Create a variable for storage_availability_zone

Bug #1399415 reported by Ian Cordasco
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenStack-Ansible
Fix Released
Medium
Kevin Carter

Bug Description

Opened by ihavegerms on 2014-11-01 02:47:40+00:00 at https://github.com/rcbops/ansible-lxc-rpc/issues/458

------------------------------------------------------------

X will be doing most of their work from the dashboard and want to be able to specify which of their Cinder nodes they'll be creating block devices on. With how things are currently configured, all cinder nodes are part of the default "Nova" availability zone. The variable could/would be able to hold a few values to define different availability zones.

At this point, I've manually modified /etc/cinder/cinder.conf on each (2) of the cinder nodes to the first part of the hostname (prior to the underscore) as seen below (the change does require the cinder-volume containers to be restarted);

| cinder-scheduler | 604001-cinder1_cinder_volumes_container-bd45f3e0 | 604001-cinder1 | enabled | up | 2014-11-01T02:35:18.000000 | None |
| cinder-scheduler | 604009-cinder2_cinder_volumes_container-4bb88f1e | 604009-cinder2 | enabled | up | 2014-11-01T02:35:02.000000 | None |
| cinder-volume | 604001-cinder1_cinder_volumes_container-bd45f3e0@lvm | 604001-cinder1 | enabled | up | 2014-11-01T02:35:10.000000 | None |
| cinder-volume | 604009-cinder2_cinder_volumes_container-4bb88f1e@lvm | 604009-cinder2 | enabled | up | 2014-11-01T02:35:04.000000 | None |

With an obvious pitfall... if the containers were for whatever reason destroyed, re-creation would erase anything statically defined.

Tags: Backport Potential, enhancement, Pending Backport Merge, prio:3

====================== COMMENTS ============================

Comment created by JCallicoat on 2014-11-03 01:20:40+00:00

We can do this with cinder multi-backends, which are already supported by the playbooks[1], though I have no idea if/how well it works with Horizon.

I seem to recall @cloudnull telling me how to stick arbitrary variables in configs, but I don't remember the details. I'm helpful.

[1] https://github.com/rcbops/ansible-lxc-rpc/blob/master/etc/README.rst#adding-options-to-containers-within-targeted-hosts
https://wiki.openstack.org/wiki/Cinder-multi-backend
http://docs.openstack.org/admin-guide-cloud/content/multi_backend.html

------------------------------------------------------------

Comment created by cloudnull on 2014-11-03 05:20:54+00:00

@ihavegerms @JCallicoat Please review https://github.com/rcbops/ansible-lxc-rpc/pull/459 to make sure it fits the use case for the customer. If all is well then we may need to think about backporting to 9.0.x

------------------------------------------------------------

Comment created by cfarquhar on 2014-11-03 21:26:28+00:00

@cloudnull #459 fits the use case for this customer. Please backport to 9.0.x.

------------------------------------------------------------

Comment created by sigmavirus24 on 2014-11-04 21:07:52+00:00

Fixed in #459, backported in #467 and #466

------------------------------------------------------------

Comment created by cloudnull on 2014-11-17 20:43:20+00:00

This issue is remaining open as there is a doc impact per issue: https://github.com/rcbops/privatecloud-docs/issues/721

Revision history for this message
Kevin Carter (kevin-carter) wrote :

fix was released. prior to stackforge migration.

Changed in openstack-ansible:
status: New → Fix Released
importance: Undecided → Medium
assignee: nobody → Kevin Carter (kevin-carter)
milestone: none → 9.0.4
Changed in openstack-ansible:
status: Fix Released → Fix Committed
Changed in openstack-ansible:
milestone: 9.0.4 → 10.1.0
Changed in openstack-ansible:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.