Random modem hangups AX88U

[w0ks2]

My internet connection:
Fiber -> SFP -> Media Convertor -> RJ45 Cable -> Asus RT-AX88U WAN Port
Fiber was installed by ISP.
SFP I bought from my ISP.
Media Convertor bought from local store.
RJ45 Cable is CAT8.
Asus RT-AX88U running Latest Official Asus Firmware.

My issue:
My router Asus RT-AX88U has this weird and random issue where my WAN connection is dropped (and while it's dropped I can't connect to WIFI) and only a reboot to the router fixes it.
It can happen between 5 minutes and 4 days, each time at a random time and it seems that nothing I can do can solve it.
WAN light on the router stays white (connected) but on the WebUI it says "Disconnected".
The log:
pppd: serial link apperas to be disconnected
WAN connection: fail to connect with some issues
pppd: connection terminated
pppd: modem hangup
pppd: timeout waiting for pado packets

Trying to solve it:
Turn OFF the WAN connection on WebUI and turning it back ON.
I switched the SFP from my ISP - 3 times!
I switched the media convertor from a TP-Link one to a D-Link one.
I switched the RJ45 cable from CAT8 to CAT5e.
I switched the Asus RT-AX88U in a new unit of the same model.
I tried using AsusWRT-Merlin latest firmware.
I tried factory reset the router and not changing anything.
I contacted my ISP and switched the port I am connected to on the cabinet outside my house.
I contacted my ISP and switched the port I am connected to on the main ISP cabinet in my city.
I contacted Asus in my region which fowarded my issue to Asus directly and they found no issues with the router.
I contacted my ISP and a technician showed me the issue called "BRAS disconnections".

THIS ISSUE HAS BEEN GOING ON FOR 6 MONTHES!

I've searched the entire web trying to learn about this issue and I did not find a solution.

People in my country has the same equipment (ISP, SFP, Media Convertor and Asus Router) and they have no issues.

 

[Tech9]

I contacted my ISP and a technician showed me the issue called "BRAS disconnections".

What was the technician's suggestion to do in this case?

 

[w0ks2]

What was the technician's suggestion to do in this case?

Replace the media convertor (after he suggested it - I replaced), but it didn't solve the issue.

 

[Tech9]

I believe you have to work with your ISP on this issue. Did you call them again?

 

[w0ks2]

I believe you have to work with your ISP on this issue. Did you call them again?

"call them" is an understatement... I sent an inquiry to the ministry of communication about this issue and the response from the ISP was that their "media team" (which is their specialized team) is on this issue and they "working to track the issue".
Of course I don't expect they will solve it.

Everyone I contact regarding this issue (ISP "media team", ISP tech support, ISP technicians) all of them keep telling me to use ISP provided router - which is not a solution, why would I buy Asus RT-AX88U, SFP and Media convertor only for them to tell me to use the ISP provided one.
I forgot to mention but I own Asus RT-AX56U as my AiMesh Node, so you can see why I would rather use my equipment than the ISP's one.
(One more reason I refuse to use ISP router is that the terms of use of the router clearly state some privacy concerns, they track everything on their router + serve you ads, and I cannot control it)

 

[Tech9]

all of them keep telling me to use ISP provided router

Do you have ISP provided router and does it have the same issue? You can use your own equipment behind it like many folks around do.

 

[w0ks2]

Do you have ISP provided router and does it have the same issue? You can use your own equipment behind it like many folks around do.

I don't have the ISP provided router.
It doesn't have a modem (bridge) mode, only a router mode and as I said - cannot control it, but if I do use it it will be as:
ISP router -> AX88U (router) -> AX56U (AiMesh Node)
So a router behind a router.

 

[Tech9]

Yes, but if it doesn't work your ISP has to fix it. Now the problem is on you because you insist on own equipment. I personally have all my equipment behind ISP provided routers in three different residential places. There is no problem with your router behind the ISP router.

 

[drinkingbird]

I don't have the ISP provided router.
It doesn't have a modem (bridge) mode, only a router mode and as I said - cannot control it, but if I do use it it will be as:
ISP router -> AX88U (router) -> AX56U (AiMesh Node)
So a router behind a router.

