The objectizer process
The objectization process converts files ingested from the file interface on unified file and access enabled container path to be available from the object interface. The name of that service is ibmobjectizer.
When new files are added from the file interface, they need to be visible to the Swift database to show correct container listing and container or account statistics.
The ibmobjectizer service provides synchronization between the file metadata and the object metadata at a predefined time interval (that assists with accurate container and account listing). The ibmobjectizer service identifies new files added from the file interface and adds the Swift system metadata to them so that they are objectized. The ibmobjectizer service then determines its container and account databases and adds an object entry to them. It also identifies files deleted from file interface and deletes their corresponding entries from container and account databases.
These functions are useful in the setups where data is ingested using legacy file interface-based devices such as medical and scientific devices. They are helpful when data needs to be stored and accessed over cloud using the object interface.
The ibmobjectizer service is a singleton and it is started when object is enabled and the file-access object capability is set. However, the ibmobjectizer service starts objectization only when there are containers with unified file and object access storage policies configured and the file-access object capability is set.
There are use cases in which objects are ingested using the object interface and the file interface is used only for reading them. For these use cases, the ibmobjectizer service is not needed. It can be disabled using the mmobj file-access command. For more information, see Starting and stopping the ibmobjectizer service.
mmces service list --verbose
gpfsnode3
as the
object singleton
node:mmces service list --verbose -a | grep ibmobjectizer
gpfsnode3: OBJ:ibmobjectizer is running
For information on limitations on the objectizer process, see Limitations of unified file and object access.