MPLS ldp flapping

[ElshanMammadli5597]

We have faced mpls ldp flapping issue. For example in random time during day log appears that  mpls ldp session gone down and up..

MPLS runs over OSFP as underlay network. I have not observed ospf flapping. All router are same ASR901 seriees

I have checked timers,physical connection  and it seems that everthing is OK.  

[MHM Cisco World]

show tcp brief 
show mpls ldp neighbor 

[ElshanMammadli5597]

[MHM Cisco World]

show mpls ldp discovery detail 

do this in routers check holdtime config 

[ElshanMammadli5597]

Holdtime configuration in all routers is default configuration

LEN_MEATS-5_10G#sh mpls ldp discovery detail
Local LDP Identifier:
10.46.32.150:0
Discovery Sources:
Interfaces:
Vlan4066 (ldp): xmit/recv
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
LDP Id: 10.46.32.25:0
Src IP addr: 10.46.31.150; Transport IP addr: 10.46.32.25
Hold time: 15 sec; Proposed local/peer: 15/15 sec
Reachable via 10.46.32.25/32
Password: not required, none, in use
Clients: IPv4

Vlan4070 (ldp): xmit/recv
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
LDP Id: 10.46.32.6:0
Src IP addr: 10.46.31.106; Transport IP addr: 10.46.32.6
Hold time: 15 sec; Proposed local/peer: 15/15 sec
Reachable via 10.46.32.6/32
Password: not required, none, in use
Clients: IPv4

Vlan4080 (ldp): xmit/recv
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
LDP Id: 10.46.32.60:0
Src IP addr: 10.46.31.66; Transport IP addr: 10.46.32.60
Hold time: 15 sec; Proposed local/peer: 15/15 sec
Reachable via 10.46.32.60/32
Password: not required, none, in use
Clients: IPv4

Vlan4082 (ldp): xmit
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
Clients: IPv4

Vlan4084 (ldp): xmit/recv
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
LDP Id: 10.46.32.1:0
Src IP addr: 10.46.0.90; Transport IP addr: 10.46.32.1
Hold time: 15 sec; Proposed local/peer: 15/15 sec
Reachable via 10.46.32.1/32
Password: not required, none, in use
Clients: IPv4

Vlan4088 (ldp): xmit/recv
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
LDP Id: 10.46.32.5:0
Src IP addr: 10.46.31.161; Transport IP addr: 10.46.32.5
Hold time: 15 sec; Proposed local/peer: 15/15 sec
Reachable via 10.46.32.5/32
Password: not required, none, in use
Clients: IPv4

Vlan4090 (ldp): xmit/recv
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
LDP Id: 172.20.19.42:0
Src IP addr: 10.46.0.254; Transport IP addr: 172.20.19.42
Hold time: 15 sec; Proposed local/peer: 15/15 sec
Reachable via 172.20.19.42/32
Password: not required, neighbor, in use
Clients: IPv4

Vlan4061 (ldp): xmit/recv
Enabled: Interface config
Hello interval: 5000 ms; Transport IP addr: 10.46.32.150
LDP Id: 10.46.32.27:0
Src IP addr: 10.46.0.138; Transport IP addr: 10.46.32.27
Hold time: 15 sec; Proposed local/peer: 15/15 sec
Reachable via 10.46.32.27/32
Password: not required, none, in use
Clients: IPv4
Targeted Hellos:
10.46.32.150 -> 192.168.254.4 (ldp): active, xmit
Enabled by: LDP Config,
Hello interval: 5000 ms; Transport IP addr: 0.0.0.0

10.46.32.150 -> 10.46.32.2 (ldp): active/passive, xmit/recv
Enabled by: MPLS AToM Circuit,
Hello interval: 10000 ms; Transport IP addr: 10.46.32.150
LDP Id: 10.46.32.2:0
Src IP addr: 10.46.32.2; Transport IP addr: 10.46.32.2
Hold time: 90 sec; Proposed local/peer: 90/90 sec
Reachable via 10.46.32.2/32
Password: not required, none, in use
LEN_MEATS-5_10G#

[MHM Cisco World]

The neighbor smallest  up time is 12 hrs

And ldp send/receive not 100% equal but near value.
and I see BGP and LDP reset?
can you check MTU size and drop in CPU queue.

[ElshanMammadli5597]

Hello @MHM Cisco World 

I have chechked mtu size. Mtu size is same under interface. But I see "output drops" when show interfase. Maybe it is related output queue. Because uplink interface is 10G and downlinks are 1G. It is expected situation.