ASA Failover and Redundancy

The following must be identical:

Model (5505 or 5510 or 5520, etc)
Amount of RAM
Number of Interfaces (should be the same type as well)
External Modules (CSC-SSM or IPS-SSM)
Activation key with the same features

Failover mode
Encryption Level
Number of VPN peers

*** Same size Flash is not required ****

Definitions:

Active -> Responsibel for creating the state and translation tables, transferring the data packets and monitoring the other units.

Standby -> reponsible for monitoring the status of the active unit

The active and standby units are connected through a dedicated network link and send failover-related messages to each other. This connection, known as theFAILOVER CONTROL LINK is established over a dedicated failover LAN interface.

*** When failover occurs, the standby unit takes over the IP and MAC address that were used by the previous unit

FAILOVER CONTROL LINK communicates:

Unit state (active or standby)
Network link status
Hello or keepalive messages (which are sent on all interfaces. This protocol uses IP 105)
MAC address exchange
Configuration replication from the active to standby

Conditions that trigger failover:

– Administrator has manually switched over from active to standby

-> “no failover active” on the active unit

-> “failover active” on the standby unit

– Active unit has lost power or crashed due to a hardware or software defect
– Standby unit has stopped receiving hello (or keepalive) packets on the failover interface
If 3 consecutive hello packets are missed, additional testing packets are sent to the remaining data-passing interfaces. If it still does not receive a response from the active unit, it assumes that a failure has occured and takes over
– The failover control link is down
The ASA sends additional testing packets to the remaining interfaces to determine where the peer’s control interface is also down. If the peer’s control interface is also down, then failover does not occure and the failed interface is marked as “Failed”. However, if the peer’s control interface is not down, then failover occurs because the standby unit is deemed healthier than the current active.
– The link state of the data-passing interface is down
The ASA marks the interface as “Failed” and initiates the failover process. Additionally, if the standby unit does not receive the hello packets for two consecutive polling period on an interface, the appliance goes through a series of additional tests on the interface to determine the root cause of the problem

FAILOVER INTERFACE TESTS:

Link up/down tests -> status of the NIC. For example hardware port failure, unplugged cable and a failure on the hub or switch to which the interfaces are connected
Network Activity Test -> All packets received are counted for up to 5 seconds. If any packets are received, the interface is marked as operational
ARP Test -> ARP table is read for the last 10 entries. It sends an ARP request to those machines one at a time and then counts packets for up to 5 seconds. If any packets are received, the interface is marked as operational
Broadcast Ping Test -> Sends a broadcast ping request and then counts all received packets for up to 5 seconds. If any packets are received, the interface is operational.

*** If both active and standby interfaces fail all tests, then both interfaces go into the “unknown state”. The interface with the unknown state do not count toward the monitored inteface failover limit ***

CONNECTION TABLE: Connection entries include the source and destination IP address, protocol used, current state of the connection, the interface to which it is tied and the number of bytes transferred.

In stateful failover, the active unit sends updates to the standby unit whenever these is a change in the state table. In this mode, the active unit sends stateful updates over a dedicated link to the the standby unit. When the standby unit becomes active, it does not need to build any connection entries because all the entries already exist in its database.

*** You can use the same physical interface for both failover control and stateful link. However it is not recommended if your appliances generate a lot of state updates. Additionally, it is recommended that you use the fastest interface as the stateful link and that the latency for the link should be less than 10 ms to avoid performance degradation. ****

What is not replicated in stateful failover:

Uauth cache
URL Filtering cache
TCP Intercept
SNMP Firewall MIB
Routing Table
State Info for SSM
Phone Proxy sessions
IPv6 sessions

Types of failover -> Device-level failover and interface-level failover

Device-level failover -> Active/Standby and Active/Active

Active/Standby Failover:

Active unit passes traffic
Standby unit monitors active unit
Both units send hello messages to monitor each other’s state

ELECTION PROCESS

When both are up and running, one unit assumes the role of active while the other unit assumes the standby role
If both units boot up simultenously, the primary unit takes over the active role and the secondary goes into standby. The primary unit uses the active IP address and its MAC address as the Layer 3 and Layer 2 respectively. If failover occurs, the secondary keeps using the IP address and the primary’s MAC addrss as the active
If one of the units boots up and detects an active failover unit, it goes into standby regardless of the primary or secondary designation
If one of the units boots up and doesn’t detect an active failover unit, it goes into the active state regardless of its primary or secondary designation
In case both appliances become active, the secondary changes its state to standby as soon as it discovers another active primary firewall while the primary remains active
In case both units become standby, the primary changes its state to active, while the secondary remains standby after they detect each other’s state

ACTIVE/ACTIVE FAILOVER

Both units pass traffic actively
Only supported in multimode
Needs at least two user context to work properly
With stateful failover, both stand connection tables are replicated
Uses per-failover groups instead of per-context

**** It is recommended that you don’t oversubscribe the firewalls in active/active failover as during failover, one of the units has to pass all traffic

In active/active failover, it is possible that packets can leave one active unit and can return to the other active unit. The ASA implements a feature known as “asymmetric routing” to guide packets back to the context from which they originated.

ACTIVE/ACTIVE FAILOVER AND ASYMMETRIC ROUTING

Multiple ISPs or remote locations
Load-balancing or backup
** only supported in multimode

If the asymmetric routing feature is enabled, the units restore asymmetric routed packets to the correct interface. The unit that sent the original SYN packet will replicate the connection table entry for the SYN packet to the standby unit over the stateful failover link. When the standby unit receives a packet but does not have an active connection, it checks other interfaces that are in the same asymmetric routing group for the corresponding connection. The Layer 2 information is written and forwarded to the original unit.

