Description of problem: 1. setup xenial node to use rhcs packages ceph_origin: repository ceph_repository: rhcs ceph_repository_type: cdn ceph_rhcs_cdn_debian_repo: https://xxxxx:xxxx@rhcs.download.redhat.com ceph_rhcs_version: 3 2. hosts config is [mons] ceph-clacroix-run400-node1-mon monitor_interface=ens3 ceph-clacroix-run400-node3-mon monitor_interface=ens3 ceph-clacroix-run400-node2-mon monitor_interface=ens3 [mgrs] ceph-clacroix-run400-node7-mgr monitor_interface=ens3 [osds] ceph-clacroix-run400-node5-osd monitor_interface=ens3 devices='["/dev/vdb", "/dev/vdc", "/dev/vdd"]' ceph-clacroix-run400-node4-osd monitor_interface=ens3 devices='["/dev/vdb", "/dev/vdc", "/dev/vdd"]' ceph-clacroix-run400-node6-osd monitor_interface=ens3 devices='["/dev/vdb", "/dev/vdc", "/dev/vdd"]' [rgws] ceph-clacroix-run400-node9-rgw radosgw_interface=ens3 [clients] ceph-clacroix-run400-node10-client client_interface=ens3 After cluster is setup, mgr is not active, debugging further shows that mgr doesn't have repo setup properly unlike mon/osd etc and mgr is using upstream packages. more check should be added to ceph-ansible to ensure rhcs packages are properly installed. cephuser@ceph-clacroix-run400-node1-mon:~$ sudo ceph --version ceph version 12.2.1-42redhat1xenial (c6d85fd953226c9e8168c9abe81f499d66cc2716) luminous (stable) cephuser@ceph-clacroix-run400-node1-mon:~$ sudo ceph -s cluster: id: 19dc7064-49b0-42e0-beef-f8e14b163927 health: HEALTH_WARN no active mgr clock skew detected on mon.ceph-clacroix-run400-node1-mon, mon.ceph-clacroix-run400-node3-mon services: mon: 3 daemons, quorum ceph-clacroix-run400-node2-mon,ceph-clacroix-run400-node1-mon,ceph-clacroix-run400-node3-mon mgr: no daemons active osd: 9 osds: 9 up, 9 in data: pools: 0 pools, 0 pgs objects: 0 objects, 0 bytes usage: 0 kB used, 0 kB / 0 kB avail pgs:
This has to be fixed in z4 folks not z5.
Does that mean there is a downstream repo for ceph-mgr? Thanks. Something like rhel-7-server-rhceph-3-mgr-rpms?
Sebastein, This issue is on xenial using rhcs_cdn packages , Manohar found this to be a doc issue and we are missing following in docs, I think a better approach is to link 3-release folder to 3-updates, I will check with Ken or RCM on this. 1. #ceph_rhcs_cdn_debian_repo: https://customername:customerpasswd@rhcs.download.redhat.com 2. #ceph_rhcs_cdn_debian_repo_version: "/3-release/" # for GA, later for updates use /3-updates/
I'm not sure I'm following, but let me know what to do and I'll do it or submit a PR on ceph-ansible :). Thanks.
ok changing from 3-release to 3-udpates didnt work, although we got new ceph version, sebastein to explain this, i think you are assuming rhel nodes thats why you are saying rhel-7-server-rhceph-3-mgr-rpms, but for xenial its in /etc/apt/sources.list.d/ where ceph-ansible creates the repo based on vars on other nodes eg: osd, mon it creates the repo $ ls -lt /etc/apt/sources.list.d/rhcs_download_redhat_com_3_updates_MON.list -rw-r--r-- 1 root root 116 Jun 26 17:12 /etc/apt/sources.list.d/rhcs_download_redhat_com_3_updates_MON.list $ sudo ceph --version ceph version 12.2.4-14redhat1xenial (03fd19535b3701f3322c68b5f424335d6fc8dd66) luminous (stable) But on mgr, we can see there is no repo and it ends up using upstream packages mgr:~$ ls -lt /etc/apt/sources.list.d/ total 0 $ sudo ceph --version ceph version 10.2.9 (2ee413f77150c0f375ff6f10edd6c8f9c7d060d0)
I gave this a try here https://github.com/ceph/ceph-ansible/pull/2814
@Gregory & @Sebastien, is this going to be fixed in z4? If not, can you please change the target release?
I'll be in z4. We are going to make a new release upstream today which will correspond to z4. Thanks.
https://github.com/ceph/ceph-ansible/releases/tag/v3.0.37
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/RHSA-2018:2177