Description of problem: I know this already that starting RHCS 3 and later releases provide additional protection for pools to prevent unauthorized access by enabling application type. This is not a Blocker, please consider this BZ as a feature request for RHHI4C 13 ( OSP 13 + Ceph 3 ) Request Description: OSPd (tripleo) triggers ceph-ansible to deploy Ceph Cluster. Tripleo later configures cinder, glance, nova to use Ceph as backend. By default, tripleo creates all the required pools like images, vms, volumes, metrics & backup. What it does not do currently is "it doesn't assign an application type to the Ceph pool" as a result cluster health stays in WARN state, in a freshly deployed OSP-13 & RHCS 3 system. We know application type is RBD for all the default pools created by triples. So tripleo should set RBD as application type by default. [heat-admin@controller-0 ~]$ ceph -s cluster: id: 1ed62898-b2ad-11e8-916e-2047478ccfaa health: HEALTH_WARN application not enabled on 2 pool(s) too few PGs per OSD (8 < min 30) services: mon: 1 daemons, quorum controller-0 mgr: controller-0(active) osd: 60 osds: 60 up, 60 in data: pools: 5 pools, 160 pgs objects: 1298 objects, 1692 MB usage: 7939 MB used, 218 TB / 218 TB avail pgs: 160 active+clean [heat-admin@controller-0 ~]$ [heat-admin@controller-0 ~]$ [heat-admin@controller-0 ~]$ ceph df GLOBAL: SIZE AVAIL RAW USED %RAW USED 218T 218T 7939M 0 POOLS: NAME ID USED %USED MAX AVAIL OBJECTS images 1 45056k 0 67022G 12 metrics 2 0 0 67022G 0 backups 3 0 0 67022G 0 vms 4 1648M 0 67022G 1286 volumes 5 0 0 67022G 0 [heat-admin@controller-0 ~]$ Version-Release number of selected component (if applicable): ceph-ansible-3.1.2-1.el7.noarch ansible-tripleo-ipsec-8.1.1-0.20180308133440.8f5369a.el7ost.noarch openstack-tripleo-common-containers-8.6.1-23.el7ost.noarch openstack-tripleo-ui-8.3.1-3.el7ost.noarch openstack-tripleo-puppet-elements-8.0.0-2.el7ost.noarch openstack-tripleo-heat-templates-8.0.2-43.el7ost.noarch puppet-tripleo-8.3.2-8.el7ost.noarch openstack-tripleo-common-8.6.1-23.el7ost.noarch openstack-tripleo-validations-8.4.1-5.el7ost.noarch python-tripleoclient-9.2.1-13.el7ost.noarch openstack-tripleo-image-elements-8.0.1-1.el7ost.noarch How Reproducible Always (100%) Steps to Reproduce: 1. Deploy OSP-13 with RHCS 3 2. Login to overcloud controller node 3. Do ceph -s Actual results: After tripleo deploys Ceph cluster, ceph cluster remains in health_warn state because of no application type flage not set on Ceph pool Expected results: After tripleo deploys Ceph cluster, application type must be automatically set on Ceph pools Additional info:
Looks like out of 5 default pools (images, vms, volumes, metrics and backups) application is not set on 2 pools images and volumes. Which mean, tripleo already has required bits to enable application on Ceph pools, we just need to instruct tripleo (by default) to set application type on images and volumes pools. Or am i missing something ?
Hi John Here is the output from the overcloud deployed very recently. https://paste.fedoraproject.org/paste/MYNMx5x9OJrL~Qxp7pU-uw/raw [heat-admin@controller-0 ~]$ ceph -s cluster: id: 1ed62898-b2ad-11e8-916e-2047478ccfaa health: HEALTH_WARN application not enabled on 1 pool(s) too few PGs per OSD (8 < min 30) services: mon: 1 daemons, quorum controller-0 mgr: controller-0(active) osd: 60 osds: 60 up, 60 in data: pools: 5 pools, 160 pgs objects: 1286 objects, 10240 MB usage: 37260 MB used, 218 TB / 218 TB avail pgs: 160 active+clean [heat-admin@controller-0 ~]$ [heat-admin@controller-0 ~]$ ceph osd dump | grep -i pool pool 1 'images' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 last_change 88 flags hashpspool stripe_width 0 expected_num_objects 1 pool 2 'metrics' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 last_change 76 flags hashpspool stripe_width 0 expected_num_objects 1 pool 3 'backups' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 last_change 78 flags hashpspool stripe_width 0 expected_num_objects 1 pool 4 'vms' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 last_change 80 flags hashpspool stripe_width 0 expected_num_objects 1 pool 5 'volumes' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 last_change 82 flags hashpspool stripe_width 0 expected_num_objects 1 [heat-admin@controller-0 ~]$
Fixed in openstack-tripleo-heat-templates-8.0.4-3.el7ost *** This bug has been marked as a duplicate of bug 1583333 ***