Aggregator failover settings
=============================

The `Aggregator failover <../aggregators/aggregator-failover.html>`__ service
has a number of settings that controls how often aggregators are checked
and how quickly bonds fail to their secondary aggregators when a primary
aggregator fails.

In most cases the defaults are sufficient.
The defaults can be changed from this administration page, however,
and aggregators can be configured to use values different from the defaults configured here.

.. NOTE::
    Known hosts and Suspension are not configurable on a per-aggregator basis.
    These features are configured and enabled here for all aggregators
    in the administration page.

For more information on Aggregator failover, please review the
`documentation <../aggregators/aggregator-failover.html>`__.


Global settings
^^^^^^^^^^^^^^^

Maintenance mode
-----------------

When enabled, aggregator failover is temporarily disabled and bonds will
not be moved to their secondary aggregator if a failure is detected.


Known host enabled
-------------------

When checked, before moving bonds to a secondary aggregator, the
failover process will ping a list of hosts in the known hosts field to
verify if the management server is online or not. If any of the hosts do
not respond, the management server will assume its network is
interrupted and will not move bonds.


Known hosts
------------

A list of IP addresses of reliable hosts that the management server
should ping before moving bonds during a possible aggregator outage. Either IPv4
or IPv6 addresses can be entered. Hostnames will be sent pings over IPv4.


Allow suspension mode
---------------------

When checked, aggregator failover is temporarily disabled if the specified
percentage of aggregators is detected as failed.

Enabling aggregator failover suspension mode can help avoid false positives in
scenarios where a brief networking outage causes a large number of aggregators
to simultaneously become unreachable from the management server, even though
they are still online and passing traffic.


Suspension percentage
---------------------

The percentage of aggregators needed to be detected as failed before failover
is suspended. Once the percentage of failed aggregators drops below this value,
regular aggregator failover will resume.


Aggregator defaults
^^^^^^^^^^^^^^^^^^^

Default Interval
----------------

The number of seconds between each check sent from the management server
to the aggregator. *Recommended value is 30 seconds.*


Default Receive timeout
-----------------------

The number of seconds to wait to receive a response from the aggregator.
*Recommended value is 5 seconds.*


Default Failure threshold
-------------------------

How many sequential checks must fail before an aggregator is considered
to have failed. *Recommended value is 3.*


Default Recovery checks
-----------------------

How many sequential checks must succeed before an aggregator is
considered to have recovered. *Recommended value is 2.*


Default Max flap checks
-----------------------

The maximum number of checks that the flap damping technique will
require the aggregator to respond to. *Recommended value is 20.*
