TiCDC GC Issue 2

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: ticdc gc问题2

| username: jeff

[TiDB Usage Environment] Production Environment

[Encountered Problem: Problem Phenomenon and Impact]
When I backed up data in TiDB and restored it in downstream MySQL, I received a prompt indicating that the checkpoint is earlier than the GC save point when creating a CDC task. The GC-TTL for CDC is set to 48 hours, but it did not take effect.

| username: 扬仔_tidb | Original post link

Don’t say anything, just increase tidb_gc_life_time.
Last time, a database backup took 24 hours. You need to increase this parameter directly.

| username: TiDBer_oHSwKxOH | Original post link

Do you use CloudCanal?

| username: TiDBer_oHSwKxOH | Original post link

Core capabilities

  1. Provide TiDB → Tidb
| username: TiDBer_oHSwKxOH | Original post link

You see, it only provides replication between TiDBs.

| username: tidb菜鸟一只 | Original post link

Just change tidb_gc_life_time. If your current tidb_gc_life_time is less than 24 hours, and you set the gc-ttl to 48 hours when creating CDC, specifying data from 24 hours ago will still be invalid because the data from 24 hours ago in the cluster has already expired. The gc-ttl of CDC is based on the premise that you can create it successfully. After that, the GC advancement time of the TiDB cluster will follow the gc-ttl. If you haven’t created it successfully yet, it will definitely follow tidb_gc_life_time

| username: Soysauce520 | Original post link

The default value of tidb_gc_life_time is 10 minutes. If there is a downstream or flashback, it needs to be increased.

| username: TiDBer_vfJBUcxl | Original post link

Increase the value of tidb_gc_life_time

| username: ajin0514 | Original post link

Try increasing tidb_gc_life_time and see if it helps.