🧯BuckUp

This section documents backup strategies for each middleware/app deployed in the 72602 cluster.

Backup Quick Reference

Service	Method	Data	Frequency	Restore
PostgreSQL	`pg_dump`	All databases	Daily crons	`psql` restore
Redis	RDB/AOF	Cache data	On-demand	Auto on restart
Minio	`mc mirror`	Object storage	Periodic	`mc mirror` reverse
N8N	Web UI export	Workflows + credentials	Manual	Web UI import
Uptime Kuma	Web UI Backup	Monitors + settings	Manual	Web UI upload
Open WebUI	SQLite copy	Chat history + config	Manual	File replace
Homepage	Git-managed	Dashboard config	On commit	`kubectl apply`
ElasticSearch	Snapshot API	Indexes	Scheduled	Snapshot restore
Git repos	GitHub hosting	Source code	Automatic	Git clone

PostgreSQL

Prerequisites

alias pgk='kubectl -n database exec deployment/postgresql --'

全量备份所有数据库

pgk pg_dumpall -U postgres > postgresql-full-$(date +%Y%m%d).sql

备份单个数据库

pgk pg_dump -U postgres <dbname> > <dbname>-$(date +%Y%m%d).sql

备份到 PVC 内

pgk sh -c 'pg_dumpall -U postgres > /bitnami/postgresql/backup-$(date +%Y%m%d).sql'

从备份文件恢复

# 全量恢复
cat postgresql-full-20260512.sql | pgk psql -U postgres

# 单库恢复（先创建空库）
pgk psql -U postgres -c "CREATE DATABASE <dbname>;"
cat <dbname>-20260512.sql | pgk psql -U postgres -d <dbname>

定时备份（CronJob）

apiVersion: batch/v1
kind: CronJob
metadata:
  name: postgresql-backup
  namespace: database
spec:
  schedule: "0 3 * * *"
  jobTemplate:
    spec:
      template:
        spec:
          containers:
          - name: backup
            image: m.daocloud.io/docker.io/bitnami/postgresql:latest
            command:
            - sh
            - -c
            - |
              pg_dumpall -U postgres -h postgresql \
                > /backup/postgresql-$(date +%Y%m%d%H%M).sql
              find /backup -name "*.sql" -mtime +7 -delete
            volumeMounts:
            - name: backup
              mountPath: /backup
          restartPolicy: Never
          volumes:
          - name: backup
            hostPath:
              path: /data/backups/postgresql

Redis

Manual RDB backup

# Trigger RDB save
kubectl -n storage exec redis-shared-master-0 -- redis-cli -a <password> SAVE

# Copy RDB file
kubectl -n storage cp redis-shared-master-0:/data/dump.rdb ./redis-backup-$(date +%Y%m%d).rdb

Manual AOF backup

# Trigger AOF rewrite
kubectl -n storage exec redis-shared-master-0 -- redis-cli -a <password> BGREWRITEAOF

# Copy AOF file
kubectl -n storage cp redis-shared-master-0:/data/appendonly.aof.1.base.rdb ./redis-aof-$(date +%Y%m%d).rdb

Restore

# Copy backup file to pod
kubectl -n storage cp ./redis-backup-20260512.rdb redis-shared-master-0:/data/dump.rdb

# Restart pod to reload RDB
kubectl -n storage delete pod redis-shared-master-0
# Redis will load dump.rdb on startup

⚠️ Redis data is ephemeral cache — for most apps, losing Redis is inconvenient but not catastrophic. Prioritize PostgreSQL backups.

Minio

Prerequisites

alias mc='kubectl -n storage exec deployment/minio -- mc'
mc config host add myminio http://minio.storage.svc:9000 <access-key> <secret-key>

Backup a bucket

mc mirror myminio/my-bucket ./backups/my-bucket/

Restore a bucket

mc mirror ./backups/my-bucket/ myminio/my-bucket

Backup all buckets (script)

for bucket in $(mc ls myminio | awk '{print $NF}'); do
  echo "Backing up $bucket ..."
  mc mirror myminio/$bucket ./backups/$bucket/
done

