Enable_rados/Osnailyfacter::Wait_for_backend[object-storage]/Haproxy_backend_status[object-storage]/ensure) change from down to up failed

[Egor Kotko]

Deployment of cluster failed with enabled default storages and Rados GW (https://snag.gy/RYSLBP)

        Scenario:
            1. Create cluster with RadosGW enabled
            2. Add 3 nodes with controller role
            3. Add 2 nodes with compute and ceph-osd role
            4. Add 2 nodes with ceph-osd role
            5. Deploy the cluster
            6. Run OSTF tests
            7. Check ceph status
            8. Check the radosgw daemon is started
            9. Create custom image via glance
            10.Compare custom image IDs
            via glance and swift client
            11.Check S3 API

http://paste.openstack.org/show/494816/

[Kyrylo Galanov (kgalanov)]

Could not reproduce. Could you please name the test failed?

[Timur Nurlygayanov (tnurlygayanov)]

Hi Egor, could you please provide more information and attach diagnostic snapshot from the environment?

Status changed to Incomplete because we still need additional info to fix the issue.

Thank you!

[Egor Kotko (ykotko)]

9.0 ISO #250
1. Create cluster with RadosGW enabled (https://snag.gy/daU71Q.jpg)
2. Add 3 nodes with controller role
3. Add 2 nodes with compute and ceph-osd role
4. Add 2 nodes with ceph-osd role
5. Deploy cluster

Actual result:
Deployment finished with errors in puppet logs of the controllers

[Egor Kotko (ykotko)]

You can find reproduced bug on the environment:
Host: srv84-bud.infra.mirantis.net
Then start the VMs:
9.0.custom_system_test.381.381_admin running
9.0.custom_system_test.381.381_slave-01
9.0.custom_system_test.381.381_slave-02
9.0.custom_system_test.381.381_slave-03
9.0.custom_system_test.381.381_slave-04
9.0.custom_system_test.381.381_slave-05
9.0.custom_system_test.381.381_slave-06
9.0.custom_system_test.381.381_slave-07
Master ip: 10.109.10.2

[Egor Kotko (ykotko)]

I have also tried to reproduce it on 10.0, the deployment failed with errors in puppet log, but not sure that it is connected with the feature.
10.0 ISO#185

[Egor Kotko (ykotko)]

Deployment has failed. All nodes are finished. Failed tasks: Task[upload_cirros/1], Task[ceph-compute/5], Task[ceph-compute/4] Stopping the deployment process!

[Denis Egorenko (degorenko)]

I faced with this bug too.

I have env with just one controller+ceph-osd, enabled radosgw and murano - same
error during executing modular task enable_radosgw.

305 ISO, 9.0-all.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/mitaka
Review: https://review.openstack.org/318533

[OpenStack Infra (hudson-openstack)]

Change abandoned by Kyrylo Galanov (<email address hidden>) on branch: stable/mitaka
Review: https://review.openstack.org/318533

[Kyrylo Galanov (kgalanov)]

I could reproduce it on 305 ISO, 9.0-all, but the bug is not reproduced on #3​50.
Could you please confirm?

[Kyrylo Galanov (kgalanov)]

No reply for a week. Please reopen the issue if the bug is reproduced with recent Fuel build.