在CNDP部署中對基於RCM的AIO伺服器進行RMA的步驟
下載選項
已更新: 2022 年 7 月 20 日
文件 ID:217620
無偏見用語
本產品的文件集力求使用無偏見用語。針對本文件集的目的，無偏見係定義為未根據年齡、身心障礙、性別、種族身分、民族身分、性別傾向、社會經濟地位及交織性表示歧視的用語。由於本產品軟體使用者介面中硬式編碼的語言、根據 RFP 文件使用的語言，或引用第三方產品的語言，因此本文件中可能會出現例外狀況。深入瞭解思科如何使用包容性用語。
關於此翻譯
思科已使用電腦和人工技術翻譯本文件，讓全世界的使用者能夠以自己的語言理解支援內容。請注意，即使是最佳機器翻譯，也不如專業譯者翻譯的內容準確。Cisco Systems, Inc. 對這些翻譯的準確度概不負責，並建議一律查看原始英文文件（提供連結）。
目錄
簡介
本檔案介紹在雲原生部署平台(CNDP)部署中針對基於冗餘配置管理器(RCM)的多功能一體(AIO)伺服器的返回材料授權(RMA)的詳細程式，以解決任何硬體問題或維護相關活動。
必要條件
需求
思科建議您瞭解以下主題：
- RCM
- 庫伯內特斯
採用元件
本檔案中的資訊是根據RCM版本 — rcm.2021.02.1.i18
本文中的資訊是根據特定實驗室環境內的裝置所建立。文中使用到的所有裝置皆從已清除（預設）的組態來啟動。如果您的網路運作中，請確保您瞭解任何指令可能造成的影響。
瞭解RCM IP方案
本檔案將說明RCM設計，包括兩個AIO節點，其中兩個RCM Openters和一個RCM CEE，每個AIO節點一個。
本文中RMA的目標RCM AIO節點是AIO-1(AI0301)，它同時包含處於主狀態的RCM運算子。
POD_NAME |
NODE_NAME |
IP_ADDRESS |
DEVICE_TYPE |
OS_TYPE |
UP0300 |
RCE301 |
10.1.2.9 |
RCM_CEE_AIO_1 |
opscenter |
UP0300 |
RCE302 |
10.1.2.10 |
RCM_CEE_AIO_2 |
opscenter |
UP0300 |
AI0301 |
10.1.2.7 |
RCM_K8_AIO_1 |
linux |
UP0300 |
AI0302 |
10.1.2.8 |
RCM_K8_AIO_2 |
linux |
UP0300 |
RM0301 |
10.1.2.3 |
RCM1_ACTIVE |
opscenter |
UP0300 |
RM0302 |
10.1.2.4 |
RCM1_STANDBY |
opscenter |
UP0300 |
RM0303 |
10.1.2.5 |
RCM2_ACTIVE |
opscenter |
UP0300 |
RM0304 |
10.1.2.6 |
RCM2_STANDBY |
opscenter |
備份過程
備份配置
首先，從目標AIO節點上運行的RCM操作程式收集運行配置的配置備份。
# show running-config | nomore
從目標AIO節點上運行的RCM CEE操作程式收集運行配置。
# show running-config | nomore
預檢查過程
對AIO進行預檢查
收集兩個AIO節點的命令輸出並驗證所有Pod是否均處於Running狀態。
# kubectl get ns
# kubectl get pods -A -o wide
Prechecks輸出示例
請注意，兩個RCM操作員和一個在AIO-1節點上運行的RCM CEE操作中心
cloud-user@up0300-aio-1-master-1:~$ kubectl get ns
NAME STATUS AGE
cee-rce301 Active 110d <--
default Active 110d
istio-system Active 110d
kube-node-lease Active 110d
kube-public Active 110d
kube-system Active 110d
nginx-ingress Active 110d
rcm-rm0301 Active 110d <--
rcm-rm0303 Active 110d <--
registry Active 110d
smi-certs Active 110d
smi-node-label Active 110d
smi-vips Active 110d
cloud-user@up0300-aio-1-master-1:~$
登入到AIO-1的兩個RCM操作中心並驗證狀態。
[up0300-aio-1/rm0301] rcm# rcm show-status
message :
{"status":[" Fri Oct 29 07:21:11 UTC 2021 : State is MASTER"]}
[up0300-aio-1/rm0301] rcm#
[up0300-aio-1/rm0303] rcm# rcm show-status
message :
{"status":[" Fri Oct 29 07:22:18 UTC 2021 : State is MASTER"]}
[up0300-aio-1/rm0303] rcm#
在AIO-2節點上重複相同步驟，其中其他兩個RCM運算子對應於AIO-1節點。
cloud-user@up0300-aio-2-master-1:~$ kubectl get ns
NAME STATUS AGE
cee-rce302 Active 105d <--
default Active 105d
istio-system Active 105d
kube-node-lease Active 105d
kube-public Active 105d
kube-system Active 105d
nginx-ingress Active 105d
rcm-rm0302 Active 105d <--
rcm-rm0304 Active 105d <--
registry Active 105d
smi-certs Active 105d
smi-node-label Active 105d
