作者: 数据源的TiDB学习之路
背景
金融机构越来越多的选择将多套业务系统融合到一套分布式数据库集群来支撑,一方面可以节约硬件成本,另一方面可以简化运维复杂性。多租户能力及资源管控技术是数据库中实现多业务融合的最常用及有效手段。
容灾是金融机构 IT 系统满足高可用性必需的一个能力,金融机构的容灾能力根据高可用程度划分为多个等级,较常用的一种容灾部署方式是采用主备库的方案,在分布式数据库中也称为主备集群方案。
金融机构每年都有容灾切换演练的场景,根据系统重要等级不同,切换演练的频次也有所不同。对于 A 类业务系统,一般是每年覆盖一次;对于 B 类或 C 类系统,则可能是 2 年或更长时间循环覆盖一次。
在多业务融合的场景中,为了缩小切换演练的范围,避免切换演练对所有业务产生影响,金融企业期望底层数据库产品能支持单个租户 (应用) 的容灾切换能力。
构建基于 资源管控+TiCDC 的多业务融合容灾方案
TiDB 支持基于RU (Request Unit,资源单元) 的资源管控能力,方便解决多业务融合共享集群资源的场景。TiDB 也支持 TiCDC 实时同步工具,可以实现 TiDB 集群间的容灾方案。基于 TiDB 的资源管控及TiCDC,理论上可以实现一套多业务整合的容灾方案,并满足单个租户 (应用) 容灾切换而不影响其他租户的要求。
以下通过实际测试步骤来验证方案的可行性。
- 准备集群环境。
搭建 2 套相同节点数的 TiDB 测试环境,命名为 tidb-A 和 tidb-B,并分别安装部署 TiCDC 组件。
[tidb@host-10-17-3-151 packages]$ tiup cluster display tidb-A
tiup is checking updates for component cluster ...
Starting component `cluster`: /home/tidb/.tiup/components/cluster/v1.14.1/tiup-cluster display tidb-A
Cluster type: tidb
Cluster name: tidb-A
Cluster version: v7.5.1
Deploy user: tidb
SSH type: builtin
Dashboard URL: http://10.17.3.152:2379/dashboard
Grafana URL: http://10.17.3.151:3000
ID Role Host Ports OS/Arch Status Data Dir Deploy Dir
-- ---- ---- ----- ------- ------ -------- ----------
10.17.3.151:9093 alertmanager 10.17.3.151 9093/9094 linux/aarch64 Up /data1/tidb-data/alertmanager-9093 /data1/tidb-deploy/alertmanager-9093
10.17.3.151:8300 cdc 10.17.3.151 8300 linux/aarch64 Up /data1/cdc-free /data1/tidb-deploy/cdc-8300
10.17.3.152:8300 cdc 10.17.3.152 8300 linux/aarch64 Up /data1/cdc-free /data1/tidb-deploy/cdc-8300
10.17.3.151:3000 grafana 10.17.3.151 3000 linux/aarch64 Up - /data1/tidb-deploy/grafana-3000
10.17.3.151:2379 pd 10.17.3.151 2379/2380 linux/aarch64 Up /data1/tidb-data/pd-2379 /data1/tidb-deploy/pd-2379
10.17.3.152:2379 pd 10.17.3.152 2379/2380 linux/aarch64 Up|UI /data1/tidb-data/pd-2379 /data1/tidb-deploy/pd-2379
10.17.3.153:2379 pd 10.17.3.153 2379/2380 linux/aarch64 Up|L /data1/tidb-data/pd-2379 /data1/tidb-deploy/pd-2379
10.17.3.151:9095 prometheus 10.17.3.151 9095/12020 linux/aarch64 Up /data1/tidb-data/prometheus-9095 /data1/tidb-deploy/prometheus-9095
10.17.3.151:4000 tidb 10.17.3.151 4000/10080 linux/aarch64 Up - /data1/tidb-deploy/tidb-4000
10.17.3.152:4000 tidb 10.17.3.152 4000/10080 linux/aarch64 Up - /data1/tidb-deploy/tidb-4000
10.17.3.153:4000 tidb 10.17.3.153 4000/10080 linux/aarch64 Up - /data1/tidb-deploy/tidb-4000
10.17.3.151:20160 tikv 10.17.3.151 20160/20180 linux/aarch64 Up /data1/tidb-data/tikv-20160 /data1/tidb-deploy/tikv-20160
10.17.3.152:20160 tikv 10.17.3.152 20160/20180 linux/aarch64 Up /data1/tidb-data/tikv-20160 /data1/tidb-deploy/tikv-20160
10.17.3.153:20160 tikv 10.17.3.153 20160/20180 linux/aarch64 Up /data1/tidb-data/tikv-20160 /data1/tidb-deploy/tikv-20160
Total nodes: 14
[tidb@host-10-17-3-151 packages]$ tiup cluster display tidb-B
tiup is checking updates for component cluster ...
