Hello,

Might not be related but when we upgraded from Queens to Rocky the Cinder API stopped
working when requests with query parameters in the URL were passed, the API just responded
with 400 bad request, but without query parameters the request worked.

We ended up having to add the resource_filters.json file to /etc/cinder directory which was not
in the cinder RDO package, have a patch up for that [1].

We fixed it manually in our upgrade workflow with Ansible.

- name: check cinder resource_filters.json file
  stat: path=/etc/cinder/resource_filters.json
  register: resource_filters_file

- name: download cinder resource_filters.json if not exist
  get_url:
    url: "https://git.openstack.org/cgit/openstack/cinder/plain/etc/cinder/resource_filters.json?h=stable/{{ openstack_version | lower }}"
    dest: /etc/cinder/resource_filters.json
    owner: root
    group: cinder
    mode: 0644
  when: not resource_filters_file.stat.exists

Best regards
Tobias

[1] https://review.rdoproject.org/r/#/c/16448/

On 10/15/2018 02:31 PM, Tzach Shefi wrote:
Hey Cody, 

Sorry for delay, just getting back up to speed.  
I just checked on my Rocky (OPSD) deployment same issue with Cinder/Horizon. 
In my case I created a new project/user (as _member_)
Creating an empty volume fails with same errors you reported:
Error: Unable to retrieve volumes
Error: Unable to retrieve volume snapshots

Checking with folks here, I'll get back to you with an answer. 

Tzach 


On Fri, Oct 12, 2018 at 2:57 AM Cody <codeology.lab@gmail.com> wrote:
Hello everyone,

Lately I run into two problems with using TripleO to deploy OpenStack (Queens). One of them is that users with _member_ or member roles appear to be unable to use cinder-volume service via the Dashboard (Horizon).

For any projects, users with either _member_ or member role received following pop-up error messages while trying to launch instances:

Error: Unable to retrieve volumes
Error: Unable to retrieve volume snapshots

There is also no volume option to choose from the drop-down menu in the Source sector, although the bootable volume is present and publicly accessible.

It doesn't affect the same users using CLI - they can create volumes or boot instances from volume normally from the command line.

The issue is reproducible and consistent with varied cluster configurations and cinder backends (nfs, file, rbd).  

Has anyone else encountered the same experience?


Best regards,
Cody










_______________________________________________
users mailing list
users@lists.rdoproject.org
http://lists.rdoproject.org/mailman/listinfo/users

To unsubscribe: users-unsubscribe@lists.rdoproject.org


--

Tzach Shefi

Senior Quality Engineer, RHCSA

Red Hat

tshefi@redaht.com    M: +972-54-4701080     IM: tshefi