Try   HackMD

Learning and Practicing Harvester

Preface

本篇文章會主要會介紹以下主題 :

  1. Harvester 的架構
  2. 如何安裝 Harvester，並將額外的節點加入 Harveter Cluster
  3. 將 Harvester 與 Rancher 整合，做出 Nested Kubernetes Clusters 以及 Node Pool

可以透過點擊以下目錄，選擇想看的內容，跳轉至特定章節

目錄

Harvester Architecture

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

Harvester 架構由以下開源技術組成：

  1. Linux 作業系統 : Elemental for SLE-Micro 5.3 是 Harvester 的核心，它是一個不可變的 Linux 發行版，旨在盡可能消除 Kubernetes 叢集中的作業系統維護工作。
  2. 建構於 Kubernetes(RKE2) 之上 : Kubernetes 已成為所有形式因素中最主要的基礎架構語言，而 Harvester 就是一個在引擎蓋下裝有 Kubernetes 的 HCI 解決方案。
  3. 利用 Kubevirt 進行虛擬化管理 : Kubevirt 在 Kubernetes 的基礎上使用 KVM 提供虛擬化管理。
  4. 利用 Longhorn 進行儲存管理 : Longhorn 提供分散式區塊儲存。

Harvester Features

Harvester 是一個 enterprise-ready、易於使用的基礎架構平台，它利用本地直接連接的 Storage，而不是複雜的外部 SAN。 它利用 Kubernetes API 作為跨 Container 和 虛擬機器 workloads 的統一自動化語言。 Harvester 的一些主要功能包括 :

  1. 易於上手 : 由於 Harvester 是以可啟動裝置鏡像的形式提供的，因此您可以使用 ISO 映像直接將其安裝到裸機伺服器上，也可以使用 iPXE 腳本自動安裝。
  2. 虛擬機器生命週期管理 : 輕鬆建立、編輯、複製和刪除虛擬機，包括 SSH 金鑰注入、Cloud Init 以及圖形化界面 (VNC) 和 serial port (22 port) 控制台。
  3. 虛擬機器即時遷移 : 零停機時間將虛擬機器遷移到不同的主機或節點。
  4. 虛擬機器備份、快照和還原 : 從 NFS、S3 伺服器或 NAS 設備備份虛擬機器。 使用備份好的虛擬硬碟檔還原故障的虛擬機，或在不同的叢集上建立新的虛擬機。
  5. 儲存管理 : Harvester 支援分散式區塊的儲存和 tiering (分層)。 Volumes 代表 storage；你可以輕鬆建立、編輯、複製或匯出磁碟區。
  6. 網路管理 : 支援使用虛擬 IP（VIP）和多個網路介面卡（NIC）。 如果虛擬機器需要連接外部網絡，可建立 VLAN 或 untagged (無標記) 網路。
  7. 與 Rancher 整合 : 透過 Rancher 的虛擬化管理頁面直接存取 Rancher 中的 Harvester，並與 Kubernetes 叢集一起管理 VM workloads (虛擬機器工作負載)，實現 Node Pool 的功能。

PreRequest

Hardware and Network Requirements

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

詳細資訊，請以下參考 Harvester 官網 Docs :
Hardware and Network Requirements

Install Single-Node Harvester Clusters with USB

1. 下載 ISO 並透過 Rufus 將 USB 做成可開機裝置

1.1. 下載 ISO

可以使用 ISO 鏡像直接將 Harvester 安裝到裸機伺服器上，若要取得 ISO 映像，請從 Harvester releases 頁面下載 💿 harvester-v1.x.x-amd64.iso。

$ wget --show-progress https://releases.rancher.com/harvester/v1.2.1/harvester-v1.2.1-amd64.iso

螢幕輸出 :

--2024-02-16 10:59:20--  https://releases.rancher.com/harvester/v1.2.1/harvester-v1.2.1-amd64.iso
Resolving releases.rancher.com (releases.rancher.com)... 13.35.35.42, 13.35.35.123, 13.35.35.33, ...
Connecting to releases.rancher.com (releases.rancher.com)|13.35.35.42|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 5497749504 (5.1G) [application/octet-stream]
Saving to: ‘harvester-v1.2.1-amd64.iso’

harvester-v1.2.1-amd64.iso        100%[===========================================================>]   5.12G  35.8MB/s    in 4m 14s

2024-02-16 11:03:34 (20.6 MB/s) - ‘harvester-v1.2.1-amd64.iso’ saved [5497749504/5497749504]

