blob: a64a660b5600f323d1ee8bc854c03e353ba3f599 [file] [log] [blame]
Gary Wucd47a012018-11-30 07:18:36 -08001.. This work is licensed under a Creative Commons Attribution 4.0
2 International License. http://creativecommons.org/licenses/by/4.0
3 Copyright 2018 Huawei Technologies Co., Ltd. All rights reserved.
4
5.. _docs_vcpe:
6
Gary Wue4a2df82018-11-29 12:49:09 -08007vCPE Use Case
8----------------------------
9
10Description
11~~~~~~~~~~~
12vCPE use case is based on Network Enhanced Residential Gateway architecture specified in Technical Report 317 (TR-317), which defines how service providers deploy residential broadband services like High Speed Internet Access. The use case implementation has infrastructure services and customer service. The common infrastructure services are deployed first and shared by all customers. The use case demonstrates ONAP capabilities to design, deploy, configure and control sophisticated services.
13
14More details on the vCPE Use Case can be found on wiki page https://wiki.onap.org/pages/viewpage.action?pageId=3246168
15
16Source Code
17~~~~~~~~~~~
Yang Xucbf3f2d2019-06-24 08:31:16 -040018vcpe test scripts: https://git.onap.org/integration/tree/test/vcpe?h=dublin
Gary Wue4a2df82018-11-29 12:49:09 -080019
20How to Use
21~~~~~~~~~~
22Most part of the use case has been automated by vcpe scripts. For the details on how to run the scripts, please refer to the use case tutorial on https://wiki.onap.org/display/DW/vCPE+Use+Case+Tutorial%3A+Design+and+Deploy+based+on+ONAP.
23
Yang Xuaa085032019-06-17 09:44:34 -040024Here are the main steps to run the use case in Integration lab environment, where vCPE script is pre-installed on Rancher node under /root/integration/test/vcpe:
25
261. Run Robot script from Rancher node to onboard VNFs, create and distribute models for vCPE four infrastructure services, i.e. infrastructure, brg, bng and gmux
27
28::
29
30 demo-k8s.sh onap init
31
Yang Xu587704d2019-06-26 08:44:10 -0400322. Add customer SDN-ETHERNET-INTERNET (see the use case tutorial wiki page for detail)
Yang Xuaa085032019-06-17 09:44:34 -040033
Yang Xu587704d2019-06-26 08:44:10 -0400343. Add identity-url to RegionOne data in A&AI. First use POSTMAN to GET cloud-region RegionOne data, then add identity-url and PUT back to A&AI
Yang Xuaa085032019-06-17 09:44:34 -040035
36::
37
38 GET https://{{aai}}:{{port}}/aai/v14/cloud-infrastructure/cloud-regions/cloud-region/CloudOwner/RegionOne
39
40::
41
42 PUT https://{{aai}}:{{port}}/aai/v14/cloud-infrastructure/cloud-regions/cloud-region/CloudOwner/RegionOne
43 {
44 "cloud-owner": "CloudOwner",
45 "cloud-region-id": "RegionOne",
46 "cloud-type": "SharedNode",
47 "owner-defined-type": "OwnerType",
48 "cloud-region-version": "v1",
49 "identity-url": "http://10.12.25.2:5000/v2.0",
50 "cloud-zone": "CloudZone",
51 "resource-version": "1559336510793",
52 "relationship-list": {
53 ... ...
54
Yang Xu587704d2019-06-26 08:44:10 -0400554. Add route on sdnc cluster VM node, which is the cluster VM node where pod sdnc-sdnc-0 is running on. This will allow ONAP SDNC to configure BRG later on.
