WAN goes down, Wan light still white. renew DHCP lease brings it back

[slvrscoobie]

Im trying to eliminate an issue Im having where my internet will stop working randomly. It usually occurs during work hours which for WFH is extremely annoying.
Im on FIOS with ethernet from the ONT -> AC3100->MacBook via Lan. This setup has worked flawlessly for ~5 years, until xmas when I decided to upgrade to a AX router. Ended up returning it, but when I went back to my AC3100 I now get these drops in internet.
It will be working fine one minute and often on the hour, it just stops. Asus says it disconnected and I cant ping anything external. the WAN light is still white though.. unplugging the WAN or renewing the DHCP lease fixes it, even if there is hours left on the DHCP lease according to the router. This occurs on both ethernet connected and wifi devices. I did unplug the router but this only resolved it for a day until it did it again.
Im using the Merlin 384.4 from Jan 1.
I did have fios support on the phone while it dropped and they said there was no issues with their side. I replaced the end of the ethernet cable as the AX port was upside down from my AC so I thought maybe the wire was faulty. This fixed things for about the past week.
Ive run ping plotter the last week and have seen 0 drops on pinging both their site, and twitter. until this morning, when it just went down again. DNS is set as Fios, I do have about 2 dozen IOT devices but again nothing major different since Nov when things were working well.

 

[Tech Junky]

I had a period of time when Comcast was doing something similar where for whatever reason the half-life renewal of the lease wasn't kicking in to prevent the drop. I setup a script to kill dhclient / renew the IP when it occurred. It hasn't been an issue lately but, for about a month or so it seemed to happen quite a bit.

I'm not using a run of the mill router though and haven't in quite awhile now since I built my own instead for higher bandwidth on the LAN side.

I've seen some mention here that people are scheduling daily reboots of their routers.... There might be something buggy in Merlin causing it?

 

[slvrscoobie]

I had a period of time when Comcast was doing something similar where for whatever reason the half-life renewal of the lease wasn't kicking in to prevent the drop. I setup a script to kill dhclient / renew the IP when it occurred. It hasn't been an issue lately but, for about a month or so it seemed to happen quite a bit.

I'm not using a run of the mill router though and haven't in quite awhile now since I built my own instead for higher bandwidth on the LAN side.

I've seen some mention here that people are scheduling daily reboots of their routers.... There might be something buggy in Merlin causing it?

yeah, that could be a real PITA, as ive been on a fairly static IP for some time, and use a domain to connect to my VPN and rebooting the router every night would likely cause a new IP every day. not that I need one, but I usually had a IP for months on end before this all started. I have to assume its a misconfiguration or issue im missing. Might go back to 382.2 since that was stable just to see if its that. never had an issue with a stable merlin release before but

 

[Tech Junky]

This is why I made my own. I just update the linux kernel weekly and keep on truckin'.

IP assignments w/ providers don't usually change all that often unless they're making scope changes to the DHCP system. My typical DHSP period is 7 days which in normal circumstances should auto renew when the timer hits 50%.

I'm more inclined to say it's a provider issue that reps don't know about but, if they're keeping a list of their calls / problems it should show a trend or it's not being reported often enough to trigger a flag in someone's head.

The DNS to the Domain as you put it shouldn't really be an issue connecting. If you're trying to keep track of your IP for connecting to the house though it can be a PITA unless you invoke some DDNS or use a torrent tracker to announce it to a server. There's a ton of ways to keep track but, the dropping of connectivity randomly makes things frustrating.

It's possible the FIOS gateway might be having an issue. From when I was looking into FIOS as an option and liking to run my own equipment it seems the handoff from VZ just requires a VLAN tagging to get routed out using a standard router to bypass their device. Could be as simple as a managed switch doing the handoff if you want.

 

[slvrscoobie]

Im just trying to rule out other issue, as they ran their tests while I was 'down' and didnt see anything.
replacing the end of that cable then resolved the issue for 7 days so Im wondering if the cables gone bad. not sure how to test that other than replacing 75ft of cable running outside around the house.
it is weird ive never seen it go down (and I run a couple of devices that should report an outage) overnight. its only like work hours. of course if there was some DHCP lease issue with the ONT it would have started around then since thats when I was mucking with replacing the router.

I'm more inclined to say it's a provider issue that reps don't know about but, if they're keeping a list of their calls / problems it should show a trend or it's not being reported often enough to trigger a flag in someone's head.