10-11-2018 10:38 AM
Another 35 went off last night, 85 radios affected at this time. We can not be calling up clients everytime this happens, they tend to get annoyed. 85 truck rolls will cost me $10,000 if I have to do that.
I thought at first it was power related, and it seems it is in some cases, there have been rolling line upgrades and it tends to happen after that a lot. But last night there were no rolling line upgrades.
Radios are in router mode, private ip's with 1:1 nat to public. I will block discovery at the main router and see how it goes.
It comes in waves, 3 weeks between the last batch and this one. They do not all go down at the same time, seems a few at a time, over a few hours. Reboot solves it up. Clients do not notice, it just affects Web UI and SSH.
10-12-2018 02:46 PM
Blocked TCP and UDP port 10001 from public. I was seeing 30k-50k dropped packets per hour UDP.
Today we have more showing offline that are really still online, no web or ssh access, still connected to the tower, still moving data.
4 weeks ago
So since I blocked 10001 from public web, I have 1 PB M5 620, 1 PB M5 400, 1 LB 5AC, 1 PB 5AC 400 cpe that have no ssh or web ui access but are still passing data and working from the clients perspective. Firmware is 6.1.3, 6.1.7, 8.4.3, 8.5.4
I have cpe's showing offline in AC2 as far back as 157 days ago, and 108 days ago.