Hi Allison,

I was able to test the upgrade twice today (2 control nodes, 1 compute 
node). The first attempt was from A to C (slurp). I had several issues 
along the way. I haven't been able to look closely for any root causes 
yet, and I also didn't check for existing bugs yet, but I wanted to 
share anyway:

1. nova-manage db online_data_migrations: one entry doesn't migrate (1 
rows matched query populate_instance_compute_id, 0 migrated)

2. cinder-manage db online_data_migrations:
Running batches of 50 until complete.
2025-04-22 11:01:30.558 837433 WARNING py.warnings [None 
req-daa04f7f-7daf-4b46-ac78-eff68794cae5 - - - - - -] 
/usr/lib/python3/dist-packages/cinder/db/sqlalchemy/api.py:8620: 
SAWarning: Coercing Subquery object into a select() for use in IN(); 
please pass a select() construct explicitly
   filter(admin_meta_table.id.in_(ids_query)).\

+------------------------------------------------+----------------+-------------+
| Migration                                      |   Total Needed |   
Completed |
|------------------------------------------------+----------------+-------------|
| remove_temporary_admin_metadata_data_migration |              0 |     
        0 |
+------------------------------------------------+----------------+-------------+

3. Some neutron agents (dhcp, metadata) don't properly start until all 
control nodes are upgraded, I needed to stop and start them again. 
Need to look for more details in the logs.

4. Horizon has stopped working entirely. I only get a "400 Bad 
Request", no matter what I do. When the packages upgraded, I saw a 
warning "No local_settings file found", but it is there:

root@controller02:~# ll 
/usr/lib/python3/dist-packages/openstack_dashboard/local/local_settings.py
lrwxrwxrwx 1 root root 42 Jun  7  2024 
/usr/lib/python3/dist-packages/openstack_dashboard/local/local_settings.py -> 
/etc/openstack-dashboard/local_settings.py

The dashboard_error.log also contains the warning that there's no 
local_settings file. I suspect that it has to do with the Django 
change mentioned in the Caracal release notes:

> Django 3.2 support was dropped. Django 3.2 ends its extended support 
> in April 2024. Considering this horizon dropped Django 3.2 support 
> and uses Django 4.2 as default.


I decided to rollback the VMs to a previous snapshot (back to 
Antelope) and tried the upgrade to Bobcat. The result is better, I 
didn't see the neutron agents fail, and the dashboard is still usable. 
The only issues I still see are the 'nova-manage db 
online_data_migrations' and 'cinder-manage db online_data_migrations'.

If some or all of these issues are know, could anyone point me to the 
relevant bug reports? If these are new bugs, I can create reports for 
them. But as I mentioned, I don't have too much information yet.

Thanks,
Eugen

Zitat von Allison Price <allison@openinfra.dev>:

> Hi Eugen,
>
> Excellent! Please keep me posted on how testing goes and then we can 
> take the next steps in talking about your production environment.
>
> Cheers,
> Allison
>
>> On Apr 21, 2025, at 3:10 PM, Eugen Block <eblock@nde.ag> wrote:
>>
>> Hi and thank you for the links. We just upgraded to Antelope last 
>> week, and we plan to do the next upgrade directly to Caracal since 
>> we need one of the fixes. I’m planning to test the slurp upgrade on 
>> a lab environment quite soon, maybe this week even. And if all goes 
>> well, we’ll upgrade our production quickly after that. I‘d be happy 
>> to share my experience in this thread. :-)
>>
>> Thanks!
>> Eugen
>>
>> Zitat von Allison Price <allison@openinfra.dev>:
>>
>>> Hi everyone,
>>>
>>> We have now published two case studies from OpenStack users who 
>>> have implemented and are benefiting from the SLURP upgrade 
>>> process[1]: Cleura[2] and Indiana University[3]
>>>
>>> I wanted to follow up to see if there are other organizations who 
>>> have implemented the SLURP upgrade process who would like to tell 
>>> your story? If we can get a few more, I would love to schedule an 
>>> OpenInfra Live to deep dive into the improvements made around 
>>> OpenStack upgrades and what users stuck on older releases should 
>>> know.
>>>
>>> If you are interested, please let me know. And even if you’re not 
>>> but you’re still using OpenStack—please remember to take the 
>>> OpenStack User Survey[4] so we can learn more!
>>>
>>> Thanks!
>>> Allison
>>>
>>>
>>>
>>> [1] 
>>> https://docs.openstack.org/project-team-guide/release-cadence-adjustment.html
>>> [2]https://superuser.openinfra.org/articles/streamlining-openstack-upgrades-managing-scalability-efficiency-and-compliance-at-cleura/
>>> [3] 
>>> https://superuser.openinfra.org/articles/streamlining-openstack-upgrades-a-case-study-of-indiana-universitys-efficient-upgrade-strategy/
>>> [4] https://www.openstack.org/user-survey
>>
>>
>>