Yang Xuaa085032019-06-17 09:44:34 -040056
57::
58
59 ip route add 10.3.0.0/24 via 10.0.101.10 dev ens3
60
Yang Xuaa085032019-06-17 09:44:34 -040061
Bartek Grzybowski9018a842019-10-16 15:28:23 +0200625. Install Python and other Python libraries
Yang Xuaa085032019-06-17 09:44:34 -040063
64::
65
66 integration/test/vcpe/bin/setup.sh
67
68
Bartek Grzybowski9018a842019-10-16 15:28:23 +0200696. Change the Openstack env parameters and one customer service related parameter in vcpecommon.py
Yang Xuaa085032019-06-17 09:44:34 -040070
71::
72
73 cloud = {
74 '--os-auth-url': 'http://10.12.25.2:5000',
75 '--os-username': 'xxxxxxxxxx',
76 '--os-user-domain-id': 'default',
77 '--os-project-domain-id': 'default',
78 '--os-tenant-id': 'xxxxxxxxxxxxxxxx' if oom_mode else '1e097c6713e74fd7ac8e4295e605ee1e',
79 '--os-region-name': 'RegionOne',
80 '--os-password': 'xxxxxxxxxxx',
81 '--os-project-domain-name': 'xxxxxxxxx' if oom_mode else 'Integration-SB-07',
82 '--os-identity-api-version': '3'
83 }
84
85 common_preload_config = {
86 'oam_onap_net': 'xxxxxxxx' if oom_mode else 'oam_onap_lAky',
87 'oam_onap_subnet': 'xxxxxxxxxx' if oom_mode else 'oam_onap_lAky',
88 'public_net': 'xxxxxxxxx',
89 'public_net_id': 'xxxxxxxxxxxxx'
90 }
91
92::
93
94 # CHANGEME: vgw_VfModuleModelInvariantUuid is in rescust service csar, open service template with filename like service-VcpesvcRescust1118-template.yml and look for vfModuleModelInvariantUUID under groups vgw module metadata.
95 self.vgw_VfModuleModelInvariantUuid = 'xxxxxxxxxxxxxxx'
96
Bartek Grzybowski9018a842019-10-16 15:28:23 +0200977. Initialize vcpe
Yang Xuaa085032019-06-17 09:44:34 -040098
99::
100
101 vcpe.py init
102
Bartek Grzybowski9018a842019-10-16 15:28:23 +02001038. Run a command from Rancher node to insert vcpe customer service workflow entry in SO catalogdb. You should be able to see a sql command printed out from the above step output at the end, and use that sql command to replace the sample sql command below (inside the double quote) and run it from Rancher node:
Yang Xuaa085032019-06-17 09:44:34 -0400104
105::
106
Yang Xu587704d2019-06-26 08:44:10 -0400107 kubectl exec dev-mariadb-galera-mariadb-galera-0 -- mysql -uroot -psecretpassword catalogdb -e "INSERT INTO service_recipe (ACTION, VERSION_STR, DESCRIPTION, ORCHESTRATION_URI, SERVICE_PARAM_XSD, RECIPE_TIMEOUT, SERVICE_TIMEOUT_INTERIM, CREATION_TIMESTAMP, SERVICE_MODEL_UUID) VALUES ('createInstance','1','vCPEResCust 2019-06-03 _04ba','/mso/async/services/CreateVcpeResCustService',NULL,181,NULL, NOW(),'6c4a469d-ca2c-4b02-8cf1-bd02e9c5a7ce')"
Yang Xuaa085032019-06-17 09:44:34 -0400108
Bartek Grzybowski9018a842019-10-16 15:28:23 +02001099. Run Robot to create and distribute for vCPE customer service. This step assumes step 1 has successfully distributed all vcpe models except customer service model
Yang Xu587704d2019-06-26 08:44:10 -0400110
111::
112
113 ete-k8s.sh onap distributevCPEResCust
114
Bartek Grzybowskif816c302019-11-06 15:26:37 +010011510. Instantiate vCPE infra services
Yang Xuaa085032019-06-17 09:44:34 -0400116
117::
118
119 vcpe.py infra
120
Bartek Grzybowski19199da2019-11-07 12:44:46 +010012111. From Rancher node run vcpe healthcheck command to check connectivity from sdnc to brg and gmux, and vpp configuration of brg and gmux.
Yang Xuaa085032019-06-17 09:44:34 -0400122
123::
124
Bartek Grzybowski4baa51c2019-09-10 13:50:15 +0200125 healthcheck-k8s.py --namespace <namespace name> --environment <env name>
Yang Xuaa085032019-06-17 09:44:34 -0400126
Bartek Grzybowski19199da2019-11-07 12:44:46 +010012712. Instantiate vCPE customer service.
