gnocchi-metricd fails to start in nonha

[Ben Nemec]

In a nonha deployment gnocchi-metricd is failing to start with the following traceback:

Aug 30 15:29:50 localhost gnocchi-metricd: Traceback (most recent call last):
Aug 30 15:29:50 localhost gnocchi-metricd: File "/usr/lib64/python2.7/multiprocessing/util.py", line 268, in _run_finalizers
Aug 30 15:29:50 localhost gnocchi-metricd: finalizer()
Aug 30 15:29:50 localhost gnocchi-metricd: File "/usr/lib64/python2.7/multiprocessing/util.py", line 201, in __call__
Aug 30 15:29:50 localhost gnocchi-metricd: res = self._callback(*self._args, **self._kwargs)
Aug 30 15:29:50 localhost gnocchi-metricd: File "/usr/lib64/python2.7/multiprocessing/managers.py", line 609, in _finalize_manager
Aug 30 15:29:50 localhost gnocchi-metricd: if process.is_alive():
Aug 30 15:29:50 localhost gnocchi-metricd: File "/usr/lib64/python2.7/multiprocessing/process.py", line 155, in is_alive
Aug 30 15:29:50 localhost gnocchi-metricd: assert self._parent_pid == os.getpid(), 'can only test a child process'
Aug 30 15:29:50 localhost gnocchi-metricd: AssertionError: can only test a child process
Aug 30 15:29:50 localhost gnocchi-metricd: Error in atexit._run_exitfuncs:
Aug 30 15:29:50 localhost gnocchi-metricd: Traceback (most recent call last):
Aug 30 15:29:50 localhost gnocchi-metricd: File "/usr/lib64/python2.7/atexit.py", line 24, in _run_exitfuncs
Aug 30 15:29:50 localhost gnocchi-metricd: func(*targs, **kargs)
Aug 30 15:29:50 localhost gnocchi-metricd: File "/usr/lib64/python2.7/multiprocessing/util.py", line 319, in _exit_function
Aug 30 15:29:50 localhost gnocchi-metricd: p.join()
Aug 30 15:29:50 localhost gnocchi-metricd: File "/usr/lib64/python2.7/multiprocessing/process.py", line 143, in join
Aug 30 15:29:50 localhost gnocchi-metricd: assert self._parent_pid == os.getpid(), 'can only join a child process'
Aug 30 15:29:50 localhost gnocchi-metricd: AssertionError: can only join a child process

It then proceeds to attempt restarts indefinitely, which wastes a bunch of CPU on the controller node. I don't see these messages in the ha job.

[Pradeep Kilambi (pkilambi)]

This is fixed in cotyledon 1.2.7, what version are we using in tripleo repos? we might need to pull in new version from rdo and this should be resolved.

[Brent Eagles (beagles)]

cotyledon requirement notwithstanding, I believe this is related to or simply a duplicate of https://bugs.launchpad.net/tripleo/+bug/1618510.

I see these stack traces on my HA deployment until I reconfigure the redis access info to point to an endpoint that the redis service is bound to. After that the stack traces go away. I'll investigate further to see if this can be marked a duplicate.