Starting component `cluster`: /home/tidb/.tiup/components/cluster/v1.14.1/tiup-cluster display tidb-B
Cluster type: tidb
Cluster name: tidb-B
Cluster version: v7.5.1
Deploy user: tidb
SSH type: builtin
Dashboard URL: http://10.17.3.151:12399/dashboard
Grafana URL: http://10.17.3.151:13009
ID Role Host Ports OS/Arch Status Data Dir Deploy Dir
-- ---- ---- ----- ------- ------ -------- ----------
10.17.3.151:19089 alertmanager 10.17.3.151 19089/19084 linux/aarch64 Up /data1/tidb-re-data/alertmanager-19089 /data1/tidb-re-deploy/alertmanager-19089
10.17.3.151:8400 cdc 10.17.3.151 8400 linux/aarch64 Up /data1/cdc-recover /data1/tidb-re-deploy/cdc-8400
10.17.3.152:8400 cdc 10.17.3.152 8400 linux/aarch64 Up /data1/cdc-recover /data1/tidb-re-deploy/cdc-8400
10.17.3.151:13009 grafana 10.17.3.151 13009 linux/aarch64 Up - /data1/tidb-re-deploy/grafana-13009
10.17.3.151:12399 pd 10.17.3.151 12399/12390 linux/aarch64 Up|UI /data1/tidb-re-data/pd-12399 /data1/tidb-re-deploy/pd-12399
10.17.3.152:12399 pd 10.17.3.152 12399/12390 linux/aarch64 Up|L /data1/tidb-re-data/pd-12399 /data1/tidb-re-deploy/pd-12399
10.17.3.153:12399 pd 10.17.3.153 12399/12390 linux/aarch64 Up /data1/tidb-re-data/pd-12399 /data1/tidb-re-deploy/pd-12399
10.17.3.151:19099 prometheus 10.17.3.151 19099/22029 linux/aarch64 Up /data1/tidb-re-data/prometheus-19099 /data1/tidb-re-deploy/prometheus-19099
10.17.3.151:24000 tidb 10.17.3.151 24000/20090 linux/aarch64 Up - /data1/tidb-re-deploy/tidb-24000
10.17.3.152:24000 tidb 10.17.3.152 24000/20090 linux/aarch64 Up - /data1/tidb-re-deploy/tidb-24000
10.17.3.153:24000 tidb 10.17.3.153 24000/20090 linux/aarch64 Up - /data1/tidb-re-deploy/tidb-24000
10.17.3.151:30169 tikv 10.17.3.151 30169/30190 linux/aarch64 Up /data1/tidb-re-data/tikv-30169 /data1/tidb-re-deploy/tikv-30169
10.17.3.152:30169 tikv 10.17.3.152 30169/30190 linux/aarch64 Up /data1/tidb-re-data/tikv-30169 /data1/tidb-re-deploy/tikv-30169
10.17.3.153:30169 tikv 10.17.3.153 30169/30190 linux/aarch64 Up /data1/tidb-re-data/tikv-30169 /data1/tidb-re-deploy/tikv-30169
Total nodes: 14
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43.
