bonding, mode=1, temporary ping failure when failing over back to the configured "primary" slave
Issue
When testing failover on an active/backup bond (mode=1), with one slave configured as "primary", and running a continuous ping test to the host, from a different host on the network, it is observed that ping probes will fail for a few seconds.
This will only happen when failing over from the "secondary" interface to the configured primary, which is done automatically by bonding when the configured primary comes back online (without the current active slave, the secondary interface, going down). When bringing down (unplugging) the primary slave while it is the active slave, the failover will work nicely and no probes (or at most one probe) will fail.
short: ping failure on failover, ping loss on failover, packet loss on failover
Environment
- RHEL, any version
- bonding, active/backup (mode=1)
- primary slave explicitly configured
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.