ePMP425 Weird Configuration Revert

Products ePMP
1

I really hope we are doing something wrong or there is something that we are not aware of. So we are using ePMP425 as PTP Backhauls between sites. Don’t really ever have issues with them.

However, over the past couple days, we’ve run into a couple of power outages that caused some weird things to happen with the 425.

Issue 1 arose from a site we were doing maintenance on unrelated to the 425. Crew was working on troubleshooting a different link to a different tower. Tower crew unplugged the 425 to do some work and plugged power back in after some time… no big deal.

Now for reference I will use our topology.

Tower Site A is 32
Tower Site B is 109

Tower Site A was the site we were performing maintenance on and where the 425 was unplugged at. As soon as the 425 was plugged back in we could no longer communicate with it. Upon further inspection we had found the Ip address scheme had completely changed.

So the Ip should have been: 10.32.109.32 signifying that this ptp is going from Tower Site A to Tower Site B.

The Ip it had changed to is another active segment in our network but the fact that it even changed is all the weirder. The Ip, subnet mask, and gateway changed itself to match network of 10.123.161.123.

Not only did it change IP information but it somehow inherited all other wireless, system, and misc network information from that other active link.

Fast forward to today we had a power outage on Tower Site C.

Tower Site C is 123
Tower Site D is 161

Crews arrived on site and found a power issue which caused total outage for site… fine… no big deal.

Crews restored the power issue and everything started coming back up as normal. All except the 425 on that site. Now this site is supposed to have a link from Site C to Site D with Ip address 10.123.161.123. Every last bit of information reverted and changed to another link in our network which happened to be the other link we had just fixed a couple days ago from Site A to Site B.

The Ip and all corresponding wireless, system, and misc network information all got changed to match the other link from Site A to Site B.

These sites have never been configured with other system parameters. They’ve been fresh out of the box installs. These sites, in routing terms, are several hops away from each other (8 or so) and are on complete separate sides of the network and have no correlation whatsoever.

I want to make sure I’m not missing something but we have NEVER had this happen before in the years of using them. It’s just a very odd and peculiar issue. Willing to provide any extra information needed.

2

What version of firmware are you running on these radios?

Have you tried disabling the reset power sequence?

image
image1053×476 45.2 KB

3

We’ve always kept the reset via power sequence off even on CPEs. I double checked to be sure and they’re disabled on both ends of each link. We are currently running 5.1.3 on all of those radios

4

That’s a pretty old release for the e4k series. I’d update to at least the newest stable release of 5.7.0, or even 5.7.1-RC9 beta.

5

I reported the same thing a couple years ago - a big power outage and a number of our AP’s ‘swapped configurations’. My theory was that it was cnMaestro uploaded the wrong AP config to the wrong APs somehow… I know that might not make any sense, but I can’t fathom any other way that could have happened.

6

It looks very odd!
Is there DHCP IP assignment or static?
CnMaestro not supposed to push any configuration when a radio reconnects.
5.1.3 is very very old firmware. I suggest to upgrade as soon as possible.
I believe it is better to do a step upgrade for already deployed radios. Because there was a lot of new features and some default values changes. Something like 5.1.3 → 5.4.2 → 5.6.1 → 5.7.1