The volume_backup_unset message is a red herring, as it comes from python-openstackclient itself. Ron, you may want to reinstall your client (but it won't solve your placement issues). On Tue, 24 Sept 2024 at 15:28, Eugen Block <eblock@nde.ag> wrote:

Hi,

I'd start with debug logs for placement and see where it refers to "volume_backup_unset". I don't have an idea how the volume backup gets into the way here. You could also run the command with the debug flag:

openstack --debug --os-placement-api-version 1.2 resource class list

and see if anything stands out.

What does a curl return when you query the endpoint directly? Here's my output:

root@controller02:~# curl http://controller.domain:8778 {"versions": [{"id": "v1.0", "max_version": "1.36", "min_version": "1.0", "status": "CURRENT", "links": [{"rel": "self", "href": ""}]}]}

Regards, Eugen

Zitat von Ron Gage <ron@rongage.org>:

...

Hi all:

I must have done something wrong in the installation of Openstack Placement. I have been through the steps to install placement multiple times now. I am using the Caracal release. I am on CentOS Stream 9.

Everything seems to be installed correctly. Indeed, the upgrade check returns no errors:

[root@cloud ~]# placement-status upgrade check +-------------------------------------------+ | Upgrade Check Results | +-------------------------------------------+ | Check: Missing Root Provider IDs | | Result: Success | | Details: None | +-------------------------------------------+ | Check: Incomplete Consumers | | Result: Success | | Details: None | +-------------------------------------------+ | Check: Policy File JSON to YAML Migration | | Result: Success | | Details: None | +-------------------------------------------+ [root@cloud ~]#

Verifying operation however returns nothing:

[root@cloud ~]# openstack --os-placement-api-version 1.2 resource class list Could not load 'volume_backup_unset': module 'openstackclient.volume.v2.volume_backup' has no attribute 'UnsetVolumeBackup' Expecting value: line 1 column 1 (char 0)

The placement service has been defined (the host name is "cloud", not "controller").

[root@cloud ~]# openstack endpoint list Could not load 'volume_backup_unset': module 'openstackclient.volume.v2.volume_backup' has no attribute 'UnsetVolumeBackup'

+----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+

...

| ID | Region | Service Name | Service Type | Enabled | Interface | URL |

+----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+

...

| 0e81501847db4aa38c864a2329a43052 | RegionOne | glance | image | True | admin | http://cloud:9292 | | 171c34b267674e798f625a811e87fcf8 | RegionOne | keystone | identity | True | admin | http://cloud:5000/v3/ | | 209386e7cddd44bbbbd5d5b3b9aa4ff5 | RegionOne | nova | compute | True | internal | http://cloud:8774/v2.1 | | 2f9d477115924b24a12204f54b1d6ddb | RegionOne | placement | placement | True | internal | http://cloud:8778 | | 58eb375281724e7998c0b6c8015302c4 | RegionOne | keystone | identity | True | public | http://cloud:5000/v3/ | | 618b400dc67c4038a6f62540f97d4246 | RegionOne | glance | image | True | internal | http://cloud:9292 | | 77843d0ddbac4cacb805cefd1c029544 | RegionOne | glance | image | True | public | http://cloud:9292 | | 8a8cca4fee32461fbe7836e756751df5 | RegionOne | keystone | identity | True | internal | http://cloud:5000/v3/ | | a01d89f6ea624e09af9d08d85e9bef40 | RegionOne | nova | compute | True | admin | http://cloud:8774/v2.1 | | d06d6b75e9bf4014b4de07df0bce8dd1 | RegionOne | placement | placement | True | public | http://cloud:8778 | | dfd618e52cc54af6bcf7afbf1a3a32cd | RegionOne | nova | compute | True | public | http://cloud:8774/v2.1 | | f6c41b2d7ba946908da5f0ed4d591c2a | RegionOne | placement | placement | True | admin | http://cloud:8778 |

+----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+

...

What on earth could I have messed up to prevent the Placement API from working. Yes, I have restarted httpd.

How do I know that Placement isn't working? Nova Conductor and Nova Scheduler can't connect to it (firewall is off and iptables rules are flushed) and therefore won't start properly. I can telnet to port 8778 yet the Nova services claim it can't connect.

Thank you for your help!

