Flow pipeline testing framework

This page describes how to set up a workstation and configure low-level parameters for contributors who are improving Flow itself and running its integration test suite. If you're a pipeline author testing your code, see Testing (C++), Testing (Java), Testing (Go), and Testing (Python) — these pages include the launch parameters and techniques you need for pipeline tests.

Environment setup

The Flow test suite captures Flow process crashes and extracts backtraces from coredumps. For this to work, you must enable coredump collection on your machine:

# Temporary (it’s better to add this to ~/.bashrc).
ulimit -c unlimited

Also, add the process PID to the coredump filename:

# Temporary — valid until reboot:
sudo sysctl -w kernel.core_uses_pid=1

# Permanent:
echo "kernel.core_uses_pid=1" | sudo tee -a /etc/sysctl.conf
sudo sysctl -p

Without core_uses_pid=1, you might see odd behavior, such as extracting multiple backtraces from a single coredump.

Flow test parameters

In addition to the general parameters (see Testing (C++)), these parameters are useful when you’re working on Flow itself — they affect the framework’s internal behavior:

Parameter	Default	Values	Action
`DUMP_BACKTRACES_IF_HANGS`	`0`	`0`, `1`	Write backtraces of threads and fibers from running Flow processes to disk when an unhandled `WaitFailed` occurs during shutdown.
`FLOW_BINARY_PATH`		Absolute path	Use a fixed pipeline binary — for example, one built with a thread sanitizer.
`ERROR_BACKTRACE_ENRICHER_LEVEL`	`enabled_for_not_native_errors`	see enum	Controls whether backtraces are added to errors in the logs.

Examples:

ya make -A --test-param ERROR_BACKTRACE_ENRICHER_LEVEL=enabled_for_all
ya make -A --test-param DUMP_BACKTRACES_IF_HANGS=1

Authentication internals

Overview