1.2. 開啟 Rufus，將乾淨的 USB 碟插入電腦，Rufus 會自動偵測 USB。 從 "裝置 "下拉式選單中選擇要使用的 USB 裝置，在 "啟動選擇 "中選擇 "選擇"，然後找到要燒錄到 USB 上的 Harvester 安裝 ISO 映像。

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

2. 安裝步驟

2.1. 使用 USB 開機，並透過選擇 Harvester Installer 選項啟動伺服器。

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

使用預設選項，直接按 Enter 按鍵。

2.2. 選擇建立一個新的 Harvester Cluster

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

使用預設選項，直接按 Enter 按鍵。

2.3. 選擇要安裝的 Harvester 系統碟和要儲存虛擬機器資料的資料碟。

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

按 Tab 鍵可展開下拉式選單，按 Enter 鍵儲存設定值。

2.4. 設定主機的 Hostname

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

輸入 h1

2.5. 設定管理網路

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

網路卡的部分，按 Tab 展開下拉式選單，按空白鍵選擇網路卡，確定之後按 Enter 鍵
IPv4 Address 欄位，輸入 : 172.20.0.34/24
Gateway 欄位，輸入 : 172.20.0.254

2.6. 設定 DNS Server

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

DNS Server 欄位 : 172.20.0.254

2.7. 設定 VIP

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

按 Tab 展開下拉式選單，按向下方向鍵選擇 Static，確認後按 Enter 鍵。
VIP 欄位 : 172.20.0.36

2.8. 設定 Cluster Token

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

Cluster token 欄位 : lab
Cluster token 用於將其他節點新增至 Harvester 叢集。

2.9. 設定使用者密碼

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

預設的使用者是 rancher。

2.10. 設定 NTP Server

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

NTP Server 欄位 : time.google.com。

2.11. 設定 HTTP Proxy Server ( 此步驟不一定要設定，取決於當下的環境 )

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

如果不設定就空白，直接按 Enter 鍵。

2.12. 匯入 SSH Keys ( 此步驟不一定要設定 )

image

如果不設定就空白，直接按 Enter 鍵。

2.13. 設定 Harvester 設定檔自訂主機 ( 此步驟不一定要設定 )

image

如果不設定就空白，直接按 Enter 鍵。

2.14. 查看並確認安裝選項

image

確認安裝選項沒問題後，按 Enter 鍵，Harvester 將會安裝到主機上。

2.15. 安裝完成後，節點將重新啟動，以下是啟動完成畫面

image

重新啟動後，Harvester 控制台會顯示管理 URL 和狀態。網路介面的預設 URL 是 https://your-virtual-ip。 可以使用 F12 按鍵從 Harvester 控制台切換到 Shell，然後鍵入 exit 返回 Harvester 控制台。

2.16. 第一次登入的時候，系統會提示您設定預設管理員使用者的密碼。

image

使用者預設是 admin

Update Harvester Configuration After Installation

DNS servers

  1. SSH 連線至 Harvester Node
$ ssh rancher@172.20.0.36
  1. 備份基本的 cloud-init 檔案 /oem/90_custom.yaml
$ sudo cp /oem/90_custom.yaml /oem/install/90_custom.yaml.$(date --iso-8601=minutes)
  1. 修改 /oem/90_custom.yaml 檔案，並且更新以下 Yaml 檔中路徑的值 stages.initramfs[0].commands
$ sudo vim /oem/90_custom.yaml

修改範例內容如下 :

name: Harvester Configuration
stages:
  initramfs:
  - commands:
    - modprobe kvm
    - modprobe vhost_net
    - sed -i 's/^NETCONFIG_DNS_STATIC_SERVERS.*/NETCONFIG_DNS_STATIC_SERVERS="172.20.0.37"/'
      /etc/sysconfig/network/config

將原本的 172.20.0.254 改成 172.20.0.37

  1. 重新開機
$ sudo reboot
  1. 確認 DNS Server 修改正確
$ cat /etc/resolv.conf | grep -v '#'

螢幕輸出如下 :

nameserver 172.20.0.37

設定 Storage Class

  1. 預設的 StorageClass 他的 Replicas 會被設為 3 ，但因為只有單一台 Node，所以要額外新增一個 StorageClass 並將 Replicas 設為 1，否則 VM 會無法建立。
  2. 點選左側選單 Advanced -> Storage Classes -> Create

image

  1. 輸入以下資訊 :
  • Name: new-sc
  • Number of Replicas 設為 1

確認資訊輸入正確後，點選 Create

image

  1. 設成 Default

image

透過 URL 下載 Cloud Images

  1. 左側選單選擇 Images -> Create

image

  1. 透過 URL 下載 OpenSUSE Leap 15.5 的 Cloud Images

image

URL 如下 :
https://download.opensuse.org/repositories/Cloud:/Images:/Leap_15.5/images/openSUSE-Leap-15.5.x86_64-NoCloud.qcow2