Scheduled backup (CronJob)

apiVersion: batch/v1
kind: CronJob
metadata:
  name: minio-backup
  namespace: storage
spec:
  schedule: "0 4 * * *"
  jobTemplate:
    spec:
      template:
        spec:
          containers:
          - name: backup
            image: m.daocloud.io/docker.io/minio/mc:latest
            command:
            - sh
            - -c
            - |
              mc alias set myminio http://minio.storage.svc:9000 $ACCESS_KEY $SECRET_KEY
              mc mirror myminio/ ./backup/$(date +%Y%m%d)/
            env:
            - name: ACCESS_KEY
              valueFrom:
                secretKeyRef:
                  name: minio
                  key: root-user
            - name: SECRET_KEY
              valueFrom:
                secretKeyRef:
                  name: minio
                  key: root-password
            volumeMounts:
            - name: backup
              mountPath: /backup
          restartPolicy: Never
          volumes:
          - name: backup
            hostPath:
              path: /data/backups/minio

N8N

Method 1: Web UI Export (Recommended)

N8N 内置了工作流导入导出功能：

登录 https://n8n.72602.online
进入 Workflows 列表页面
勾选需要备份的工作流
点击 Download → 导出为 JSON 文件
如需恢复，点击 Import from File

Method 2: API Export

# 获取所有工作流
curl -s -u <email>:<password> https://n8n.72602.online/rest/workflows \
  | jq . > n8n-workflows-$(date +%Y%m%d).json

Method 3: PostgreSQL Database Backup

N8N 的工作流定义存储在 PostgreSQL 中。备份 PG 即可包含 N8N 数据：

pg_dump -U postgres -h postgresql.database n8n > n8n-db-$(date +%Y%m%d).sql

N8N 的凭证（Credentials）包含 API key 等敏感信息，建议在 Export 时选择 Include Credentials 并妥善保管导出的文件。

Open WebUI

Backup Chat History (SQLite)

Open WebUI 的数据默认存储在 SQLite 数据库中：

# 备份数据库文件
kubectl -n ai exec open-webui-0 -- cat /app/backend/data/webui.db > open-webui-$(date +%Y%m%d).db

Restore

# 复制备份文件到 pod
kubectl -n ai cp open-webui-20260512.db open-webui-0:/app/backend/data/webui.db

# 重启 pod 以重新加载数据库
kubectl -n ai delete pod open-webui-0

注意事项

备份文件包含：聊天记录、用户设置、外部连接配置
不包含：对话中的文件附件（存储在 static/uploads/ 目录）

如需完整恢复，还需备份上传文件：

kubectl -n ai exec open-webui-0 -- tar czf - /app/backend/data/uploads > open-webui-uploads-$(date +%Y%m%d).tar.gz

Uptime Kuma

Backup via Web UI

Uptime Kuma 内置了备份功能：

登录 https://uptime.72602.online
进入 Settings → Backup
点击 Export → 下载 JSON 文件
备份文件包含：所有监控项、状态页配置、通知设置

Restore

登录 Uptime Kuma
进入 Settings → Backup
点击 Import → 选择之前下载的 JSON 文件
确认后自动恢复所有配置

注意事项

备份文件不包含监控历史数据，只包含配置
历史数据存储在 SQLite 数据库中（/app/data/kuma.db）

如需保留历史，可以备份 SQLite 文件：

kubectl -n monitor exec deploy/uptime-kuma -- cat /app/data/kuma.db > kuma-backup-$(date +%Y%m%d).db

ElasticSearch

ES [Local Disk]

Preliminary

ElasticSearch has installed, if not check link

The elasticsearch.yml has configed path.repo, which should be set the same value of settings.location (this will be handled by helm chart, don’t worry)

