Re: [rhelv5-list] eth device naming
Besides hardcoding the MAC address in ifcfg-ethX config, I'd say look into creating custom udev rules for your interfaces. -Original Message- From: rhelv5-list-boun...@redhat.com [mailto:rhelv5-list-boun...@redhat.com] On Behalf Of Marco Shaw Sent: Thursday, June 28, 2012 9:56 AM To: Red Hat Enterprise Linux 5 (Tikanga) discussion mailing-list Subject: [rhelv5-list] eth device naming I'm having problems getting the eth devices numbered as I'd like them so 2 nodes of a cluster will match. One server had a 2nd NIC added after the OS was installed, and that seems to have wreaked havoc. This seems to have a bunch of different suggestions: http://www.science.uva.nl/research/air/wiki/LogicalInterfaceNames What concerns me is that I'm going to have 2 interfaces that will be part of a bond, so I don't know if I can work reliably using the MAC address. I noticed modprobe.conf was wrong, so fixed that, but I think I'm still seeing the numbering being different between host A and B. Any ideas/comments? Can I rely on the MAC address to hardcode the device number when I'm using bonding? Marco ___ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list ___ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list
Re: [rhelv5-list] eth device naming
I'm still playing around with things. I'm seeing some weird stuff like multiple identical MAC addresses on a system. I'll figure it out... If you bond two devices they will both use the same. This email communication and any files transmitted with it may contain confidential and or proprietary information and is provided for the use of the intended recipient only. Any review, retransmission or dissemination of this information by anyone other than the intended recipient is prohibited. If you receive this email in error, please contact the sender and delete this communication and any copies immediately. Thank you. http://www.encana.com ___ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list
Re: [rhelv5-list] eth device naming
On Thu, Jun 28, 2012 at 3:44 PM, Kinzel, David david.kin...@encana.com wrote: I'm still playing around with things. I'm seeing some weird stuff like multiple identical MAC addresses on a system. I'll figure it out... If you bond two devices they will both use the same. So the ifcfg-ethX file may have one MAC, but ifconfig might report another? Marco ___ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list
Re: [rhelv5-list] eth device naming
Jun 28, 2012 at 3:44 PM, Kinzel, David david.kin...@encana.com wrote: I'm still playing around with things. I'm seeing some weird stuff like multiple identical MAC addresses on a system. I'll figure it out... If you bond two devices they will both use the same. So the ifcfg-ethX file may have one MAC, but ifconfig might report another? Marco This is correct. Both eth* devices in the bond pair will share the same MAC in ifconfig output. The Ethernet device which becomes the first setup in the bond will be used for both (so likely eth0's MAC if you have sane naming). This email communication and any files transmitted with it may contain confidential and or proprietary information and is provided for the use of the intended recipient only. Any review, retransmission or dissemination of this information by anyone other than the intended recipient is prohibited. If you receive this email in error, please contact the sender and delete this communication and any copies immediately. Thank you. http://www.encana.com ___ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list
Re: [rhelv5-list] eth device naming
On 2012-06-28, at 3:30 PM, Kinzel, David wrote: Jun 28, 2012 at 3:44 PM, Kinzel, David david.kin...@encana.com wrote: I'm still playing around with things. I'm seeing some weird stuff like multiple identical MAC addresses on a system. I'll figure it out... If you bond two devices they will both use the same. So the ifcfg-ethX file may have one MAC, but ifconfig might report another? Marco This is correct. Both eth* devices in the bond pair will share the same MAC in ifconfig output. The Ethernet device which becomes the first setup in the bond will be used for both (so likely eth0's MAC if you have sane naming). This behaviour can be disabled in the bonding module with the fail_over_mac option. However, it usually makes sense to let the system duplicate the HW address of the first slave too all the other slave interfaces, in an active-backup configuration at least. This way, when the second link becomes primary, there is no traffic interruption or delay caused by ARP caches needing to be updated or cleared, or gratuitous ARP needing to be sent to force an update.. Said otherwise, the other machines in the same network will not know that a failover has occurred, since the MAC address that they use to send traffic to your machine has not changed. As others have said, it would be highly advisable to provide the HWADDR variable in your /etc/sysconfig/network-scripts/ifcfg-eth* files, to ensure that the naming of your interfaces is consistent. /sbin/modinfo bonding can be used to see the various options to the bonding module. You definitely want to adjust the mode parameter, and may also want to adjust additional parameters such as updelay, primary and primary_reselect, for example. If you wish to see the real HW addresses of your interfaces that are already enslaved into a bond master, poke around /proc/net/bonding/bond*, look for lines that say Permanent HW addr. HTH, Antoine Antoine Reid Conseiller Principal Principal Consultant Infrastructures Technologiques Gallium 1250, Boul. René-Lévesque Ouest Bureau 2200 Montréal (Québec) Canada H3B 4W8 M : +1 514.515.3314 F : +1 514.394.0420 E : ar...@gallium-it.com www.Gallium-IT.com ___ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list