Glance file error on HA model and default configration

Bug #1681301 reported by Gavin
22
This bug affects 4 people
Affects Status Importance Assigned to Milestone
kolla-ansible
Fix Released
Medium
Paul Bourke

Bug Description

I Just config a OpenStack HA environment with Kolla.
All global config is default.
After I run init-runonce and try to init a new machine,the create job is faild.

And I found that there is only one control node have the image file.

So.I think the images should not found in create machine job.I think that is a bug...

Tags: glance
Revision history for this message
Eduardo Gonzalez (egonzalez90) wrote :

Hi,
the issue here is using glance with backend file.
All glance nodes need the same image uploaded since file is not really meant to be HA.
There is a change under review at this moment to only start one glance process when backend is file.
https://review.openstack.org/#/c/448654/

The recommended method to deploy a HA environment is using ceph as backend for storage services.
You can quickly fix the issue stopping all glance services having only one node with glance running, or scp the images to the other nodes.

Regards

affects: kolla → kolla-ansible
Changed in kolla-ansible:
status: New → Confirmed
importance: Undecided → Medium
assignee: nobody → Paul Bourke (pauldbourke)
Changed in kolla-ansible:
status: Confirmed → In Progress
Changed in kolla-ansible:
milestone: none → pike-2
Revision history for this message
Eduardo Gonzalez (egonzalez90) wrote :
Changed in kolla-ansible:
milestone: pike-2 → pike-3
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to kolla-ansible (master)

Reviewed: https://review.openstack.org/448654
Committed: https://git.openstack.org/cgit/openstack/kolla-ansible/commit/?id=4fde486dc8b29b8d087ab4bfff0e626b2479abd2
Submitter: Jenkins
Branch: master

commit 4fde486dc8b29b8d087ab4bfff0e626b2479abd2
Author: Paul Bourke <email address hidden>
Date: Wed Mar 22 13:14:37 2017 +0000

    Only start one glance_api when using file backend

    The default backend storage for Glance in Kolla is file, as this is the
    easiest to deploy. However, this leads to frustration in a multinode
    scenario as the volumes aren't shared.

    This change sets glance_api to only start on one control node when using
    the file backend.

    DocImpact
    Closes-Bug: #1681301
    Change-Id: I2c167bae6f34a102cd7fe66401d7defe13bf2bf7

Changed in kolla-ansible:
status: In Progress → Fix Released
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix included in openstack/kolla-ansible 5.0.0.0b3

This issue was fixed in the openstack/kolla-ansible 5.0.0.0b3 development milestone.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

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