Graceful Restarts

When BGP restarts on a router, all of the router’s BGP peers detect that the BGP session transitioned from up to down. The transition causes a routing flap throughout the network as the peers recalculate their best routes in light of the loss of routes from that peering session.

The BGP graceful restart capability reduces the network disruption that normally results from a peer session going down. If the session is with a peer that had previously advertised the graceful restart capability, the receiving BGP speaker marks all routes from that peer in the BGP routing table as stale. BGP keeps these stale routes for a limited time and continues to use these routes to forward traffic. Any existing stale routes from that peer are deleted to account for consecutive restarts.

When the restarting peer reestablishes the session, the receiving BGP speaker replaces the stale routes with the fresh routes it receives from the peer. The restarted peer sends an End-of-RIB marker to signal when it has finished sending all its routes to the BGP speaker. Until this point, BGP has still been using the stale routes to forward traffic. Upon receipt of the End-of-RIB marker, the BGP speaker flushes any remaining stale routes from the restarted peer.

The End-of-RIB marker is an update message that contains no advertised or withdrawn prefixes; it is sent only to BGP speakers that have previously advertised the graceful restart capability.

The receiving speaker also sends its own routes to the restarted speaker, and sends an End-of-RIB marker when it completes the update. The restarted peer defers reinitiating the BGP best-path selection process until it has received this marker from all peers with which it had a session in the established state and from which it had received an End-of-RIB marker before it restarted.

After running the selection process to pick the best route to all prefixes using the fresh routes, BGP then installs the best routes in the IP routing table on the restarted peer. Any of these that are best overall routes to a prefix are then pushed by the router to the forwarding tables on the line modules.

By waiting for all restarted peers to send the End-of-RIB marker, BGP risks delaying the initiation of the best path decision process indefinitely due to a single very slow peer. For a specific peer, you can avoid this delay by hard clearing the peer or issuing the clear ip bgp wait-end-of-rib command. Either method removes that peer from the set of peers for which BGP is awaiting an End-of-RIB marker. Alternatively, you can minimize this effect by using the bgp graceful-restart path-selection-defer-time-limit command to specify a maximum period that the restarted peer waits for the marker from its peers.

Note that the receiving peer does not defer its best-path selection process while waiting for a restarted peer to reestablish a session. The receiving peer continues to use the stale routes from the restarted peer in the decision process. When it flushes stale routes, the receiving peer then uses the freshly updated routes.

A restarting peer must bring the session back up and refresh its routes within a limited period, or BGP on the receiving peer will flush all the stale routes. When a BGP speaker advertises the graceful restart capability, it also advertises how long it expects to take to reestablish a session if it restarts. If the session is not reestablished within this restart period, the speaker’s peers flush the stale routes from the speaker. You can use the bgp graceful-restart restart-time command to modify the restart period advertised to all peers; the neighbor graceful-restart restart-time command modifies the restart period advertised to specific peers or peer groups. A receiving peer starts the timer as soon as it recognizes that the session with the restarting peer has transitioned to down.

The receiving peer also has a configurable timer that starts when it recognizes that the session with the restarting peer has gone down. The bgp graceful-restart stalepaths-time command determines how long a receiving peer is willing to use stale paths from any restarted peer; the neighbor graceful-restart stalepaths-time command does the same for a specified restarted peer or peer group. If the receiving peer does not receive an End-of-RIB marker from the restarted peer before the stalepaths timer expires, the receiving peer flushes all stale routes from the peer.

bgp graceful-restart

• Use to enable the BGP graceful restart capability.
• Advertisement of the graceful restart capability is enabled by default.
• The no neighbor capability negotiation command prevents the advertisement of all BGP capabilities, including graceful restart, to the specified peers.
• This command takes effect immediately and automatically bounces the session.
• Example

  host1(config-router)#bgp graceful-restart

• Use the no version to disable advertisement of the graceful restart capability. Use the default version to restore the default condition, advertising this capability.
• See bgp graceful-restart.

