AEM Datastore

AEM’s datastores implement an append-only model of data persistence. While this allows for faster writes to the datastore, it also means that when files are modified or deleted that the old version is not removed from disk. To prevent a datastore from continually growing, AEM provides a datastore garbage collection feature to prune old unreferenced blobs from the datastore.

     When planning for garbage collection, it is important to understand the relationship between the node store and the data store. Nodes in the node store reference the blobs in the data store. If a node references a blob that is missing, the system will be unstable and potentially unusable. This is normally not a problem but can become one if an appropriate backup strategy has not been implemented. Should garbage collection be executed and then a backup of the node store from before garbage collection was executed needs to be restored, it is likely that there will be nodes in the node store pointing to missing blob files.


Additional Resources

1. https://cqdump.wordpress.com/2016/02/23/tarmk-on-nas/ 
2. https://docs.adobe.com/docs/en/aem/6-2/deploy/platform/data-store-config.html 
3. https://helpx.adobe.com/experience-manager/6-3/sites/deploying/using/data-store-config.html
4. https://docs.adobe.com/docs/en/aem/6-2/administer/content/assets/best-practices-for-assets/assetssizing-guide.html 
5. https://docs.adobe.com/docs/en/aem/6-2/administer/content/assets/best-practices-forassets/performance-tuning-guidelines.html 
6. https://docs.adobe.com/docs/en/aem/6-2/administer/content/assets/best-practices-for-assets/assetsoffloading-best-practices.html