A lot of people post about similar issues when using a media converter to trick the ISP ONT into thinking their router is connected. Their router may very well have some tweaks to either the SFP port or WAN connection/PPP protocol, or the media converter (cheap ones are often problematic) is dropping certain frames that it doesn't recognize, etc. Maybe their SFP isn't quite compatible with the media converter you have. Maybe it is an MTU issue on your WAN? Are you set to 1500 - if so you may need to adjust it down for PPP. I would first put their router in with your router behind it and see if the issue goes away. If it does, you can either leave it that way, or then you know that you need to focus on tweaking something with the asus or a different brand media converter (if it is really important to you to eliminate their router). As @Tech9 says, make it their problem, right now you're giving them an excuse to blow you off as you are running an unsupported configuration. If their router does not support bridge mode, doesn't matter. You can put the Asus in DMZ of their router and essentially have the same thing. Or take it one step further and disable NAT in the Asus and add a couple static routes to eliminate the double NAT. But honestly neither of those are necessary, the tiny latency difference is not going to be perceptible and the DMZ is just to make inbound port forwarding easier if you are using that. If you aren't using that, having the dual layers of NAT without DMZ actually gives you more security.

Putting their router in is going to be the easiest way to prove out where the problem lies, or at lest start narrowing it down. If you still have the problem, great, now you've proven it is their device and they have to fix it, after which you can try removing it again.

I've also seen mention of the WIFI not connecting when the WAN goes down on some of these routers. That is a bit odd, some sort of glitch, or whatever disconnect situation you're getting into may be spiking the CPU or locking up something. Don't really have any good theories on that one.

 

[w0ks2]

Yes, but if it doesn't work your ISP has to fix it. Now the problem is on you because you insist on own equipment. I personally have all my equipment behind ISP provided routers in three different residential places. There is no problem with your router behind the ISP router.

You are right.

Maybe their SFP isn't quite compatible with the media converter you have.

To be honest, I don't think that's the case since I'm not the only one with this equipment and no-one else is having issues, only me.
There's only 3 media convertors to choose from, a TP-Link one, a D-Link one and a UCY once, I already used 2 out of 3, I don't want to use the UCY one.
SFP - I swithced a few times already, so not even that seems to be the issue.

Maybe it is an MTU issue on your WAN? Are you set to 1500 - if so you may need to adjust it down for PPP.

"Maybe it is an MTU issue on your WAN?" - No, if it's PPPOE it defaults to 1492, I don't need to touch it, it's already 1492.

right now you're giving them an excuse to blow you off as you are running an unsupported configuration.

You are right. and after reading both comments I'll probably try to use ISP router -> AX88U (router).
But without DMZ.

Putting their router in is going to be the easiest way to prove out where the problem lies, or at lest start narrowing it down. If you still have the problem, great, now you've proven it is their device and they have to fix it, after which you can try removing it again.

After proving the issue still persists on their router I'll return it to them.

I've also seen mention of the WIFI not connecting when the WAN goes down on some of these routers. That is a bit odd, some sort of glitch, or whatever disconnect situation you're getting into may be spiking the CPU or locking up something. Don't really have any good theories on that one.

This issue that I can't connect to WIFI when WAN is down is super weird to me and I couldn't find any information about it.

 

[drinkingbird]

"Maybe it is an MTU issue on your WAN?" - No, if it's PPPOE it defaults to 1492, I don't need to touch it, it's already 1492.

PPPoE is MAX 1492. Not all PPPoE is 1492. If they are using VLANs, that becomes 1488. If they have other limitations or technologies in their network, it can go as low as 1400.

You have to test to find what your exact number is.

In some rare cases ISPs are now supporting 1508 or even larger to allow you to run 1500 over PPPoE but that is uncommon.

 

[w0ks2]

PPPoE is MAX 1492. Not all PPPoE is 1492. If they are using VLANs, that becomes 1488. If they have other limitations or technologies in their network, it can go as low as 1400.

You have to test to find what your exact number is.

In some rare cases ISPs are now supporting 1508 or even larger to allow you to run 1500 over PPPoE but that is uncommon.

I doubt they are using VLANs, but, better safe than sorry - will test it next time at 1400, just to be safe, either way I don't expect it to solve it and the issue will return in 24hours.
Will keep you updated.

 

[drinkingbird]

I doubt they are using VLANs, but, better safe than sorry - will test it next time at 1400, just to be safe, either way I don't expect it to solve it and the issue will return in 24hours.
Will keep you updated.

