EIGRP sessions won't stay connected, even with static neighbor stateme

[spfister336]

We have several sites connected over ATT's ASE on demand switched metro ethernet service. We use EIGRP as our routing protocol, since we're an all Cisco environment. Since we suspect multicast might be hitting a rate limit, and our EIGRP sessions keep dropping and reconnecting, I have put about half of the sites on static neighbor statements. Either I haven't set them up correctly, or its still using multicast anyway. Shouldn't EIGRP be using unicast only on these sites?

[balaji.bandi]

read some FAQ in related to EIGRP :

https://www.cisco.com/c/en/us/support/docs/ip/enhanced-interior-gateway-routing-protocol-eigrp/13681-eigrpfaq.html#ten

if you can share the device model and IOS code running on it.

configuration of both the side which failing frequenty, and what Log you see that time ?

are you able to ping each other at the time of EIGRP Lost peering ?

what kind of Links and how is utilisation of the Links and and CPU ? at the time of Eigrp lost ?

there are many factors cauing this issue,.

[paul driver]

Hello
You can switch to using unicast peering, but they need to be P2P peering connections as MC cannot coexist with unicast on multiaccess networks however your issue suggestively sounds like may have some congestion on the network which is impeding the adjacencies .

[spfister336]

Each remote site is on a separate vlan with only that site and the central site in it. Direct communications between remote sites is not really needed or wanted.

[balaji.bandi]

Each remote site is on a separate vlan with only that site and the central site in it

we understand that, you looking to do p2p - so you like to use unicast to establish eigrp neighbourship.

Since we suspect multicast might be hitting a rate limit, and our EIGRP sessions keep dropping and reconnecting, I have put about half of the sites on static neighbor statements. Either I haven't set them up correctly, or its still using multicast anyway. Shouldn't EIGRP be using unicast only on these sites?

we are addressing the issue you encountering up and down,

So we have asked some questions above. do you have any input to understand the issue?

if you can share the device model and IOS code running on it.

configuration of both the side which failing frequenty, and what Log you see that time ?

are you able to ping each other at the time of EIGRP Lost peering ?

what kind of Links and how is utilisation of the Links and and CPU ? at the time of Eigrp lost ?

there are many factors cauing this issue,.

[spfister336]

The central site is running a pair of 9500s, IOS XE 17.3.4. A typical remote site is running a pair of 4500Xs, IOS XE 15.2(7)E4.

Interface on central site is TwentyFiveGigabit (all sites reachable through Twe1/0/4), remote sites are TenGigabit

EIGRP sessions are down for only a few seconds, I'm not sure I can do a ping during that time because I don't know when it will happen.

Central site configs:

interface TwentyFiveGigE1/0/4
description AT&T(ASEoD) Fiber Network Trunk
switchport trunk allowed vlan 88-94,212-297,322,905,979,980
switchport mode trunk
ip access-group 188 out

!

router eigrp 1
distribute-list 2 out
network 10.0.0.0
network 172.18.0.0 0.0.0.3
redistribute connected
redistribute static
redistribute bgp 65150 metric 1000 1 255 1 1500
neighbor 10.210.48.2 TwentyFiveGigE1/0/4
neighbor 10.210.75.2 TwentyFiveGigE1/0/4
neighbor 10.210.23.2 TwentyFiveGigE1/0/4
neighbor 10.210.77.2 TwentyFiveGigE1/0/4
neighbor 10.210.43.2 TwentyFiveGigE1/0/4
neighbor 10.210.76.2 TwentyFiveGigE1/0/4
neighbor 10.210.44.2 TwentyFiveGigE1/0/4
neighbor 10.210.15.2 TwentyFiveGigE1/0/4
neighbor 10.210.22.2 TwentyFiveGigE1/0/4
neighbor 10.210.74.2 TwentyFiveGigE1/0/4
neighbor 10.210.47.2 TwentyFiveGigE1/0/4
neighbor 10.210.37.2 TwentyFiveGigE1/0/4
neighbor 10.210.17.2 TwentyFiveGigE1/0/4
neighbor 10.210.36.2 TwentyFiveGigE1/0/4

!

interface Vlan231
description AT&T ASEoD for Thurgood
ip address 10.210.31.1 255.255.255.0
no ip redirects
no ip unreachables
no ip proxy-arp
ip pim sparse-dense-mode
no logging event link-status

!

remote site:

interface TenGigabitEthernet1/1/1
description Trunk to 99-c65-clan1(10.99.94.1)
switchport trunk allowed vlan 88,93,231
switchport mode trunk
speed nonegotiate
service-policy output QOS-EGRESS-POLICY-DBL

router eigrp 1
distribute-list 1 in
network 10.0.0.0
network 192.168.31.0
network 192.168.131.0
auto-summary
passive-interface Vlan31
passive-interface Vlan131
passive-interface Vlan313
passive-interface Vlan314
passive-interface Vlan316
passive-interface Vlan317
passive-interface Vlan931
no eigrp log-neighbor-changes
!

interface Vlan231
description AT&T ASEoD
ip address 10.210.31.2 255.255.255.0
no ip redirects
no ip unreachables
no ip proxy-arp
ip pim sparse-dense-mode

log lines, central site:

