Guidelines for Configuring Stateful Line Module Switchover

Keep the following points in mind when you configure stateful switchover for line modules:

• Line module high availability, similar to dual SRP stateful switchover, does not to prevent the root cause of the reload or restart. This functionality is designed to return the system to the online, active state as soon as possible. The secondary line module takes the role of the primary module to preserve subscriber sessions in the up state with minimal subscriber data outage.
• The architecture of line modules supports switchover simultaneously across multiple 1:1 sets of line modules that participate in line module high availability.
• Line module high availability is available only on the operational image that runs on the interface controller (IC). The behavior of the forwarding controller (FC) image, and the IC boot and diagnostic images provide the same functionality as the behavior that existed before line module high availability support was implemented.
• Line module high availability in a 1:1 redundancy model is supported for ES2 4G LMs and Service IOAs on E120 and E320 routers. The architecture of line module high availability ensures that it does not depend on high availability for SRP modules to be enabled and operational for stateful line module switchover to work. Similarly, any modifications made to the dual SRP stateful switchover settings do not require or depend on line module high availability to be enabled and operational.
• Unified ISSU is supported on the primary line module in a high availability pair of line modules. The secondary line module is disabled during the unified ISSU operation and cold boots after the unified ISSU operation is complete. Line module high availability mode is active after the secondary line module is up, provided that line module HA configuration is enabled.
• Applications that are configured on the router ensure that their defined settings and memory requirements are handled on E120 and E320 routers. The primary and secondary line modules in a high availability pair are determined using the slot information specified using the mode high-availability slot command in Redundancy Configuration mode.
• Packets that are transmitted between the FC and IC and between the FC and system controller (SC) are not preserved during a stateful line module recovery.
• 1:N hot standby mode is not supported for stateful line module switchover. Automatic switchover of the serial connection to the line module that is designated as the primary module after switchover is also not supported. Similarly, cold switchover is not supported if line module high availability is not configured.
• Recovery of routers from double failures, such as simultaneous switchover of SRP and line modules, is not supported. Application-specific statistical details are not retained across a stateful line module switchover.
• Subscriber sessions that constantly move between up and down states are not maintained across a stateful switchover.
• If the line module that contains a downlink interface (connecting to the LAC device) reloads, owing to hardware or software failures, subscriber sessions are not maintained, even if the LM and Service IOA are HA-protected. Also, subscriber sessions are not retained if the line module that connects to the LAC device reloads, when the LM and Service IOA are part of a redundancy group.
• ES2 10G LMs cannot be used as downlink modules in an LNS device. These LMs cannot be used as access modules in a LNS device that contains a Service IOA that is HA-enabled.
• Certain statistics might be lost during the period of the stateful line module switchover. PPP and policy statistics are polled and collected every 10 minutes and sent to the standby line module. The statistics that were last collected before the switchover occurred are used as the baseline for statistics on the newly configured primary module. At a maximum, statistics for around 10 minutes might be lost. This scenario normally happens when polling is about to happen and the primary module switched over.
• A historical record of information about the forwarding and drop events and forwarding and drop rates on egress queues is not retained across a stateful line module switchover. The queue statistics for subscriber interfaces are calculated afresh after a stateful switchover of line modules.
• Sequence number checking for data packets received on all L2TP tunnels in the router is not maintained and supported during a stateful line module switchover. We recommend that you set up the router to ignore sequence numbers in data packets received on L2TP tunnels by entering the l2tp ignore-receive-data-sequencing command on an LNS device to prevent requests from a LAC device to enable insertion of sequence numbers into data packets.
• Some performance impact might occur when a new secondary module is provisioned or inserted, with the primary module containing maximum tunneled PPP sessions. In this case, data synchronization consumes a portion of the backplane bandwidth, which might have some impact on call setup rate (CSR) during this time. Under peak load conditions, it might take about 20 minutes for the system to become HA-active for Service IOAs.
