This documents udev integration Debian specifics. Please see man udev(7) and its referenced manpages for general documentation. Network interface naming ~~~~~~~~~~~~~~~~~~~~~~~~ Since version 197 udev has a builtin persistent name generator which checks firmware/BIOS provided index numbers or slot names (similar to biosdevname), falls back to slot names (PCI numbers, etc., in the spirit of /dev/disks/by-path/), and then optionally falls back to MAC address, and generates names based on these properties. This provides "location oriented" names for PCI cards such as "enp0s1" for ethernet, or wlp1s0" for a WIFI card so that replacing a broken network card does not change the name (as long as the new card is fitted into the bus in the old card's slot.) As location based naming does not work well for USB devices, these use a MAC based naming schema (see /lib/systemd/network/73-usb-net-by-mac.link). This has been enabled by default since udev 220-7, which affects new installations/hardware. Existing installations/hardware which already got covered by the old 75-persistent-net-generator.rules may keep their existing interface names until the release of Debian 10 / Ubuntu 18.04 LTS; see below. You can disable these stable names and go back to the kernel-provided ones (which don't have a stable order) in one of two ways: - Put "net.ifnames=0" into the kernel command line (e. g. in /etc/default/grub's GRUB_CMDLINE_LINUX_DEFAULT, then run "update-grub"). - Disable the default *.link rules with "ln -s /dev/null /etc/systemd/network/99-default.link" "ln -s /dev/null /etc/systemd/network/73-usb-net-by-mac.link" and rebuild the initrd with "update-initramfs -u". See this page for more information: http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/ Legacy persistent network interface naming ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Debian releases up to 8 ("Jessie") and Ubuntu up to 15.04 had an udev rule /lib/udev/rules.d/75-persistent-net-generator.rules which fixed the name of a network interface that it got when its MAC address first appeared in a dynamically created /etc/udev/rules.d/70-persistent-net.rules file. This had inherent race conditions (which sometimes caused collisions and interface names like "rename1"), required having to write state into /etc (which isn't possible for read-only root), and did not work in virtualized environments. This old schema is deprecated in Debian 9 ("Stretch"), and will not be supported any more in Debian 10. Migration to the current network interface naming scheme ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Interface names must be be manually migrated to the new naming scheme before upgrading to Debian 10 / Ubuntu 18.04 LTS. If you rely on the old names in custom ifupdown stanzas, firewall scripts, or other networking configuration, these will eventually need to be updated to the new names. WARNING: This process may render your machine inaccessible through ssh. Be sure to have physical or serial console access to the machine or a way to revert to your existing configuration. First, determine all relevant network interface names: those in /etc/udev/rules.d/70-persistent-net.rules, or if that does not exist (in the case of virtual machines), in "ip link" or /sys/class/net/. Then for every interface name use a command like grep -r eth0 /etc to find out where it is being used. Then on "real hardware" machines, rename the file to 70-persistent-net.rules.old; alternately, if you have multiple interfaces, instead of renaming you may wish to comment out specific lines to convert a single interface at a time. On VMs remove the files /etc/systemd/network/99-default.link and /etc/systemd/network/50-virtio-kernel-names.link (the latter only exists on VMs that use virtio network devices). Rebuild the initrd with update-initramfs -u and reboot. Then your system should have a new network interface name (or names). Adjust configuration files as discovered with the grep above, and test your system. Repeat for each network interface name, as necessary. Custom net interface naming ~~~~~~~~~~~~~~~~~~~~~~~~~~~ In some cases it is convenient to define your own specific names for network interfaces. These can be customized in two different ways: * You can create your own names via *.link files (see systemd.link(5)) based on hardware properties. For example, /etc/systemd/network/10-dmz.link: ------------ snip ------------ [Match] MACAddress=11:22:aa:bb:cc:33 [Link] Name=eth-dmz ------------ snip ------------ * If you need attributes that link files don't expose, or you need more powerful pattern matching, you can create udev rules (see udev(7)) like /etc/udev/rules.d/76-netnames.rules: ------------ snip ------------ # identify by vendor/model ID SUBSYSTEM=="net", ACTION=="add", ENV{ID_VENDOR_ID}=="0x8086", \ ENV{ID_MODEL_ID}=="0x1502", NAME="eth-intel-gb" # USB device by path # get ID_PATH if not present yet ENV{ID_PATH}=="", IMPORT{builtin}="path_id" SUBSYSTEM=="net", ACTION=="add", ENV{ID_PATH}=="*-usb-0:3:1*", NAME="eth-blue-hub" ------------ snip ---------- The name of the rules file needs to have a prefix smaller than "80" so that it runs before /lib/udev/rules.d/80-net-setup-link.rules, and should have a prefix bigger than "75" so that it runs after 75-net-description.rules and thus you can use matches on ID_VENDOR and similar properties. * Unless you disabled net.ifnames, you can change the policy (kernel/bios/path/MAC based naming) in an /etc/systemd/network/*.link file, for individual devices or entire device classes. See man systemd.link(5) for details about this. /lib/systemd/network/99-default.link is the default policy. Note that /lib/systemd/network/73-usb-net-by-mac.link uses MAC based names for USB devices. Any of the above changes require an initrd update with "update-initramfs -u" to get effective. Using udev with LDAP or NIS ~~~~~~~~~~~~~~~~~~~~~~~~~~~ If the rules files reference usernames or groups not present in the /etc/{passwd,group} files and the system is configured to use a network-based database like LDAP or NIS then udev may fail at boot time because users and groups are looked up well before the network has been initialized. A possible solution is to configure /etc/nsswitch.conf like this: passwd: files ldap [UNAVAIL=return] group: files ldap [UNAVAIL=return] The nsswitch.conf syntax is documented in the glibc manual.