diff options
author | 2018-11-08 19:06:46 +0100 | |
---|---|---|
committer | 2019-01-10 15:22:59 +0100 | |
commit | 6eaeafea81f30ee66f2c187596815a0166f1132b (patch) | |
tree | ab296858aac68b09b2a905486aa072dc8c304e0c /mcp/reclass/classes/cluster/mcp-fdio-noha/openstack/compute.yml.j2 | |
parent | 920d7ef405e9fed9af7b0576b94f6d19fe265c81 (diff) |
[noha] Bring in FDIO (VPP+DPDK) scenario
- cmp, gtw: bump RAM allocation to accomodate hugepages/VPP;
for now we overcommit, gtw01 resources can probably be lowered;
- submodule: add salt-formula-neutron so we can locally patch it;
- repo:
* FD.IO repos for VPP packages;
* networking-vpp PPA for python-networking-vpp Neutron driver;
- use vpp-router for L3, disable neutron-l3-agent;
- baremetal_init: apply repo config before network (otherwise UCA
repo is missing when trying to install DPDK on baremetal nodes);
- arm64: iommu.passthrough=1 is required on ThunderX for VPP on
newer kernels;
Design quirks:
- vpp service runs as 'neutron' user, which does not exist at the
time VPP is installed and initially started, hence the need to
restart it before starting the vpp-agent service;
- gtw01 node has DPDK, yet to configure it via IDF we use the
compute-specific OVS-targeted parameters like
`compute_ovs_dpdk_socket_mem`, which is a bit misleading;
- vpp-agent requires ml2_conf.ini on ALL compute AND network nodes
to parse per-node physnet-to-real interface names;
- vpp process is bound to core '1' (not parameterized via IDF);
Change-Id: I659f7dbebcab7b154e7b1fb829cd7159b4372ec8
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
(cherry picked from commit 455b46a6be4bca145c047ed6957727c119285796)
Diffstat (limited to 'mcp/reclass/classes/cluster/mcp-fdio-noha/openstack/compute.yml.j2')
-rw-r--r-- | mcp/reclass/classes/cluster/mcp-fdio-noha/openstack/compute.yml.j2 | 47 |
1 files changed, 37 insertions, 10 deletions
diff --git a/mcp/reclass/classes/cluster/mcp-fdio-noha/openstack/compute.yml.j2 b/mcp/reclass/classes/cluster/mcp-fdio-noha/openstack/compute.yml.j2 index 2f1b8c39e..825d9d550 100644 --- a/mcp/reclass/classes/cluster/mcp-fdio-noha/openstack/compute.yml.j2 +++ b/mcp/reclass/classes/cluster/mcp-fdio-noha/openstack/compute.yml.j2 @@ -6,27 +6,54 @@ # http://www.apache.org/licenses/LICENSE-2.0 ############################################################################## {%- import 'net_map.j2' as nm with context %} -{%- import 'net_macros.j2' as ma with context %} -{%- set vlan_private_start = (nm.vlan_private | string).rsplit('-')[0] %} --- classes: - service.neutron.compute.single + - system.nova.compute.nfv.hugepages - cluster.mcp-common-noha.openstack_compute - cluster.mcp-fdio-noha parameters: + neutron: + compute: + backend: + router: 'vpp-router' + tenant_network_types: "${_param:neutron_tenant_network_types}" + ~mechanism: + vpp: + driver: vpp + etcd_port: ${_param:node_port} + etcd_host: ${_param:node_address} + l3_hosts: ${_param:openstack_gateway_node01_hostname} + physnets: + physnet1: + vpp_interface: ${_param:external_vpp_tap} + physnet2: + vpp_interface: ${_param:dpdk0_vpp} linux: + system: + kernel: + isolcpu: 1 # NOTE: Hardcoded for now + boot_options: + - spectre_v2=off + - nopti + - intel_iommu=on + - iommu=pt + - nohz_full=${linux:system:kernel:isolcpu} + - rcu_nocbs=${linux:system:kernel:isolcpu} + - iommu.passthrough=1 network: interface: + dpdk0: + name: ${_param:dpdk0_name} + pci: ${_param:dpdk0_pci} + driver: ${_param:dpdk0_driver} + enabled: true + type: dpdk_vpp_port + mtu: ${_param:interface_mtu} + {{ nm.cmp001.nic_private }}: + type: dpdk # Not a meaningful type, just match 'dpdk' for filtering pxe_admin_int: # For scenarios without public network on cmp, set admin gw gateway: {{ nm.net_admin_gw }} name_servers: - {{ nm.net_admin_gw }} - br-mesh: - enabled: true - type: bridge - proto: static - address: ${_param:tenant_address} - netmask: ${_param:opnfv_net_private_mask} - use_interfaces: - - {{ ma.interface_str(nm.cmp001.nic_private, vlan_private_start) }} |