Frequent Primary Uplink Changes

[SuperM]

Hi Community,

Firstly, I want to express my appreciation for the quality of support within this community and the willingness of members to assist. So, thank you to all the contributors.

Now, onto business 😉

I have a customer who is experiencing frequent "Primary uplink status changes," and from the ISP perspective, all looks good.

I have reviewed the Meraki documentation on "Connection Monitoring for WAN Failover" and would like to make an informed assumption about which specific test is causing the failure event to occur. Previously, Meraki support could not provide this answer.

Using these logs as an example:

Based on the time difference of 1 minute and 50 seconds from uplink: 0 >> uplink: 1, I assume it was not ARP that failed, as if it were ARP, it would have transitioned from uplink: 0 >> uplink: 1 within 15 seconds. I came to this assumption from this extract from the Connection Monitoring Test Process: "if a physical link is not detected, the failover would take place immediately."

Also, when I run an ARP table query on the MX, I sometimes notice two entries under Internet 1 ARP entries. Is this a possible issue?

Basically, the connection is as follows: MX -> NTU -> PE router. ARP Entry 1 is the public IP of the PE router interface, and ARP Entry 2 is the management IP.

I know @PhilipDAth has mentioned to check DNS previously on "MX250 WAN1 and WAN2 uplink changes for unknown reason."

Any thoughts are welcome.

Thank you!

[alemabrahao]

It’s hard to pinpoint the exact cause without more detailed logs or network analysis.

"The MX uses heuristic detection to determine if an internet link is down. If the Ethernet link state is up but the MX is unable to ping internet hosts, resolve DNS names of Internet hosts or ARP entries for the default gateway, it will trigger a link failover."

Primary Uplink Status and Ethernet Port Carrier Change Events - Cisco Meraki Documentation

Connection Monitoring for WAN Failover - Cisco Meraki Documentation

[PhilipDAth]

Also make sure you are using a currently stable or better firmware release.

[Edu_Chico]

@SuperM how do you resolved this problem?