Description of problem: Cinder fails to connect to an internal Ceph cluster with customize name. At first try cinder-volume logs show the messages: 2018-10-01 19:53:06.141 60 ERROR cinder.volume.drivers.rbd [req-f7b133c8-f050-45e5-9bd9-a5d0f19d956f - - - - -] Error connecting to ceph cluster.: OSError: [errno 95] error connecting to the cluster 2018-10-01 19:53:06.141 60 ERROR cinder.volume.drivers.rbd Traceback (most recent call last): 2018-10-01 19:53:06.141 60 ERROR cinder.volume.drivers.rbd File "/usr/lib/python2.7/site-packages/cinder/volume/drivers/rbd.py", line 349, in _do_conn 2018-10-01 19:53:06.141 60 ERROR cinder.volume.drivers.rbd client.connect() 2018-10-01 19:53:06.141 60 ERROR cinder.volume.drivers.rbd File "rados.pyx", line 886, in rados.Rados.connect (/builddir/build/BUILD/ceph-12.2.5/build/src/pybind/rados/pyrex/rados.c:9784) 2018-10-01 19:53:06.141 60 ERROR cinder.volume.drivers.rbd OSError: [errno 95] error connecting to the cluster 2018-10-01 19:53:06.141 60 ERROR cinder.volume.drivers.rbd As an account of it, cinder is not functional at all, though Glance and Nova are working fine with the same cluster and the cluster itself is healthy. The Ceph directory on the container is: # docker exec -it openstack-cinder-volume-docker-0 ls /etc/ceph CephSite1.client.admin.keyring CephSite1.mgr.controller-0.keyring CephSite1.client.manila.keyring CephSite1.mgr.controller-1.keyring CephSite1.client.openstack.keyring CephSite1.mgr.controller-2.keyring CephSite1.client.radosgw.keyring CephSite1.mon.keyring CephSite1.conf rbdmap The backend configuration in cinder.conf is [tripleo_ceph] backend_host=hostgroup volume_backend_name=tripleo_ceph volume_driver=cinder.volume.drivers.rbd.RBDDriver rbd_ceph_conf=/etc/ceph/CephSite1.conf rbd_user=openstack rbd_pool=volumes rbd_secret_uuid=<secret key> Version-Release number of selected component (if applicable): rhosp14/openstack-cinder-volume:2018-09-26.1 $ docker exec -it mistral_executor rpm -qa | grep ceph ceph-ansible-3.1.5-1.el7cp.noarch How reproducible: Steps to Reproduce: 1. Deploy a ceph cluster with customized name parameter_defaults: CephClusterName: CephSite1 2. create a volume Actual results: Cinder fails to connect to the Ceph cluster using the rbd driver Expected results: Cinder connects to the Ceph cluster with RBD driver with the given parameters from tripleo Additional info:
We test custom Ceph cluster names upstream ... this looks more like an issue with rbd_secret_uuid this is empty?
(In reply to Giulio Fidente from comment #1) > We test custom Ceph cluster names upstream ... this looks more like an issue > with rbd_secret_uuid this is empty? No, the secret key is set and set properly, cause Nova is able to run VMs with ephemeral disks on the Ceph cluster
verified
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2019:0045