doris ccr 数据同步工具（跨集群同步）

官方文档

注意：本章使用的工具为二开工具, 官方工具不支持endpoint_mapping, 需要使用二进制部署的方式部署, 但具体步骤差不太多

前提条件

  • FE / BE 配置都需要打开：enable_feature_binlog = true

    • 在 fe.conf / be.conf 文件中声名

  • 库/表需要打开 binlog：ALTER DATABASE db_name SET properties ("binlog.enable" = "true");

    • 使用脚本一键开启 binlog

      • ./enable_db_binlog.sh -h ip -p sql_port -u root -d db_name

  • 源集群和备份集群每个 fe, be 都需要定义对应 nodeport svc 用于数据同步

使用 ccr 工具进行数据同步

确定源和目标 FE/BE 信息、记录返回 Host 用于之后做映射

SHOW FRONTENDS;
SHOW BACKENDS;

分别固定 FE/BE 端口

源集群 NodePort

FE

Pod

QueryPort 9030

RpcPort 9020

doriscluster-helm-fe-0

32580

32581

doriscluster-helm-fe-1

32582

32583

doriscluster-helm-fe-2

32584

32585

BE

Pod

BePort 9060

HttpPort 8040

BrpcPort 8060

doriscluster-helm-be-0

32060

32040

32061

doriscluster-helm-be-1

32062

32042

32063

doriscluster-helm-be-2

32064

32044

32065

目标集群 NodePort

FE

Pod

QueryPort 9030

RpcPort 9020

doriscluster-helm-fe-0

32580

32581

doriscluster-helm-fe-1

32582

32583

doriscluster-helm-fe-2

32584

32585

BE

Pod

BePort 9060

HttpPort 8040

BrpcPort 8060

doriscluster-helm-be-0

32060

32040

32061

doriscluster-helm-be-1

32062

32042

32063

doriscluster-helm-be-2

32064

32044

32065

NodePort Service 模板

apiVersion: v1
kind: Service
metadata:
  name: doris-fe-0-nodeport
spec:
  type: NodePort
  selector:
    statefulset.kubernetes.io/pod-name: doriscluster-helm-fe-0
  ports:
    - name: mysql
      port: 9030
      targetPort: 9030
      nodePort: 32580
    - name: thrift
      port: 9020
      targetPort: 9020
      nodePort: 32581
---
apiVersion: v1
kind: Service
metadata:
  name: doris-fe-1-nodeport
spec:
  type: NodePort
  selector:
    statefulset.kubernetes.io/pod-name: doriscluster-helm-fe-1
  ports:
    - name: mysql
      port: 9030
      targetPort: 9030
      nodePort: 32582
    - name: thrift
      port: 9020
      targetPort: 9020
      nodePort: 32583
---
apiVersion: v1
kind: Service
metadata:
  name: doris-fe-1-nodeport
spec:
  type: NodePort
  selector:
    statefulset.kubernetes.io/pod-name: doriscluster-helm-fe-2
  ports:
    - name: mysql
      port: 9030
      targetPort: 9030
      nodePort: 32584
    - name: thrift
      port: 9020
      targetPort: 9020
      nodePort: 32585
---
apiVersion: v1
kind: Service
metadata:
  name: doris-be-0-nodeport
spec:
  type: NodePort
  selector:
    statefulset.kubernetes.io/pod-name: doriscluster-helm-be-0
  ports:
    - name: be
      port: 9060
      targetPort: 9060
      nodePort: 32060
    - name: http
      port: 8040
      targetPort: 8040
      nodePort: 32040
    - name: brpc
      port: 8060
      targetPort: 8060
      nodePort: 32061
---
apiVersion: v1
kind: Service
metadata:
  name: doris-be-1-nodeport
spec:
  type: NodePort
  selector:
    statefulset.kubernetes.io/pod-name: doriscluster-helm-be-1
  ports:
    - name: be
      port: 9060
      targetPort: 9060
      nodePort: 32062
    - name: http
      port: 8040
      targetPort: 8040
      nodePort: 32042
    - name: brpc
      port: 8060
      targetPort: 8060
      nodePort: 32063
---
apiVersion: v1
kind: Service
metadata:
  name: doris-be-2-nodeport
spec:
  type: NodePort
  selector:
    statefulset.kubernetes.io/pod-name: doriscluster-helm-be-2
  ports:
    - name: be
      port: 9060
      targetPort: 9060
      nodePort: 32064
    - name: http
      port: 8040
      targetPort: 8040
      nodePort: 32044
    - name: brpc
      port: 8060
      targetPort: 8060
      nodePort: 32065

