[TiDB Cloud] ch workload report mpp error after upgrading to v8.5.3 from v7.5

[jolynejo]

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. Minimal reproduce step (Required)

1. start a v7.5.7 cluster with tiflash
2. prepare ch dataset
3. run ch workload with ap query
4. upgrade to v8.5.3
5. ch workload report mpp error

2. What did you expect to see? (Required)

Succeed in executing query

3. What did you see instead (Required)

q6/q12/q20/... Each rerun fails on a different query.
Error 1105 (HY000): other error for mpp stream: Code: 0, e.displayText() = DB::Exception: Region epoch not match after retries: Region {19517,7,877} not in region cache., e.what() = DB::Exception,

4. What is your TiFlash version? (Required)

[root@ip-172-16-4-38 tiflash]# ./tiflash version
TiFlash
Release Version: v8.5.3
Edition: Enterprise
Git Commit Hash: c192b4c
Git Branch: HEAD
UTC Build Time: 2025-07-11 14:44:41
Enable Features: jemalloc sm4(GmSSL) mem-profiling asimd unwind thinlto hnsw.l2=neon hnsw.cosine=neon vec.l2=neon vec.cos=neon
Profile: RELWITHDEBINFO
Compiler: clang++ 17.0.6

Raft Proxy
Git Commit Hash: c91242c77e1854e07c8d7273c9fc0b1ff153cb64
Git Commit Branch: HEAD
UTC Build Time: ""
Rust Version: rustc 1.77.0-nightly (89e2160c4 2023-12-27)
Storage Engine: tiflash
Prometheus Prefix: tiflash_proxy_
Profile: release
Enable Features: external-jemalloc pprof-fp portable test-engine-kv-rocksdb test-engine-raft-raft-engine openssl-vendored pprof-fp