Administrator Guide

client to another client. If disabled, it is assumed that clients are in a full mesh and there is no need to advertise prexes to the other
clients.
• High CPU utilization may be observed during an SNMP walk of a large BGP Loc-RIB.
• To avoid SNMP timeouts with a large-scale conguration (large number of BGP neighbors and a large BGP Loc-RIB), Dell EMC
Networking recommends setting the timeout and retry count values to a relatively higher number. For example, t = 60 or r = 5.
• To return all values on an snmpwalk for the f10BgpM2Peer sub-OID, use the -C c option, such as snmpwalk -v 2c -C c -c
public<IP_address><OID>.
• An SNMP walk may terminate pre-maturely if the index does not increment lexicographically. Dell EMC Networking recommends using
options to ignore such errors.
• Multiple BPG process instances are not supported. Thus, the f10BgpM2PeerInstance eld in various tables is not used to locate a peer.
• Multiple instances of the same NLRI in the BGP RIB are not supported and are set to zero in the SNMP query response.
• The f10BgpM2NlriIndex and f10BgpM2AdjRibsOutIndex elds are not used.
• Carrying MPLS labels in BGP is not supported. The f10BgpM2NlriOpaqueType and f10BgpM2NlriOpaquePointer elds are set to zero.
• 4-byte ASN is supported. The f10BgpM2AsPath4byteEntry table contains 4-byte ASN-related parameters based on the conguration.
• If a received update route matches with a local prex, then that route is discarded. This behavior results from an incorrect BGP
conguration. To overcome this issue, you can trigger a route refresh after you properly congure BGP.
• If all the IP interfaces are in non-default VRFs, then you must have at least one interface in default VRF in order to congure a routing
process that works with non-default VRFs.
Traps (notications) specied in the BGP4 MIB draft <draft-ietf-idr-bgp4–mibv2–05.txt> are not supported. Such traps
(bgpM2Established and bgpM2BackwardTransition) are supported as part of RFC 1657.
Conguration Information
The software supports BGPv4 as well as the following:
• deterministic multi-exit discriminator (MED) (default)
• a path with a missing MED is treated as worst path and assigned an MED value of (0x)
• the community format follows RFC 1998
• delayed conguration (the software at system boot reads the entire conguration le prior to sending messages to start BGP peer
sessions)
The following are not yet supported:
• auto-summarization (the default is no auto-summary)
• synchronization (the default is no synchronization)
Conguring a basic BGP network
Conguring a basic BGP network includes a few mandatory tasks and many optional tasks. The initial step is to congure a BGP routing
and BGP peers, using the address family conguration. You can also congure BGP peers using the IPv4 VRF address family.
Prerequisite for conguring a BGP network
You should be familiar with the overview of BGP before proceeding with conguring a basic BGP network. For more information about the
BGP overview, see Border Gateway Protocol IPv4 (BGPv4) Overview.
Border Gateway Protocol (BGP)
187