IPExpert V3 Lab8

Troubleshooting

Ticket 6

Classical EIGRP<->OSPF loop :

- Filter with tagging
brings with suboptimal that could be resolved by
- increasing ospf distance for other 'mutual redistribution router' to somethig more than external eigrp.


Ticket 9 L3 VPN

The Workbook comes with a problem I don't have because I corrected suboptimal routing :

A PE loopback is learned from ospf on a partial-mesh NBMA network hub which prevent label switching for it because the next-hop is the spoke and the ldp neighbor advertising the label is the hub router.
show ip cef, show no label for the PE loopbakc prefix.
Make the prefix learn by eigrp is a workaround.
Another if we want it work on NBMA : goes with point-to-multipoin ospf type network. The hub will be the next-hop, so it will match the ldp neighbor address.


Ticket 10 RP

The BSR/RP is not learned. Still a problem of RPF. The pim enabled interface is not the RPF interface for the RP address. Change it by advertising the RP in EIGRP or play with distance to make eigrp prefered for the RP address.


Configuration
1.6 IPhones

The purpose was to configure both access vlan and voice vlan in one command.
This was a good time to use pre-defined cisco macro command listed with

sh parser macro

The one to use is cisco-phone :
sh parser macro name cisco-phone to discover how to use it :
And to apply on it do :
macro apply cisco-phone $access_vlan X $voice_vlan Y

1.7 Private vlan

Don't forget to map secondary vlans on SVI and to set routers on the vlan as promiscious :


vlan P
private-vlan primary
vlan I
private-vlan isolated
vlan C
private-vlan community
vlan P
private-vlan association C,I
int vlan P
private-vlan mapping C,I
int I
sw mode private-vlan host
sw private-vlan host-association P I
int C
sw mode private-vlan host
sw private-vlan host-association P C
Int R
sw mode private-vlan promiscious
sw private-vlan mapping P add C,I

2.0 MPPPoFR

A Reminder on all kind of Multilink :

MLPPP with Multilink interface

int s0/0
encap ppp
ppp multilink group 1

int s0/0
encap ppp
ppp multilink group 1

Int Multilink 1
encap ppp
ppp multilink group 1
Where goes IP and stuff.

MLPPP with Virtual-template

multilink virtual-template 1
int s0/0
encap ppp
ppp multilink

int s0/0
encap ppp
ppp multilink

Int virtual-template
encap ppp
ppp multilink
Where goes IP and stuff.

The difference with Multilink is that virtual-template will be cloned in virtual-access and can create multiple bundle all with same IP !!! a kind of multipoint PPP.
Example 2x2 serials arriving on a router, the 4 serial on the hub router call one VI interface, the 3 routers will use one subnet between them.
PPP will installed a /32 for each peer.

Multilink frame Relay FR16

Permits to use multiple physical link as one Frame-relay bundle

int mfr1
IP stuffs and frame-relay mapping goes there
int s0/0
encap frame mfr 1
int s0/1
encap frame mfr 1

MPPPoFR with virtual-interface
permits to use multiple DLCI on one interface

Like for PPP with virtual-interface but the VI is called with

frame-relay interface-dlci DLCIx ppp virtual-template 1
frame-relay interface-dlci DLCIy ppp virtual-template 1

Point-to-multipoin is still available as the IP of the VI will be replicated on each bundle.

int virtual-template 1
encap ppp
Where ip stuff goes

WRONG WAY :

MPPPoFR with multilink

Not very interesting as configuration is much the same as for Vi but only permits point-to-point :

frame-relay interface-dlci DLCIx ppp virtual-template 1
frame-relay interface-dlci DLCIy ppp virtual-template 1

int virtual-template 1
encap ppp
ppp multilink group X
Where ip stuff goes

Information of Bandwidth for routing protocol will goes on physical if Multilink interface is used or on virtual-template, the bundle will be the addition of the physical or the vi interface.

IP on a virtual-link could prevent OSPF to make adjacency or exchange routes, use unnumbered on the VI and the real IP on a loopback.

3.2 Interconnecting 2 EIGRP AS

Easy task but missed it : Gre tunnel

3.5 redistribution

- Mutual redistribution between OSPF and EIGRP, make one router better for output and other for input. Play on redistribution metric on both. Be carefull a router also learn route by RIP making it better than external eigrp. need to change metric of external eigrp to 109.

5.1 BGP Redistribution of ospf.

- Redistribute only our network into bgp by filtering with ACL on the RR. Be carefull to also include external routes
- Border router will summarize to external neighbor (preventing to redistribute also connected routes of border). Those summarization should be filtered back to inside.
- Border router will import external routes with no-export community.
- To choose default prefered route. Set local-preference in order of preference and make distance of BGP for default prefered over others protocols.

7.0 NAT

Nat task Was ok, but forget to redistribute the Pool into IGP for the return routes be knonw by neighbor.