New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Alfresco content model packaging #931
Comments
Ik ben al vrij ver hiermee, maar het leek me netjes om een story aan te maken voor de wijziging. |
Dankje voor het werk tot nu toe, zou mooi zijn als we dit in 1.4.0 meekrijgen (die staat gepland voor release over ongeveer een maand). Ik wacht je nieuwe PR af ;) |
I took a quick look at the PR nevertheless, and have some questions and statements Open Zaak does not claim to provide/demonstrate how you should deploy Alfresco (for production). I believe the content model is included for two reasons:
Now, we also don't want to imply that what Open Zaak does is the correct or best way, as evidenced by this user story and the PR. However, I'd say that detailed instructions and tooling on how to correctly deploy Alfresco with the custom content model are not in the scope of the Open Zaak application/repository. Is there a way that we can do the following:
@joeribekker @TjerkVaags @shazada @aboerma @ericherman - do you have any opinions/input on this? |
Creating a separate repository is a possibility. The reason why I created a SDK 4 alfresco project and placed the contentmodel in it, was to make it easy to use for testing. It's now possible to create a small docker container stack for testing with just one command. If we only include the model and a context file, it will require more effort to run a quick test. I can remove the information on how to deploy it in a production environment and the docker components for creating alfresco enterprise docker images with the model jar included. This would decrease the number of changes in the repository. |
Yes, the intent was clear to me, and I appreciate the contribution!
I think you mean manual testing rather than automated testing in this case?
I'm pretty sure you have the better picture of how things should be - I believe only Tahir & Tjerk in this project have deep knowledge of Alfresco. So, I'm totally okay with following your advice/suggestions. If we can make a set up without bloat that's easy to use for both manual testing and automated testing, that would be perfect. I think the following scenario's is what we should aim for: Trying out Open Zaak without extensions
Trying out Open Zaak with CMIS adapter and Alfresco
Running tests on CI with Alfresco This is covered by https://github.com/open-zaak/open-zaak/blob/main/.github/workflows/ci.yml#L64, so in essence:
This would of course require a couple tweaks on the non-Alfresco side of things as well. And perhaps it would be ideal if edit: imo it's fine to pin to 6.2 for example inside this repo, and periodically update that version number. That would probably allow us to drop some build tooling/scripts for which the main purpose appeared to be templating out the |
We at Exxellence deploy our Alfresco models via amp files (which zip or jar files). We place these in the alfresco/amps directory after which Alfresco copies the amps content to the relevant folders/files. This setup should translate well to Docker I think. |
@aboerma Deploying as jar is similar. It is also easier to use when deploying on a manually installed environment directly on windows. In that case it is only required to place the jar in a modules/platform folder. With amps you would have to run installation scripts. Both ways are correct for deploying alfresco extensions. In the end it is just about what is prefered. In the sdk4 project it is possible to change the jar packaging to amp packaging by changing one line in the pom.xml. @sergei-maertens next Tuesday I will have time to see if I can make a few modifications to satisfy your requests |
@timijntema this is because of the URL validation requiring a domain + TLD. If you define the service as It's an edge case and not something likely to change, as we require URLs to actually look like URLs :-) (principle of least surprise) |
I've spoken to @TjerkVaags and our opinion is to have a separate project within the Open-Zaak group just containing the Alfresco SDK4 project. From that project we could create the artifcacts and even an image and publish that one to docker-hub as well which then can be included in the compose file here. |
I assume that means someone who has access should create that repository for me to move the Alfresco model packaging to that repository. @shazada or @sergei-maertens is this something one of you can do? A few additional questions:
|
hi @timijntema - we just discussed this issue in the Technical Steering Group Meeting, and here's the summary + answers to the questions:
Ideal situation would be if we have a Docker image ready for use in this (open-zaak/open-zaak) repository for the CI purposes. We test using docker-compose. The new repository can be configured to run github actions, so we can automate publishing the images to Docker Hub and/or Github Image registry. We will make it clear from the naming on those repositories that it's only meant for testing (manually locally or automated in CI).
No - Alfresco production images are really out of scope for the Open Zaak project. Open Zaak implements the "API's voor Zaakgericht werken" standard, and CMIS is an optional feature as Documenten API storage backend. Open Zaak works perfectly fine without CMIS too. Anything we do with Alfresco in this repository only serves to validate that the feature works as intended, and we basically assume that Alfresco (or another CMIS provider) has been installed and configured by a more suitable party. In short: the Open Zaak project/community does not want to maintain or oversee production Alfresco things. |
I have moved everything to the other repository. Would @aboerma mind giving it a review? The repository only requires two secrets to be set up in the openzaak repository for it to push test docker images to the openzaak repository. These are:
Since I don't have access, somebody else should set that up. If the below PR is merged, I will modify the pipeline in this repository to use the new images. Until then, thery are not available |
Namens: OPEN.satisfaction
Contact persoon: Silvion Moesan
Op het moment word het content model als XML in een folder in Alfresco gezet. Dit is niet de juiste manier om een Alfresco extensie te packagen. Daarbij mist er Share configuratie die controle geeft over welke velden wel en niet bewerkbaar zijn in Alfresco Share.
Development
The text was updated successfully, but these errors were encountered: