Then 0.4 or 0.5 or 0.253 or anything that isn’t the same as the router’s address or in its DHCP pool. Or shorten the LAN subnet mask on the router by one bit so the 1.x addresses are included in the subnet (though you’ll need to restart all your equipment after for that to fully take). You need to pick an address that the router recognizes as being on the LAN subnet.

Change it to 0.3.

Did something get plugged in that connected switch 1 and 2 directly, i.e. other than through the router’s switch? If you unplug one of the switches from the router does everything, including the stuff on the unplugged switch, start to work?

I have seen your symptom when switches not running STP are connected in a loop. Since ethernet loops never stop shutting everything down is the only fix.

If it has 2 interfaces, say Ethernet and WiFi, it will have two different MAC addresses, one for each. I guess they only wrote one on the box?

A very long time ago I worked for a US telephone company helping to operate their Internet service. One day we had an outage that took out a bunch of backbone circuits in the northeast (that we thought were diversely routed but weren’t). The problem was in West Orange, New Jersey. Someone had cut out and taken a big chunk of 96 strand fiber optic cable where it came out of the ground and crossed a ravine attached to the bottom of a railway bridge. Apparently they thought it was copper, which has value as scrap, and were probably very disappointed they did all that work for something that had no value.

So I’m going to guess what you are seeing might be the result of someone starting to rip out a length of copper cable but stopping when they realized it wasn’t. Maybe putting a sign saying “This is fiber optic cable” on the cable would help.

I’m going to assume it might not be HomePNA running on the coax, then, but is that the only thing you’ve got attached to the coax? Try outside where the telephone drop cable ends up at your house if there’s nothing else inside.

That thing is a passive balun, which simply converts telephone signalling to or from the balanced twisted pair from or to the unbalanced coax, but it also claims to be a DSL filter and there is no way your modem should be downstream from that. What that suggests is that the balun is taking signal from the coax and putting it on the twisted pair, after filtering the DSL, for use by a POTS telephone. What is putting the signal on the coax is still a mystery, but there might be another balun, without the filter, at or near the telephone demarcation.

In any case it is becoming clear that the carrier is asking you to install a DSL modem when you don’t have the right wiring to plug it into. You may need to ask them to come and fix their mess.

The spacing of the contracts on an RJ45 is 1mm, so when you see the 8 wire holes all in a row like the one you link to you know that will only fit wires with an insulated diameter of no more than 1mm. The diameter of 24 AWG wires with insulation is generally at or just below 1mm so those should fit in that connector. The diameter of 23 AWG wires with insulation almost always exceeds 1mm by a bit, however, so those won’t fit (easily) in those holes. You instead need to find a connector that has the holes offset up-down-up-down so the larger conductors can be centered over the connector pins.

Almost all solid conductor cat5 and cat5e cable has 24 AWG wires, but most cat6 and cat6a cables claim to have 23 AWG conductors and, if they actually are that size, need connectors with the offset holes. Unfortunately, sometimes “23 AWG” is a lie and the actual conductor diameter is closer to 24 AWG, so the connectors with the straight-across holes sometimes fit on cat6. The only real way to know which connectors are best is to measure the actual diameter of the wires you are installing them on to see if they exceed 1mm.

Avoid ordering from the US, the UK or EU might be okay. The US has regulations requiring that WiFi equipment sold (or used) there be hardwired to only support the US frequency bands. Equipment sold for use in most other countries have a menu or other configuration to let you select your country. You want the latter.

What else is connected to the coax somewhere else in the house? I think the coax connector on that router is HomePNA which, if I’m remembering correctly, is like MoCA that doesn’t share with cable TV. Since the router isn’t connected to the DSL it must be using the coax to connect to a DSL modem that is connected to the DSL. If you can find that other box it should be plugged into the appropriate RJ11.

I see lots of T-3 timeouts and other junk in my log too, but that is DOCSIS protocol stuff that seems not to effect too much. The error counts that would be interesting would be in the “Corrected” and “Uncorrected” columns on the screen shot you posted, where all the zeros are now. They get zeroed when the modem restarts, so you need to let it run for a while to find out if there are times when the signal isn’t so nice.

I think that ancient Buffalo router has 100 Mbps LAN ports and old, slow WiFi. Anything modern would be much better.

200’ of wire isn’t long at 1Gbps, and wire doesn’t usually work sometimes and fail others, but WiFi does that. My money is on the link through the eero beacon as the problem.

It may drop when something ugly turns on, like your refrigerator’s compressor or your furnace’s blower or your heat pump, or maybe even something at your neighbor’s house that is fed from the same transformer. You can’t really fix that.

If you rebooted the modem not long before you took that screenshot you might want to let it run a few days to see if you accumulate errors over time. Otherwise I agree that everything looks very good there, in fact I’ve never had a cable connection that nice, so don’t fiddle with anything on the coax side of the modem.

. While that could have a number of causes, if it really only happens to Internet connections (the “connection to raids” comment is confusing me) then what you are seeing may be consistent with a NAT table overflow in one of the routers. The NAT table in a router tracks what it thinks are your currently active connections to Internet destinations but tends to accumulate a lot of cruft when it can’t tell, or doesn’t process, when a connection closes and is no longer in use. When the table fills and the router needs to find space for a new connection it responds by blowing out some of the existing stuff in the NAT table. If some of the stuff it dumps is currently active this causes incoming packets for those connections to be dropped until the internal host sends something that reestablishes the table entry, which may take some seconds.

To determine if this might be the cause one would look at the state of the NAT table in the router to see how close to full it is running, but many consumer routers don’t show this. To make it work better one would increase the size of the NAT table and maybe tune its other parameters, but most consumer routers don’t support that. Failing this one might buy a router with a bigger NAT table, but most consumer routers don’t tell you how big theirs is. There are fairly inexpensive routers that do let you manage the NAT table but they tend to be ones that can be deadly difficult to configure.

Note that if this is your problem the problem could be either in your router or the router that is in “bridge mode” (what it is actually doing can be a mystery; the routers AT&T makes everyone use famously keep their NAT table in use not only in the “IP passthrough” modem emulation but also when you buy extra IP addresses so no NAT is necessary). If you can replace the modem/router with a real modem you could eliminate that as an issue. As for the other router, with no NAT configuration all you could do is replace it with another that you hope works better. All this is a lot of hardware to buy if the current routers don’t give you a way to determine that the NAT table, and not some else, is in fact the problem.

This seems confused. Category 8 is defined in ANSI/TIA-568-D and ANSI/TIA-568-C.2, as are Categories 5e, 6 and 6a. There is no such thing as 7 or 8.1 in that standard.

Category 8.1 and 8.2 cables and connectors are specified in another standard, ISO/IEC 11801. This one also includes Category 1, 2, 3, 5e, 6, 6a and, yes, 7 and 7a cables and connectors.

If you hence believe Category 8.1 exists but Category 7 doesn’t you are being inconsistent. Only those that think Category 8 exists and 8.1 is a fantasy can be consistent in the belief that there is no such thing as 7.

Cat8 won’t improve the speed of anything you are likely to ever have to plug into it, so I wouldn’t install it with the expectation it’ll help with that. That said, I like cat8 cables (and cat7 for that matter) for another reason: you can get cables with 22 AWG conductors that are sometimes actually 22 AWG in size and this has a significant benefit for POE where size (i.e. DC resistance) matters. I hence use it for POE devices that run hot, or might someday run hot, in particular WiFi AP locations and links between HDBaseT modems.

So I can tell you what I would buy. 22 AWG bare copper wires are between 0.64 and 0.65 mm in diameter, so I would buy cat8 (or cat7) 22 AWG cables that have conductors that size. You can’t rely on the “22 AWG” label to tell you that, however, so prefer vendors that have a spec sheet for their cable that lists the actual conductor diameter in millimeters (but measure it when you get it anyway). As for a link to cable like that, however, I can’t give you one. I bought a whole bunch of Monoprice cat8 some years ago that had 0.65 mm conductors but the same part number now lists the size as 0.60 mm, which is now, unfortunately, quite common. I was also given some cat7 cable with full size conductors but I don’t remember the brand.

That accomplished, you now need keystone jacks that actually fit on 22 AWG wires, and again you can’t trust “22 AWG” (not to pick on them, but Monoprice was at one point listing “22-24 AWG” keystones with a max specified conductor diameter of 0.58 mm). The only ones I can vouch for are these

https://www.amazon.com/Cable-Matters-5-Pack-Shielded-Keystone/dp/B074HH9RHW

which are fantastically pricey but which fit on the wires really well.

If this seems like a lot of hassle and cost for something that is realistically unlikely to provide a speed benefit you might be right. Cat6a stuff is good and not too expensive these days.

I haven’t got much of an opinion about the choice but I do know that the Deco configuration app has a configuration option to put the mesh into AP mode (it is somewhere in the Advanced configuration). That means that if you can’t get the 5G router to act like a modem you can always leave it to act as a router and configure the Deco’s to provide the WiFi alone.

Try to get a probe from an ohm meter on the solid orange wire and use the other probe to find the pin it is attached to. If it is on pin 6 it is 568A, if it is on pin 2 it is 568B.

There’s an awful lot of NATing going on there, but clearly whatever did the a.a.a.a:1000->g.g.g.g:1500 translation on the way out should be doing a g.g.g.g:1500->a a.a.a:1000 translation on the responses coming back in. It would be better to eliminate that NAT operation entirely and just put the a.a.a.a:1000 packets into the tunnel, but whether that works depends on how the VPS works.