AEM TO S3 Blocker Case ->
Environment – AEM As a cloud service
Dependency not resolved –
We are using the AEM as a cloud service environment and we are finding issues in using the AWS SDK libraries using as a maven dependencies. We found the way to update the couple of configurations in AEM quickstart folder which is not possible in AEM as cloud service straight forward as we don't have access to unix box to update the same. Developer console also give access for Felix configs.
Failed as we are looking for system configs to get added and as per suggestion from Adobe Added the same configs in Runmode but this doesn’t support the technical usecase.
Failed as AEM as a cloud service won’t allow to install any package under /libs as it is only having read access and needs to be build this package through bundle.
ZIP building for project codebase and deploy as a package in direct instance as Approach 2 failed.
Failed as AEM as a cloud service won’t allow to install any package under /apps as it is only having read access and needs to be build this package through bundle.
ZIP building from project codebase as Approach 2 and 3 failed.
This will allow to add S3 connector zip from project bundle as per given reference. But that way of implementation also didn’t work as it is giving same error as,
14:00:46,076 [main] [ERROR] Error resolving dependencies for project com.pacificlifeinsurancecompanydam:aem-pacificlifeinsurancecompanydam-project.all
14:00:46,077 [main] [ERROR] Failed to collect dependencies at com.adobe.granite:com.adobe.granite.oak.s3connector:zip:1.10.10
Failed as a given reference of updating embedded packages for Vendor packagesof s3connector, adding s3 dependencies and <allowIndexDefinitions>true</allowIndexDefinitions> still not worked.
Creating the non-osgi aws-java-sdk-osgi bundle to osgi compatible manner. Added the dependencies also still didn't work.
Failed as loads of dependencies not resolved with the osgi package in felix.
@kiranparab - Trying to understand your use case, are you trying to create a S3 nodestore or want to programmatically connect with S3 to upload some files?
In former case, since the entire infrastructure is managed by Adobe and there's already a repository service holding all the repository data, I don't think it would be possible to have a separate S3 datastore. If it's latter, did you try adding the S3 dependencies in pom.xml?