smi-vips Active 105d
cloud-user@up0300-aio-2-master-1:~$
登入到AIO-2的兩個RCM操作中心並驗證狀態。
[up0300-aio-2/rm0302] rcm# rcm show-status
message :
{"status":[" Fri Oct 29 09:32:54 UTC 2021 : State is BACKUP"]}
[up0300-aio-2/rm0302] rcm#
[up0300-aio-2/rm0304] rcm# rcm show-status
message :
{"status":[" Fri Oct 29 09:33:51 UTC 2021 : State is BACKUP"]}
[up0300-aio-2/rm0304] rcm#
執行過程
關閉AIO節點之前在RCM上執行的步驟
- 由於AIO-1上的兩個RCM都是主控制器，因此您可以將其遷移到BACKUP。
a.為此，您必須先在活動RCM上執行rcm migrate primary命令，然後再關閉AIO-1伺服器。
[up0300-aio-1/rm0301] rcm# rcm migrate primary
[up0300-aio-1/rm0303] rcm# rcm migrate primary
b.驗證AIO-1上的狀態現在為BACKUP。
[up0300-aio-1/rm0301] rcm# rcm show-status
[up0300-aio-1/rm0303] rcm# rcm show-status
c.驗證AIO-2上的狀態現在是否為MASTER，並確保其為MASTER。
[up0300-aio-1/rm0302] rcm# rcm show-status
[up0300-aio-1/rm0304] rcm# rcm show-status
d.對rm0301和rm0303執行RCM關閉。
[up0300-aio-2/rm0301] rcm# config
Entering configuration mode terminal
[up0300-aio-2/rm0301] rcm(config)# system mode shutdown
[up0300-aio-1/rce301] rcm(config)# commit comment <CRNUMBER>
[up0300-aio-2/rm0303] rcm# config
Entering configuration mode terminal
[up0300-aio-2/rm0303] rcm(config)# system mode shutdown
[up0300-aio-1/rce303] rcm(config)# commit comment <CRNUMBER>
2.我們還必須關閉AIO-1上運行的CEE操作（使用的命令）。
[up0300-aio-1/rce301] cee# config
Entering configuration mode terminal
[up0300-aio-1/rce301] cee(config)# system mode shutdown
[up0300-aio-1/rce301] cee(config)# commit comment <CRNUMBER>
[up0300-aio-1/rce301] cee(config)# exit
等待幾分鐘，然後檢查系統以顯示0.0%。
[up0300-aio-1/rce301] cee# show system
3.驗證RCM和CEE名稱空間沒有Pod，除了文檔、智慧代理、ops-center-rcm和ops-center-cee Pod
# kubectl get pods -n rcm-rm0301 -o wide
# kubectl get pods -n rcm-rm0303 -o wide
# kubectl get pods -n cee-rce302 -o wide
關閉AIO節點之前在Kubernetes節點上執行的步驟
排出Kubernetes節點，以便正常終止關聯的池和服務。排程程式不再選擇此Kubernetes節點並從該節點中移除Pod。請一次清空一個節點。
登入到SMI群集管理器。
cloud-user@bot-deployer-cm-primary:~$ kubectl get svc -n smi-cm
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
cluster-files-offline-smi-cluster-deployer ClusterIP 10.102.108.177 <none> 8080/TCP 78d
iso-host-cluster-files-smi-cluster-deployer ClusterIP 10.102.255.174 192.168.0.102 80/TCP 78d
iso-host-ops-center-smi-cluster-deployer ClusterIP 10.102.58.99 192.168.0.100 3001/TCP 78d
netconf-ops-center-smi-cluster-deployer ClusterIP 10.102.108.194 10.244.110.193 3022/TCP,22/TCP 78d
ops-center-smi-cluster-deployer ClusterIP 10.102.156.123 <none> 8008/TCP,2024/TCP,2022/TCP,7681/TCP,3000/TCP,3001/TCP 78d
squid-proxy-node-port NodePort 10.102.73.130 <none> 3128:31677/TCP 78d
cloud-user@bot-deployer-cm-primary:~$ ssh -p 2024 admin@<Cluster IP of ops-center-smi-cluster-deployer>