说明：

  • selector 必须精确到单个 Pod

  • nodePort 必须落在集群允许范围，本次环境为 30000-32767

  • FE / BE 都需要逐 Pod 暴露

同步数据

启动 ccr 同步工具

# 启动ccr
ccr_syncer -host 127.0.0.1 -port 19192 -db_dir /tmp/ccr-syncer-test.db -log_level info
## 后台启动
bash start_syncer.sh  --host 127.0.0.1 --port 19192 --db_dir /tmp/db --log_level info --daemon

同步任务

说明：

  • src.host / src.port / src.thrift_port 填 Syncer 当前能直连的源端 FE 外部入口

  • dest.host / dest.port / dest.thrift_port 填 Syncer 当前能直连的目标端 FE 外部入口

  • endpoint_mapping 用于把 Doris 内部地址改写成外部 NodePort

  • 本次同步的是整个库，因此 table 为空字符串

  • host 为 fe master 地址、port 为 9030 端口、thrift_port 为 9020 端口

curl -X POST -H "Content-Type: application/json" -d '{
  "name": "demo_ccr",
  "src": {
    "host": "10.0.0.30",
    "port": "32580",
    "thrift_port": "32581",
    "user": "root",
    "password": "",
    "database": "demo",
    "table": "",
    "endpoint_mapping": {
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32580",
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32581",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32582",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32583",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32584",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32585",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32060",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32040",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32061",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32062",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32042",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32063",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32064",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32044",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32065"
    }
  },
  "dest": {
    "host": "10.0.1.30",
    "port": "32580",
    "thrift_port": "32581",
    "user": "root",
    "password": "",
    "database": "demo",
    "table": "",
    "endpoint_mapping": {
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32580",
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32581",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32582",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32583",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32584",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32585",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32060",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32040",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32061",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32062",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32042",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32063",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32064",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32044",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32065"
    }
  }
}' http://127.0.0.1:19192/create_ccr

其他常用 api

查看任务状态：

curl -s -X POST -H "Content-Type: application/json" -d '{"name":"demo_ccr"}' http://127.0.0.1:19192/job_status

查看延迟：

curl -s -X POST -H "Content-Type: application/json" -d '{"name":"demo_ccr"}' http://127.0.0.1:19192/get_lag

查看任务详情：

curl -s -X POST -H "Content-Type: application/json" -d '{"name":"demo_ccr"}' http://127.0.0.1:19192/job_detail

删除任务

curl -s -X POST -H "Content-Type: application/json" -d '{"name":"demo_ccr"}'   http://127.0.0.1:19192/delete

Doris 中查看恢复进度

SHOW RESTORE FROM demo;


重点关注：
 `State`
 `DownloadFinishedTime`
 `FinishedTime`
 `UnfinishedTasks`
 `TaskErrMsg`

最终结果

  • 源端 demo 全量同步成功

  • restore 完成后，正式表可见且有数据

  • 任务状态切换为 DBIncrementalSync

  • 增量插入数据后，目标端可正常看到同步结果

使用 ccr 工具从目标集群恢复数据到源集群

删除现有任务

curl -s -X POST -H "Content-Type: application/json" -d '{"name":"demo_ccr"}'   http://127.0.0.1:19192/delete

创建方向同步任务

curl -X POST -H "Content-Type: application/json" -d '{
  "name": "demo_ccr_reverse",
  "src": {
    "host": "10.0.1.30",
    "port": "32580",
    "thrift_port": "32581",
    "user": "root",
    "password": "",
    "database": "demo",
    "table": "",
    "endpoint_mapping": {
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32580",
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32581",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32582",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32583",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32584",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32585",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32060",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32040",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32061",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32062",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32042",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32063",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32064",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32044",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32065"
    }
  },
  "dest": {
    "host": "10.0.0.30",
    "port": "32580",
    "thrift_port": "32581",
    "user": "root",
    "password": "",
    "database": "demo",
    "table": "",
    "endpoint_mapping": {
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32580",
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32581",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32582",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32583",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32584",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32585",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32060",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32040",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32061",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32062",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32042",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32063",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32064",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32044",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32065"
    }
  }
}' http://127.0.0.1:19192/create_ccr