bgp graceful-restart path-selection-defer-time-limit

• Use to set the maximum time a restarted BGP speaker defers reinitiating the best-path selection process.
• This command takes effect immediately and automatically bounces the session.
• Example

  host1(config-router)#bgp graceful-restart path-selection-defer-time-limit 180

• Use the no version to restore the default value, 120 seconds.
• See bgp graceful-restart path-selection-defer-time-limit.

bgp graceful-restart restart-time

• Use to set the time BGP advertises to all peers within which it expects to reestablish a session after restarting. Peers flush stale routes from the speaker if the session is not restarted within this period.
• Specify an interval shorter than the stalepaths time.
• This command takes effect immediately and automatically bounces the session.
• Example

  host1(config-router)#bgp graceful-restart restart-time 240

• Use the no version to restore the default value, 120 seconds.
• See bgp graceful-restart restart-time.

bgp graceful-restart stalepaths-time

• Use to set the maximum time BGP waits to receive an End-of-RIB marker from any restarted peer before flushing all remaining stale routes from that peer. The timer begins when BGP recognizes that the peer session has gone down.
• This command prevents an excessive delay in BGP reconvergence due to a peer that brings a session back up but is slow to send fresh routes.
• Specify an interval longer than the restart time.
• This command takes effect immediately and automatically bounces the session.
• Example

  host1(config-router)#bgp graceful-restart stalepaths-time 480

• Use the no version to restore the default value, 360 seconds.
• See bgp graceful-restart stalepaths-time.

clear ip bgp wait-end-of-rib

• Use to clear a peer or peer group from the set of peers for which BGP is waiting to receive an End-of-RIB marker after a peer restart.
• Alternatively, performing a hard clear of a peer without this keyword has the same effect.
• This command takes effect immediately.
• Example

  host1#clear ip bgp 192.168.1.158 wait-end-of-rib

• There is no no version.
• See clear ip bgp wait-end-of-rib.

neighbor graceful-restart

• Use to control the advertisement of the BGP graceful restart capability for specified peers or peer groups.
• Advertisement of the graceful restart capability is enabled by default.
• The no neighbor capability negotiation command prevents the advertisement of all BGP capabilities, including graceful restart, to the specified peers, but does not affect global advertisement of the graceful restart capability.
• This command takes effect immediately and automatically bounces the session.
• Example

  host1(config-router)#no neighbor 10.21.3.5 graceful-restart

• Use the no version to disable advertisement of the graceful restart capability. Use the default version to remove the explicit configuration from the peer or peer group and reestablish inheritance of the capability configuration.
• See neighbor graceful-restart.

neighbor graceful-restart restart-time

• Use to set the time BGP advertises to specified peers or peer groups within which it expects to reestablish a session after restarting. Peers flush stale routes from the speaker if the session is not restarted within this period.
• Specify an interval shorter than the stalepaths time.
• This command takes effect immediately and automatically bounces the session.
• Example

  host1(config-router)#neighbor graceful-restart restart-time 240

• Use the no version to restore the default value, 120 seconds.
• See neighbor graceful-restart restart-time.

neighbor graceful-restart stalepaths-time

• Use to set the maximum time BGP waits to receive an End-of-RIB marker from the specified restarted peer or peer group before flushing all remaining stale routes from that peer. The timer begins when BGP recognizes that the peer session has gone down.
• This command prevents an excessive delay in BGP reconvergence due to a peer that brings a session back up but is slow to send fresh routes.
• Specify an interval longer than the restart time.
• This command takes effect immediately and automatically bounces the session.
• Example

  host1(config-router)#neighbor graceful-restart stalepaths-time 480

• Use the no version to restore the default value, 360 seconds.
• See neighbor graceful-restart stalepaths-time.

Copyright © 2008, Juniper Networks, Inc. All rights reserved. Trademark Notice.