Welcome to the Cisco SMI Cluster Deployer on bot-deployer-cm-primary
Copyright © 2016-2020, Cisco Systems, Inc.
All rights reserved.
admin connected from 192.168.0.100 using ssh on ops-center-smi-cluster-deployer-686b66d9cd-nfzx8
[bot-deployer-cm-primary] SMI Cluster Deployer#
[bot-deployer-cm-primary] SMI Cluster Deployer# show clusters
LOCK TO
NAME VERSION
----------------------------
cp0100-smf-data -
cp0100-smf-ims -
cp0200-smf-data -
cp0200-smf-ims -
up0300-aio-1 - <--
up0300-aio-2 -
up0300-upf-data -
up0300-upf-ims -
排出主節點：
[bot-deployer-cm-primary] SMI Cluster Deployer# clusters up0300-aio-1 nodes master-1 actions sync drain remove-node true
This would run drain on the node, disrupting pods running on the node. Are you sure? [no,yes] yes
message accepted
將master-1節點標籤為維護模式：
[bot-deployer-cm-primary] SMI Cluster Deployer# config
Entering configuration mode terminal
[bot-deployer-cm-primary] SMI Cluster Deployer(config)# clusters up0300-aio-1
[bot-deployer-cm-primary] SMI Cluster Deployer(config-clusters-up0300-aio-1)# nodes master-1
[bot-deployer-cm-primary] SMI Cluster Deployer(config-nodes-master1)# maintenance true
[bot-deployer-cm-primary] SMI Cluster Deployer(config-nodes-master1)# commit
Commit complete.
[bot-deployer-cm-primary] SMI Cluster Deployer(config-nodes-master1)# end
運行群集同步並監視同步操作的日誌：
[bot-deployer-cm-primary] SMI Cluster Deployer# clusters up0300-aio-1 nodes master-1 actions sync
This would run sync. Are you sure? [no,yes] yes
message accepted
[bot-deployer-cm-primary] SMI Cluster Deployer# clusters up0300-aio-1 nodes master-1 actions sync logs
群集同步日誌的輸出示例：
[installer-master] SMI Cluster Deployer# clusters kali-stacked nodes cmts-worker1-1 actions sync logs
Example Cluster Name: kali-stacked
Example WorkerNode: cmts-worker1
logs 2020-10-06 20:01:48.023 DEBUG cluster_sync.kali-stacked.cmts-worker1: Cluster name: kali-stacked
2020-10-06 20:01:48.024 DEBUG cluster_sync.kali-stacked.cmts-worker1: Node name: cmts-worker1
2020-10-06 20:01:48.024 DEBUG cluster_sync.kali-stacked.cmts-worker1: debug: false
2020-10-06 20:01:48.024 DEBUG cluster_sync.kali-stacked.cmts-worker1: remove_node: true
PLAY [Check required variables] ************************************************
TASK [Gathering Facts] *********************************************************
Tuesday 06 October 2020 20:01:48 +0000 (0:00:00.017) 0:00:00.017 *******
ok: [master3]
ok: [master1]
ok: [cmts-worker1]
ok: [cmts-worker3]
ok: [cmts-worker2]
ok: [master2]
TASK [Check node_name] *********************************************************
Tuesday 06 October 2020 20:01:50 +0000 (0:00:02.432) 0:00:02.450 *******
skipping: [master1]
skipping: [master2]
skipping: [master3]