Ron Gage

All: I apologize but I decided to "start over" with this project. Wiped it all out and try again.  That said, I can offer up some answers here. The Openstack client - never caught the version number but it is whatever CentOS is distributing from their Caracal repo. The installation is the results of these instructions: https://docs.openstack.org/install-guide/openstack-services.html#minimal-dep.... Yes, I know it's not Caracal but since there is no such thing posted anywhere I can find.  I figured the differences should be minimal.  The install is an all-in-one install with no load balancers or anything fun like that. I had found the problem with the install and it was what led to my decision to start over.  The WSGI linkage between the placement code and httpd wasn't right.  Doing a 'curl http://cloud:8778/' was returning the Apache Welcome page and not the expected JSON answer.  I removed the welcome.conf from apache, restarted and was getting a permissions denied page.  Apache logs were showing 403 error codes on all requests. I know it's not much to go on, but maybe it will help somehow.  I should have the rebuild done tonight and if you want I can provide access (maybe a Jitsi Screen Share or a Google Meet). Ron Gage On 9/25/2024 2:03 AM, Franciszek Przewoźny wrote:

Hi Ron,

Please share: - openstack client version you are using - openrc contents (without secrets) - is it all-in-one installation? Are you using some kind of load balancer in front of services?

Regards, Franciszek

...

On 24 Sep 2024, at 16:17, Pierre Riteau <pierre@stackhpc.com> wrote:

The volume_backup_unset message is a red herring, as it comes from python-openstackclient itself. Ron, you may want to reinstall your client (but it won't solve your placement issues).

On Tue, 24 Sept 2024 at 15:28, Eugen Block <eblock@nde.ag> wrote:

Hi,

I'd start with debug logs for placement and see where it refers to "volume_backup_unset". I don't have an idea how the volume backup gets into the way here. You could also run the command with the debug flag:

openstack --debug --os-placement-api-version 1.2 resource class list

and see if anything stands out.

What does a curl return when you query the endpoint directly? Here's my output:

root@controller02:~# curl http://controller.domain:8778 {"versions": [{"id": "v1.0", "max_version": "1.36", "min_version": "1.0", "status": "CURRENT", "links": [{"rel": "self", "href": ""}]}]}

Regards, Eugen

Zitat von Ron Gage <ron@rongage.org>:

> Hi all: > > I must have done something wrong in the installation of Openstack > Placement.  I have been through the steps to install placement > multiple times now.  I am using the Caracal release.  I am on CentOS > Stream 9. > > Everything seems to be installed correctly. Indeed, the upgrade > check returns no errors: > > [root@cloud ~]# placement-status upgrade check > +-------------------------------------------+ > | Upgrade Check Results                     | > +-------------------------------------------+ > | Check: Missing Root Provider IDs          | > | Result: Success                           | > | Details: None                             | > +-------------------------------------------+ > | Check: Incomplete Consumers               | > | Result: Success                           | > | Details: None                             | > +-------------------------------------------+ > | Check: Policy File JSON to YAML Migration | > | Result: Success                           | > | Details: None                             | > +-------------------------------------------+ > [root@cloud ~]# > > Verifying operation however returns nothing: > > [root@cloud ~]# openstack --os-placement-api-version 1.2 resource class list > Could not load 'volume_backup_unset': module > 'openstackclient.volume.v2.volume_backup' has no attribute > 'UnsetVolumeBackup' > Expecting value: line 1 column 1 (char 0) > > The placement service has been defined (the host name is "cloud", > not "controller"). > > [root@cloud ~]# openstack endpoint list > Could not load 'volume_backup_unset': module > 'openstackclient.volume.v2.volume_backup' has no attribute > 'UnsetVolumeBackup' > +----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+ > | ID                               | Region | Service Name | > Service Type | Enabled | Interface | URL                    | > +----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+ > | 0e81501847db4aa38c864a2329a43052 | RegionOne | glance       | > image        | True    | admin     | http://cloud:9292 | > | 171c34b267674e798f625a811e87fcf8 | RegionOne | keystone     | > identity     | True    | admin     | http://cloud:5000/v3/ | > | 209386e7cddd44bbbbd5d5b3b9aa4ff5 | RegionOne | nova         | > compute      | True    | internal  | http://cloud:8774/v2.1 | > | 2f9d477115924b24a12204f54b1d6ddb | RegionOne | placement    | > placement    | True    | internal  | http://cloud:8778 | > | 58eb375281724e7998c0b6c8015302c4 | RegionOne | keystone     | > identity     | True    | public    | http://cloud:5000/v3/ | > | 618b400dc67c4038a6f62540f97d4246 | RegionOne | glance       | > image        | True    | internal  | http://cloud:9292 | > | 77843d0ddbac4cacb805cefd1c029544 | RegionOne | glance       | > image        | True    | public    | http://cloud:9292 | > | 8a8cca4fee32461fbe7836e756751df5 | RegionOne | keystone     | > identity     | True    | internal  | http://cloud:5000/v3/ | > | a01d89f6ea624e09af9d08d85e9bef40 | RegionOne | nova         | > compute      | True    | admin     | http://cloud:8774/v2.1 | > | d06d6b75e9bf4014b4de07df0bce8dd1 | RegionOne | placement    | > placement    | True    | public    | http://cloud:8778 | > | dfd618e52cc54af6bcf7afbf1a3a32cd | RegionOne | nova         | > compute      | True    | public    | http://cloud:8774/v2.1 | > | f6c41b2d7ba946908da5f0ed4d591c2a | RegionOne | placement    | > placement    | True    | admin     | http://cloud:8778 | > +----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+ > > What on earth could I have messed up to prevent the Placement API > from working.  Yes, I have restarted httpd. > > How do I know that Placement isn't working? Nova Conductor and Nova > Scheduler can't connect to it (firewall is off and iptables rules > are flushed) and therefore won't start properly.  I can telnet to > port 8778 yet the Nova services claim it can't connect. > > Thank you for your help! > > Ron Gage