Yang Xuaa085032019-06-17 09:44:34 -0400128
Yang Xu587704d2019-06-26 08:44:10 -0400129::
130
131 vcpe.py customer
132
Bartek Grzybowskif816c302019-11-06 15:26:37 +010013313. Update libevel.so in vGMUX VM and restart the VM. This allows vGMUX to send events to VES collector in close loop test. See tutorial wiki for details
Yang Xu587704d2019-06-26 08:44:10 -0400134
Bartek Grzybowskif816c302019-11-06 15:26:37 +010013514. Run heatbridge. The heatbridge command usage: demo-k8s.sh <namespace> heatbridge <stack_name> <service_instance_id> <service> <oam-ip-address>, please refer to vCPE tutorial page on how to fill in those paraemters. See an example as following:
Yang Xuaa085032019-06-17 09:44:34 -0400136
137::
138
139 ~/integration/test/vcpe# ~/oom/kubernetes/robot/demo-k8s.sh onap heatbridge vcpe_vfmodule_e2744f48729e4072b20b_201811262136 d8914ef3-3fdb-4401-adfe-823ee75dc604 vCPEvGMUX 10.0.101.21
140
Bartek Grzybowskif816c302019-11-06 15:26:37 +010014115. Start closed loop test by triggering packet drop VES event, and monitor if vGMUX is restarting. You may need to run the command twice if the first run fails
Yang Xuaa085032019-06-17 09:44:34 -0400142
143::
144
145 vcpe.py loop
146
147
Gary Wue4a2df82018-11-29 12:49:09 -0800148Test Status
149~~~~~~~~~~~~~~~~~~~~~
Yang Xuaa085032019-06-17 09:44:34 -0400150The use case has been tested for Dublin release, the test report can be found on https://wiki.onap.org/display/DW/vCPE+%28Heat%29+-+Dublin+Test+Status
Gary Wue4a2df82018-11-29 12:49:09 -0800151
152Known Issues and Workaround
153~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Yang Xu4c9c4712019-06-19 01:56:35 -04001541) NATs are installed on BRG and vBNG. In order to allow SDNC to send BRG configuration message through vBNG, SDNC host VM IP address is preloaded on BRG and vBNG during VM instantiation, and provisioned into the NATs. If SDNC changes its host VM, SDNC host VM IP changes and we need to manually update the IP in /opt/config/sdnc_ip.txt. Then run:
Gary Wue4a2df82018-11-29 12:49:09 -0800155
156::
157
158 root>vppctl tap delete tap-0
159 root>vppctl tap delete tap-1
160 root>/opt/nat_service.sh
161 root>vppctl restart
162
Yang Xu5ee1c382019-06-26 22:46:31 -04001632) During vCPE customer service instantiation, though vGW should come up successfully BRG vxlan tunnel configuration is likely to fail in SDNC cluster environment due to SDNC unreachable to BRG. See more detail in JIRA INT-1127. One workaround is to run vCPE use case with SDNC cluster disabled.
Michal Ptacek7a008fd2019-09-11 12:04:35 +0000164
1653) In some Openstack environments (e.g. Ocata version), there is an issue with DHCP anti-spoofing rules preventing BRG to receive DHCP reply (Option 82) from DHCP. By default Openstack neutron is using *IptablesFirewallDriver*, which is actively inserting *Prevent DHCP Spoofing by VM* rules into linuxbridge firewall rules. This feature should prevent mailicious traffic from rogue VM inside Openstack, however it's affecting also vCPE usecase. Manual tweaking of fw rules is not persistent and those rules are automatically regenerated, but one can disable this logic by switching to *neutron.agent.firewall.NoopFirewallDriver*. More details can be found on https://codesomniac.com/2017/07/how-to-run-a-dhcp-server-as-openstack-instance/
166
Bartek Grzybowskif1d0a8b2019-09-12 09:51:34 +0200167 **NOTE:** To propagate change in firewall_driver one needs to restart neutron-linuxbridge-agent and also openstack-nova-compute services.
Michal Ptacek7a008fd2019-09-11 12:04:35 +0000168
Bartek Grzybowskif1d0a8b2019-09-12 09:51:34 +0200169 Additionally Neutron's Port Security Extension Driver is by default preventing any routing functions of an instance (be it a router or VNF). Hence for smoothest vCPE experience it's advised to either disable the packet filtering by setting port_security_enabled flag for a network/port to "False" or alternatively add allowed_address_pairs to relevant VNFs ports with appropriate network prefixes. Port security driver can be also disabled globally, for more insight into Port Security Extension Driver please visit https://wiki.openstack.org/wiki/Neutron/ML2PortSecurityExtensionDriver