Administrator Guide

High Availability (HA)
High availability (HA) is a collection of features that preserves system continuity by maximizing uptime and minimizing packet
loss during system disruptions.
Topics:
• High Availability on Chassis
• High Availability in a PE Stack
• Online Insertion and Removal
• Hitless Behavior
• Graceful Restart
• Software Resiliency
• Control Plane Redundancy
High Availability on Chassis
The primary RPM (Route Processor Module) performs the routing, switching, and control operations while the standby RPM
monitors the primary RPM. If the primary RPM fails, the standby RPM can assume control of the system without requiring a
chassis reboot.
NOTE:
Although the C9010 switch can operate with one RPM, Dell Networking recommends two RPMs for redundancy and
to provide more bandwidth to each line card. One RPM provides 120 Gigabits of switch fabric bandwidth to each line card;
two RPMs provide 240 Gigabits of bandwidth to each line card.
The primary/standby role of an RPM is indicated by the mastership LED indicator. For the primary RPM, the LED color is solid
green. For the standby RPM, the LED color is solid amber.
For information about how to install RPM in a C9010 chassis, see the C9010 Getting Started Guide or C9010 Installation Guide.
NOTE:
FEFD is not an HA-aware protocol. Due to this, the protocol states and corresponding data are not available on the
standby system. While inter-operating with a third-party switch, the FEFD might move into an unknown state on the new
RPM after failover. Due to this, the line protocol might go down.
NOTE: Dell Networking OS supports high availability (HA) on virtual link trunking (VLT). For information on HA support on
VLT, see the VLT Chapter.
High Availability in a PE Stack
A port extender (PE) stack has a master and standby management unit that provide redundancy in a similar way to redundant
route processor modules (RPMs).
If the master stack unit fails or is removed, the standby unit becomes the stack manager. The stack elects a new standby unit
and resets the failed master unit. The failed master becomes online as a member unit; the remaining members remain online.
For more information about the failover process in a PE stack, see the Port Extender (PE) Stacking chapter.
Online Insertion and Removal
You can add, replace, or remove chassis components (RPMs, line cards, fan modules, power supplies) while the switch is
operating. C9010 RPMs and line cards are hot-swappable. Use the information in this section when inserting an RPM or line card
in the C9010 chassis. For more information about how to install an RPM or line card in the C9010, see the C9010 Getting
Started Guide or C9010 Installation Guide.
20
High Availability (HA) 391