Switch down after some time - ZyXEL GS1900-24E rtl838x

Installing and Using OpenWrt
1

Hello!

I installed 2 month ago OpenWrt 23.05.3 on a ZyXEL GS1900-24E (rtl838x) switch and all worked perfectly. Since a couple of days (I didn't change anything), the switch seems stuck after ~13 minutes. No Luci, no ssh, no switching/routing functionality anymore. After restarting it works for another 10-15 mins.

I also updated to latest OpenWrt version: 23.05.05 - same behavior.

I think I have to localize the problem a bit first. What is the best way to get some proper error messages or something similar?

2

Try to find pin layouts and get your self a console adapter.

I have no clue how reliable that is, but the following site suggests there's an UART interface on something that is labeled "JP1":
https://svanheule.net/switches/gs1900-24ep

Assuming your 24E is the same as the 24EP described.

Connect a computer to that serial console and let it run until it crashes.

3

https://svanheule.net/switches/gs1900-24e#pinouts

4

Ah, cool! Thanks!

I opened the case, and connected via serial.

I connected the following (-> my serial-usb-adapter) :
JP2.1: 3.3V -> 3V3
JP2.2: TX -> TXD
JP2.3: RX -> RXD
JP2.4: GND -> GND

Should it be correct?

sudo su
screen /dev/ttyUSB0 115200 #=115200, 8N1

 ?��MH2
       ���������������

(At least for another switch that procedure was fine)

Do I need another 115200, 8N1 ?

EDIT: Sorry, with the following it worked:
JP2.2: TX -> RXD
JP2.3: RX -> TXD

5

Unfortunately there are no logs over serial connection regarding the issue.

[   51.818071] switch: port 24(lan24) entered blocking state
[   51.824282] switch: port 24(lan24) entered disabled state
[   51.832252] device lan24 entered promiscuous mode
[   51.861484] rtl83xx_fib_event: FIB_RULE ADD/DEL for IPv6 not supported
[   54.618123] RTL8380 Link change: status: 1, ports 400000
[   54.973209] rtl83xx-switch switch@1b000000 lan24: Link is Up - 1Gbps/Full - flow control rx/tx
[   54.983065] switch: port 24(lan24) entered blocking state
[   54.989158] switch: port 24(lan24) entered forwarding state
[   55.113729] RTL8380 Link change: status: 1, ports 100000
[   55.615210] rtl83xx-switch switch@1b000000 lan22: Link is Up - 1Gbps/Full - flow control rx/tx
[   55.625161] switch: port 22(lan22) entered blocking state
[   55.631403] switch: port 22(lan22) entered forwarding state
[  515.345073] RTL8380 Link change: status: 1, ports 100000
[  515.352007] rtl83xx-switch switch@1b000000 lan22: Link is Down
[  515.358679] switch: port 22(lan22) entered disabled state
[  515.367812] rtl83xx-switch switch@1b000000 lan22: Link is Up - 1Gbps/Full - flow control rx/tx
[  515.379328] switch: port 22(lan22) entered blocking state
[  515.385561] switch: port 22(lan22) entered forwarding state
[  516.334686] rtl83xx-switch switch@1b000000 lan22: Link is Down
[  516.341462] switch: port 22(lan22) entered disabled state

[ 55.631403] - is the last one after it fully started
Then the problem happens.
[ 515.345073] - I just unplugged one lan cable.

What else can I do to find out the issue?

6

I would like to push the topic again =)

Are there any other debugging option to find out or to limit the potential problem?

7

Is the console stuck too when the switch stops working? If not, then you have lots of debugging options there. Are links down? Can you do a link down/up? Does that change anything? Etc.

So far no one knows what you have tried. Which makes it a bit difficult for us to answer the question about what else you can do

8

Thank you!

The console seems to still work...

If I replug a lan cabel, I see it via console but doesnt work.

All what I tried I posted in this topic. I dont have deep knowledge about OpwnWrt and finding out the root cause. Restarting seems to fix it for the next 10 minutes.

9

No, there is no information about what you have tried and what the results were.

Only your conclusion:

Based on this, we can only assume that you have tested everything there is to test and that the conclusion is correct..