Hi,

just  a quick update. I deployed a rocky 9 vm in my openstack kolla-ansible test environment, ran bootstrap and deployed it as a compute node but there is one container that is listed as unhealthy:

(yoga) [vagrant@seed ~]$ ssh compute05 -l vagrant "sudo docker ps -a"
vagrant@compute05's password:
CONTAINER ID   IMAGE                                                                              COMMAND                  CREATED        STATUS                    PORTS     NAMES
3701de0c77e7   172.28.7.140:4000/openstack.kolla/centos-source-openvswitch-vswitchd:yoga          "dumb-init --single-…"   46 hours ago   Up 46 hours (healthy)               openvswitch_vswitchd
083e04ae0fcb   172.28.7.140:4000/openstack.kolla/centos-source-openvswitch-db-server:yoga         "dumb-init --single-…"   46 hours ago   Up 46 hours (unhealthy)             openvswitch_db
17a156ff5352   172.28.7.140:4000/openstack.kolla/centos-source-nova-compute:yoga                  "dumb-init --single-…"   46 hours ago   Up 46 hours (healthy)               nova_compute
12d308adf9ce   172.28.7.140:4000/openstack.kolla/centos-source-nova-libvirt:yoga                  "dumb-init --single-…"   46 hours ago   Up 46 hours (healthy)               nova_libvirt
cc35631ec420   172.28.7.140:4000/openstack.kolla/centos-source-nova-ssh:yoga                      "dumb-init --single-…"   46 hours ago   Up 46 hours (healthy)               nova_ssh
ad688dec24c8   172.28.7.140:4000/openstack.kolla/centos-source-prometheus-libvirt-exporter:yoga   "dumb-init --single-…"   46 hours ago   Up 46 hours                         prometheus_libvirt_exporter
02e52983458c   172.28.7.140:4000/openstack.kolla/centos-source-prometheus-cadvisor:yoga           "dumb-init --single-…"   46 hours ago   Up 46 hours                         prometheus_cadvisor
878a6eb1bb42   172.28.7.140:4000/openstack.kolla/centos-source-prometheus-node-exporter:yoga      "dumb-init --single-…"   46 hours ago   Up 46 hours                         prometheus_node_exporter
25faaf319f8a   172.28.7.140:4000/openstack.kolla/centos-source-cron:yoga                          "dumb-init --single-…"   46 hours ago   Up 46 hours                         cron
efd1cc64967c   172.28.7.140:4000/openstack.kolla/centos-source-kolla-toolbox:yoga                 "dumb-init --single-…"   46 hours ago   Up 46 hours                         kolla_toolbox
822876acf7c2   172.28.7.140:4000/openstack.kolla/centos-source-fluentd:yoga                       "dumb-init --single-…"   46 hours ago   Up 46 hours                         fluentd

To rule out that there is a general issue I deployed another rocky 8 and this worked just fine.

Also the compute services look just fine:

(yoga) [vagrant@seed ~]$ openstack compute service list
+--------------------------------------+----------------+-----------+----------+---------+-------+----------------------------+
| ID                                   | Binary         | Host      | Zone     | Status  | State | Updated At                 |
+--------------------------------------+----------------+-----------+----------+---------+-------+----------------------------+
| 3acd5340-4c79-4844-a321-ef3ac73602c5 | nova-scheduler | control01 | internal | enabled | up    | 2023-01-13T14:46:10.000000 |
| e289aef6-a3fa-402f-80e3-6bbd7807ae43 | nova-scheduler | control03 | internal | enabled | up    | 2023-01-13T14:46:12.000000 |
| 06d6ce7f-907d-4fe1-92bc-611511d7ce01 | nova-scheduler | control02 | internal | enabled | up    | 2023-01-13T14:46:08.000000 |
| b5051a46-ca04-4dbb-9067-75091fe26cfe | nova-conductor | control01 | internal | enabled | up    | 2023-01-13T14:46:08.000000 |
| e3e0a17f-b445-44b7-8355-972754ab397f | nova-conductor | control02 | internal | enabled | up    | 2023-01-13T14:46:09.000000 |
| 293b8e26-48de-4656-97a6-2c7020d19e44 | nova-conductor | control03 | internal | enabled | up    | 2023-01-13T14:46:04.000000 |
| c007e4a5-bee9-4428-8494-5f2cdddaec91 | nova-compute   | compute01 | nova     | enabled | up    | 2023-01-13T14:46:08.000000 |
| 62615300-dbee-4337-bc81-c956377abee9 | nova-compute   | compute02 | nova     | enabled | up    | 2023-01-13T14:46:07.000000 |
| a497d03a-6de8-418f-94a2-8b610babf48b | nova-compute   | compute03 | nova     | enabled | up    | 2023-01-13T14:46:12.000000 |
| cbdbd319-77e2-4d98-aeee-f3e65d7d8d00 | nova-compute   | compute04 | nova     | enabled | up    | 2023-01-13T14:46:10.000000 |
| 56e126e5-ddd5-4b4d-94d6-dff2cf850ddb | nova-compute   | compute05 | nova     | enabled | up    | 2023-01-13T14:46:06.000000 |
+--------------------------------------+----------------+-----------+----------+---------+-------+----------------------------+

I manually restarted the unhealthy container but it stays unhealthy. So I just spun up a couple cirros instances. Then I associated a FIP to the one that was scheduled on the rocky 9 compute node and tried to ping it and it doesn't work.

Would be good to get some info on how to perform the upgrade.

Cheers,
Oliver







On Jan 8, 2023, at 9:56 PM, Oliver Weinmann <oliver.weinmann@me.com> wrote:


Hi,

That is a good question. I’m also running yoga on rocky 8 and due to some problems with yoga I would like to upgrade to zed too soon. I have created a very simple staging deployment on a single ESXi host with 3 controllers and 2 compute nodes with the same config that I use in the production cluster. This lets me try the upgrade path. I assume while there is the possibility to upgrade from rocky 8 to 9, I wouldn’t do that. Instead I would do a fresh install of rocky9. I can only think of the docs not being 100% accurate and you can run yoga on rocky9 too. I will give it a try.

Cheers,
Oliver

Von meinem iPhone gesendet

Am 08.01.2023 um 10:25 schrieb wodel youchi <wodel.youchi@gmail.com>:


Hi,

Reading the kolla documentation, I saw that Yoga is supported on Rocky 8 only and Zed is supported on Rokcy 9 only, how to do the upgrade from Yoga to Zed since we have to do OS upgrade also???

Regards.