ES argocd-app yaml

apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
  name: elastic-search
spec:
  syncPolicy:
    syncOptions:
    - CreateNamespace=true
  project: default
  source:
    repoURL: https://charts.bitnami.com/bitnami
    chart: elasticsearch
    targetRevision: 19.11.3
    helm:
      releaseName: elastic-search
      values: |
        global:
          kibanaEnabled: true
        clusterName: elastic
        image:
          registry: m.zjvis.net/docker.io
          pullPolicy: IfNotPresent
        security:
          enabled: false
        service:
          type: ClusterIP
        extraConfig:
          path:
            repo: /tmp
        ingress:
          enabled: true
          annotations:
            cert-manager.io/cluster-issuer: self-signed-ca-issuer
            nginx.ingress.kubernetes.io/rewrite-target: /$1
          hostname: elastic-search.dev.72602.online
          ingressClassName: nginx
          path: /?(.*)
          tls: true
        master:
          masterOnly: false
          replicaCount: 1
          persistence:
            enabled: false
          resources:
            requests:
              cpu: 2
              memory: 1024Mi
            limits:
              cpu: 4
              memory: 4096Mi
          heapSize: 2g
        data:
          replicaCount: 0
          persistence:
            enabled: false
        coordinating:
          replicaCount: 0
        ingest:
          enabled: true
          replicaCount: 0
          service:
            enabled: false
            type: ClusterIP
          ingress:
            enabled: false
        metrics:
          enabled: false
          image:
            registry: m.zjvis.net/docker.io
            pullPolicy: IfNotPresent
        volumePermissions:
          enabled: false
          image:
            registry: m.zjvis.net/docker.io
            pullPolicy: IfNotPresent
        sysctlImage:
          enabled: true
          registry: m.zjvis.net/docker.io
          pullPolicy: IfNotPresent
        kibana:
          elasticsearch:
            hosts:
              - '{{ include "elasticsearch.service.name" . }}'
            port: '{{ include "elasticsearch.service.ports.restAPI" . }}'
        esJavaOpts: "-Xmx2g -Xms2g"        
  destination:
    server: https://kubernetes.default.svc
    namespace: application

diff from oirginal file :

extraConfig:
    path:
      repo: /tmp

Methods

Elasticsearch 做备份有两种方式，

是将数据导出成文本文件，比如通过elasticdump、esm等工具将存储在 Elasticsearch 中的数据导出到文件中。
是使用snapshot接口实现快照功能，增量备份文件

第一种方式相对简单，在数据量小的时候比较实用，但当应对大数据量场景时，更推荐使用snapshot api 的方式。

Steps

buckup

asdadas

创建快照仓库repo -> my_fs_repository

curl -k -X PUT "https://elastic-search.dev.72602.online:32443/_snapshot/my_fs_repository?pretty" -H 'Content-Type: application/json' -d'
{
  "type": "fs",
  "settings": {
    "location": "/tmp"
  }
}
'

你也能使用storage-class 挂载一个路径在pod中，将snapshot文件存放在外挂路径上

验证集群各个节点是否可以使用这个快照仓库repo

curl -k -X POST "https://elastic-search.dev.72602.online:32443/_snapshot/my_fs_repository/_verify?pretty"

查看快照仓库repo

curl -k -X GET "https://elastic-search.dev.72602.online:32443/_snapshot/_all?pretty"

查看某一个快照仓库repo的具体setting

curl -k -X GET "https://elastic-search.dev.72602.online:32443/_snapshot/my_fs_repository?pretty"

分析一个快照仓库repo

curl -k -X POST "https://elastic-search.dev.72602.online:32443/_snapshot/my_fs_repository/_analyze?blob_count=10&max_blob_size=1mb&timeout=120s&pretty"

手动打快照

curl -k -X PUT "https://elastic-search.dev.72602.online:32443/_snapshot/my_fs_repository/ay_snap_02?pretty"

使用SLM自动打快照(没生效)

Thank you!

查看指定快照仓库repo 可用的快照

curl -k -X GET "https://elastic-search.dev.72602.online:32443/_snapshot/my_fs_repository/*?verbose=false&pretty"

测试恢复

# Delete an index
curl -k -X DELETE "https://elastic-search.dev.72602.online:32443/books?pretty"

# restore that index
curl -k -X POST "https://elastic-search.dev.72602.online:32443/_snapshot/my_fs_repository/ay_snap_02/_restore?pretty" -H 'Content-Type: application/json' -d'
{
  "indices": "books"
}
'

# query
curl -k -X GET "https://elastic-search.dev.72602.online:32443/books/_search?pretty" -H 'Content-Type: application/json' -d'
{
  "query": {
    "match_all": {}
  }
}
'

ES [S3 Compatible]

