On Mon, Sep 21, 2020 at 11:11 AM Ruslanas Gžibovskis <ruslanas@lpic.lt> wrote:
Yes, I do not have vdb... I have sda sdb sdc sde sdd... and I believe it might have come from journal_size: 16384 ? here is a part of conf file... CephAnsibleDiskConfig: devices: - /dev/sdc - /dev/sde - /dev/sdd osd_scenario: lvm osd_objectstore: bluestore journal_size: 16384 # commented this out now.
If you used the above, perhaps in foo.yaml, but got the error message you shared, then I suspect you are deploying with your parameters in the wrong order. You should use the following order: openstack overcloud deploy --templates \ -e /usr/share/openstack-tripleo-heat-templates/environments/ceph-ansible/ceph-ansible.yaml \ -e foo.yaml If the order of arguments is such that foo.yaml precedes /usr/share/openstack-tripleo-heat-templates/environments/ceph-ansible/ceph-ansible.yaml then the CephAnsibleDisksConfig will override what was set in foo.yaml and instead use the default in /usr/share/openstack-tripleo-heat-templates/environments/ceph-ansible/ceph-ansible.yaml which uses a disk you don't have. Also, please do not use journal_size it is deprecated and that parameter doesn't make sense for bluestore. As linked from the documentation ceph-volume batch mode (https://docs.ceph.com/en/latest/ceph-volume/lvm/batch/) should do the right thing if you modify the above (and just drop journal size). John
Yes, undercloud node cleaning is the first option I enable/configure in undercloud.conf ;) after that I configure IP addresses/subnets :)
On Mon, 21 Sep 2020 at 16:55, John Fulton <johfulto@redhat.com> wrote:
In [2] I see Error: Could not stat device /dev/vdb - No such file or directory.
/dev/vdb is the default and as per the logs it doesn't exist on your HCI node. For your HCI node you need to have a block device (usually a dedicated disk) which can be configured as an OSD and you need to pass the path to it as described in the following section of the doc.
https://docs.openstack.org/project-deploy-guide/tripleo-docs/latest/features...
Also, ensure your disk is factory clean or the ceph tools won't initialize it as an OSD. The easiest way to do this is to configure ironic's automatic node cleaning.
John
On Mon, Sep 21, 2020 at 9:45 AM Ruslanas Gžibovskis <ruslanas@lpic.lt> wrote:
Hi all,
using tripleo 13.4.0-1.el8 and ceph-ansible 4.0.25-1.el8 on CentOS8 from @centos-openstack-ussuri repo.
but get [1] error. I thought it was due to not found python exec, but I saw later, when added verbosity 4, that it is able to find python3.
It looks like output in paste.openstack.org is too short, not sure what was an issue, but here is a second place [2] for same full output
[1] http://paste.openstack.org/show/i2XpSBiSVjuL69Ahm1sl/ [2] https://proxy.qwq.lt/ceph-ansible.html
-- Ruslanas Gžibovskis +370 6030 7030
-- Ruslanas Gžibovskis +370 6030 7030