch logical cpus
vswitch_cpus = self._get_host_cpu_list(host,
constants.VSWITCH_FUNCTION,
threads=True)
vswitch_cpuset = set([c.cpu for c in vswitch_cpus])
# non-platform logical cpus
rcu_nocbs_cpuset = host_cpuset - platform_cpuset
rcu_nocbs_ranges = utils.format_range_set(rcu_nocbs_cpuset)
# isolated logical cpus
app_isolated_cpus = self._get_host_cpu_list(
host, constants.ISOLATED_FUNCTION, threads=True)
app_isolated_cpuset = set([c.cpu for c in app_isolated_cpus])
# application cpus
app_cpus = self._get_host_cpu_list(host,
constants.APPLICATION_FUNCTION,
threads=True)
app_cpuset = set([c.cpu for c in app_cpus])
app_ranges = utils.format_range_set(app_cpuset)
cpu_options = ""
cpu_ranges = {}
if constants.LOWLATENCY in host.subfunctions:
# set PM QoS latency that achieves C1 state for all cpus
config.update({
'platform::compute::pmqos::low_wakeup_cpus':
"\"%s\"" % host_ranges,
'platform::compute::pmqos::hight_wakeup_cpus':
"\"%s\"" % "",
})
cpu_ranges.update({"nohz_full": rcu_nocbs_ranges})
isolcpus_ranges = utils.format_range_set(
vswitch_cpuset.union(app_isolated_cpuset))
cpu_ranges.update({
"isolcpus": isolcpus_ranges,
"rcu_nocbs": rcu_nocbs_ranges,
"kthread_cpus": platform_ranges
})
# Put IRQs on application cores if they are configured.
# Note that PCI IRQs for platform interfaces are reaffined to
# platform cores at runtime.
if app_cpuset:
cpu_ranges.update({"irqaffinity": app_ranges})
else:
cpu_ranges.update({"irqaffinity": platform_ranges})
for key, value in cpu_ranges.items():
if str(value).strip() != "":
cpu_options += "%s=%s " % (key, value)
config.update({
'platform::compute::params::worker_cpu_list':
"\"%s\"" % host_ranges,
'platform::compute::params::platform_cpu_list':
"\"%s\"" % platform_ranges,
'platform::compute::params::reserved_vswitch_cores':
reserved_vswitch_cores,
'platform::compute::params::reserved_platform_cores':
reserved_platform_cores,
'platform::compute::grub::params::n_cpus':
n_cpus,
'platform::compute::grub::params::cpu_options':
cpu_options,
})
return config
def _get_host_cpu_config(self, host):
config = {}
if constants.WORKER in utils.get_personalities(host):
host_cpus = self._get_host_cpu_list(host, threads=True)
if not host_cpus:
return config
# Define the full range of CPUs for the compute host
max_cpu = max(host_cpus, key=operator.attrgetter('cpu'))
worker_cpu_list = "\"0-%d\"" % max_cpu.cpu
platform_cpus_no_threads = self._get_platform_cpu_list(host)
vswitch_cpus_no_threads = self._get_vswitch_cpu_list(host)
platform_cpu_list_with_quotes = \
"\"%s\"" % ','.join([str(c.cpu) for c in platform_cpus_no_threads])
platform_numa_cpus = utils.get_numa_index_list(
platform_cpus_no_threads)
vswitch_numa_cpus = utils.get_numa_index_list(
vswitch_cpus_no_threads)
# build a list of platform reserved cpus per numa node
platform_cores = []
for node, cpus in platform_numa_cpus.items():
cpu_list = ','.join([str(c.cpu) for c in cpus])
platform_node = "\"node%d:%s\"" % (node, cpu_list)
platform_cores.append(platform_node)
# build a list of vswitch reserved cpu counts per numa node
vswitch_cores = []
for node, cpus in vswitch_numa_cpus.items():
cpu_count = len(cpus)
vswitch_node = "\"node%d:%d\"" % (node, cpu_count)
vswitch_cores.append(vswitch_node)
reserved_platform_cores = "(%s)" % ' '.join(platform_cores)
reserved_vswitch_cores = "(%s)" % ' '.join(vswitch_cores)
host_cpus = sorted(host_cpus, key=lambda c: c.cpu)
n_cpus = len(host_cpus)
host_cpu_list = [c.cpu for c in host_cpus]
platform_cpus = self._get_host_cpu_list(
host, function=constants.PLATFORM_FUNCTION, threads=True)
platform_cpus = sorted(platform_cpus, key=lambda c: c.cpu)
platform_cpu_list = \
"%s" % ','.join([str(c.cpu) for c in platform_cpus])
vswitch_cpus = self._get_host_cpu_list(host,
constants.VSWITCH_FUNCTION,
threads=True)
vswitch_cpus = sorted(vswitch_cpus, key=lambda c: c.cpu)
vswitch_cpu_list = \
"%s" % ','.join([str(c.cpu) for c in vswitch_cpus])
# rcu_nocbs = all cores - platform cores
rcu_nocbs = copy.deepcopy(host_cpu_list)
for i in [int(s) for s in platform_cpu_list.split(',')]:
rcu_nocbs.remove(i)
# change the CPU list to ranges
rcu_nocbs_ranges = ""
for key, group in itertools.groupby(enumerate(rcu_nocbs),
lambda xy: xy[1] - xy[0]):
group = list(group)
rcu_nocbs_ranges += "%s-%s," % (group[0][1], group[-1][1])
rcu_nocbs_ranges = rcu_nocbs_ranges.rstrip(',')
# non-vswitch CPUs = all cores - vswitch cores
non_vswitch_cpus = host_cpu_list
for i in [c.cpu for c in vswitch_cpus]:
non_vswitch_cpus.remove(i)
# change the CPU list to ranges
non_vswitch_cpus_ranges = ""
for key, group in itertools.groupby(enumerate(non_vswitch_cpus),
lambda xy: xy[1] - xy[0]):
group = list(group)
non_vswitch_cpus_ranges += "\"%s-%s\"," % (group[0][1],
group[-1][1])
cpu_options = ""
if constants.LOWLATENCY in host.subfunctions:
vswitch_cpu_list_with_quotes = \
"\"%s\"" % ','.join([str(c.cpu) for c in vswitch_cpus])
config.update({
'platform::compute::pmqos::low_wakeup_cpus':
vswitch_cpu_list_with_quotes,
'platform::compute::pmqos::hight_wakeup_cpus':
non_vswitch_cpus_ranges.rstrip(',')
})
vswitch_cpu_list = rcu_nocbs_ranges
cpu_options += "nohz_full=%s " % vswitch_cpu_list
cpu_options += "isolcpus=%s rcu_nocbs=%s kthread_cpus=%s " \
"irqaffinity=%s" % (vswitch_cpu_list,
rcu_nocbs_ranges,
platform_cpu_list,
platform_cpu_list)
config.update({
'platform::compute::params::worker_cpu_list':
worker_cpu_list,
'platform::compute::params::platform_cpu_list':
platform_cpu_list_with_quotes,
'platform::compute::params::reserved_vswitch_cores':
reserved_vswitch_cores,
'platform::compute::params::reserved_platform_cores':
reserved_platform_cores,
'platform::compute::grub::params::n_cpus':
n_cpus,
'platform::compute::grub::params::cpu_options':
cpu_options,
})
return config