Ultra-M元件管理員叢集的高可用性還原- vEPC
簡介
本文檔介紹在託管StarOS虛擬網路功能(VNF)的Ultra-M設定的元素管理器(EM)集群中恢復高可用性(HA)所需的步驟。
背景資訊
Ultra-M是經過預先封裝和驗證的虛擬化移動資料包核心解決方案，旨在簡化VNF的部署。Ultra-M解決方案包含上述虛擬機器(VM)型別：
- 自動IT
- 自動部署
- Ultra自動化服務(UAS)
- 元素管理員(EM)
- 彈性服務控制器(ESC)
- 控制功能(CF)
- 會話功能(SF)
Ultra-M的高階架構及相關的元件如圖所示：
UltraM架構
本文檔適用於熟悉Cisco Ultra-M平台的思科人員。
縮寫
| HA | 高可用性 |
| VNF | 虛擬網路功能 |
| CF | 控制功能 |
| SF | 服務功能 |
| ESC | 彈性服務控制器 |
| MOP | 程式方法 |
| OSD | 物件儲存磁碟 |
| 硬碟 | 硬碟 |
| SSD | 固態硬碟 |
| VIM | 虛擬基礎架構管理員 |
| VM | 虛擬機器器 |
| EM | 元素管理員 |
| UAS | Ultra自動化服務 |
| UUID | 通用唯一辨識碼 |
MoP的工作流程
EM HA還原程式的高階工作流程
檢查叢集狀態
登入到活動EM並檢查HA狀態。有兩種情況：
1. HA模式為none：
ubuntu@vnfd1deploymentem-0:~$ ncs_cli -u admin -C
admin@scm# show ncs-state ha
ncs-state ha mode none
admin@scm# show ems
%no entries found%
2. EM集群只有一個節點（EM集群包含3個虛擬機器）：
ubuntu@vnfd1deploymentem-0:~$ ncs_cli -u admin -C
admin@scm# show ncs-state ha
ncs-state ha mode master
ncs-state ha node-id 2-1528893823
admin@scm# show ems
EM VNFM
ID SLA SCM PROXY
----------------------
2 up down down
在這兩種情況下，HA狀態都可以透過下一節中提到的步驟來恢復。
HA還原程式
從nova清單中辨識屬於叢集的EM的VM名稱。EM群集中將包含三個VM。
[stack@director ~]$ nova list | grep vnfd1
| e75ae5ee-2236-4ffd-a0d4-054ec246d506 | vnfd1-deployment_c1_0_13d5f181-0bd3-43e4-be2d-ada02636d870 | ACTIVE | - | Running | tmo-autovnf2-uas-orchestration=172.18.180.22; DI-INTERNAL2=192.168.2.17; DI-INTERNAL1=192.168.1.14; tmo-autovnf2-uas-management=172.18.181.23 |
| 33c779d2-e271-47af-8ad5-6a982c79ba62 | vnfd1-deployment_c4_0_9dd6e15b-8f72-43e7-94c0-924191d99555 | ACTIVE | - | Running | tmo-autovnf2-uas-orchestration=172.18.180.13; DI-INTERNAL2=192.168.2.14; DI-INTERNAL1=192.168.1.4; tmo-autovnf2-uas-management=172.18.181.21 |
| 65344d53-de09-4b0b-89a6-85d5cfdb3a55 | vnfd1-deployment_s2_0_b2cbf15a-3107-45c7-8edf-1afc5b787132 | ACTIVE | - | Running | SERVICE-NETWORK1=192.168.10.4, 192.168.10.9; SERVICE-NETWORK2=192.168.20.17, 192.168.20.6; tmo-autovnf2-uas-orchestration=172.18.180.12; DI-INTERNAL2=192.168.2.6; DI-INTERNAL1=192.168.1.12 |
| e1a6762d-4e84-4a86-a1b1-84772b3368dc | vnfd1-deployment_s3_0_882cf1ed-fe7a-47a7-b833-dd3e284b3038 | ACTIVE | - | Running | SERVICE-NETWORK1=192.168.10.22, 192.168.10.14; SERVICE-NETWORK2=192.168.20.5, 192.168.20.14; tmo-autovnf2-uas-orchestration=172.18.180.14; DI-INTERNAL2=192.168.2.7; DI-INTERNAL1=192.168.1.5 |
| b283d43c-6e0c-42e8-87d4-a3af15a61a83 | vnfd1-deployment_s5_0_672bbb00-34f2-46e7-a756-52907e1d3b3d | ACTIVE | - | Running | SERVICE-NETWORK1=192.168.10.21, 192.168.10.24; SERVICE-NETWORK2=192.168.20.21, 192.168.20.24; tmo-autovnf2-uas-orchestration=172.18.180.20; DI-INTERNAL2=192.168.2.13; DI-INTERNAL1=192.168.1.16 |
| 637547ad-094e-4132-8613-b4d8502ec385 | vnfd1-deployment_s6_0_23cc139b-a7ca-45fb-b005-733c98ccc299 | ACTIVE | - | Running | SERVICE-NETWORK1=192.168.10.13, 192.168.10.19; SERVICE-NETWORK2=192.168.20.9, 192.168.20.22; tmo-autovnf2-uas-orchestration=172.18.180.16; DI-INTERNAL2=192.168.2.19; DI-INTERNAL1=192.168.1.21 |
