changefeed resolved ts can not go on after restart all pd、tikv and tidb

[Lily2025]

What did you do?

1、create changefeed with simple protocol
sink_url_options: simple-test?protocol=simple&max-message-bytes=1024000&partition-num=3&replication-factor=3
config: "[integrity]\nintegrity-check-level = "correctness"\ncorruption-handle-level = "warn"\n[sink]\nsend-bootstrap-interval-in-sec = 120\nsend-bootstrap-in-msg-count = 10000\n[sink.kafka-config.codec-config]\nencoding-format = "avro"\n[sink.kafka-config.large-message-handle]\nlarge-message-handle-compression = "lz4"\nlarge-message-handle-option = "claim-check"\nclaim-check-storage-uri = "s3://tmp?access-key=minioadmin&secret-access-key=minioadmin&endpoint=http%3a%2f%2fminio-peer%3a9000""
2、run sysbench and cdc boundary workload
3、kill all pd、tikv and tidb

What did you expect to see?

after fault recover, changefeed resolved ts can go on

What did you see instead?

changefeed resolved ts can not go on after restart all pd、tikv and tidb

Versions of the cluster

./cdc version
Release Version: v8.2.0-alpha
Git Commit Hash: 8d22340
Git Branch: heads/refs/tags/v8.2.0-alpha
UTC Build Time: 2024-05-08 11:36:18
Go Version: go version go1.21.6 linux/amd64
Failpoint Build: false
2024-05-09T08:21:41.925+0800

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

No response

[Lily2025]

/remove-area dm
/area ticdc
/severity major

[Lily2025]

/assign hongyunyan