As a first step towards migration to a lshw based inventory gathering system [1], I have first collected the data that we are currently gathering. Here they are with the occasional note:
CPU ===
- Arch - Vendor - Model - Model name - Speed - Processors (number of CPUs) - Cores - Sockets - CPU Flags - Family - Stepping
Notes: same fields for all archs are recorded. Some of them are not relevant to non-x86 architectures and are hard coded to 0. On the other hand, these archs have some fields of potential interest which are hence not being recorded.
Devices =======
- Vendor ID - Device ID - subsys Vendor ID - subsys Device ID - bus - driver - type - description
Misc others ===========
- Running under a hypervisor or not - Memory - Number of NUMA nodes - System vendor - System model - System form factor
Disk ====
- Disks: Model, sector size, geometry, etc.
Key value data ==============
- Uses the data above to create a collection of key value fields - Creates some additional fields (such as PCIID and USBID for devices)
Data collected but not stored =============================
- Some of the data collected are not stored currently: kernel modules, network interfaces, using multipath or not, disk controller with the disk
If there is anything I may have missed, please add to it. If there is anything which we should gather, please do the same.
[1] http://beaker-project.org/dev/proposals/inventory-lshw-migration.html
Best, Amit.
beaker-devel@lists.fedorahosted.org