I have had to deal with a strange behaviour in one of my SLES 11 SP3 virtual machines that is hosted in an ESXi 5.5 hypervisor. After adding a new network interface, the operating system started a process to rename the ethernet interfaces. You can check the case below:

<6>[ 24.883996] udev: renamed network interface eth4 to rename6
<6>[ 24.899864] udev: renamed network interface eth5 to rename7
<6>[ 24.911977] udev: renamed network interface rename6 to eth5
<6>[ 24.935901] udev: renamed network interface rename2 to eth2
<6>[ 24.951781] udev: renamed network interface eth6 to rename8
<6>[ 24.971782] udev: renamed network interface rename7 to eth4
It seems that "udevadm settle" command receives timeout for the following interface (maybe it kept for some reason both names):

udevadm settle - timeout of 30 seconds reached, the event queue contains:
/sys/devices/pci0000:00/0000:00:18.0/0000:1b:00.0/net/eth6 (1126)
/sys/devices/pci0000:00/0000:00:18.0/0000:1b:00.0/net/rename8 (1383)

I have configured 70-persistent-net.rules by specifying the MAC address of the NIC and by removing the KERNEL filter in order to be able, to change again the interface "rename8" to eth1.

cat /etc/udev/rules.d/70-persistent-net.rules | egrep 'eth6|eth1|rename8'
SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="00:50:56:91:12:eb", ATTR{dev_id}=="0x0", ATTR{type}=="1", NAME="eth6"
SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="00:50:56:91:4b:f3", ATTR{dev_id}=="0x0", ATTR{type}=="1", NAME="eth1"
Finally, the result is that I have the correct interface configuration , but I would like to neither to wait for the "udevadm settle" timeout nor to rename back the interface back to the correct name. Can you explain me where this "event queue" information is kept and how it can be clean again? Just to clarify also that after the restart the ethernet mapped to this port is eth1, so the device existing in the filesystem is the following:

Thanks for your help!