Hi Ron, Glad that you were able to find the problem and thank you for sharing the solution. That being said, whether you are a beginner or an experienced OpenStack operator, I would always recommend using a deployment tool, such as Kolla Ansible for example: https://docs.openstack.org/kolla-ansible/latest/ Pierre Riteau On Wed, 25 Sept 2024 at 20:30, Ron Gage <ron@rongage.org> wrote:

I found the problem and have it fixed.

The base of the problem is that there is no granted permissions on the wsgi script.

I got the placement api working by adding the following to the 00-placement-api.conf file:

<Directory /usr/bin> <IfVersion >= 2.4> Require all granted </IfVersion> <IfVersion < 2.4> Order allow,deny Allow from all </IfVersion> </Directory> </VirtualHost>

Without this stanza in the conf file, attempting to get the resource class list would return nothing. With this stanza in place, then the resource class list returns expected data.

Someone might want to consider this stanza for inclusion in the .conf file.

Ron Gage

On 9/25/24 7:32 AM, Ron Gage wrote:

All:

I apologize but I decided to "start over" with this project. Wiped it all out and try again. That said, I can offer up some answers here.

The Openstack client - never caught the version number but it is whatever CentOS is distributing from their Caracal repo.

The installation is the results of these instructions: https://docs.openstack.org/install-guide/openstack-services.html#minimal-dep.... Yes, I know it's not Caracal but since there is no such thing posted anywhere I can find. I figured the differences should be minimal. The install is an all-in-one install with no load balancers or anything fun like that.

I had found the problem with the install and it was what led to my decision to start over. The WSGI linkage between the placement code and httpd wasn't right. Doing a 'curl http://cloud:8778/' was returning the Apache Welcome page and not the expected JSON answer. I removed the welcome.conf from apache, restarted and was getting a permissions denied page. Apache logs were showing 403 error codes on all requests.

I know it's not much to go on, but maybe it will help somehow. I should have the rebuild done tonight and if you want I can provide access (maybe a Jitsi Screen Share or a Google Meet).

Ron Gage

On 9/25/2024 2:03 AM, Franciszek Przewoźny wrote:

Hi Ron,

Please share: - openstack client version you are using - openrc contents (without secrets) - is it all-in-one installation? Are you using some kind of load balancer in front of services?

Regards, Franciszek

On 24 Sep 2024, at 16:17, Pierre Riteau <pierre@stackhpc.com> <pierre@stackhpc.com> wrote:

The volume_backup_unset message is a red herring, as it comes from python-openstackclient itself. Ron, you may want to reinstall your client (but it won't solve your placement issues).