- 44.
- 45.
- 46.
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53.
- 54.
- 55.
- 创建用户并绑定资源组。
分别在两套集群中创建 3 个用户 (userA、userB、userC) 及 3 个资源组 (rgA、rgB、rgC),用户和资源组一一绑定。分别使用不同用户在数据库中创建各自的 database,分别为 sbtest_a、sbtest_b、sbtest_c,它们代表不同业务系统的后台数据库。
mysql> create user userA;
Query OK, 0 rows affected (0.06 sec)
mysql> create user userB;
Query OK, 0 rows affected (0.05 sec)
mysql> create user userC;
Query OK, 0 rows affected (0.03 sec)
mysql> create resource group rgA RU_PER_SEC=1000;
Query OK, 0 rows affected (0.54 sec)
mysql> create resource group rgB RU_PER_SEC=2000;
Query OK, 0 rows affected (0.55 sec)
mysql> create resource group rgC RU_PER_SEC=3000;
Query OK, 0 rows affected (0.54 sec)
mysql> alter user userA resource group rgA;
Query OK, 0 rows affected (0.04 sec)
mysql> alter user userB resource group rgB;
Query OK, 0 rows affected (0.04 sec)
mysql> alter user userC resource group rgC;
Query OK, 0 rows affected (0.04 sec)
mysql> select user, user_attributes from mysql.user;
+-------+---------------------------+
| user | user_attributes |
+-------+---------------------------+
| root | NULL |
| userA | {"resource_group": "rga"} |
| userB | {"resource_group": "rgb"} |
| userC | {"resource_group": "rgc"} |
+-------+---------------------------+
4 rows in set (0.01 sec)
mysql> alter user userA identified by 'userA123';
Query OK, 0 rows affected (0.05 sec)
mysql> alter user userB identified by 'userB123';
Query OK, 0 rows affected (0.04 sec)
mysql> alter user userC identified by 'userC123';
Query OK, 0 rows affected (0.04 sec)
mysql> create database sbtest_a;
Query OK, 0 rows affected (0.54 sec)
mysql> grant all on sbtest_a.* to userA;
Query OK, 0 rows affected (0.06 sec)
mysql> create database sbtest_b;
Query OK, 0 rows affected (0.54 sec)
mysql> grant all on sbtest_b.* to userB;
Query OK, 0 rows affected (0.03 sec)
mysql> create database sbtest_c;
Query OK, 0 rows affected (0.54 sec)
mysql> grant all on sbtest_c.* to userC;
Query OK, 0 rows affected (0.04 sec)
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43.
- 44.
- 45.
- 46.
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53.
- 54.
- 55.
- 56.
- 57.
- 58.
- 59.
- 60.
- 61.
- 62.
- 63.
- 64.
- 创建同步链路。
创建 3 条 TiCDC 同步链路 (changefeed),分别为 sbtest_a (集群 A) -> sbtest_a (集群 B) 、sbtest_b (集群 A) -> sbtest_b (集群 B) 、sbtest_c (集群 A) -> sbtest_c (集群 B) 。
[tidb@host-10-17-3-151 packages]$ cat changfeed_sbtest_a.toml
[filter]
rules = ['sbtest_a.*']
[tidb@host-10-17-3-151 packages]$ cat changfeed_sbtest_b.toml
[filter]
rules = ['sbtest_b.*']
[tidb@host-10-17-3-151 packages]$ cat changfeed_sbtest_c.toml
[filter]
rules = ['sbtest_c.*']
tiup cdc cli changefeed create --server=http://10.17.3.151:8300 --sink-uri="mysql://userA:userA123@10.17.3.151:24000/" --changefeed-id="sbtest-a-atob" --config=changfeed_sbtest_a.toml
tiup cdc cli changefeed create --server=http://10.17.3.151:8300 --sink-uri="mysql://userB:userB123@10.17.3.151:24000/" --changefeed-id="sbtest-b-atob" --config=changfeed_sbtest_b.toml
tiup cdc cli changefeed create --server=http://10.17.3.151:8300 --sink-uri="mysql://userC:userC123@10.17.3.151:24000/" --changefeed-id="sbtest-c-atob" --config=changfeed_sbtest_c.toml
[tidb@host-10-17-3-151 packages]$ tiup cdc cli changefeed list --server=http://10.17.3.151:8300
tiup is checking updates for component cdc ...
