Bonding on Blade servers does not fail over when a link to the core switch is unplugged
This document (7006507) is provided subject to the disclaimer at the end of this document.
Environment
Situation
The problem is the that the OS running on the Blade server does not get notified that the external link is down.
This is because the OS running on the Blade server is checking the link status of it's NIC which is connected to a switch inside the blade enclosure, and that link is not down.
Resolution
Example 1: mode=active-backup arp_ip_target=x.x.x.x
In this example the Bonding method is active-backup. In this mode only one interface will be active at a time.
The arp_target is set to an external device that you expect will always be there. There can be more than one arp target. If you have more than one target and any arp target responds then the link is seen by Bonding as up.
Example 2: mode=active-backup arp_interval=100 arp_ip_target=x.x.x.x arp_validate=3
In this example the Bonding method is also active-backup.
The arp_target is set to an external device that you expect will always be there.
The arp_interval is set to 100 which is the interval for checking to see if slaves have received ARP traffic.
The default = 0, which means disabled.
The arp_validate is set to 3 which means validate all slaves.
The default value is 0 which means don't validate any slaves.
Disclaimer
This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.
- Document ID:7006507
- Creation Date: 26-Jul-2010
- Modified Date:03-Mar-2020
-
- SUSE Linux Enterprise Server
For questions or concerns with the SUSE Knowledgebase please contact: tidfeedback[at]suse.com