skipping: [cmts-worker1]
skipping: [cmts-worker2]
skipping: [cmts-worker3]
PLAY [Wait for ready and ensure uncordoned] ************************************
TASK [Cordon and drain node] ***************************************************
Tuesday 06 October 2020 20:01:51 +0000 (0:00:00.144) 0:00:02.594 *******
skipping: [master1]
skipping: [master2]
skipping: [master3]
skipping: [cmts-worker2]
skipping: [cmts-worker3]
TASK [upgrade/cordon : Cordon/Drain/Delete node] *******************************
Tuesday 06 October 2020 20:01:51 +0000 (0:00:00.205) 0:00:02.800 *******
changed: [cmts-worker1 -> 172.22.18.107]
PLAY RECAP *********************************************************************
cmts-worker1 : ok=2 changed=1 unreachable=0 failed=0 skipped=1 rescued=0 ignored=0
cmts-worker2 : ok=1 changed=0 unreachable=0 failed=0 skipped=2 rescued=0 ignored=0
cmts-worker3 : ok=1 changed=0 unreachable=0 failed=0 skipped=2 rescued=0 ignored=0
master1 : ok=1 changed=0 unreachable=0 failed=0 skipped=2 rescued=0 ignored=0
master2 : ok=1 changed=0 unreachable=0 failed=0 skipped=2 rescued=0 ignored=0
master3 : ok=1 changed=0 unreachable=0 failed=0 skipped=2 rescued=0 ignored=0
Tuesday 06 October 2020 20:02:29 +0000 (0:00:38.679) 0:00:41.479 *******
===============================================================================
2020-10-06 20:02:30.057 DEBUG cluster_sync.kali-stacked.cmts-worker1: Cluster sync successful
2020-10-06 20:02:30.058 DEBUG cluster_sync.kali-stacked.cmts-worker1: Ansible sync done
2020-10-06 0:02:30.058 INFO cluster_sync.kali-stacked.cmts-worker1: _sync finished. Opening lock
伺服器維護程式
從CIMC正常關閉伺服器。按照硬體MoP中的定義繼續硬體相關維護活動，並確保在伺服器加電後通過所有運行狀況檢查。
注意：本文不包括伺服器的硬體或維護活動MoP，因為它們與問題陳述不同
Kubernetes還原過程
在AIO節點上加電後Kubernetes節點上執行的步驟
登入到SMI群集管理器：
cloud-user@bot-deployer-cm-primary:~$ kubectl get svc -n smi-cm
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
cluster-files-offline-smi-cluster-deployer ClusterIP 10.102.108.177 <none> 8080/TCP 78d
iso-host-cluster-files-smi-cluster-deployer ClusterIP 10.102.255.174 192.168.0.102 80/TCP 78d
iso-host-ops-center-smi-cluster-deployer ClusterIP 10.102.58.99 192.168.0.100 3001/TCP 78d
netconf-ops-center-smi-cluster-deployer ClusterIP 10.102.108.194 10.244.110.193 3022/TCP,22/TCP 78d
ops-center-smi-cluster-deployer ClusterIP 10.102.156.123 <none> 8008/TCP,2024/TCP,2022/TCP,7681/TCP,3000/TCP,3001/TCP 78d
squid-proxy-node-port NodePort 10.102.73.130 <none> 3128:31677/TCP 78d
cloud-user@bot-deployer-cm-primary:~$ ssh -p 2024 admin@<ClusterIP of ops-center-smi-cluster-deployer>
Welcome to the Cisco SMI Cluster Deployer on bot-deployer-cm-primary
Copyright © 2016-2020, Cisco Systems, Inc.
