Nova Helm Chart Override Generation - Part 2
Add more helm chart overrides for stx-nova. These are based on
nova.conf options previously created by nova puppet manifests on the
controllers and computes.
- DEFAULT/reserved_host_memory_mb: Nova sets default value to 512. In
stx, we handle memory overheads via compute per numa size config
options so we can set this option to 0.
- DEFAULT/compute_monitors: enable cpu monitor
- scheduler/periodic_task_interval: Temporarily disable stx-nova
scheduler audit that attempts to contact patching api (to determine
which hosts are patched) and sysinv api (to determine which hosts
are upgraded). This currently doesn't work in the containers setup
and is spewing exceptions in the nova scheduler log every minute.
Disable it for now and we'll come back to this later.
- vnc/novncproxy_host: Set to floating management address as this is
where haproxy directs the requests.
- vnc/novncproxy_base_url: Set based on floating oam address.
- vnc/vncserver_proxyclient_address: Set per host to primary management
address. Currently this is superseded by upstream openstack-helm
via file /tmp/pod-shared/nova-console.conf. However, that doesn't
work for standby AIO so we'll potentially disable it and use this
setting.
Story: 2003909
Task: 27081
Change-Id: I3d165a2ca19f02de31ca1c6a53f7de27c3d8625d
Signed-off-by: Gerry Kopec <Gerry.Kopec@windriver.com>
This commit is contained in:
@@ -80,6 +80,8 @@ class NovaHelm(openstack.OpenstackBaseHelm):
|
||||
'nova': {
|
||||
'DEFAULT': {
|
||||
'default_mempages_size': 2048,
|
||||
'reserved_host_memory_mb': 0,
|
||||
'compute_monitors': 'cpu.virt_driver',
|
||||
'running_deleted_instance_poll_interval': 60,
|
||||
'mkisofs_cmd': '/usr/bin/genisoimage',
|
||||
'network_allocate_retries': 2,
|
||||
@@ -139,11 +141,18 @@ class NovaHelm(openstack.OpenstackBaseHelm):
|
||||
'soft_affinity_weight_multiplier': 0.0,
|
||||
'soft_anti_affinity_weight_multiplier': 0.0
|
||||
},
|
||||
'scheduler': {
|
||||
'periodic_task_interval': -1
|
||||
},
|
||||
'metrics': {
|
||||
'required': False,
|
||||
'weight_setting_multi': 'vswitch.multi_avail=100.0',
|
||||
'weight_setting': 'vswitch.max_avail=100.0'
|
||||
},
|
||||
'vnc': {
|
||||
'novncproxy_host': self._get_management_address(),
|
||||
'novncproxy_base_url': self._get_novncproxy_base_url(),
|
||||
},
|
||||
'upgrade_levels': 'None'
|
||||
},
|
||||
'overrides': {
|
||||
@@ -230,6 +239,11 @@ class NovaHelm(openstack.OpenstackBaseHelm):
|
||||
|
||||
return overrides
|
||||
|
||||
def _get_novncproxy_base_url(self):
|
||||
oam_addr = self._get_oam_address(),
|
||||
url = "http://%s:6080/vnc_auto.html" % oam_addr
|
||||
return url
|
||||
|
||||
def _get_virt_type(self):
|
||||
if utils.is_virtual():
|
||||
return 'qemu'
|
||||
@@ -324,10 +338,14 @@ class NovaHelm(openstack.OpenstackBaseHelm):
|
||||
primary_mgmt = iface
|
||||
|
||||
migration_iface = primary_infra or primary_mgmt
|
||||
if migration_iface is None:
|
||||
return
|
||||
for addr in addresses:
|
||||
if addr.interface_uuid == migration_iface.uuid:
|
||||
migration_ip = addr.address
|
||||
ip_family = addr.family
|
||||
if addr.interface_uuid == primary_mgmt.uuid:
|
||||
mgmt_ip = addr.address
|
||||
|
||||
default_config.update({'my_ip': migration_ip})
|
||||
if ip_family == 4:
|
||||
@@ -336,6 +354,7 @@ class NovaHelm(openstack.OpenstackBaseHelm):
|
||||
vnc_config.update({'vncserver_listen': '::0'})
|
||||
|
||||
libvirt_config.update({'live_migration_inbound_addr': str(host.hostname) + '-infra'})
|
||||
vnc_config.update({'vncserver_proxyclient_address': mgmt_ip})
|
||||
|
||||
def _update_host_memory(self, host, default_config):
|
||||
vswitch_2M_pages = []
|
||||
|
||||
Reference in New Issue
Block a user