Description of problem: There is no documentation on rhc deploy usage. [0] is the first reference of this feature (when it was released to online). It was also covered in the 2.0 Beta (2). Because of these its referenced in the Release Notes [2], however I do not see any thing on this in the users guide [3]. We should file a doc bug for this! [0] https://www.openshift.com/blogs/introduction-to-deployments-and-rollbacks-on-openshift [1] https://access.redhat.com/site/articles/531473#Deployments_and_Rollbacks [2] https://access.redhat.com/site/documentation/en-US/OpenShift_Enterprise/2/html-single/2.0_Release_Notes/index.html [3] https://access.redhat.com/site/documentation/en-US/OpenShift_Enterprise/2/html-single/User_Guide/index.html#sect-Deployment Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
https://access.redhat.com/site/documentation/en-US/OpenShift_Enterprise/2/html-single/User_Guide/index.html#sect-Deployment looks like the section of documentation that needs an update.
Looks as though the information in the user Guide may be enough. Awaiting feedback on more info from devs.
Ok, I added the tarball structure to the Introduction to Snapshot section in the User Guide. I'm going to tag Andy because he was the one who provided the info. Please, if there's anything else anyone thinks I should add then please let me know.
Brice, Instead of calling it a binary snapshot, I would call it either a deployment snapshot, or a binary deployment snapshot. I think the word "deployment" is key. I wonder if it might make sense to move this section to the section on deployments (somewhere around 11.3.2.4). Or at least have the 2 sections reference each other perhaps? I didn't see an explanation of what goes in the "repo" directory - that's the application source code; i.e. what you'd find when you clone the git repository. I did want to clarify that there are 2 possible entries in managed_files.yml that matter: dependency_dirs and build_dependency_dirs. If you look at one of the JBoss cartridges, you'll see an example (e.g. https://github.com/openshift/origin-server/blob/master/cartridges/openshift-origin-cartridge-jbossas/metadata/managed_files.yml#L19-L22). Also, a little more clarification on entries starting/not starting with ~. An entry in dependency_dirs such as ~/foo will result in the following: dependencies/foo And entry that does not start with ~ such as just "foo" will be scoped to the cartridge, based on the cartridge's name: dependencies/mycartridge/foo where mycartridge could be nodejs, jbossas, jbosseap, ruby, php, etc. The same rules apply for build_dependency_dirs (which I believe are only used by the JBoss cartridges at present). So an entry of ~/.m2 under build_dependency_dirs would result in build_dependencies/.m2. And entry of foo under build_dependency_dirs would result in build_dependencies/mycartridge/foo, where "mycartridge" would be the name of the cartridge.
Andy, thanks for your suggestions. Would you prefer a binary package? I'd also be wary of using the word deployment as that may not be 100% of the reason why people might use this. Or at least, it may change up in the future. I added the info about the repo directory, and tried to clarify the way the dependencies directory is made up. Also added a little bit on the build_dependencies directory being JBoss specific. Also put links to the deployment topics at the end. I do think the layout of the topics is fine for now. We try to have a basic workflow, and I think as it is works. Please let me know if you have any more suggestions, or if I've misunderstood something here. Thanks, again. (the link to the topic in my previous comment above still works)
re "I'd also be wary of using the word deployment as that may not be 100% of the reason why people might use this." -- I'm not sure what you mean? The only prebuilt thing that can be deployed via "rhc deploy" is a tar.gz file that was either created via "rhc snapshot save --deployment" or crafted by hand to match the format we've specified. Maybe we should call it something like an OpenShift Deployment Artifact (get the name OpenShift in there to help distinguish it from just a binary deployment artifact, which could be a .war file). I looked at your changes and they look good to me!
Andy, thanks for the thumbs up. I don't know what I was trying to say in my last comment, but I understand you, re what people use tar.gz files for. I went with Binary Deployment File. I think if we add in the OpenShift it takes it to a formal level. I think this adequately clears it up. As for now, I think this has had enough eyes on it to sufficiently put it to QA. Thanks, all.