s2i-dotnetcore
s2i-dotnetcore copied to clipboard
redhat-developer
Diff imagestreams and templates with openshift/openshift-ansible and openshift/origin
The imagestreams and templates from this repos are copied to the openshift/openshift-ansible and openshift/origin repos.
We need to check these files are the same. And it would be good to have a CI job that detects when they are no longer in sync.
CC @omajid @aslicerh @RheaAyase
I was looking at this earlier, and it looks like ansible creates it's example repos from a script that grabs the newest imagestream files directly from here.
See: https://github.com/openshift/openshift-ansible/blob/2af1f4992c52559015297562f9dc897821f0488c/roles/openshift_examples/examples-sync.sh#L45
So, I don't know if it is necessary to check the files there.
The openshift origin one is probably valid to check.
I also found imagestreams in openshift/library. Those are also 'auto-pulled' by a script, but unlike the ansible stuff above, the library stuff seems to only keep one version around, not 'packs' of version numbered examples.
I was thinking about creating a python script to put into the base of s2i-dotnetcore to just md5 these things. We could then use that as a jenkins job to check for this stuff.
Thoughts?
@aslicerh see this comment: https://github.com/openshift/library/pull/68#issuecomment-377533964
From chatting with @jerboaa I learned that updates of openshift/openshift-ansible go through a RHBZ workflow. After a ticket is created, OpenShift qa does verification. Once that is ok, Scott Dodson is asked via mail to update those files. For openshift/origin, we are responsible for keeping things in sync.
I was thinking about creating a python script to put into the base of s2i-dotnetcore to just md5 these things. We could then use that as a jenkins job to check for this stuff.
This is the remaining action for this issue. I personally prefer a simple bash script. And perhaps we can call 'diff' to visually show differences between the files.