You can test by setting it to 1500, finding out the largest successful ping with "do not fragment" set, add 28, and that is your MTU. Note it is a good idea to reboot the router after changing and applying MTU as I've seen it not update completely until reboot. You can start with a ping size of 1464 which would be a 1492 MTU, if that succeeds but 1465 doesn't, then 1492 is correct. If not, you need to go lower until you find your magic number.

 

[w0ks2]

You can test by setting it to 1500, finding out the largest successful ping with "do not fragment" set, add 28, and that is your MTU. Note it is a good idea to reboot the router after changing and applying MTU as I've seen it not update completely until reboot. You can start with a ping size of 1464 which would be a 1492 MTU, if that succeeds but 1465 doesn't, then 1492 is correct. If not, you need to go lower until you find your magic number.

Update: Changing MTU to 1400 (and reboot) didn't solve the issue. After 3 days and a few hours (as I said - at random times) the WAN went down and so does the WIFI, so now I changed it back to 1492.
By the way, I used TCP Analyzer from speedguide and my MTU is 1492, so it's correct from the ISP side and from my side.

 

[Tech9]

THIS ISSUE HAS BEEN GOING ON FOR 6 MONTHES!

You can try fixing it yourself for 6 more months or you can order an ISP device and transfer the issue to the ISP. It's up to you.

 

[w0ks2]

You can try fixing it yourself for 6 more months or you can order an ISP device and transfer the issue to the ISP. It's up to you.

Yes, I already contacted my ISP and their router will arrive on 15th this month (3 more days), and I will use it as follows:
ISP router -> Asus RT-AX88U (router mode)

But untill then - Currently I'm testing it with a spare router as follows:
D-Link DSL224 (router mode, PPPOE dual stack) -> Asus RT-AX88U (router mode, automatic IP and IPv6 passthrough)
If this issue will repeat with this combination, then I can rest assured it will repeat with the ISP router aswell.

 

[drinkingbird]

Yes, I already contacted my ISP and their router will arrive on 15th this month (3 more days), and I will use it as follows:
ISP router -> Asus RT-AX88U (router mode)

But untill then - Currently I'm testing it with a spare router as follows:
D-Link DSL224 (router mode, PPPOE dual stack) -> Asus RT-AX88U (router mode, automatic IP and IPv6 passthrough)
If this issue will repeat with this combination, then I can rest assured it will repeat with the ISP router aswell.

But you still won't know if the issue is the router or the ISP.

If ISP router ->Asus router still has the issue you're going to have to run just ISP router for a few days (can probably use Asus as Access Point if you want). If it still happens then, you know it is something out in your ISP's network and having their router there will allow them to finally take blame and look into it.

 

[w0ks2]

But you still won't know if the issue is the router or the ISP.

If ISP router ->Asus router still has the issue you're going to have to run just ISP router for a few days (can probably use Asus as Access Point if you want). If it still happens then, you know it is something out in your ISP's network and having their router there will allow them to finally take blame and look into it.

Mmmm? Am I missing something?
Why would I just use ISP router? Do you believe they will blame the connected devices to the ISP router in order to isolate the issue?

 

[drinkingbird]

Mmmm? Am I missing something?
Why would I just use ISP router? Do you believe they will blame the connected devices to the ISP router in order to isolate the issue?

Depends. If you can access meaningful logs in the ISP router and it is clear that it is now the device dropping the connection, maybe you can skip that step. But for prudence I would want to prove beyond any doubt (to myself and to them).

And yes, they will probably try to blame any and everything, have you ever worked with an ISP's support before? I definitely would not volunteer the info that you have a second router behind theirs. If theirs is the only router and you have at least two devices losing internet that are directly connected to it (wired and/or wireless, it leaves very little they could attempt to blame on you.

If it works fine with their router with the Asus behind it, then there is something in the Asus (possibly a bug, possibly a setting) to blame.

 

[w0ks2]

If you can access meaningful logs in the ISP router

No. Beyond any doubt - NO.
No logs, no DNS settings, I have zero control.
The things I can change (and I tested it with my neighbor which is a family member who has the same ISP and their router) - SSID and Password (WIFI name, WIFI password), ISP username, ISP password, port forwarding.
That's it, nothing more, nothing less.

And yes, they will probably try to blame any and everything, have you ever worked with an ISP's support before? I definitely would not volunteer the info that you have a second router behind theirs. If theirs is the only router and you have at least two devices losing internet that are directly connected to it (wired and/or wireless, it leaves very little they could attempt to blame on you.

Seems reasonable, I see, thank you for the heads up.