注意：每次创建任务都为全量同步，最终会覆盖掉原本的数据

源集群查看数据状态、等待数据同步完成

SHOW RESTORE FROM demo;

删除方向同步任务

curl -s -X POST -H "Content-Type: application/json" -d '{"name":"demo_ccr_reverse"}'   http://127.0.0.1:19192/delete

创建源到目标的任务

curl -X POST -H "Content-Type: application/json" -d '{
  "name": "demo_ccr",
  "src": {
    "host": "10.0.0.30",
    "port": "32584",
    "thrift_port": "32585",
    "user": "root",
    "password": "",
    "database": "demo",
    "table": "",
    "endpoint_mapping": {
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32580",
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32581",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32582",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32583",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.0.30:32584",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.0.30:32585",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32060",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32040",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32061",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32062",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32042",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32063",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.0.30:32064",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.0.30:32044",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.0.30:32065"
    }
  },
  "dest": {
    "host": "10.0.1.30",
    "port": "30670",
    "thrift_port": "31185",
    "user": "root",
    "password": "",
    "database": "demo",
    "table": "",
    "endpoint_mapping": {
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32580",
      "doriscluster-helm-fe-0.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32581",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32582",
      "doriscluster-helm-fe-1.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32583",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9030": "10.0.1.30:32584",
      "doriscluster-helm-fe-2.doriscluster-helm-fe-internal.default.svc.cluster.local:9020": "10.0.1.30:32585",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32060",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32040",
      "doriscluster-helm-be-0.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32061",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32062",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32042",
      "doriscluster-helm-be-1.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32063",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:9060": "10.0.1.30:32064",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8040": "10.0.1.30:32044",
      "doriscluster-helm-be-2.doriscluster-helm-be-internal.default.svc.cluster.local:8060": "10.0.1.30:32065"
    }
  }
}' http://127.0.0.1:19192/create_ccr

等待完成即可

参数优化

## 备份/恢复任务的总超时时间
- backup_job_default_timeout_ms             
   源集群 FE、目标集群 FE 都要配，单位是毫秒。默认 1 天。
   先看当前值：

   SHOW FRONTEND CONFIG LIKE '%backup_job_default_timeout_ms%';
   临时在线生效：

   ADMIN SET FRONTEND CONFIG ("backup_job_default_timeout_ms" = "259200000");
    是 3 天。这个参数是 MasterOnly，SQL 会转发到当前 Master，但这类改法重启后会失效。

   259200000 是 3 天。这个参数是 MasterOnly，SQL 会转发到当前 Master，但这类改法重启后会失效。
   
## 上游 binlog 的保留时间
 - binlog.ttl_seconds
   配在上游库属性，单位是秒。
   例如把 hk_active 的 binlog 保留 7 天：

   ALTER DATABASE hk_active SET PROPERTIES ("binlog.ttl_seconds" = "604800");
   
## 下游单个 BE 上，单个下载线程的限速
 - max_download_speed_kbps
   配在下游 BE，单位是 KB/s。默认 50000，也就是 50MB/s 左右。
   查看：

   SHOW BACKEND CONFIG LIKE '%max_download_speed_kbps%';

   临时在线改某个 BE：

   curl -X POST "http://<be_host>:8040/api/update_config?max_download_speed_kbps=204800"

   204800 约等于 200MB/s。

## 下游单个 BE 的下载线程数
 - download_worker_count
   也配在下游 BE。
   查看：

   SHOW BACKEND CONFIG LIKE '%download_worker_count%';

   临时在线改某个 BE：

   curl -X POST "http://<be_host>:8040/api/update_config?download_worker_count=4"

   这个参数通常比单纯调限速更有效。官网也说如果把它调上去，很多时候就不用再改 max_download_speed_kbps 了。

 你这个环境里怎么落地

 - 先用上面的 SQL / HTTP API 做临时验证。
 - 验证有效后，再写进 operator/helm 管理的 FE/BE 配置里做持久化。
 - 不建议只 kubectl exec 进 Pod 改 fe.conf / be.conf，Pod 重建后大概率会丢。
 - backup_job_default_timeout_ms 最好写到所有 FE 的配置里，不然以后 FE 切主后，新 Master 可能又回默认值。
 - max_download_speed_kbps 和 download_worker_count 要下发到所有下游 BE，不是只改一个。

#云原生
doris ccr 数据同步工具（跨集群同步）
https://www.ysdy-zla.top//archives/doris-ccr-syncer
作者
nuoli
发布于
2026年03月23日
许可协议