I see this error when running pool_mgm in daemon mode, but not with the -n argument.
The directory is not specified in any config file / it doesn't seem an obvious choice as the pool_mgm doesn't directly manipulate haproxy, but rather spins nodes and registers them with the api server.
Not necessarily a bug, but documenting this as it seems like odd behavior.
libra_pool_mgm -c /etc/libra.cfg
Traceback (most recent call last):
File "/usr/local/bin/libra_pool_mgm", line 9, in <module>
load_entry_point('libra==1.0', 'console_scripts', 'libra_pool_mgm')()
File "/usr/local/lib/python2.7/dist-packages/libra-1.0-py2.7.egg/libra/mgm/mgm.py", line 448, in main
with context:
File "/usr/local/lib/python2.7/dist-packages/daemon/daemon.py", line 354, in __enter__
self.open()
File "/usr/local/lib/python2.7/dist-packages/daemon/daemon.py", line 329, in open
change_working_directory(self.working_directory)
File "/usr/local/lib/python2.7/dist-packages/daemon/daemon.py", line 477, in change_working_directory
raise error
daemon.daemon.DaemonOSEnvironmentError: Unable to change working directory ([Errno 2] No such file or directory: '/etc/haproxy')
Pool mgr should be setting a different working directory in mgm.py:
context = daemon.
)