Hi Alan. Thanks for your help. As I don't see the problem, as I can't find the solution in the cinder-scheduler logs, and as it's a lab with physical servers (openstack) and virtual machines (ceph), I'm going to start all over again. I’m a beginner with Ceph. Maybe there is something I did wrong. I will also change the version of Ceph. Thanks again. Franck
Le 15 nov. 2023 à 15:43, Alan Bishop <abishop@redhat.com> a écrit :
On Wed, Nov 15, 2023 at 6:32 AM Franck VEDEL <franck.vedel@univ-grenoble-alpes.fr <mailto:franck.vedel@univ-grenoble-alpes.fr>> wrote:
Hi Pierre Maybe I found something: In the container « cinder_volume », there is a file /etc/ceph/ceph.conf But, no /etc/ceph/ceph.conf file in « cinder_scheduler » container. I think it’s my problem !!
Sorry, but the scheduler does not require access to /etc/ceph files. I think you need to review the scheduler logs to see why it concluded there are no available c-vol backends. Since you are able to create a volume, I assume the ceph backend is "up," but it would also be good to verify it remains up, especially during the time when you're trying to create a volume from an image.
Alan
but why ? What is the option in globals.yml ? Or maybe there is something to add in /etc/kolla/config/cinder/
I have this (the last try):
enable_cinder: "yes" enable_cinder_backup: "no" enable_cinder_backend_lvm: "no" external_ceph_cephx_enabled: "yes"
# Glance ceph_glance_keyring: "ceph.client.glance.keyring" ceph_glance_user: "glance" ceph_glance_pool_name: "images" # Cinder ceph_cinder_keyring: "ceph.client.cinder.keyring" ceph_cinder_user: "cinder" ceph_cinder_pool_name: "volumes"
#ceph_cinder_backup_keyring: "ceph.client.cinder-backup.keyring" #ceph_cinder_backup_keyring: "ceph.client.cinder.keyring" #ceph_cinder_backup_user: "cinder" #ceph_cinder_backup_user: "cinder-backup" #ceph_cinder_backup_pool_name: "backups" # Nova #ceph_nova_keyring: "{{ ceph_cinder_keyring }}" ceph_nova_keyring: "ceph.client.nova.keyring" ceph_nova_user: "nova" ceph_nova_pool_name: "vms"
# Configure image backend. glance_backend_ceph: "yes"
# Enable / disable Cinder backends cinder_backend_ceph: "yes"
# Nova - Compute Options ######################## nova_backend_ceph: "yes"
Thanks a lot
Franck
Le 15 nov. 2023 à 09:54, Pierre Riteau <pierre@stackhpc.com <mailto:pierre@stackhpc.com>> a écrit :
Hi Franck,
It would help if you could share more details about the error (check both cinder-scheduler and cinder-volume logs).
Do you have capabilities to access the images pool (presumably what is used by Glance) on your client.cinder user?
Best wishes, Pierre Riteau (priteau)
On Wed, 15 Nov 2023 at 09:50, Franck VEDEL <franck.vedel@univ-grenoble-alpes.fr <mailto:franck.vedel@univ-grenoble-alpes.fr>> wrote:
Good morning, I'll come back for help, thanks in advance. I am testing an Openstack 2023.1/Ceph Pacific environment. The cluster works, and from Horizon for example, if I create a simple volume, it works normally, it is placed in the "volumes" pool in the Ceph cluster, exactly as expected.
if I create a volume from an image, ERROR cinder-schedule. In the cinder_volume and cinder_scheduler containers, same settings in cinder.conf.
Before continuing to find out what the error could be, could there be a problem between Antelope (2023.1) and Ceph Pacific? (I'm on Ubuntu 22.04).
Franck