  1. 確認下載完成

image

新增 Cloud Config Templates

User Data

  1. 點選左側選單 Cloud Config Templates -> Create

image

  1. 輸入以下資訊 :
  • Name: opensuse

































#cloud-config
package_update: true
packages:
  - qemu-guest-agent
write_files:
  - path: /etc/sysctl.conf
    permissions: 0644
    owner: root
    content: |
      net.ipv6.conf.all.disable_ipv6 = 1
      net.ipv6.conf.default.disable_ipv6 = 1
      net.ipv6.conf.lo.disable_ipv6 = 1
runcmd:
  - - systemctl
    - enable
    - '--now'
    - qemu-guest-agent.service
  - yast dns edit nameserver1=172.20.0.37
  - timedatectl set-timezone Asia/Taipei
  - systemctl restart network
  - sysctl -p
ssh_pwauth: true
ssh_authorized_keys:
  - ssh-rsa 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 rancher@rancher
users:
  - default
  - name: rancher
    gecos: rancher
    groups: users
    lock_passwd: false
    sudo: ALL=(ALL) NOPASSWD:ALL
    plain_text_passwd: 'rancher'
    shell: /bin/bash
  • package_update，會先更新系統的套件
  • packages，代表要安裝的套件
  • write_files，針對特定檔案寫入特定內容，這邊的範例是關 ipv6
  • runcmd，系統開機以後要執行的命令，這邊有設定啟動 qemu-guest-agent 服務，並設為開機自動啟動，還有設定 DNS Server 的位置，還有設定時區，特別設定 DNS Server 的原因是因為要可以跟 Rancher 溝通
  • ssh_authorized_keys，讓跳板機可以免密碼登入使用這個 Cloud init config 做出來的 VM
  • users，新增使用者，並設定使用者的群組、sudo 權限和密碼。

確認資訊輸入正確後，點選 Create

image

Network Data

Static













network:
  version: 1
  config:
    - type: physical
      name: eth0
      subnets:
        - type: static
          address: 172.20.0.43/24
          gateway: 172.20.0.253
    - type: nameserver
      address:
        - 172.20.0.40

DHCP








network:
  version: 1
  config:
    - type: physical
      name: eth0
      subnets:
        - type: dhcp

建立 VM Network

  1. 點選左側選單 Networks -> VM Networks -> Create

image

  1. 編輯以下設定資訊
  • Name: vlan
  • Vlan ID: 100
  • Cluster Netwrok: mgmt

確認沒問題後，點選 Create

image

  1. 確認是否建立成功

State 和 Route Connectivity 的狀態都要顯示 Active

image

將其他節點加入 Harvester 叢集

1. 將燒有 Harvester ISO 的 USB 可開機裝置插入伺服器，並使用 USB 開機。

2. 安裝步驟

2.1. 使用 USB 開機，並透過選擇 Harvester Installer 選項啟動伺服器。

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

使用預設選項，直接按 Enter 按鍵。

2.2. 選擇加入一個已經存在的 Harvester Cluster

image

選擇 Join an existing Harvester cluster 後，按 Enter 按鍵。

2.3. 選擇要安裝的 Harvester 系統碟和要儲存虛擬機器資料的資料碟。

image

按 Tab 鍵可展開下拉式選單，按 Enter 鍵儲存設定值。

2.4. 設定主機的 Hostname

image

輸入 h2

2.5. 設定網路

image

網路卡的部分，按 Tab 展開下拉式選單，按空白鍵選擇網路卡，確定之後按 Enter 鍵
IPv4 Address 欄位，輸入 : 172.20.0.35/24
Gateway 欄位，輸入 : 172.20.0.254

2.6. 設定 DNS Server

image

DNS Server 欄位 : 172.20.0.37

2.7. 設定 管理 IP Address

image

Management address 欄位 : 172.20.0.36

2.8. 設定 Cluster Token

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

Cluster token 欄位 : lab
Cluster token 用於將其他節點新增至 Harvester 叢集。

2.9. 設定使用者密碼

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

預設的使用者是 rancher。

2.10. 設定 NTP Server

image

NTP Server 欄位 : time.google.com。

2.11. 設定 HTTP Proxy Server ( 此步驟不一定要設定，取決於當下的環境 )

