Posted on January 7, 2020 8:06 pm
 |  Asked by Martin Wolf
Print Friendly, PDF & Email

I was wondering, what the recommended way is how to set up heartbeat between MLAG peers.
At older EOS versions, such as in 4.19 I was able to separate Keep-alive traffic by running it via VRF Mgmt interface. This version did support the heartbeat entry under MLAG – see example below:

peer-address heartbeat vrf mgmt
dual-primary detection delay 10 action errdisable all-interfaces
errdisable recovery cause mlagdualprimary

In newer version, such as 4.23F, it seems that I am not allowed to run heartbeat via VRF Mmgmt anymore; this heartbeat entry is not available under MLAG config:

leaf1(config-mlag)#peer-address ?

Therefore, I was wondering, what the recommended way is how to route keep-alive[heartbeat messages] between MLAG peers and what there advantages/disadvantages are.



Posted by Victor
Answered on January 7, 2020 11:34 pm

Hello Martin,

It is still possible to configure heartbeats via VRF on 4.23F as seen below.

From your email, I realise you did not enter the heartbeat keyword after the IP address.

al301...23:26:34(config-mlag)#show version
Arista DCS-7170-64C-F
Hardware version: 21.00
Serial number: SSJ18111738
System MAC address: 7483.ef8d.a48c

Software image version: 4.23.1F-DPE
al301...23:26:34(config-mlag)#peer-address heartbeat ?
vrf Specify VRF name

al301...23:26:13(config-mlag)#peer-address ?


Victor Njoroge

Posted by Sreelekha
Answered on January 8, 2020 9:10 am

Hi Martin,

Happy New Year!

From the lab test, I see that the exact command is supported on 4.23 EOS code. Could you please use a "?" after peer-address as below:

bri457...09:03:49(config-mlag)#peer-address ?
A.B.C.D Peer's IP address
heartbeat Optional separate peer-address for heartbeat


Software image version: 4.23.0F
Architecture: i686
Internal build version: 4.23.0F-13614065.4230F
Internal build ID: 158ef907-fbb0-49af-950b-a2a8e86f3d07

Uptime: 7 weeks, 1 days, 6 hours and 18 minutes
Total memory: 16365380 kB
Free memory: 8056544 kB

If you are still not seeing the option, can you please provide the exact EOS version and the platform of the device.


Thank you all and Happy New Year as well: Best on my testing there are two ways [where running heartbeat via VRF MGT is the more stable one]. a) UDP Heartbeat via VRF Management b)UDP Heart beat via MLAG Link peer. Testing with vEOS 4.230F: Note: Usinfg MLAG Peer Link for UDP Heartbeat: After MLAG Peer Reconnect: According to MLAG Detail output, both Leaf switches remained in Primary status and UDP Heartbeat did not resume. Howver, according to Wireshark, TCP Keepalive as well as UDP Heartbeat has been resumed after MLAG Peer Link was reestablished. Therefore, I was wondering, whether you monitored MLAG status detail and capture packets amid simulating a loss of MLAG Peer Link with both settings: Here: 10.0.0.X MLAG Peer LINK SVI, VLAN 4094 20.0.0.X/30 VRF MGT 1) mlag configuration domain-id mlagDomain heartbeat-interval 2500 local-interface Vlan4094 peer-address peer-address heartbeat vrf mgmt peer-link Port-Channel10 and 2) mlag configuration domain-id mlagDomain heartbeat-interval 2500 local-interface Vlan4094 peer-address peer-address heartbeat peer-link Port-Channel10 Regards, Martin
(Martin Wolf at January 13, 2020 5:59 am)
Posted by Victor
Answered on January 21, 2020 12:34 am

Hello Martin,

Thanks for your question. However your text is not very readable as the configurations have all been pasted on one line. Is it possible to rephrase / resend this in a more readable form.


Post your Answer

You must be logged in to post an answer.