Deadlock in txn_status_cache with Pipelined DML #18384
Description
Bug Report
What version of TiKV are you using?
master
What operating system and CPU are you using?
Steps to reproduce
Run a long Pipelined DML and it has an opportunity to happen
What did you expect?
No deadlock.
What did happened?
There is a deadlock in the scheduler pool. According to the metrics, the threads seem to be hanging.
Pending tasks in scheduler and resolved-ts both pile up