NIC MAC Address change after upgrade to PVE 7.x

[GabrieleV]

I've followed the guidelines on the wiki, installing ifupdown2 prior to upgrade my hosts from 6.x to 7.x and prevent MAC address change of the bridge and bond interfaces.

The MAC addresses of bridge and bond interfaces remained the same.
But the MAC addressess of ALL the PHYSICAL interfaces has been changed.

Have a look.

Spoiler: Prior to upgrade

1: lo: mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: ens18: mtu 1500 qdisc pfifo_fast master bond0 state UP mode DEFAULT group default qlen 1000 link/ether ca:4e:91:93:96:25 brd ff:ff:ff:ff:ff:ff
3: ens19: mtu 1500 qdisc pfifo_fast master bond0 state UP mode DEFAULT group default qlen 1000 link/ether ca:4e:91:93:96:25 brd ff:ff:ff:ff:ff:ff
4: ens20: mtu 1500 qdisc pfifo_fast master bond1 state UP mode DEFAULT group default qlen 1000 link/ether c6:12:67:f1:20:0f brd ff:ff:ff:ff:ff:ff
5: ens21: mtu 1500 qdisc pfifo_fast master bond1 state UP mode DEFAULT group default qlen 1000 link/ether c6:12:67:f1:20:0f brd ff:ff:ff:ff:ff:ff
6: ens22: mtu 1500 qdisc pfifo_fast master bond2 state UP mode DEFAULT group default qlen 1000 link/ether e2:3c:23:a7:25:af brd ff:ff:ff:ff:ff:ff
7: ens23: mtu 1500 qdisc pfifo_fast master bond2 state UP mode DEFAULT group default qlen 1000 link/ether e2:3c:23:a7:25:af brd ff:ff:ff:ff:ff:ff
8: bond1: mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000 link/ether c6:12:67:f1:20:0f brd ff:ff:ff:ff:ff:ff
9: bond2: mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000 link/ether e2:3c:23:a7:25:af brd ff:ff:ff:ff:ff:ff
10: bond0: mtu 1500 qdisc noqueue master vmbr0 state UP mode DEFAULT group default qlen 1000 link/ether ca:4e:91:93:96:25 brd ff:ff:ff:ff:ff:ff
11: vmbr0: mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000 link/ether ca:4e:91:93:96:25 brd ff:ff:ff:ff:ff:ff

Spoiler: After upgrade:

1: lo: mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: ens18: mtu 1500 qdisc pfifo_fast master bond0 state UP mode DEFAULT group default qlen 1000 link/ether f6:05:89:19:ad:4c brd ff:ff:ff:ff:ff:ff permaddr ca:4e:91:93:96:25 altname enp0s18
3: ens19: mtu 1500 qdisc pfifo_fast master bond0 state UP mode DEFAULT group default qlen 1000 link/ether f6:05:89:19:ad:4c brd ff:ff:ff:ff:ff:ff permaddr 2a:e6:b4:53:13:d5 altname enp0s19
4: ens20: mtu 1500 qdisc pfifo_fast master bond1 state UP mode DEFAULT group default qlen 1000 link/ether 12:99:dd:c4:a1:29 brd ff:ff:ff:ff:ff:ff permaddr c6:12:67:f1:20:0f altname enp0s20
5: ens21: mtu 1500 qdisc pfifo_fast master bond1 state UP mode DEFAULT group default qlen 1000 link/ether 12:99:dd:c4:a1:29 brd ff:ff:ff:ff:ff:ff permaddr 5e:72:58:d8:e3:99 altname enp0s21
6: ens22: mtu 1500 qdisc pfifo_fast master bond2 state UP mode DEFAULT group default qlen 1000 link/ether 46:49:50:7a:28:20 brd ff:ff:ff:ff:ff:ff permaddr e2:3c:23:a7:25:af altname enp0s22
7: ens23: mtu 1500 qdisc pfifo_fast master bond2 state UP mode DEFAULT group default qlen 1000 link/ether 46:49:50:7a:28:20 brd ff:ff:ff:ff:ff:ff permaddr ae:77:28:cc:a2:da altname enp0s23
8: bond1: mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000 link/ether 12:99:dd:c4:a1:29 brd ff:ff:ff:ff:ff:ff
9: bond2: mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000 link/ether 46:49:50:7a:28:20 brd ff:ff:ff:ff:ff:ff
10: bond0: mtu 1500 qdisc noqueue master vmbr0 state UP mode DEFAULT group default qlen 1000 link/ether f6:05:89:19:ad:4c brd ff:ff:ff:ff:ff:ff
11: vmbr0: mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000 link/ether f6:05:89:19:ad:4c brd ff:ff:ff:ff:ff:ff

It seems that a random mac has been generated for the physical interfaces, overwriting the permaddr.

No changes has occured to

Spoiler: /etc/network/interfaces

source /etc/network/interfaces.d/*

auto lo
iface lo inet loopback

auto ens18
iface ens18 inet manual

auto ens19
iface ens19 inet manual

auto ens20
iface ens20 inet manual

auto ens21
iface ens21 inet manual

auto ens22
iface ens22 inet manual

auto ens23
iface ens23 inet manual

auto bond0
iface bond0 inet manual
bond-slaves ens18 ens19
bond-miimon 100
bond-mode balance-rr
bond_downdelay 200
bond_updelay 200
xmit_hash_policy layer3+4

auto bond1
iface bond1 inet static
address 10.20.30.133/24
bond-slaves ens20 ens21
bond-miimon 100
bond-mode balance-rr
network 10.20.30.0
bond_downdelay 200
bond_updelay 200

auto bond2
iface bond2 inet static
address 10.20.40.133/24
bond-slaves ens22 ens23
bond-miimon 100
bond-mode balance-rr
network 10.20.40.0
bond_downdelay 200
bond_updelay 200

auto vmbr0
iface vmbr0 inet static
address 192.168.10.133/24
gateway 192.168.10.254
bridge-ports bond0
bridge-stp off
bridge-fd 0
network 192.168.10.0
bridge_ageing 0
bridge_maxwait 0

There is nothing in

Spoiler: ls /etc/network/interfaces.d/

total 8
drwxr-xr-x 2 root root 4096 Jul 7 2021 .
drwxr-xr-x 8 root root 4096 Aug 5 20:26 ..

How can i get the REAL MAC addressess back to physical interfaces ?

I know that I've not assigned an IP address to these nterfaces, but I don't want to have the RANDOM MAC addresses to be generated.

 

[shrdlicka]

Hi,

to me it looks like the old MAC addresses where also generated. Is the ip l output from from a bare metal host or from inside a VM?

 

[GabrieleV]

to me it looks like the old MAC addresses where also generated.

No, they are the same as the virtual NIC mac address (see next)

Is the ip l output from from a bare metal host or from inside a VM?

Yes, they are VM, It's a lab with nested virtualization to test PVE upgrades.

 

[shrdlicka]

Which VM did you upgrade? The nested VM or the host VM?

 

[GabrieleV]

Which VM did you upgrade? The nested VM or the host VM?

The problem is in the nested PVE host.