Debugging concurrent programs¶

Category: Testing and tools · Status: stub · Lessons: chapter 09, Testing and tools (planned)

One line: Thread dumps, deadlock detectors and tracing that show what every thread or task is waiting for at a given moment.

Also called: thread dump, deadlock detection, goroutine dump, stack dump.

How it connects¶


Go	`SIGQUIT` (Ctrl-backslash) makes a program exit with a stack dump ↗; `runtime/pprof` ↗ has `goroutine`, `block` and `mutex` profiles
Java	`jcmd` `Thread.print` ↗ prints all threads with stack traces; `ThreadMXBean.findDeadlockedThreads` ↗ finds cycles of deadlocked threads
Python	`faulthandler.dump_traceback` ↗ dumps the tracebacks of all threads, also after a timeout or on a signal
C#	`dotnet-stack` ↗ inspects managed stack traces
Elsewhere	GDB ↗ `thread apply all bt` applies `bt` to every thread, printing a backtrace for each

In a sibling library: Rust: Instrumenting async code ↗
In a sibling library: Go: All goroutines are asleep ↗
In the books: Effective Concurrency in Go, Burak Serdar — ch. 10, 'Troubleshooting Concurrency Issues'
In the books: Programming with POSIX Threads, David R. Butenhof — ch. 8, 'Hints to Avoid Debugging'
In the books: C++ Concurrency in Action, Anthony Williams — ch. 11, 'Testing and debugging multithreaded applications'
In the books: Concurrent Programming on Windows, Joe Duffy — ch. 11, 'Concurrency Hazards'
In the books: Mastering C++ Multithreading, Maya Posch — ch. 6, 'Debugging Multithreaded Code'
In the books: Learning Concurrent Programming in Scala, Aleksandar Prokopec — ch. 9, 'Concurrency in Practice' → 'Debugging concurrent programs'
Notes: tools for concurrent programming ↗