Preliminary

ElasticSearch has installed, if not check link

ES argocd-app yaml

apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
  name: elastic-search
spec:
  syncPolicy:
    syncOptions:
    - CreateNamespace=true
  project: default
  source:
    repoURL: https://charts.bitnami.com/bitnami
    chart: elasticsearch
    targetRevision: 19.11.3
    helm:
      releaseName: elastic-search
      values: |
        global:
          kibanaEnabled: true
        clusterName: elastic
        image:
          registry: m.zjvis.net/docker.io
          pullPolicy: IfNotPresent
        security:
          enabled: true
        service:
          type: ClusterIP
        extraEnvVars:
        - name: S3_ACCESSKEY
          value: admin
        - name: S3_SECRETKEY
          value: ZrwpsezF1Lt85dxl
        extraConfig:
          s3:
            client:
              default:
                protocol: http
                endpoint: "http://<PRIVATE_STORAGE_HOST>:9090"
                path_style_access: true
        initScripts:
          configure-s3-client.sh: |
            elasticsearch_set_key_value "s3.client.default.access_key" "${S3_ACCESSKEY}"
            elasticsearch_set_key_value "s3.client.default.secret_key" "${S3_SECRETKEY}"
        hostAliases:
        - ip: <PRIVATE_STORAGE_HOST>
          hostnames:
          - minio-api.dev.72602.online
        ingress:
          enabled: true
          annotations:
            cert-manager.io/cluster-issuer: self-signed-ca-issuer
            nginx.ingress.kubernetes.io/rewrite-target: /$1
          hostname: elastic-search.dev.72602.online
          ingressClassName: nginx
          path: /?(.*)
          tls: true
        master:
          masterOnly: false
          replicaCount: 1
          persistence:
            enabled: false
          resources:
            requests:
              cpu: 2
              memory: 1024Mi
            limits:
              cpu: 4
              memory: 4096Mi
          heapSize: 2g
        data:
          replicaCount: 0
          persistence:
            enabled: false
        coordinating:
          replicaCount: 0
        ingest:
          enabled: true
          replicaCount: 0
          service:
            enabled: false
            type: ClusterIP
          ingress:
            enabled: false
        metrics:
          enabled: false
          image:
            registry: m.zjvis.net/docker.io
            pullPolicy: IfNotPresent
        volumePermissions:
          enabled: false
          image:
            registry: m.zjvis.net/docker.io
            pullPolicy: IfNotPresent
        sysctlImage:
          enabled: true
          registry: m.zjvis.net/docker.io
          pullPolicy: IfNotPresent
        kibana:
          elasticsearch:
            hosts:
              - '{{ include "elasticsearch.service.name" . }}'
            port: '{{ include "elasticsearch.service.ports.restAPI" . }}'
        esJavaOpts: "-Xmx2g -Xms2g"        
  destination:
    server: https://kubernetes.default.svc
    namespace: application

diff from oirginal file :

extraEnvVars:
- name: S3_ACCESSKEY
  value: admin
- name: S3_SECRETKEY
  value: ZrwpsezF1Lt85dxl
extraConfig:
  s3:
    client:
      default:
        protocol: http
        endpoint: "http://<PRIVATE_STORAGE_HOST>:9090"
        path_style_access: true
initScripts:
  configure-s3-client.sh: |
    elasticsearch_set_key_value "s3.client.default.access_key" "${S3_ACCESSKEY}"
    elasticsearch_set_key_value "s3.client.default.secret_key" "${S3_SECRETKEY}"
hostAliases:
- ip: <PRIVATE_STORAGE_HOST>
  hostnames:
  - minio-api.dev.72602.online

Methods

Elasticsearch 做备份有两种方式，

是将数据导出成文本文件，比如通过elasticdump、esm等工具将存储在 Elasticsearch 中的数据导出到文件中。
是使用snapshot接口实现快照功能，增量备份文件

第一种方式相对简单，在数据量小的时候比较实用，但当应对大数据量场景时，更推荐使用snapshot api 的方式。

Steps

buckup

asdadas

创建快照仓库repo -> my_s3_repository

curl -k -X PUT "https://elastic-search.dev.72602.online:32443/_snapshot/my_s3_repository?pretty" -H 'Content-Type: application/json' -d'
{
  "type": "s3",
  "settings": {
    "bucket": "local-test",
    "client": "default",
    "endpoint": "http://<PRIVATE_STORAGE_HOST>:9000"
  }
}
'

你也能使用storage-class 挂载一个路径在pod中，将snapshot文件存放在外挂路径上

验证集群各个节点是否可以使用这个快照仓库repo

curl -k -X POST "https://elastic-search.dev.72602.online:32443/_snapshot/my_s3_repository/_verify?pretty"

查看快照仓库repo

curl -k -X GET "https://elastic-search.dev.72602.online:32443/_snapshot/_all?pretty"

查看某一个快照仓库repo的具体setting

curl -k -X GET "https://elastic-search.dev.72602.online:32443/_snapshot/my_s3_repository?pretty"

分析一个快照仓库repo

curl -k -X POST "https://elastic-search.dev.72602.online:32443/_snapshot/my_s3_repository/_analyze?blob_count=10&max_blob_size=1mb&timeout=120s&pretty"

手动打快照

curl -k -X PUT "https://elastic-search.dev.72602.online:32443/_snapshot/my_s3_repository/ay_s3_snap_02?pretty"

使用SLM自动打快照(没生效)

Thank you!