Starting component `cdc`: /home/tidb/.tiup/components/cdc/v7.5.1/cdc cli changefeed list --server=http://10.17.3.151:8300
[
{
"id": "sbtest-a-atob",
"namespace": "default",
"summary": {
"state": "normal",
"tso": 451948847132835850,
"checkpoint": "2024-08-19 14:13:51.360",
"error": null
}
},
{
"id": "sbtest-b-atob",
"namespace": "default",
"summary": {
"state": "normal",
"tso": 451948847132835850,
"checkpoint": "2024-08-19 14:13:51.360",
"error": null
}
},
{
"id": "sbtest-c-atob",
"namespace": "default",
"summary": {
"state": "normal",
"tso": 451948847132835850,
"checkpoint": "2024-08-19 14:13:51.360",
"error": null
}
}
]
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43.
- 44.
- 45.
- 46.
- 47.
- 48.
- 49.
- 模拟业务。
运行 3套 sysbench 测试,模拟 3 个业务场景。第 1 套 sysbench 测试使用 userA 用户连接到集群 A 的 sbtest_a,第 2 套 sysbench 测试使用 userB 用户连接到集群 A 的 sbtest_b,第 3套 sysbench 测试使用 userC 用户连接到集群 A 的 sbtest_c。
查看各资源组 RU 使用情况,可以看到都按照之前设置的1000 RU、2000 RU、3000 RU 使用到最大上限,且未出现超用的情况。
- 验证数据一致性。
停止 sysnbech 模拟测试,使用 sync-inspector 检验两边集群数据是否一致。
对比配置文件
[tidb@host-10-17-3-151 packages]$ cat sync-diff.toml
# Diff Configuration.
######################### Global config #########################
# 检查数据的线程数量,上下游数据库的连接数会略大于该值
check-thread-count = 4
# 如果开启,若表存在不一致,则输出用于修复的 SQL 语句。
export-fix-sql = true
# 只对比表结构而不对比数据
check-struct-only = false
# 如果开启,会跳过校验上游或下游不存在的表。
skip-non-existing-table = false
######################### Datasource config #########################
[data-sources]
[data-sources.tidba] # mysql1 是该数据库实例唯一标识的自定义 id,用于下面 task.source-instances/task.target-instance 中
host = "10.17.3.151"
port = 4000
user = "userA"
password = "userA123" # 设置连接上游数据库的密码,可为明文或 Base64 编码。
#(可选)使用映射规则来匹配上游多个分表,其中 rule1 和 rule2 在下面 Routes 配置栏中定义
route-rules = ["rule1"]
[data-sources.tidbb]
host = "10.17.3.151"
port = 24000
user = "userA"
password = "userA123" # 设置连接下游数据库的密码,可为明文或 Base64 编码。
#(可选)使用 TLS 连接 TiDB
# security.ca-path = ".../ca.crt"
# security.cert-path = ".../cert.crt"
# security.key-path = ".../key.crt"
#(可选)使用 TiDB 的 snapshot 功能,如果开启的话会使用历史数据进行对比
# snapshot = "386902609362944000"
# 当 snapshot 设置为 "auto" 时,使用 TiCDC 在上下游的同步时间点,具体参考 <https://github.com/pingcap/tidb-tools/issues/663>
# snapshot = "auto"
########################### Routes ###########################
# 如果需要对比大量的不同库名或者表名的表的数据,或者用于校验上游多个分表与下游总表的数据,可以通过 table-rule 来设置映射关系
# 可以只配置 schema 或者 table 的映射关系,也可以都配置
[routes]
[routes.rule1] # rule1 是该配置的唯一标识的自定义 id,用于上面 data-sources.route-rules 中
schema-pattern = "sbtest_a" # 匹配数据源的库名,支持通配符 "*" 和 "?"
