TiDB v4.0 DM任务启动失败

为提高效率,提问时请提供以下信息,问题描述清晰可优先响应。

  • 【TiDB 版本】:V4.0.4
  • 【问题描述】:

[root@r-tidb4cleaner-DM-m220-213 dm-v1.0.6-linux-amd64]# bin/dmctl -master-addr 10.108.220.213:8261
Welcome to dmctl
Release Version: v1.0.6
Git Commit Hash: eaf2683c05ab44143bfb286bfbbc3ba157c555cc
Git Branch: release-1.0
UTC Build Time: 2020-06-17 10:22:01
Go Version: go version go1.13 linux/amd64

» check-task conf/task_basic.yaml
{
“result”: true,
“msg”: “check pass!!!”
}

» start-task conf/task_basic.yaml
{
“result”: true,
“msg”: “”,
“workers”: [
{
“result”: false,
“worker”: “10.108.220.213:8262”,
“msg”: “[code=38033:class=dm-master:scope=internal:level=high] request to dm-worker 10.108.220.213:8262 is timeout, but request may be successful, please execute query-status to check status\ngithub.com/pingcap/dm/pkg/terror.(*Error).Generate\ \t/home/jenkins/agent/workspace/build_dm_master/go/src/github.com/pingcap/dm/pkg/terror/terror.go:236\ngithub.com/pingcap/dm/dm/master.(*Server).waitOperationOk\ \t/home/jenkins/agent/workspace/build_dm_master/go/src/github.com/pingcap/dm/dm/master/server.go:1923\ngithub.com/pingcap/dm/dm/master.(*Server).handleOperationResult\ \t/home/jenkins/agent/workspace/build_dm_master/go/src/github.com/pingcap/dm/dm/master/server.go:1946\ngithub.com/pingcap/dm/dm/master.(*Server).StartTask.func1\ \t/home/jenkins/agent/workspace/build_dm_master/go/src/github.com/pingcap/dm/dm/master/server.go:276\ngithub.com/pingcap/dm/dm/master.(*AgentPool).Emit\ \t/home/jenkins/agent/workspace/build_dm_master/go/src/github.com/pingcap/dm/dm/master/agent_pool.go:117\ runtime.goexit\ \t/usr/local/go/src/runtime/asm_amd64.s:1357”
}
]
}

» query-status online
{
“result”: true,
“msg”: “”,
“workers”: [
{
“result”: true,
“worker”: “10.108.220.213:8262”,
“msg”: “”,
“subTaskStatus”: [
{
“name”: “online”,
“stage”: “Paused”,
“unit”: “InvalidUnit”,
“result”: {
“isCanceled”: false,
“errors”: [
{
“Type”: “UnknownError”,
“msg”: “”,
“error”: {
“ErrCode”: 10006,
“ErrClass”: 1,
“ErrScope”: 2,
“ErrLevel”: 3,
"Message": “fail to initial unit Load of subtask online : execute statement failed: CREATE SCHEMA IF NOT EXISTS dm_meta: context deadline exceeded”,
“RawCause”: “context deadline exceeded”
}

[root@r-tidb4cleaner-DM-m220-213 dm-v1.0.6-linux-amd64]# telnet 10.108.220.213 8262
Trying 10.108.220.213…
Connected to 10.108.220.213.
Escape character is ‘^]’.

Telnet端口没有问题

若提问为性能优化、故障排查类问题,请下载脚本运行。终端输出的打印结果,请务必全选并复制粘贴上传。

  1. 麻烦再确认下端口和ip都是放通的,如果有防火墙,关闭防火墙
  2. 检查task中配置的用户是否有足够的权限,是用root用户测试的吗?

1,没有防火墙
2,root用户

发现是 pump + drainer的问题 ,又配了 Binlog.inogre-error = false,导致 tidb整个挂住了,同步也进行不了,现在同步恢复了。
感觉TiDB binlog的故障点比较多,考虑用TiCDC 替换 TiDB binlog,靠谱吗 ?

可以的,如果对 HA 要求比较高,建议使用 TICDC 来同步数据的。TiCDC 应该在接下来最近的小版本 GA,欢迎使用。