Mac OS crash: "[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead."

[jwmatthews]

We've introduced a regression recently in main that impacts Mac OS

objc[79384]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug.

Thank you @jmontleon for identifying the workaround:
export OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES
make run-server

With above, we can run on Mac.

More reason for the workaround here:

• https://www.wefearchange.org/2018/11/forkmacos.rst.html

Below is what you will see without the above workaroudn set on Mac:

File logging for 'kai' is set to level 'DEBUG' writing to file: '/Users/jmatthews/git/jwmatthews/kai/logs/kai_server.log'
INFO - 2024-09-18 09:52:42,606 - kai.service.kai_application.kai_application - [  kai_application.py:54   -             __init__()] - Tracing enabled.
INFO - 2024-09-18 09:52:42,609 - kai.service.kai_application.kai_application - [  kai_application.py:63   -             __init__()] - Selected provider: ChatIBMGenAI
INFO - 2024-09-18 09:52:42,609 - kai.service.kai_application.kai_application - [  kai_application.py:64   -             __init__()] - Selected model: meta-llama/llama-3-70b-instruct
objc[79384]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called.
objc[79384]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug.
[2024-09-18 09:52:42 -0400] [79192] [ERROR] Worker (pid:79384) was sent SIGKILL! Perhaps out of memory?
objc[79385]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called.
objc[79385]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug.
[2024-09-18 09:52:42 -0400] [79192] [ERROR] Worker (pid:79385) was sent SIGKILL! Perhaps out of memory?
[2024-09-18 09:52:42 -0400] [79387] [INFO] Booting worker with pid: 79387
Config loaded: KaiConfig(log_level='info', file_log_level='debug', log_dir='$pwd/logs', demo_mode=False, trace_enabled=True, gunicorn_workers=8, gunicorn_timeout=3600, gunicorn_bind='0.0.0.0:8080', incident_store=KaiConfigIncidentStore(solution_detectors=<SolutionDetectorKind.NAIVE: 'naive'>, solution_producers=<SolutionProducerKind.TEXT_ONLY: 'text_only'>, args=KaiConfigIncidentStorePostgreSQLArgs(provider=<KaiConfigIncidentStoreProvider.POSTGRESQL: 'postgresql'>, host='127.0.0.1', database='kai', user='kai', password='dog8code', connection_string=None, solution_detection=<SolutionDetectorKind.NAIVE: 'naive'>)), models=KaiConfigModels(provider='ChatIBMGenAI', args={'model_id': 'meta-llama/llama-3-70b-instruct', 'parameters': {'max_new_tokens': 2048}}, template=None, llama_header=None, llm_retries=5, llm_retry_delay=10.0), solution_consumers=[<SolutionConsumerKind.DIFF_ONLY: 'diff_only'>, <SolutionConsumerKind.LLM_SUMMARY: 'llm_summary'>])
Console logging for 'kai' is set to level 'INFO'
File logging for 'kai' is set to level 'DEBUG' writing to file: '/Users/jmatthews/git/jwmatthews/kai/logs/kai_server.log'
INFO - 2024-09-18 09:52:42,686 - kai.service.kai_application.kai_application - [  kai_application.py:54   -             __init__()] - Tracing enabled.
INFO - 2024-09-18 09:52:42,689 - kai.service.kai_application.kai_application - [  kai_application.py:63   -             __init__()] - Selected provider: ChatIBMGenAI
INFO - 2024-09-18 09:52:42,690 - kai.service.kai_application.kai_application - [  kai_application.py:64   -             __init__()] - Selected model: meta-llama/llama-3-70b-instruct
objc[79386]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called.
objc[79386]: +[__NSCFConstantString initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug.
[2024-09-18 09:52:42 -0400] [79192] [ERROR] Worker (pid:79386) was sent SIGKILL! Perhaps out of memory?
[2024-09-18 09:52:42 -0400] [79388] [INFO] Booting worker with pid: 79388
[2024-09-18 09:52:42 -0400] [79389] [INFO] Booting worker with pid: 79389

For those curious there is an Ansible issue filed which helped us to learn of the workaround: ansible/ansible#76322

[jmontleon]

This seems to have been introduced with:
fa08eb4

urllib3 and requests are implicated in the Ansible PR and this PR updated urllib3, so my guess is this is what brough this on.