All rights reserved.
admin connected from 192.168.0.100 using ssh on ops-center-smi-cluster-deployer-686b66d9cd-nfzx8
[bot-deployer-cm-primary] SMI Cluster Deployer#
[bot-deployer-cm-primary] SMI Cluster Deployer# show clusters
LOCK TO
NAME VERSION
----------------------------
cp0100-smf-data -
cp0100-smf-ims -
cp0200-smf-data -
cp0200-smf-ims -
up0300-aio-1 - <--
up0300-aio-2 -
up0300-upf-data -
up0300-upf-ims -
關閉要重新新增到集群中的master-1的維護標誌。
[bot-deployer-cm-primary] SMI Cluster Deployer# config
Entering configuration mode terminal
[bot-deployer-cm-primary] SMI Cluster Deployer(config)# clusters up0300-aio-1
[bot-deployer-cm-primary] SMI Cluster Deployer(config-clusters-up0300-aio-1)# nodes master-1
[bot-deployer-cm-primary] SMI Cluster Deployer(config-nodes-master-1)# maintenance false
[bot-deployer-cm-primary] SMI Cluster Deployer(config-nodes-master-1)# commit
Commit complete.
[bot-deployer-cm-primary] SMI Cluster Deployer(config-nodes-master-1)# end
使用群集同步操作恢復主節點Pod和服務。
[bot-deployer-cm-primary] SMI Cluster Deployer# clusters up0100-aio-1 nodes master-1 actions sync run debug true
This would run sync. Are you sure? [no,yes] yes
message accepted
監視同步操作的日誌。
[bot-deployer-cm-primary] SMI Cluster Deployer# clusters up0100-aio-1 nodes master-1 actions sync logs
檢查AIO-1主機的群集狀態。
[bot-deployer-cm-primary] SMI Cluster Deployer# clusters up0300-aio-1 actions k8s cluster-status
輸出示例：
[installer-] SMI Cluster Deployer# clusters kali-stacked actions k8s cluster-status
pods-desired-count 67
pods-ready-count 67
pods-desired-are-ready true
etcd-healthy true
all-ok true
RCM還原過程
在CEE和RCM運營中心執行恢復應用程式的步驟
將CEE opscenter和RCM opscenter更新為運行模式。
配置rce301的運行模式。
[up0300-aio-1/rce301] cee# config
Entering configuration mode terminal
[up0300-aio-1/rce301] cee(config)# system mode running
[up0300-aio-1/rce301] cee(config)# commit comment <CRNUMBER>
[up0300-aio-1/rce301] cee(config)# exit
等待幾分鐘，檢查系統是否為100.0%。
[up0300-aio-1/rce301] cee# show system
配置rm0301的運行模式。
[up0300-aio-2/rm0301] rcm# config
Entering configuration mode terminal
[up0300-aio-2/rm0301] rcm(config)# system mode running
[up0300-aio-1/rce301] rcm(config)# commit comment <CRNUMBER>
等待幾分鐘，驗證系統是否為100.0%。
[up0300-aio-1/rm0301] cee# show system
配置rm0303的運行模式。
[up0300-aio-2/rm0303] rcm# config
Entering configuration mode terminal
[up0300-aio-2/rm0303] rcm(config)# system mode running
[up0300-aio-1/rce303] rcm(config)# commit comment <CRNUMBER>
等待幾分鐘，檢查系統是否為100.0%。
[up0300-aio-1/rm0303] cee# show system
驗證程式
使用以下命令驗證兩個AIO節點上的Pod是否均處於UP和Running狀態。
on AIO nodes:
kubectl get ns
kubectl get pods -A -o wide
on RCM ops-centers:
rcm show-status
修訂記錄
| 修訂 | 發佈日期 | 意見 |
|---|---|---|
2.0 |
20-Jul-2022 |
新增了cluster sync命令並更改了恢復過程中的步驟。 |
1.0 |
11-Jan-2022 |
初始版本 |
由思科工程師貢獻
- Venkata Naga Rajesh BadvetiCisco TAC Engineer
意見