On Tue, 24 Sept 2024 at 15:28, Eugen Block <eblock@nde.ag> wrote:

...

Hi,

I'd start with debug logs for placement and see where it refers to "volume_backup_unset". I don't have an idea how the volume backup gets into the way here. You could also run the command with the debug flag:

openstack --debug --os-placement-api-version 1.2 resource class list

and see if anything stands out.

What does a curl return when you query the endpoint directly? Here's my output:

root@controller02:~# curl http://controller.domain:8778 {"versions": [{"id": "v1.0", "max_version": "1.36", "min_version": "1.0", "status": "CURRENT", "links": [{"rel": "self", "href": ""}]}]}

Regards, Eugen

Zitat von Ron Gage <ron@rongage.org>:

...

Hi all:

I must have done something wrong in the installation of Openstack Placement. I have been through the steps to install placement multiple times now. I am using the Caracal release. I am on CentOS Stream 9.

Everything seems to be installed correctly. Indeed, the upgrade check returns no errors:

[root@cloud ~]# placement-status upgrade check +-------------------------------------------+ | Upgrade Check Results | +-------------------------------------------+ | Check: Missing Root Provider IDs | | Result: Success | | Details: None | +-------------------------------------------+ | Check: Incomplete Consumers | | Result: Success | | Details: None | +-------------------------------------------+ | Check: Policy File JSON to YAML Migration | | Result: Success | | Details: None | +-------------------------------------------+ [root@cloud ~]#

Verifying operation however returns nothing:

[root@cloud ~]# openstack --os-placement-api-version 1.2 resource class list Could not load 'volume_backup_unset': module 'openstackclient.volume.v2.volume_backup' has no attribute 'UnsetVolumeBackup' Expecting value: line 1 column 1 (char 0)

The placement service has been defined (the host name is "cloud", not "controller").

[root@cloud ~]# openstack endpoint list Could not load 'volume_backup_unset': module 'openstackclient.volume.v2.volume_backup' has no attribute 'UnsetVolumeBackup'

+----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+

...

| ID | Region | Service Name | Service Type | Enabled | Interface | URL |

+----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+

...

| 0e81501847db4aa38c864a2329a43052 | RegionOne | glance | image | True | admin | http://cloud:9292 | | 171c34b267674e798f625a811e87fcf8 | RegionOne | keystone | identity | True | admin | http://cloud:5000/v3/ | | 209386e7cddd44bbbbd5d5b3b9aa4ff5 | RegionOne | nova | compute | True | internal | http://cloud:8774/v2.1 | | 2f9d477115924b24a12204f54b1d6ddb | RegionOne | placement | placement | True | internal | http://cloud:8778 | | 58eb375281724e7998c0b6c8015302c4 | RegionOne | keystone | identity | True | public | http://cloud:5000/v3/ | | 618b400dc67c4038a6f62540f97d4246 | RegionOne | glance | image | True | internal | http://cloud:9292 | | 77843d0ddbac4cacb805cefd1c029544 | RegionOne | glance | image | True | public | http://cloud:9292 | | 8a8cca4fee32461fbe7836e756751df5 | RegionOne | keystone | identity | True | internal | http://cloud:5000/v3/ | | a01d89f6ea624e09af9d08d85e9bef40 | RegionOne | nova | compute | True | admin | http://cloud:8774/v2.1 | | d06d6b75e9bf4014b4de07df0bce8dd1 | RegionOne | placement | placement | True | public | http://cloud:8778 | | dfd618e52cc54af6bcf7afbf1a3a32cd | RegionOne | nova | compute | True | public | http://cloud:8774/v2.1 | | f6c41b2d7ba946908da5f0ed4d591c2a | RegionOne | placement | placement | True | admin | http://cloud:8778 |

+----------------------------------+-----------+--------------+--------------+---------+-----------+------------------------+

...

What on earth could I have messed up to prevent the Placement API from working. Yes, I have restarted httpd.

How do I know that Placement isn't working? Nova Conductor and Nova Scheduler can't connect to it (firewall is off and iptables rules are flushed) and therefore won't start properly. I can telnet to port 8778 yet the Nova services claim it can't connect.

Thank you for your help!

Ron Gage