Apr 13 14:37:54.381: %DUAL-5-NBRCHANGE: EIGRP-IPv4 1: Neighbor 10.210.31.2 (Vlan231) is down: Interface PEER-TERMINATION received
Apr 13 14:37:57.666: %DUAL-5-NBRCHANGE: EIGRP-IPv4 1: Neighbor 10.210.31.2 (Vlan231) is up: new adjacency

log lines, remote site:

Apr 14 13:12:12.780: %PIM-5-NBRCHG: neighbor 10.210.31.1 DOWN on interface Vlan231 non DR
Apr 14 13:12:26.889: %PIM-5-NBRCHG: neighbor 10.210.31.1 UP on interface Vlan231

[spfister336]

The central site is a pair of 9500s, IOS XE 17.3.4. Typical remote site is a pair of 4500Xs, IOS XE 15.2(7)E4.

central site config:
interface TwentyFiveGigE1/0/4
description AT&T(ASEoD) Fiber Network Trunk
switchport trunk allowed vlan 88-94,212-297,322,905,979,980
switchport mode trunk
ip access-group 188 out
!
router eigrp 1
distribute-list 2 out
network 10.0.0.0
network 172.18.0.0 0.0.0.3
redistribute connected
redistribute static
redistribute bgp 65150 metric 1000 1 255 1 1500
neighbor 10.210.48.2 TwentyFiveGigE1/0/4
neighbor 10.210.75.2 TwentyFiveGigE1/0/4
neighbor 10.210.23.2 TwentyFiveGigE1/0/4
neighbor 10.210.77.2 TwentyFiveGigE1/0/4
neighbor 10.210.43.2 TwentyFiveGigE1/0/4
neighbor 10.210.76.2 TwentyFiveGigE1/0/4
neighbor 10.210.44.2 TwentyFiveGigE1/0/4
neighbor 10.210.15.2 TwentyFiveGigE1/0/4
neighbor 10.210.22.2 TwentyFiveGigE1/0/4
neighbor 10.210.74.2 TwentyFiveGigE1/0/4
neighbor 10.210.47.2 TwentyFiveGigE1/0/4
neighbor 10.210.37.2 TwentyFiveGigE1/0/4
neighbor 10.210.17.2 TwentyFiveGigE1/0/4
neighbor 10.210.36.2 TwentyFiveGigE1/0/4
!
interface Vlan275
description AT&T ASEoD for Eastmont
ip address 10.210.75.1 255.255.255.0
no ip redirects
no ip unreachables
no ip proxy-arp
ip pim sparse-dense-mode
no logging event link-status

remote site config:
interface TenGigabitEthernet1/1/1
description Trunk to 99-c95-clan1(10.210.75.1)
switchport trunk allowed vlan 89,91,275
switchport mode trunk
speed nonegotiate
service-policy output QOS-EGRESS-POLICY-DBL
!
router eigrp 1
distribute-list 1 in
network 10.0.0.0
network 192.168.75.0
network 192.168.175.0
neighbor 10.210.75.1 TenGigabitEthernet1/1/1
passive-interface Vlan75
passive-interface Vlan175
passive-interface Vlan753
passive-interface Vlan754
passive-interface Vlan756
passive-interface Vlan757
passive-interface Vlan975
eigrp stub connected summary
no eigrp log-neighbor-changes
!
interface Vlan275
description AT&T ASEoD
ip address 10.210.75.2 255.255.255.0
no ip redirects
no ip unreachables
no ip proxy-arp
ip pim sparse-dense-mode
ip summary-address eigrp 1 10.175.0.0 255.255.0.0

log lines, central:
Apr 14 13:48:14.146: %DUAL-5-NBRCHANGE: EIGRP-IPv4 1: Neighbor 10.210.75.2 (Vlan275) is down: Interface PEER-TERMINATION received
Apr 14 13:48:16.917: %DUAL-5-NBRCHANGE: EIGRP-IPv4 1: Neighbor 10.210.75.2 (Vlan275) is up: new adjacency

log lines, remote:
Apr 11 15:54:17.955: %PIM-5-NBRCHG: neighbor 10.210.75.1 DOWN on interface Vlan275 non DR
Apr 11 15:56:28.269: %PIM-5-NBRCHG: neighbor 10.210.75.1 UP on interface Vlan275

(I notice that the last log lines on the central side were from today, but no lines appear on the remote side relating to this issue since static neighbor statement were put into place for this site on the afternoon of April 11)

the EIGRP session is usually down for 4-5 seconds, so ping tests are kind of hard to do because I don't know when it will happen and where.

The central side link is a TwentyFiveGigabit interface, and TenGigabit on the remote sides

About half of our remote sides have been given static neighbor statements and half do not. These were added on the afternoon of April 11. I am wondering if having half one way and half the other is a problem.

[spfister336]

Also... I think our BUM traffic through the ISP is limited to 30Mbps

[balaji.bandi]

sure if the link fully used then you will see issue, you can use QoS to reserve bandwidth for Control plane traffic if you do not like to happen this at least for 5%

[spfister336]

I keep posting answers to the above questions, but I think my post is getting deleted maybe. I'll try once more