• Removal of the Service IOA from the primary ES2 4G LM without powering it down does not trigger stateful line module switchover.
• The PPP application on ES2 4G LMs with Service IOA supports line module high availability. PPP session data is mirrored to the standby line module to attain high availability. The PPP application replicates the PPP sessions on the standby module and retains them across switchovers, in addition to accounting statistics. During line module switchover, the forwarding controller (FC) in the access module of the router that works as the LNS attempts to prevent timeouts of PPP sessions (due to the lack of PPP echo reply messages to the active subscribers) by sending echo response packets until the switchover is successfully completed.
• Policy manager is stateful line module switchover safe. Policy manager downloads the policy attachments from the SRP to the newly active line module after a switchover operation is detected. Policy statistics are preserved and made available across line module switchovers.
• The QoS application accomplishes the stateful line module switchover functionality by restoring the queues on subscriber interfaces in the newly active line module when the previously designated primary line module fails.
• During stateful line module switchover, the forwarding controller (FC) in the access module on the router functioning as the LNS device prevents timeouts of PPP sessions owing to the absence of PPP echo reply messages in response to echo requests received from clients.
• Only two pairs of primary and secondary line modules can be configured on a single chassis for stateful switchover. As a result, only two line modules can be HA-safe. If high availability is activated, when the secondary module takes over as the primary module, existing subscribers are retained. If high availability is not activated, when the primary line module fails, the standby line module processes the regular router functions, but previously active subscriber sessions are not retained.
• Stateful line module switchover can be triggered when one of the following actions is performed on the primary line module, with high availability for line modules enabled on the router:
  • Disabling the module in the specified slot using the slot disable command
  • Rebooting a module in a selected slot on the router using the reload slot command
  • Performing a graceful switchover to the secondary line module using the line-card switch command
• If both the primary and secondary modules are cold booted (for example, when a chassis is cold started), and if the primary module does not become online for 8 minutes, the secondary module takes the role of the primary module. This behavior is similar to the line module redundancy mechanism.
• If high availability for line modules is active, the switchover is stateful. Subscribers are not disconnected and none of the existing client sessions are terminated or locked out during the line module switchover. A data traffic outage of about 2 minutes occurs, although subscribers are not disconnected. PPP echo requests from the subscribers are responded by the access module itself during the switchover period. This method works properly even if LAG interfaces are configured to connect to a LAC device.
• Information related to line module switchover is not forwarded to applications such as the AAA or RADIUS servers. These modules are not requested again for any accounting or authorization information for the same subscribers that were connected during the time of switchover.
• When the unified ISSU process is in progress, you cannot configure high availability for line modules if the initialization state of the unified ISSU operation has started. You must wait until the unified ISSU procedure is completed to enable high availability for line modules.
• Line module high availability does not interfere with the configurations made for unified ISSU and stateful SRP switchover functions. The secondary module in a line module high availability pair does not participate in the unified ISSU operation and is disabled during the upgrade process. The secondary module is cold started after the unified ISSU procedure is completed. However, the primary line module takes part in the unified ISSU process and undergoes a warm restart.
• PPP-based stacks (L2TP, PPP, and IP applications) for both IPv4 and IPv6 interfaces support stateful line module switchover.
• You can manually switch between the primary and secondary modules. While the secondary module attempts to take over as the primary module during a switchover, if the secondary module fails to transition as the primary module within 5 minutes, the secondary module is cold booted.
• SNMP traps are generated after the switchover of the primary line module.
• Similar to dual SRP configuration and high availability of SRP modules, stateful line module recovery does not prevent the root cause that caused a router reload or stoppage of functioning. Stateful line module recovery enables the system to be returned to the fully functional state as soon as possible. If the conditions that caused the problem recur after a restart, an abrupt reload of the router might occur again. Stateful line module recovery minimizes forwarding impact on a restart to maximize customer uptime and causes the loss of packets during a restart to be limited to a small number of packets that are dropped in a timespan of a few seconds.