How to interpret actual and rss memory in diagnostic results?
When using diagnostics to display the system information, like in the following example:
$ nova diagnostics ubuntu
+------------------+---------------+
| Property | Value |
+------------------+---------------+
| cpu0_time | 1138410000000 |
| memory | 524288 |
| memory-actual | 524288 |
| memory-rss | 591664 |
| vda_errors | -1 |
| vda_read | 334864384 |
| vda_read_req | 13851 |
| vda_write | 2985382912 |
| vda_write_req | 177180 |
| vnet4_rx | 45381339 |
| vnet4_rx_drop | 0 |
| vnet4_rx_errors | 0 |
| vnet4_rx_packets | 106426 |
| vnet4_tx | 37513574 |
| vnet4_tx_drop | 0 |
| vnet4_tx_errors | 0 |
| vnet4_tx_packets | 162200 |
+------------------+---------------+
Why memory-rss could be greater than memory-actual? Based on my understanding, memory-actual should be the allocated memory to the guest and memory-rss should be the non-swapped memory the guest is using in the host server as a process. Anything wrong with my understanding? Or someone please help me by explaining this?