Am 02.10.19 um 18:01 schrieb Alexandra N. Kossovsky: > Package: udev > Version: 242-7 > Severity: normal > > When I create a bonding interface on an older Debian system, it > naturally inherits one of the slave's mac address. With new Debian > testing (systemd=242-7) the bonding interface gets a strange MAC > address. > > The issue is, if I connect 2 Debian systems with a bonding interface, > then the 2 ends of the link has the same MAC address, and it all breaks. > > I tried to downgrade systemd to Debian buster version 241-7~deb10u1, and > it did not help (systemd, libsystemd0, libpam-systemd, systemd-coredump > packages). Then I downgraded udev as well (udev, libudev1, libudev-dev, > libudev1:i386, libudev-dev:i386 packages). Bonding interfaces got > different MAC addresses and started to work properly. > > > I.e. it is a regression in udev. which assigns some strange MAC address > to a just-created bonding interface; with a high chance that the > addresses of 2 link ends are the same. > > > To reproduce: > sudo ip link add bond0 type bond mode 1 > cat /sys/devices/virtual/net/bond0/addr_assign_type > ip li li dev bond0 > > I expect to see NET_ADDR_RANDOM=1 as the addr_assign_type, and different > MAC addresses on different systems as a result of the "ip" command. > With the last udev, I see NET_ADDR_SET=3 as the addr_assign_type, and > the same MAC address on both ends of my bond link.
Thanks for your bug report. Could you file this upstream at https://github.com/systemd/systemd/issues and report back with the bug number? -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
signature.asc
Description: OpenPGP digital signature