Image Not Showing Possible Reasons
  • The image was uploaded to a note which you don't have access to
  • The note which the image was originally uploaded to has been deleted
Learn More →

如果不設定就空白，直接按 Enter 鍵。

2.12. 匯入 SSH Keys ( 此步驟不一定要設定 )

image

如果不設定就空白，直接按 Enter 鍵。

2.13. 設定 Harvester 設定檔自訂主機 ( 此步驟不一定要設定 )

image

如果不設定就空白，直接按 Enter 鍵。

2.14. 查看並確認安裝選項

image

確認安裝選項沒問題後，按 Enter 鍵，Harvester 將會安裝到主機上。

2.15. 安裝完成後，節點將重新啟動，以下是啟動完成畫面

image

2.16. 打開 Harvester 網站，檢查 Node 是否變 2 台

點選左側選單 Hosts

image

使用者預設是 admin

與 Rancher 整合

PreRequest

Harvester v1.2.1 肚子裡面的 Rancher 無法啟動虛擬化管理，所以需額外再安裝並啟動 Rancher Server 2.7.9，或是在 Harvester 啟動 rancher-vcluster (正在開發的功能)。

Rancher 的安裝請參考我的另一篇筆記 RKE2 & Rancher Prime Installation

Rancher 啟動 Virtualization management

  1. 打開虛擬化管理介面

點選左側選單 Virtualization Management

image

  1. Import Harvester Cluster

點選 Import Existing

image

  1. 填入叢集資訊
  • Cluster Name: harvester

image

  1. 填入叢集資訊
  • Cluster Name: harvester

image

  1. 複製讓 Harvester 與 Rancher 註冊的 URL

image

  1. 回到 Harvester 的網站，與 Rancher 註冊 (1/2)

點選左側選單 Settings -> 找到 cluster-registration-url -> 點選 Edit Setting

image

  1. 回到 Harvester 的網站，與 Rancher 註冊 (2/2)

將註冊 URL 貼上 -> 點選 Save

image

  1. 到 Rancher UI 確認

image

Creating an K3s Kubernetes Cluster

  1. 在 Home Page 點選 Create

image

  1. Node Driver 選擇 Harvester

image

  1. 建立 Cloud Credential
  • Credential Name: harvester
  • Imported Harvester Cluster 選擇 harvester
  • 點選 Continue

image

  1. 編輯 Machine Pools

VM 必須要能透過 DNS Server 解析 Rancher 的 FQDN，否則 K3S 將無法建立。

  • Cluster Name: bobo
  • CPU: 4
  • Memory: 8
  • Namespace: default
  • SSH User: Opensuse
  • Volumes:
    • Image Volume: openSUSE-Leap-15.5.x86_64-NoCloud.qcow2
    • Disk: 100

image

  • Networks:
    • Network Name: default/vlan (vlanId=100)
  • 點選 Show Advanced 按鈕

image

  • User Data:
    • User Data Template: opensuse

image

  1. 編輯 Cluster Configuration
  • 選擇 Kubernetes 版本 : k3s v1.26.13+k3s2
  • 點選 Create

image

  1. 檢查 K3S 是否建立成功

點選 Provisioning Log

image

要 Troubleshooting 的話，可以 ssh 連線到 VM 中，看以下兩個檔案 :

  • /var/log/cloud-init.log。
  • /var/log/cloud-init-output.log，這檔案裡面紀錄 cloud init 命令執行的結果，包含透過 Rancher 安裝 K3S 的過程都會記在這個檔案中。

K3s node Scale up to 2

  1. Scale up node

點選 Machine Pools -> +

image

  1. 檢查是否符合預期

image

K3s node Scale down to 1

  1. Scale down node

點選 Machine Pools -> -

image

  1. 檢查是否符合預期

image

Add Additional Disks

  1. 再新增一顆硬碟後，要先 Format 成 ext4 檔案系統再重新開機， Harvester 才認的到
  2. Go to the Hosts page.
  3. On the node you want to modify, click ⋮ > Edit Config.
  4. Select the Storage tab and click Add Disk.
  5. Select an additional raw block device to add as an additional data volume.
  6. The Force Formatted option is required if the block device has never been force-formatted.
  7. Last, you can click ⋮ > Edit Config again to check the newly added disk.