| 4169438f-6a24-4357-ad39-2a35671d29e1 | vnfd1-deployment_vnfd1-_0_02d1510d-53dd-4a14-9e21-b3b367fef5b8 | ACTIVE | - | Running | tmo-autovnf2-uas-orchestration=172.18.180.6; tmo-autovnf2-uas-management=172.18.181.8 |
| 30431294-c3bb-43e6-9bb3-6b377aefbc3d | vnfd1-deployment_vnfd1-_0_f17989e3-302a-4681-be46-f2ebf62b252a | ACTIVE | - | Running | tmo-autovnf2-uas-orchestration=172.18.180.7; tmo-autovnf2-uas-management=172.18.181.9 |
| 28ab33d5-7e08-45fe-8a27-dfb68cf50321 | vnfd1-deployment_vnfd1-_0_f63241f3-2516-4fc4-92f3-06e45054dba0 | ACTIVE | - | Running | tmo-autovnf2-uas-orchestration=172.18.180.3; tmo-autovnf2-uas-management=172.18.181.7 |
從ESC停止其中一個EM，並檢查它是否已進入SHUTOFF狀態。
[admin@vnfm1-esc-0 esc-cli]$ /opt/cisco/esc/esc-confd/esc-cli/esc_nc_cli vm-action STOP vnfd1-deployment_vnfd1-_0_02d1510d-53dd-4a14-9e21-b3b367fef5b8
[admin@vnfm1-esc-0 esc-cli]$ ./esc_nc_cli get esc_datamodel | egrep --color "<state>|<vm_name>|<vm_id>|<deployment_name>"
<snip>
<state>SERVICE_INERT_STATE</state>
vnfd1-deployment_vnfd1-_0_02d1510d-53dd-4a14-9e21-b3b367fef5b8
VM_SHUTOFF_STATE
<vm_name>vnfd1-deployment_vnfd1-_0_f17989e3-302a-4681-be46-f2ebf62b252a</vm_name>
<state>VM_ALIVE_STATE</state>
<vm_name>vnfd1-deployment_vnfd1-_0_f63241f3-2516-4fc4-92f3-06e45054dba0</vm_name>
<state>VM_ALIVE_STATE</state>
現在，當EM進入SHUTOFF STATE時，請從OpenStack Platform Director (OSPD)中重新啟動其他EM。
[stack@director ~]$ nova reboot --hard vnfd1-deployment_vnfd1-_0_f17989e3-302a-4681-be46-f2ebf62b252a
Request to reboot server <Server: vnfd2-deployment_vnfd1-_0_f17989e3-302a-4681-be46-f2ebf62b252a> has been accepted.
再次登入EM VIP並檢查HA狀態。
ubuntu@vnfd1deploymentem-0:~$ ncs_cli -u admin -C
admin@scm# show ncs-state ha
ncs-state ha mode master
ncs-state ha node-id 2-1528893823
如果HA處於「master」狀態，請從ESC啟動之前關閉的EM。否則，繼續從OSPD重新啟動下一個EM，然後再次檢查HA狀態。
[admin@vnfm1-esc-0 esc-cli]$ /opt/cisco/esc/esc-confd/esc-cli/esc_nc_cli vm-action START vnfd1-deployment_vnfd1-_0_02d1510d-53dd-4a14-9e21-b3b367fef5b8
[admin@vnfm1-esc-0 esc-cli]$ ./esc_nc_cli get esc_datamodel | egrep --color "<state>|<vm_name>|<vm_id>|<deployment_name>"
<snip>
<state>SERVICE_ACTIVE_STATE</state>
vnfd1-deployment_vnfd1-_0_02d1510d-53dd-4a14-9e21-b3b367fef5b8
VM_ALIVE_STATE
<vm_name>vnfd1-deployment_vnfd1-_0_f17989e3-302a-4681-be46-f2ebf62b252a</vm_name>
<state>VM_ALIVE_STATE</state>
<vm_name>vnfd1-deployment_vnfd1-_0_f63241f3-2516-4fc4-92f3-06e45054dba0</vm_name>
<state>VM_ALIVE_STATE</state>
從ESC啟動EM之後，請檢查EM的HA狀態。它應該被恢復了。
admin@scm# em-ha-status
ha-status MASTER
admin@scm# show ncs-state ha
ncs-state ha mode master
ncs-state ha node-id 4-1516609103
ncs-state ha connected-slave [ 2-1516609363 ]
admin@scm# show ems
EM VNFM
ID SLA SCM PROXY
---------------------
2 up up up
4 up up up
意見