target-schema = "sbtest_a" # 目标库名
######################### Task config #########################
# 配置需要对比的*目标数据库*中的表
[task]
# output-dir 会保存如下信息
# 1 sql:检查出错误后生成的修复 SQL 文件,并且一个 chunk 对应一个文件
# 2 log:sync-diff.log 保存日志信息
# 3 summary:summary.txt 保存总结
# 4 checkpoint:a dir 保存断点续传信息
output-dir = "./output2"
# 上游数据库,内容是 data-sources 声明的唯一标识 id
source-instances = ["tidba"]
# 下游数据库,内容是 data-sources 声明的唯一标识 id
target-instance = "tidbb"
# 需要比对的下游数据库的表,每个表需要包含数据库名和表名,两者由 `.` 隔开
# 使用 ? 来匹配任意一个字符;使用 * 来匹配任意;详细匹配规则参考 golang regexp pkg: https://github.com/google/re2/wiki/Syntax
target-check-tables = ["sbtest_a.*"]
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43.
- 44.
- 45.
- 46.
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53.
- 54.
- 55.
- 56.
- 57.
- 58.
- 59.
- 60.
- 61.
- 62.
- 63.
- 64.
- 65.
- 66.
- 67.
- 68.
- 69.
- 70.
- 71.
对比结果,3个 schema 的数据结果完全一致
[tidb@host-10-17-3-151 packages]$ ./tidb-community-toolkit-v7.5.1-linux-arm64/sync_diff_inspector --config=sync-diff-a.toml
A total of 5 tables need to be compared
Comparing the table structure of ``sbtest_a`.`sbtest1`` ... equivalent
Comparing the table structure of ``sbtest_a`.`sbtest2`` ... equivalent
Comparing the table structure of ``sbtest_a`.`sbtest4`` ... equivalent
Comparing the table structure of ``sbtest_a`.`sbtest3`` ... equivalent
Comparing the table structure of ``sbtest_a`.`sbtest5`` ... equivalent
Comparing the table data of ``sbtest_a`.`sbtest1`` ... equivalent
Comparing the table data of ``sbtest_a`.`sbtest3`` ... equivalent
Comparing the table data of ``sbtest_a`.`sbtest2`` ... equivalent
Comparing the table data of ``sbtest_a`.`sbtest4`` ... equivalent
Comparing the table data of ``sbtest_a`.`sbtest5`` ... equivalent
_____________________________________________________________________________
Progress [============================================================>] 100% 0/0
A total of 5 table have been compared and all are equal.
You can view the comparision details through './output3/sync_diff.log'
[tidb@host-10-17-3-151 packages]$ ./tidb-community-toolkit-v7.5.1-linux-arm64/sync_diff_inspector --config=sync-diff-b.toml
A total of 5 tables need to be compared
Comparing the table structure of ``sbtest_b`.`sbtest3`` ... equivalent
Comparing the table structure of ``sbtest_b`.`sbtest5`` ... equivalent
Comparing the table structure of ``sbtest_b`.`sbtest2`` ... equivalent
Comparing the table structure of ``sbtest_b`.`sbtest4`` ... equivalent
Comparing the table structure of ``sbtest_b`.`sbtest1`` ... equivalent
Comparing the table data of ``sbtest_b`.`sbtest3`` ... equivalent
Comparing the table data of ``sbtest_b`.`sbtest4`` ... equivalent
Comparing the table data of ``sbtest_b`.`sbtest5`` ... equivalent
Comparing the table data of ``sbtest_b`.`sbtest2`` ... equivalent
Comparing the table data of ``sbtest_b`.`sbtest1`` ... equivalent
_____________________________________________________________________________
Progress [============================================================>] 100% 0/0
A total of 5 table have been compared and all are equal.