Resize Disk (虛擬化環境)

  1. 先在虛擬化那層擴硬碟空間
  2. SSH 連線進那台 node，並切換成 root 使用者
  3. Resize Partitaion 大小
# 確認 /var/lib/harvester/defaultdisk 在哪顆硬碟
$ df -h /var/lib/harvester/defaultdisk
Filesystem      Size  Used Avail Use% Mounted on
/dev/sda6       330G   30G  283G  10% /var/lib/harvester/defaultdisk

$ fdisk /dev/sda
Command (m for help): p

Disk /dev/sda: 712 GiB, 764504178688 bytes, 1493172224 sectors
Disk model: QEMU HARDDISK
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: C1C6588D-E0EB-42EB-A68B-3676181AED1E

Device         Start        End   Sectors   Size Type
/dev/sda1       2048       4095      2048     1M BIOS boot
/dev/sda2       4096     106495    102400    50M Linux filesystem
/dev/sda3     106496   16883711  16777216     8G Linux filesystem
/dev/sda4   16883712   48340991  31457280    15G Linux filesystem
/dev/sda5   48340992  369205247 320864256   153G Linux filesystem
/dev/sda6  369205248 1073739775 704534528 335.9G Linux filesystem

Command (m for help): d
Partition number (1-6, default 6):

Partition 6 has been deleted.

Command (m for help): n
Partition number (6-128, default 6):
First sector (369205248-1493172190, default 369205248):
Last sector, +/-sectors or +/-size{K,M,G,T,P} (369205248-1493172190, default 1493172190):

Created a new partition 6 of type 'Linux filesystem' and of size 535.9 GiB.
Partition #6 contains a ext4 signature.

Do you want to remove the signature? [Y]es/[N]o: N

Command (m for help): p

Disk /dev/sda: 712 GiB, 764504178688 bytes, 1493172224 sectors
Disk model: QEMU HARDDISK
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: C1C6588D-E0EB-42EB-A68B-3676181AED1E

Device         Start        End    Sectors   Size Type
/dev/sda1       2048       4095       2048     1M BIOS boot
/dev/sda2       4096     106495     102400    50M Linux filesystem
/dev/sda3     106496   16883711   16777216     8G Linux filesystem
/dev/sda4   16883712   48340991   31457280    15G Linux filesystem
/dev/sda5   48340992  369205247  320864256   153G Linux filesystem
/dev/sda6  369205248 1493172190 1123966943 535.9G Linux filesystem

Command (m for help): w
The partition table has been altered.
Syncing disks.
  1. 執行以下命令 Resize Longhorn 的 Filesystem
$ resize2fs -p /dev/sda6
    
$ df -h /var/lib/harvester/defaultdisk
Filesystem      Size  Used Avail Use% Mounted on
/dev/sda6       527G   30G  472G   6% /var/lib/harvester/defaultdisk
  1. 再進 Longhorn UI 看就會發現空間擴出來了

Stop VM with cli

# 1. SSH into harvester node
$ ssh rancher@x.x.x.x

# 2. Get VVirtual Machine Instance
$ kubectl get vmi
NAME                      AGE     PHASE     IP            NODENAME   READY
harbor                    12m     Running                 hh         False
hehe                      6m41s   Running   10.52.0.180   hh         True
k3-pool1-2d0bd15b-4fv96   14h     Running   172.20.0.42   hh         True
rancher                   14h     Running   172.20.0.41   hh         True

# 3. Stop VM
$ virtctl stop harbor
VM harbor was scheduled to stop

參考文章

Last changed by 
 
Antony
2
453

Read more

SUSE Edge Image Builder 介紹與實作

SUSE Edge 是專為解決在 edge 部署 infrastructure (基礎架構) 和 cloud-native applications (雲原生應用程式) 的獨特挑戰，而設計、緊密整合且經過全面驗證的端對端解決方案。

Apr 15, 2025
Install NVIDIA GPU Operator on SL Micro

本文將解釋 NVIDIA GPU Operator，概述其所管理的 NVIDIA GPU 元件，並總結使用 NVIDIA GPU Operator 的好處。 1.1 NVIDIA GPU Operator 是什麼？ NVIDIA GPU Operator 是 Kubernetes 的操作者，可簡化 Kubernetes 集群中 NVIDIA GPU 資源的管理與部署。 它能自動化配置與監控 NVIDIA GPU 驅動程式，以及相關元件，例如 CUDA、容器執行時，以及其他 GPU 相關軟體。

Apr 14, 2025
RKE2 & Rancher Prime Installation

本篇文章會主要會介紹，如何安裝 Rancher Kubernetes Engine 2 或 K3S，並在 Kubernetes 之中部屬 Rancher Prime

Mar 27, 2025
How to Pull Images from Google Artifact Registry into RKE2

如 RKE2 沒有被 Rancher 管理，則叢集中的每台節點都需要做設定

Mar 26, 2025
Read more from Antony
Published on HackMD