Connectivity issue on Flint 3 when WANs are up/down or change

Routers
1

When the WAN connections reset or go offline - it seems like my LAN connectivity goes wonky. Tonight I had my entire LAN go down, the router was blinking blue - and I can't confirm if the WANs were unavailable or not, because I couldn't get in to the router. Nothing in my LAN was accessible. Couldn't ping local devices or the router itself, from any of my systems. Best I could tell my primary WAN was online (modem had no noticeable error lights)

I have dual WAN, I have done controlled failover and it worked fine. When I click "connect" on the tether option for WAN2 (it's a USB 5g dongle), my entire LAN freezes up. I'm not even trying to use WAN2, just "connect" it so it's available. I've been able to reproduce this multiple times.

I don't expect my LAN to be affected at all by WAN connections being up or down or change at all. Tonight's issue left me blind - I looked at the logs on the router after power cycling it - there was nothing for a long time before the router shows the logs from the restart. So wasn't anything tangible there.

Last night this happened TWICE - after the first one, I decided to post (on Reddit, first, but nothing going on there...) and then later it happened again. For the moment I've unplugged my WAN2 tether and stopped that. It's been okay since then - but it was okay for weeks with the WAN2 tether active before this (except for the initial disconnection, which was wonky but I didn't think it was a big enough deal to bring up at the time)

During these LAN "outages" - it seems like pre-existing connections kinda stay open. it's .. weird. I have some pre-existing web and samba connections to some local LAN that were connected/busy at the time of this disconnection... and those are still working, even though I can no longer ping that same system that I am actively talking to.

I can have .17 ping .2 and .3 and .4 but .1 (router) is nothing, at the same time, .2 can't ping .17 back, or .3 or .4 etc.. and .1 (router) is nothing. The LAN is somehow fractured in a weird way.

2

Hi,

That is so strange issue, I would like to check these info:

  1. With this method to collect the syslog, SSH to the router:
logread -f > /root/disconnect_issue.log &

Bring the issue reproduce and export the file disconnect_issue.log

  1. May I know what USB dongle model is? is WAN2 you mentioned the tethering interface?

  2. Checking from your description, it seems that the LAN of Flint 3 has a Web server and Samba server, and with a wired connection, right?
    It is best to draw a network topology so that I can also clearly understand which devices able to ping, and wired or wireless connection.

  3. Have you installed plug-ins or changed the configuration in Luci/SSH? If yes, please reset the firmware and check it. What is the firmware version of Flint 3?

4
  1. will do this soon
  2. TCL Linkport IK511
  3. all of my network is wired. the flint 3 has wireless turned off, i use something else as wireless APs. all of these devices can ping each other normally. my network is 192.168.0.0/16, with 192.168.2.0/24 dedicated for DHCP and 192.168.1.0/24 where i use static DHCP or have defined on host the static address.
  4. i have not done anything in Luci except for change the DHCP domain suffix, for some reason i couldn't find it in the UI. firmware is latest stable 4.7.14 (release 4) - i am not sure exactly when i updated it but this may be something new since then.

definitely has a disconnect issue when i click "connect" to WAN2 - i noticed my LAN gets weird. but this whole thing last night where the LAN and WANs all become unstable and the router itself begins blinking LED - that's new. but has the same symptoms when i was adding on WAN2 to be available.

5

here's an example of what happens.

the first blip was when i simply plugged the USB cable back in with the USB 5G dongle. there was a quick little blip there but quickly recovered. the interface was not activated or anything.

the second blip is when i went to the UI and for eth2 (this USB dongle) i clicked "connect"

note that this ping is from a hardwired desktop 192.168.1.2 to the router 192.168.1.1 - there is unmanaged switch in between but i've never had an internal LAN network die off like this when a WAN issue (or something else...) is occurring.

lan connectivity drop
lan connectivity drop1022×718 35.2 KB

6

also @bruce i private messaged you some of the log output. didn't want to figure out how to redact possibly private details on the public forum that wouldn't render the thing useless :slight_smile:

7

Thanks, I received your syslog, and it seems no problem for the tethering interface.

But strangely, why is there an eth2 interface? When I tested it on my Flint3, the network interface name of tethering Android was usb0.

Please try upgrading to v4.8.1 and resetting the firmware and test it again.

In addition, I would like to know what device this USB dongle is, what model name? Is it a USB NIC?!

8

no idea why it's eth2. this is pure out of the box Flint 3... the usb dongle is a TCL Linkport IK511

as this is my router I can't be making a bunch of changes during the day - my admin panel says v4.7.14 so you want me to upgrade to 4.8.1 - from here GL.iNet download center ?