查看指定快照仓库repo 可用的快照

curl -k -X GET "https://elastic-search.dev.72602.online:32443/_snapshot/my_s3_repository/*?verbose=false&pretty"

测试恢复

# Delete an index
curl -k -X DELETE "https://elastic-search.dev.72602.online:32443/books?pretty"

# restore that index
curl -k -X POST "https://elastic-search.dev.72602.online:32443/_snapshot/my_s3_repository/ay_s3_snap_02/_restore?pretty" -H 'Content-Type: application/json' -d'
{
  "indices": "books"
}
'

# query
curl -k -X GET "https://elastic-search.dev.72602.online:32443/books/_search?pretty" -H 'Content-Type: application/json' -d'
{
  "query": {
    "match_all": {}
  }
}
'

ES Auto BackUp

Preliminary

ElasticSearch has installed, if not check link
We use local disk to save the snapshots, more deatils check link

And the security is enabled.

ES argocd-app yaml

apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
  name: elastic-search
spec:
  syncPolicy:
    syncOptions:
    - CreateNamespace=true
  project: default
  source:
    repoURL: https://charts.bitnami.com/bitnami
    chart: elasticsearch
    targetRevision: 19.11.3
    helm:
      releaseName: elastic-search
      values: |
        global:
          kibanaEnabled: true
        clusterName: elastic
        image:
          registry: m.zjvis.net/docker.io
          pullPolicy: IfNotPresent
        security:
          enabled: true
          tls:
            autoGenerated: true
        service:
          type: ClusterIP
        extraConfig:
          path:
            repo: /tmp
        ingress:
          enabled: true
          annotations:
            cert-manager.io/cluster-issuer: self-signed-ca-issuer
            nginx.ingress.kubernetes.io/rewrite-target: /$1
          hostname: elastic-search.dev.72602.online
          ingressClassName: nginx
          path: /?(.*)
          tls: true
        master:
          masterOnly: false
          replicaCount: 1
          persistence:
            enabled: false
          resources:
            requests:
              cpu: 2
              memory: 1024Mi
            limits:
              cpu: 4
              memory: 4096Mi
          heapSize: 2g
        data:
          replicaCount: 0
          persistence:
            enabled: false
        coordinating:
          replicaCount: 0
        ingest:
          enabled: true
          replicaCount: 0
          service:
            enabled: false
            type: ClusterIP
          ingress:
            enabled: false
        metrics:
          enabled: false
          image:
            registry: m.zjvis.net/docker.io
            pullPolicy: IfNotPresent
        volumePermissions:
          enabled: false
          image:
            registry: m.zjvis.net/docker.io
            pullPolicy: IfNotPresent
        sysctlImage:
          enabled: true
          registry: m.zjvis.net/docker.io
          pullPolicy: IfNotPresent
        kibana:
          elasticsearch:
            hosts:
              - '{{ include "elasticsearch.service.name" . }}'
            port: '{{ include "elasticsearch.service.ports.restAPI" . }}'
        esJavaOpts: "-Xmx2g -Xms2g"        
  destination:
    server: https://kubernetes.default.svc
    namespace: application

diff from oirginal file :

security:
  enabled: true
extraConfig:
    path:
      repo: /tmp

Methods

Steps

auto buckup

创建快照仓库repo -> slm_fs_repository

curl --user elastic:L9shjg6csBmPZgCZ -k -X PUT "https://10.88.0.143:30294/_snapshot/slm_fs_repository?pretty" -H 'Content-Type: application/json' -d'
{
  "type": "fs",
  "settings": {
    "location": "/tmp"
  }
}
'

你也能使用storage-class 挂载一个路径在pod中，将snapshot文件存放在外挂路径上

验证集群各个节点是否可以使用这个快照仓库repo

curl --user elastic:L9shjg6csBmPZgCZ  -k -X POST "https://10.88.0.143:30294/_snapshot/slm_fs_repository/_verify?pretty"

