cdc owner stuck after injection downstream failure 3 mins and recover

[Lily2025]

What did you do?

1、run workload storage:"s3://benchmark/sysbench_64_7000w" subType:"oltp_read_write" db:"sysbench_64_7000w" tableNum:64 tableSize:70000000 threads:32 ignoreErrors:"2013,1213,1105,1205,8022,8028,9004,9007,1062" 2、 injection downstream failure 3 mins and recover

What did you expect to see?

changefeed status are normal after fault recover

What did you see instead?

1、the status of changefeed task keep warning for 20 minutes after fault recover

2、the log lost during status of changefeed abnormal

Versions of the cluster

./cdc version Release Version: v6.5.11 Git Commit Hash: 7ce6da9bbb2f43e0d47c2ad95357ab62f869e93b Git Branch: HEAD UTC Build Time: 2024-09-09 04:49:13 Go Version: go version go1.19.13 linux/amd64 Failpoint Build: false 2024-09-11T22:18:58.435+0800

current status of DM cluster (execute query-status <task-name> in dmctl)

No response