9

When this dongle is connected to the computer, and it should use the Remote NDIS protocol, right?

There is no issue with the router syslog when the dongle connects.

Please connect the mobile phone to router to test if the router tethering works fine, as a comparison test.

Please upgrade to v4.8.1 and reset the firmware, and only test the dongle connection first, let us to see if the interface works fine.

If no luck, you can share the Flint3 with us through GoodCloud, I will try to check it remotely.

Please connect the dongle and WAN if you shared the router with us.
Please PM me the router MAC and Wen UI password.

10

this happened last night - one time after the other.

first time, i had WAN2 still on. the LAN started becoming unstable and then it was obvious the whole thing was dead. checked router, the light was blinking. unplugged WAN2 and unplugged router, replugged it back in.

things came back online after boot, and were okay for something like 10 mins, and then it all went caput again. weird things like i could ping 192.168.1.2 FROM .24, but could not ping .24 from .2, no idea what that even is. this time the router light was solid so it wasn't detecting no connection, i guess.

had to force unplug it to restart, been okay since then. but that time WAN2 was not even plugged in or trying to be active. this is a big deal.

i will try upgrading the firmware. right now i'm trying to document all my config changes since i invested a lot of time (especially in static DHCP mapping) to wind up having to lose all of it.

had some logs still tailing and nothing showed up that seemed relevant.

11

just uploaded 4.8.1 beta firmware. i can now ping 192.168.1.1 from .2, and connect to the web interface... but none of my wired machines on my local LAN are reachable. can't ping or connect to .3, .4, on the clients page on the router they all show as being offline... why in the heck would hardwired LAN clients that have no issues now be offline even on the local LAN?

12

This isn't isolated to my .2 machine either. it's various systems throughout the network acting weird.

At the moment...

$ ping 192.168.1.24
PING 192.168.1.24 (192.168.1.24) 56(84) bytes of data.
From 192.168.1.3 icmp_seq=1 Destination Host Unreachable
From 192.168.1.3 icmp_seq=2 Destination Host Unreachable

However, 192.168.1.24 can ping 192.168.1.3 just fine.

this is a weird abnormality, I have no idea how my LAN can keep going "one sided" like this.

this is on my flint 3 which has barely been touched other than:

  1. changing my LAN to 192.168.1.0/255.255.0.0
  2. disabling all wifi
  3. changing dhcp to 192.168.1.2-254

i haven't even setup any address reservations other than 192.168.1.2 (wondering if that would fix anything) and 192.168.1.3 (which is static on the system itself anyway, it's not getting a DHCP IP)

...

back to when i set it up after upgrading firmware to latest beta, and resetting all settings...
logged in (from desktop, which is 192.168.1.2 usually) - but it had a 192.168.8.x IP since everything was fresh.

  1. disabled all wifi
  2. edit my LAN settings https://192.168.1.1/#/lanip
    router IP: 192.168.1.1
    netmask: 255.255.0.0

i had to release and renew local interface (makes sense) - it gets back a 192.168.1.185 IP (from DHCP) and has 192.168.1.1 as gateway. but now the router is no longer accessible, can't ping 192.168.1.1. somehow i got a DHCP address, but can't actually talk to what gave it to me?

my phone gets a 192.168.2.x address and CAN talk to 192.168.1.1. wtf is going on here. i get into the router on my phone, and force my desktop to 192.168.1.2 reserved IP

release, renew on desktop (.2) - still no dice. can't talk to 192.168.1.1. router says it's online. i get the DHCP address assignment. but it's ... simply not available otherwise. it has full WAN access, just a fragmented weird LAN situation for some reason.

i even restart my desktop after resetting windows network settings - nothing. still can't talk to 192.168.1.1

my phone (192.168.2.60 from DHCP) is able to hit 192.168.1.4, 192.168.1.1

other systems on network are okay at the moment. 192.168.1.24 is able to tracert 192.168.1.1

but the desktop, still:

C:\Users\mike>tracert 192.168.1.1

Tracing route to console.gl-inet.com [192.168.1.1]
over a maximum of 30 hops:

1 * * * Request timed out.
2 * * * Request timed out.

13

Hello,

For check and debug, if the router WAN/Internet is no issue, only LAN is effected, I think you can enable the GoodCloud service in router, and it is available to access the router with GoodCloud remote features.

If the remote web and SSH work ok, please share the router with us to check.