Finally have victory over a longstanding issue with Meraki MX!
The smaller of the two issues still seems to remain unsolved (though maybe will be fixed after making the changes below). When the primary Internet fails, the cellular does not seem to failover. We manually will have to remove the Internet port, and follow the “voodoo” post I made earlier. //www.yodernetwork.com/cisco-meraki-mx-magic-cellular-voodoo
The larger issue is that the cellular will not stay connected for more than 10-15min. At which point, the unit will need reset and the “voodoo” process followed again. This morning, one of our branch locations’ Internet failed. However, this time the cellular would not stay connected more than a couple minutes.
I have a 6 month old case open with Cisco Meraki about the cellular issue – they stated that it was a bug in the firmware, which supposedly is patched with the version I am now on – 13.36.
Obviously I was a bit upset by the 1-2min disconnects, so I gave support a ring. I have discussed the cellular issue with multiple support folks in the past. However, almost immediately this guy asked if my warm spare was intentionally disconnected. I told him yes. He said that that is probably causing the problem. He stated that the primary MX will try to fail over to the warm spare even though the primary is on an active cellular connection. I suppose that makes sense to a degree, but the fact the Meraki cloud will attempt to fail over to a device that it clearly cannot communicate with is bad design in my opinion. However, I was not surprised by this, as there tends to be several gotchas with the MX series. The solution was simply to remove the warm spare from the configuration. I can confirm that this works, and the cellular will stay active! All of our branch locations have cold spares sitting on the shelf, which are configured as warm spare (primarily so I can track them, and we can switch to them quickly if primary dies). Now I will just be tracking the serials in a spreadsheet with location in case I need to activate one in a failover situation.
ALWAYS be on high alert when making changes like this. Removing a spare WILL cause a brief network interruption. Meraki seems to be sensitive about some configuration changes, which will cause network outages. Always perform these after business hours.