Open gojeaqui opened 8 months ago
If you want to update the assisted installer version you can set ai_version
in the all section of your inventory.
This being said to updated the OCP versions you need these three variables. os_images release_images supported_ocp_versions
The relation between Z version of os_image
and release_image
. This can be a useful reference.
Hello, unfortunately that is not enough.
There are things that changed between versions and it is not just changing the versions.
For example the people of nmstate.io remove the word "slave" and are using the word "port" for the bonding so that needs to change in the inventory and also in the templates. (that change is easy though)
Another thing is that something is different with regards to how Assisted Installer works, because the new version does not have the ISO where the playbook expects it. In fact the README from (https://github.com/openshift/assisted-image-service) says: None of these APIs should be considered stable for end-users of assisted installer. Users should never construct URLs that match these APIs; instead users should obtain an ISO URL from an InfraEnv resource, as provided by assisted installer. These APIs represent a contract between assisted-image-service and assisted installer only.
It might be that they already have changed those URLs.
There is a fix for the nmstate issue, I haven't had chance to test it. #281
I tested v2.26.0
the other day and it worked for a fairly basic flow. We use the download_url returned by the api to fetch the iso and it doesn't look like the have removed that based on the swagger. Do you have the error you saw?
It might be worth using the agent based install method use_agent_based_installer: True
which extracts the openshift installer and uses the agent subcommand.
Do you have the error you saw?
No, I'm sorry. I should have kept it but it was lost.
The issue was in the task:
TASK [generate_discovery_iso : Download discovery ISO]
status_code: 403
uid: 0
url: http://192.168.207.9:8888/images/9a63de32-4892-472a-bd44-90d4c02d03e5?arch=x86_64&type=full-iso&version=4.12
(something like this)
And the problem was that in the URL there was nothing, no ISO.
When I went back the the current version of Assisted Installed the issue dissapeared.
I found this in my logs, but it's not the result of the ansible task:
[core@xx crucible]$ curl -vk http://192.168.xx.9:8888/byid/a54181f5-53b6-4411-bfaf-9e9b098f0eca
* Uses proxy env variable no_proxy == 'localhost,127.0.0.1,192.168.x.9,192.168.x.9'
* Trying 192.168.207.9...
* TCP_NODELAY set
* Connected to 192.168.207.9 (192.168.207.9) port 8888 (#0)
> GET /byid/a54181f5-53b6-4411-bfaf-9e9b098f0eca HTTP/1.1
> Host: 192.168.207.9:8888
> User-Agent: curl/7.61.1
> Accept: */*
>
< HTTP/1.1 404 Not Found
< Content-Type: text/plain; charset=utf-8
< X-Content-Type-Options: nosniff
< Date: Fri, 10 Nov 2023 22:24:36 GMT
< Content-Length: 19
< 404 page not found
Bug description
When trying to deploy the cluster (Openshift version 4.13.10) with the
deploy_cluster.yml
playbook I'm getting the error (in log output)I think this is because the
ai_version: v2.12.1
is from one year ago and Openshift 4.13 was GA in may this year.I tried to update to the latest in quay.io: v.26.1, but the playbooks keep using the same version, I don't know how to update the version now that the prerequisites have been deployed.
OpenShift version
other (provide in the description)
Assisted Installer version
other (provide in the description)
Relevant log output
Inventory file
No response
Required statements