I’m doing pretty-much the same thing. Prior to “prerequisites” I run the
following play:
- hosts: nodes
become: yes
tasks:
- name: Install origin39 repo
yum:
name: centos-release-openshift-origin39
state: present
The 3.9 repo appears in /etc/yum.repos.d/ but, after the prerequisites, so does
"CentOS-OpenShift-Origin.repo” and the main “deploy_cluster.yml” fails again.
The only way through this for me to add “package_version” to
“openshift_disable_check”.
Alan Christie
[email protected]
> On 20 Aug 2018, at 15:20, Peter Heitman <[email protected]> wrote:
>
> I agree with you. I've hit this same error when previous versions were
> released. I'm not sure why defining the version we want to install (and then
> using that version of the openshift ansible git) isn't sufficient. As for
> installing the repo, I do this before I run the prerequisite playbook, i.e.
> ansible all -i <inventory> -m yum -a "name=centos-release-openshift-origin39
> state=present" --become. That seems to resolve the issue.
>
> On Mon, Aug 20, 2018 at 10:10 AM Alan Christie
> <[email protected] <mailto:[email protected]>>
> wrote:
> Thanks Peter.
>
> Interestingly it looks like it’s Origin’s own “prerequisites.yml” playbook
> that’s adding the repo that’s causing problems. My instances don’t have this
> repo until I run that playbook.
>
> Why do I have to remove something that’s being added by the prerequisite
> playbook? Especially as my inventory explicitly states
> "openshift_release=v3.9”?
>
> If the answer is “do not run prerequisites.yml” what’s the point of it?
>
> I still wonder why this specific issue is actually an error? Shouldn’t it be
> installing specific version anyway? Shouldn’t it be error occur if there is
> no 3.9 package, not if there’s a 3.10 package?
>
> Incidentally, I’m using the ansible code from "openshift-ansible-3.9.40-1”.
>
> Alan Christie
> [email protected] <mailto:[email protected]>
>
>
>
>> On 18 Aug 2018, at 13:36, Peter Heitman <[email protected]
>> <mailto:[email protected]>> wrote:
>>
>> See the recent thread "How to avoid upgrading to 3.10". The bottom line is
>> to install the 3.9 specific repo. For CentOS that is
>> centos-release-openshift-origin39
>>
>> On Sat, Aug 18, 2018, 2:44 AM Alan Christie
>> <[email protected] <mailto:[email protected]>>
>> wrote:
>> HI,
>>
>> I’ve been deploying new clusters of Origin v3.9 using the official Ansible
>> playbook approach for a few weeks now, using what appear to be perfectly
>> reasonable base images on OpenStack and AWS. Then, this week, with no other
>> changes having been made, the deployment fails with this message: -
>>
>> One or more checks failed
>> check "package_version":
>> Some required package(s) are available at a version
>> that is higher than requested
>> origin-3.10.0
>> origin-node-3.10.0
>> origin-master-3.10.0
>> This will prevent installing the version you requested.
>> Please check your enabled repositories or adjust
>> openshift_release.
>>
>> I can avoid the error, and deploy what appears to be a perfectly functional
>> 3.9, if I add package_version to openshift_disable_check in the inventory
>> the deployment. But this is not the right way to deal with this sort of
>> error.
>>
>> Q1) How does one correctly address this error?
>>
>> Q2) Out of interest … why is this specific issue an error? I’ve instructed
>> the playbook to instal v3.9. I don't care if there is a 3.10 release
>> available - I do care if there is not a 3.9. Shouldn’t the error occur if
>> there is no 3.9 package, not if there’s a 3.10 package?
>>
>> Alan Christie
>> Informatics Matters Ltd.
>>
>> _______________________________________________
>> users mailing list
>> [email protected] <mailto:[email protected]>
>> http://lists.openshift.redhat.com/openshiftmm/listinfo/users
>> <http://lists.openshift.redhat.com/openshiftmm/listinfo/users>
>
_______________________________________________
users mailing list
[email protected]
http://lists.openshift.redhat.com/openshiftmm/listinfo/users