You can view the comparision details through './output4/sync_diff.log'
[tidb@host-10-17-3-151 packages]$ ./tidb-community-toolkit-v7.5.1-linux-arm64/sync_diff_inspector --config=sync-diff-c.toml
A total of 5 tables need to be compared
Comparing the table structure of ``sbtest_c`.`sbtest1`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest4`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest5`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest3`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest2`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest4`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest1`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest5`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest3`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest2`` ... equivalent
_____________________________________________________________________________
Progress [============================================================>] 100% 0/0
A total of 5 table have been compared and all are equal.
You can view the comparision details through './output5/sync_diff.log'
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43.
- 44.
- 45.
- 46.
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53.
- 切换单个同步链路。
模拟切换单个同步链路,选择 sbtest_c (集群 A) -> sbtest_c (集群 B)。首先需要停止 sbtest_c 上的业务并移除(或暂停)当前同步链路,其次重新构建反向同步链路 sbtest_c (集群 B) -> sbtest_c (集群 A),最后将模拟业务连接切换到集群 B 上。
删除正向同步链路:
此时观察到业务C对应的资源组的 RU 消耗为0,而另外两个资源组仍然在稳定运行当中。
搭建反向同步链路:
此时再查看各资源组 RU 使用情况,由于此时业务的主要压力来自于 TiCDC 同步过来的压力,所以跟直接在源库执行时会有差异,从图中看到平均使用 RU 数约为 500。
- 验证数据一致性。
重新启动 sbtest_c 上的业务并写入集群 B,运行一段时间后停止并验证切换后的 sbtest_c 两边数据是否一致,同时也验证原来的同步链路 sbtest_a 和 sbtest_b 数据是否仍然保持一致性同步。
//运行模拟业务c,并写入集群 B
sh run_sysbench_c_revers.sh auto
[tidb@host-10-17-3-151 packages]$ ./tidb-community-toolkit-v7.5.1-linux-arm64/sync_diff_inspector --config=sync-diff-c.toml
A total of 5 tables need to be compared
Comparing the table structure of ``sbtest_c`.`sbtest3`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest1`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest5`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest4`` ... equivalent
Comparing the table structure of ``sbtest_c`.`sbtest2`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest5`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest4`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest1`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest3`` ... equivalent
Comparing the table data of ``sbtest_c`.`sbtest2`` ... equivalent
_____________________________________________________________________________
Progress [============================================================>] 100% 0/0
A total of 5 table have been compared and all are equal.
You can view the comparision details through './output5/sync_diff.log'
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
问题记录
问题1:CDC 报错 Error 1044 Access denied for user xxx
解决:用户权限不足,执行 GRANT ALL PRIVILEGES ON `sbtest_a`.* TO 'userA'@'%';
mysql> show grants for userA;
+-----------------------------------+
| Grants for userA@% |
+-----------------------------------+
| GRANT USAGE ON *.* TO 'userA'@'%' |
+-----------------------------------+
1 row in set (0.00 sec)
mysql> GRANT ALL PRIVILEGES ON `sbtest_a`.* TO 'userA'@'%';
Query OK, 0 rows affected (0.05 sec)
mysql> show grants for userA;
+-----------------------------------------------------+
| Grants for userA@% |
+-----------------------------------------------------+
| GRANT USAGE ON *.* TO 'userA'@'%' |
| GRANT ALL PRIVILEGES ON `sbtest_a`.* TO 'userA'@'%' |
+-----------------------------------------------------+
2 rows in set (0.00 sec)
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
添加完权限后, changefeed 任务并不会自动更新,需要先 pause 再 resume。
问题2:sync-diff-inspector 报错 SHOW command denied to user 'userC'@'%'SHOW command denied to user 'userC'@'%'
解决:原因是 changfeed 定义的连接下游数据库用户对要同步的对象不具有 SELECT 等权限,重新删除 changefeed 并创建