Short disruption in network traffic when failing back to primary bonding slave in RHEL 5
Issue
- When using active-backup bonding with the
primaryoption when failing back to the primary interface we are seeeing greater network timeouts causing Oracle RAC to evict the node. Failing over to the non primary does not cause an issue, just when the primary interface is recovered. - If we physically unplugged the active member of the active-passive pair used for the interconnect, when we plugged it back in, we got an outage of approximately 30 seconds whilst the bonding driver made that interface the active one again
Environment
- Red Hat Enterprise Linux (RHEL) 5
- bonding, with a primary interface specified
- For RHEL 5.4 and later,
primary_reselectis set toalways(the default; prior to 5.4, this was not configurable)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase of over 48,000 articles and solutions.
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.
