Status is set to aligned even no PCI device/VF was attached to pod

[smalleni]

Even in a pod that does not have any VF attached numalign, the pod is being considered to be aligned from a NUMA perspective. I guess this is probably not very intuitive. For example,

kni@e22-h20-b01-fc640 sriov-operator]$ oc rsh podvert-pod-3
/ # ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
3: eth0@if94: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 1400 qdisc noqueue state UP 
    link/ether aa:aa:7c:82:02:0d brd ff:ff:ff:ff:ff:ff
    inet 10.130.2.12/23 brd 10.130.3.255 scope global eth0
       valid_lft forever preferred_lft forever
    inet6 fe80::90cc:14ff:fe9e:b35c/64 scope link 
       valid_lft forever preferred_lft forever
/ # numalign
STATUS ALIGNED=true
NUMA NODE=0

This was in a pod with isolated CPUs but no VF attached.

[ffromani]

Hi! From resource perspective, all the resources numalign can find, in this case CPUs, are on the same NUMA node.
Hence it considers them aligned. In general is possible for a pod to care only about CPUs.
I understand your point here but unfortunately I don't see any actionable item considering that "only CPUs, no NICs" is a valid resource set requirements we want to check for alignment.

[ffromani]

Additionally: setting the env var NUMALIGN_DEBUG=1 can give you insights about why the tool did what it did. Example from my laptop:

2020/09/10 09:20:18 CPU: allowed for "self": [0 1 2 3]
2020/09/10 09:20:18 CPU: NUMA node by id: map[0:0 1:0 2:0 3:0]
2020/09/10 09:20:18 PCI: devices: none found - SKIP
STATUS ALIGNED=true
NUMA NODE=0

the downside is debug logs and actual output get squashed on the same stream (oc logs...)