INTERFACE-LEVEL FAILOVER – REDUNDANT INTERFACES

During traditional device-level failover, traffic gets disrupted even when stateful failover is enabled.

Example

All incomplete TCP sessions
Routing tables (OSPF, RIP, EIGRP, etc)
Most Inspection Engines

Interface-level failover is used to avoid device-level switchover which is much more disruptive

Redundant interfaces are only supported in version 8.0 and higher.

GUIDELINES

All interfaces supported except Ma0/0 or Ma0/1
Up to 8 redundant interfaces
Network-related commands go on the redundant interface (nameif, security-level, IP address, etc)
Speed, duplex and shutdown go on the physical inteface
Link status for the physical is monitored by default
The standby interface drops all inbound packets and does not send any outbound packets
Interface stats on the redundant interface are a summation of the active and standby interfaces. As soon as a physical interface becomes a member of the redundant interface, its interface stats are cleared
Redundant interfaces use the MAC address of the first member physical interface unless you are manaully assigning a unique virtual MAC address on the redundant interface
*** The physical interfaces in the redundant interface must both be of the same physical type

*** The redundant interfaces can be configured with or without device-level failover

ACTIVE/STANDBY FAILOVER CONFIG

Select the failover link
Assign failover IP address
Set the failover key (optional)
Designate the primary unit
Enable the stateful failover (optional)
Enable failover globally
Configure failover on the secondary unit

*** There can’t be a “nameif” command on an interface when configuring failover on it. Also “managment-only” needs to be removed.

You can use the “failover exec” command to send commands to the correct unit.

Failover link is used for failover control messages

*** If a failover key is not used, the active appliance sends all information in clear text, including the TCP/UDP states, the user credentials and VPN-related information.

You must designate the primary and secondary status through software configureation

If you want to use the failover control interface as the stateful failover link, use the “failover link” command without specifying the physical interface

BOOTSTRAP CONFIG ON SECONDARY

Enable the failover control interface

“no shutdown”

Failover designation as secondary

“failover lan unit secondary”

Failover link interace
“failover lan interface failover Gi0/1”
Same failover interface IP addresses
“failover interface ip failover 1.2.3.4 255.255.255.0 standby 1.2.3.5”
Same failover shared key
“failover key cisco 123”
Failover enable (enabled globally)
“failover”

ACTIVE/ACTIVE FAILOVER CONFIG

Select the failover link
Assign failover interface IP addresses
Set the failover key (optional)
Desginate the primary unit
Enable stateful failover (optional)
Set up failover group
Assign failover-group membership
“join-failover-group”
Assign interface IP addresses
Set up asymmetric routing (optional)
Enable failover globally
Configure failover on the secondary

**** Connection and translation tables are sent over the stateful failover link. So is the asymmetric routing updates

OPTIONS FOR FAILOVER GROUPS

Interface-policy
MAC address
polltime
Preempt
Primary
Replication
Secondary

**** You must enable HTTP inspection to allow the HTTP traffic to work in the Active/Active setup with asymmetric routing

*** The Admin context is always a part of failover group 1. If you don’t explicitly map a context to a failover group, then it is automatically assigned to group 1.

*** The “http replication” command in the failover group overrides the global “http replication” command

OPTIONAL FAILOVER COMMANDS

Specify failover MAC address – in Active/Active failover, during failover, the IP addresses ad MAC addresses are swapped between units. The new active unit sends out a gratuitous ARP on the network. Using gratuitious ARP, the Layer 2 devices including bridges and switches, also update the CAM tables with the MAC address and the updated switch port information. The use of virtual MAC addresses is useful to avoid disruptions. With the virtual MAC address, the appliances do not need to swap the MAC address.

ACTIVE/STANDBY MAC ADDRESS PRIORITY LIST

Use the virtual MAC cofigured on the interfaces (physical, logical or subinterface)
Use the “failover mac address ” configured globally
Use the burned-in MAC address

ACTIVE/ACTIVE MAC ADDRESS PRIORITY LIST

Use the virtual MAC at the interface level (physical, logical or subinterface)
Use the “mac address” in the failover group configuration mode
Use the “mac address auto” in the System Execution Space
Use the burned-in MAC on the physical interface

**** In Active/Active setup, the failover “mac address” command has no effect if issues in the global configuration mode

INTERFACE POLICY

Status of all interfaces monitored
If one of the interfaces fails to repsond, failover occurs

If you want to change this behavior, use the “criteria” tab in the ASDM

*** This setting is applied to the failover group in Active/Active

FAILOVER TIMERS

By default, ASA sends periodic keepalives (hello) packets to check the status of the peer failover unit. If the standby unit does not receive acknowledgement packets for the keepalives sent out, it initiates a failover only if it deems itself healthier than the current unit.

Two types of failover hello messages -> Unit (sent every second)

-> Interface (every 15 seconds)

In active/active failover, unit failover timers are changed in the System Execution Space. The interface failover timers are changed in the failover group policy.

*** When an ASA is configured for failover, whether active/standby or active/active, it monitors the status of all the main physical interfaces that have a “nameif” and an IP address configured.

*** Interface monitoring is disabled by default on all subinterfaces until you explicitly enable it.

*** Stateful failover is mandatory for a zero-downtime software upgrade.