查看快照仓库repo

curl --user elastic:L9shjg6csBmPZgCZ  -k -X GET "https://10.88.0.143:30294/_snapshot/_all?pretty"

查看某一个快照仓库repo的具体setting

curl --user elastic:L9shjg6csBmPZgCZ  -k -X GET "https://10.88.0.143:30294/_snapshot/slm_fs_repository?pretty"

分析一个快照仓库repo

curl --user elastic:L9shjg6csBmPZgCZ  -k -X POST "https://10.88.0.143:30294/_snapshot/slm_fs_repository/_analyze?blob_count=10&max_blob_size=1mb&timeout=120s&pretty"

查看指定快照仓库repo 可用的快照

curl --user elastic:L9shjg6csBmPZgCZ  -k -X GET "https://10.88.0.143:30294/_snapshot/slm_fs_repository/*?verbose=false&pretty"

创建SLM admin 角色

curl --user elastic:L9shjg6csBmPZgCZ -k -X POST "https://10.88.0.143:30294/_security/role/slm-admin?pretty" -H 'Content-Type: application/json' -d'
{
  "cluster": [ "manage_slm", "cluster:admin/snapshot/*" ],
  "indices": [
    {
      "names": [ ".slm-history-*" ],
      "privileges": [ "all" ]
    }
  ]
}
'

创建自动备份cornjob

curl --user elastic:L9shjg6csBmPZgCZ -k -X PUT "https://10.88.0.143:30294/_slm/policy/nightly-snapshots?pretty" -H 'Content-Type: application/json' -d'
{
  "schedule": "0 30 1 * * ?",       
  "name": "<nightly-snap-{now/d}>", 
  "repository": "slm_fs_repository",    
  "config": {
    "indices": "*",                 
    "include_global_state": true    
  },
  "retention": {                    
    "expire_after": "30d",
    "min_count": 5,
    "max_count": 50
  }
}
'

启动自动备份

curl --user elastic:L9shjg6csBmPZgCZ -k -X POST "https://10.88.0.143:30294/_slm/policy/nightly-snapshots/_execute?pretty"

查看SLM备份历史

curl --user elastic:L9shjg6csBmPZgCZ -k -X GET "https://10.88.0.143:30294/_slm/stats?pretty"

测试恢复

# Delete an index
curl --user elastic:L9shjg6csBmPZgCZ  -k -X DELETE "https://10.88.0.143:30294/books?pretty"

# restore that index
curl --user elastic:L9shjg6csBmPZgCZ  -k -X POST "https://10.88.0.143:30294/_snapshot/slm_fs_repository/my_snapshot_2099.05.06/_restore?pretty" -H 'Content-Type: application/json' -d'
{
  "indices": "books"
}
'

# query
curl --user elastic:L9shjg6csBmPZgCZ  -k -X GET "https://10.88.0.143:30294/books/_search?pretty" -H 'Content-Type: application/json' -d'
{
  "query": {
    "match_all": {}
  }
}
'

Example Shell Script

Init ES Backup Setting

create an ES backup setting in s3, and make an snapshot after creation

#!/bin/bash
ES_HOST="http://192.168.58.2:30910"
ES_BACKUP_REPO_NAME="s3_fs_repository"
S3_CLIENT="default"
ES_BACKUP_BUCKET_IN_S3="es-snapshot"
ES_SNAPSHOT_TAG="auto"

CHECK_RESPONSE=$(curl -s -k -X POST "$ES_HOST/_snapshot/$ES_BACKUP_REPO_NAME/_verify?pretty" )
CHECKED_NODES=$(echo "$CHECK_RESPONSE" | jq -r '.nodes')


if [ "$CHECKED_NODES" == null ]; then
  echo "Doesn't exist an ES backup setting..."
  echo "A default backup setting will be generated. (using '$S3_CLIENT' s3 client and all backup files will be saved in a bucket : '$ES_BACKUP_BUCKET_IN_S3'"

