TiDB CDC can not get change data

[yujunyong]

Describe the bug(Please use English) A clear and concise description of what the bug is.

Environment :

• Flink version : 1.14
• Flink CDC version: 2.2.1
• Database and version: 5.4.0

Taskmanager Log

TiDB CDC started normal in a day, but can not get change data after a day. no error log in taskmanager. no error log in TiDB CDC Server. when 2022-06-14 16:32:32,858 INFO org.tikv.cdc.CDCClient [] - handle resolvedTs: 433899583002640387, regionId: 67814 log don't receive, change data not receive too.

[RaiToKU]

me too https://github.com/ververica/flink-cdc-connectors/issues/1252

[GOODBOY008]

@yujunyong @RaiToKU https://github.com/ververica/flink-cdc-connectors/pull/1196 you can checkout pr and compile local.

[Hugoincug]

I had the same situation. The synchronization will stop when the incremental synchronization reaches a certain value

[xieyi888]

I had the same question. maybe the problem is similar to this issue：https://github.com/tikv/client-java/issues/628 tidb-cdc don't receive ChangeDataEvent from tikv store after several hours,because tikv region transfer leader. you can verify log of tikv store,if there are logs about "transfer leader" of your region_id

......
 ["try to transfer leader"] [to_peer="id: **** store_id: ***"] [from_peer="id: **** store_id: ****"] [region_id=****]
["transfer leader"] [peer="id: **** store_id: ****"] [peer_id=****] [region_id=***]
.....

[GOODBOY008]

I had the same question. maybe the problem is similar to this issue：tikv/client-java#628 tidb-cdc don't receive ChangeDataEvent from tikv store after several hours,because tikv region transfer leader. you can verify log of tikv store,if there are logs about "transfer leader" of your region_id

......
 ["try to transfer leader"] [to_peer="id: **** store_id: ***"] [from_peer="id: **** store_id: ****"] [region_id=****]
["transfer leader"] [peer="id: **** store_id: ****"] [peer_id=****] [region_id=***]
.....

When a region was created after start cdclient, this case also cann't capture data change.

[Hugoincug]

I had the same question. maybe the problem is similar to this issue：tikv/client-java#628 tidb-cdc don't receive ChangeDataEvent from tikv store after several hours,because tikv region transfer leader. you can verify log of tikv store,if there are logs about "transfer leader" of your region_id

......
 ["try to transfer leader"] [to_peer="id: **** store_id: ***"] [from_peer="id: **** store_id: ****"] [region_id=****]
["transfer leader"] [peer="id: **** store_id: ****"] [peer_id=****] [region_id=***]
.....

When a region was created after start cdclient, this case also cann't capture data change.

it means cdcClient needs to be fix ?

[MengXiangDing]

Is there a follow-up to this question?

[MengXiangDing]

I had the same question. maybe the problem is similar to this issue：tikv/client-java#628 tidb-cdc don't receive ChangeDataEvent from tikv store after several hours,because tikv region transfer leader. you can verify log of tikv store,if there are logs about "transfer leader" of your region_id

......
 ["try to transfer leader"] [to_peer="id: **** store_id: ***"] [from_peer="id: **** store_id: ****"] [region_id=****]
["transfer leader"] [peer="id: **** store_id: ****"] [peer_id=****] [region_id=***]
.....

When a region was created after start cdclient, this case also cann't capture data change.

Will this issue be fixed in the future?

[PatrickRen]

Closing this issue because it was created before version 2.3.0 (2022-11-10). Please try the latest version of Flink CDC to see if the issue has been resolved. If the issue is still valid, kindly report it on Apache Jira under project Flink with component tag Flink CDC. Thank you!