| 6 | | |
| 7 | | Inventory AM consist of a database with a [http://mytestbed.net/projects/omf/wiki/NewInventoryDesign database schema] that can hold information for arbitrary type of resource and a set of services that are used to manage it |
| | 6 | The Inventory Aggregate Manager stores node configuration information in database with a custom [http://mytestbed.net/projects/omf/wiki/NewInventoryDesign database schema]. It can be interfaced via a rest API and is depended up by several services to answer questions like: |
| | 7 | * which disk should we boot off? |
| | 8 | * what is the name of the control interface? |
| | 9 | * which kernel and initramfs should the node boot during pxe phases? |
| | 10 | * what kind of network cards does this node have? |
| | 11 | In the Inventory AM, some of the configuration data is static and configured during the installation process (e.g. the control interface of a node). In contrast some configuration information is gathered via a live inventory collection process where the node is booted into a PRE-OS (pxelinux). While in the PRE-OS scripts on the node collection information about the nodes current configuration (e.g installed disk size, types of wireless cards). |
