Recently, I discovered a potential multithreading deadlock issue. The symptom is that one Java thread (TransactInit) is using 100% CPU, while other Java threads are stuck. After analyzing the pstack logs and the related source code, I found that there is a possibility of deadlock involving the TransactInit thread, a thread in the JNI critical section, and the VM thread. The deadlock relationship diagram and related logs are attached.