FAQ / When use S30interface script to configure an additional VIP, does RSF-1 monitor this network interface? And if this interface fails, does this trigger a fail over?

RSF-1 monitors the physical interface state (if configured to do
so) and as a by product any VIP’s bound to that interface are therefore
also monitored. When interface monitoring is configured, any services
that use that physical interface are considered dependants, and as such
should the interface fail, then those services are marked as potential
for fail over (dependant on other factors such as the secondary node
being in a position to accept the service – a dependency which also
includes the monitoring the health of it’s own network physical interface that
the service will use).

The S30interface script is used to configure additional VIP’s for a
service (on the physical interface declared in the configuration file
for that service) so yes if an
interface fails, and RSF-1 is configured to monitor network interfaces
(and the S30interface scripts places additional interfaces on the *same*
interface configured for the service) then if the interface fails RSF-1
will fail over the service and as such any additional VIP’s bound on
that interface (note! the corresponding S30interface script should exist on
the node being failed over to to ensure the additional VIP is
configured for the service).

If the S30interface scripts places
additional VIP’s on interfaces that RSF-1 has no knowledge of (for
example the service uses hme0, but S30interface plumbs in an additional
interface, bge0, then there is no way of RSF-1 knowing about (and
therefore monitoring) this additional  interface.

Posted in: Networking