I would like to know the pros & cons of having S3 as data store for AEM on Production instances where there are large binary assets that needs to be stored & used by a single Author & single publish instance topology.
If an existing AEM needs to be configured for S3 , is that just the configuration change & runmode change as on Configuring node stores and data stores in AEM 6. OR does it have to be done on a plain vanilla instance & then the repository migrated to the new instance ?
I understand that the back up and GC are not quite straight forward. Do we have any document for the best practices primarily towards maintenance, online & offline compaction, diagnosis, indexing etc.
Of course you can migrate a running instance from a disk-based datastore to an S3 Datastore. Next to reconfiguration you also have to copy all the local binaries to S3 first before you start AEM with the new configuration.
For details see the parameters for crx2oak (Using the CRX2Oak Migration Tool ), especially the --src-datastore and the --s3config parameters.
Regarding GC on S3: That's straight forward and nothing special with it. In your case you can use the documentation for GC on shared datastore: Configuring node stores and data stores in AEM 6
Of course you can migrate a running instance from a disk-based datastore to an S3 Datastore. Next to reconfiguration you also have to copy all the local binaries to S3 first before you start AEM with the new configuration.
For details see the parameters for crx2oak (Using the CRX2Oak Migration Tool ), especially the --src-datastore and the --s3config parameters.
Regarding GC on S3: That's straight forward and nothing special with it. In your case you can use the documentation for GC on shared datastore: Configuring node stores and data stores in AEM 6