  CREATE_RESPONSE=$(curl -s -k -X PUT "$ES_HOST/_snapshot/$ES_BACKUP_REPO_NAME?pretty" -H 'Content-Type: application/json' -d "{\"type\":\"s3\",\"settings\":{\"bucket\":\"$ES_BACKUP_BUCKET_IN_S3\",\"client\":\"$S3_CLIENT\"}}")
  CREATE_ACKNOWLEDGED_FLAG=$(echo "$CREATE_RESPONSE" | jq -r '.acknowledged')

  if [ "$CREATE_ACKNOWLEDGED_FLAG" == true ]; then
    echo "Buckup setting '$ES_BACKUP_REPO_NAME' has been created successfully!"
  else
    echo "Failed to create backup setting '$ES_BACKUP_REPO_NAME', since $$CREATE_RESPONSE"
  fi
else
  echo "Already exist an ES backup setting '$ES_BACKUP_REPO_NAME'"
fi

CHECK_RESPONSE=$(curl -s -k -X POST "$ES_HOST/_snapshot/$ES_BACKUP_REPO_NAME/_verify?pretty" )
CHECKED_NODES=$(echo "$CHECK_RESPONSE" | jq -r '.nodes')

if [ "$CHECKED_NODES" != null ]; then
  SNAPSHOT_NAME="meta-data-$ES_SNAPSHOT_TAG-snapshot-$(date +%s)"
  SNAPSHOT_CREATION=$(curl -s -k -X PUT "$ES_HOST/_snapshot/$ES_BACKUP_REPO_NAME/$SNAPSHOT_NAME")
  echo "Snapshot $SNAPSHOT_NAME has been created."
else
  echo "Failed to create snapshot $SNAPSHOT_NAME ."
fi

Homepage

Backup

Homepage 的配置全部存储在 Git 仓库中：

路径	内容
`manifests/homepage/config/bookmarks.yaml`	书签
`manifests/homepage/config/services.yaml`	服务 widgets
`manifests/homepage/config/settings.yaml`	布局设置
`manifests/homepage/config/widgets.yaml`	仪表盘小部件
`manifests/homepage/config/docker.yaml`	Docker 集成
`manifests/homepage/config/kubernetes.yaml`	K8s 集成
`manifests/homepage/configmap.yaml`	生成的 ConfigMap

恢复步骤

# 1. 确认配置文件正确
bash scripts/gen-homepage-configmap.sh

# 2. 更新 ConfigMap
kubectl -n monitor apply -f manifests/homepage/configmap.yaml

# 3. 触发重载
kubectl -n monitor rollout restart deploy/homepage

Git / GitLab

GitLab

Backup GitLab via Omnibus

# 创建备份（Omnibus 安装）
sudo gitlab-rake gitlab:backup:create

# 备份文件位置
ls -la /var/opt/gitlab/backups/

# 恢复
sudo gitlab-rake gitlab:backup:restore BACKUP=<timestamp>

Backup GitLab on Kubernetes

# 备份数据库
kubectl -n gitlab exec deploy/gitlab -- gitlab-backup create

# 复制备份到本地
kubectl -n gitlab cp gitlab-pod:/var/opt/gitlab/backups/<backup-file> ./

GitHub (General)

当前项目托管在 GitHub 上，数据安全由 GitHub 保障。

如需额外备份：

# 备份所有仓库
gh repo list --limit 100 --json nameWithOwner -q '.[].nameWithOwner' | \
  while read repo; do
    git clone --mirror "https://github.com/$repo" "backups/$repo.git"
  done

GitLab

https://docs.gitlab.com/charts/backup-restore/

🧯BuckUp

Backup Quick Reference

Subsections of 🧯BuckUp

PostgreSQL

Prerequisites

全量备份所有数据库

备份单个数据库

备份到 PVC 内

从备份文件恢复

定时备份（CronJob）

Redis

Manual RDB backup

Manual AOF backup

Restore

Minio

Prerequisites

Backup a bucket

Restore a bucket

Backup all buckets (script)

Scheduled backup (CronJob)

N8N

Method 1: Web UI Export (Recommended)

Method 2: API Export

Method 3: PostgreSQL Database Backup

Open WebUI

Backup Chat History (SQLite)

Restore

注意事项

Uptime Kuma

Backup via Web UI

Restore

注意事项

ElasticSearch

Subsections of ElasticSearch

ES [Local Disk]

Preliminary

Methods

Steps

ES [S3 Compatible]

Preliminary

Methods

Steps

ES Auto BackUp

Preliminary

Methods

Steps

Example Shell Script

Init ES Backup Setting

Homepage

Backup

恢复步骤

Git / GitLab

GitLab

Backup GitLab via Omnibus

Backup GitLab on Kubernetes

GitHub (General)

Subsections of Git / GitLab

GitLab

K8s