Rock 5b, 16Gb: after some runtime the network hangs

[rolsch]

Hi rock5b user 🙂

 

My sbc hangs after 8-14 hours on the network.

 

no ping from inside or outside works

 

# ip link 

- show eth0 up

 

after the commands

ip link set eth0 down
ip link set eth0 up
- the network worgs again,

 

https://paste.armbian.com/tiqabekoje
 

- any hints for this issue?

 

 

[eselarm]

already in the first dmesg part I see:

[54475.779558] r8169 0004:41:00.0 eth0: Link is Down
[54479.115749] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[54479.115748] r8169 0004:41:00.0 eth0: Link is Up - 2.5Gbps/Full - flow control off
[55390.928391] br-ce724241dce3: port 1(veth1664549) entered disabled state
[55390.928949] veth6932af4: renamed from eth0
[55391.020157] br-ce724241dce3: port 1(veth1664549) entered disabled state
[55391.021811] device veth1664549 left promiscuous mode
[55391.021831] br-ce724241dce3: port 1(veth1664549) entered disabled state
[55391.493933] br-ce724241dce3: port 1(veth1a65439) entered blocking state
[55391.493954] br-ce724241dce3: port 1(veth1a65439) entered disabled state
[55391.494085] device veth1a65439 entered promiscuous mode
[55391.500502] br-ce724241dce3: port 1(veth1a65439) entered blocking state
[55391.500519] br-ce724241dce3: port 1(veth1a65439) entered forwarding state
[55391.539757] eth0: renamed from vethb9b9ca0
[55391.561925] IPv6: ADDRCONF(NETDEV_CHANGE): veth1a65439: link becomes ready

Which means to me that you use containers/docker or so and something on that networking level goes wrong. I think it had nothing specifically to do with Armbian nor ROCK5B.

 

You should figure out what is running on your computer and what is done to networking in general. 

 

I do not use containers like seems to be done here, so cannot really help here. I have no clue about your networking setup and plans.

I use various bridges and VLANs on my ROCK5B (Armbian Trixie), but all strict manually done by myself, only own files in /etc/NetworkManager/system-connections/

[rolsch]

...or maybe an r8169 driver problem under heavy load

dmesg -w | grep -i r8169
[    5.072779] r8169 0004:41:00.0: enabling device (0000 -> 0003)
[    5.115535] r8169 0004:41:00.0 eth0: RTL8125B, 00:e0:4c:03:00:fc, XID 641, IRQ 166
[    5.115550] r8169 0004:41:00.0 eth0: jumbo features [frames: 9194 bytes, tx checksumming: ko]
[    6.328634] RTL8226B_RTL8221B 2.5Gbps PHY r8169-4-4100:00: attached PHY driver (mii_bus:phy_addr=r8169-4-4100:00, irq=MAC)
[    6.518776] r8169 0004:41:00.0 eth0: Link is Down
[    9.721015] r8169 0004:41:00.0 eth0: Link is Up - 2.5Gbps/Full - flow control off
[ 9717.405068] r8169 0004:41:00.0 eth0: NETDEV WATCHDOG: CPU: 3: transmit queue 0 timed out 5227 ms
[ 9717.405345] r8169 0004:41:00.0 eth0: ASPM disabled on Tx timeout
[ 9717.418437] r8169 0004:41:00.0 eth0: rtl_rxtx_empty_cond == 0 (loop: 42, delay: 100).
[ 9941.759110] r8169 0004:41:00.0 eth0: Link is Down
[ 9945.329364] r8169 0004:41:00.0 eth0: Link is Up - 2.5Gbps/Full - flow control off

 

I have a frigate container running on this sbc.
A second Rock 5b (same os version) with 10 docker images but very low traffic run smoth on the same switch with 2.5Gbps.

 

Could someone check which driver Radxa OS uses?

Edited 10 minutes ago by rolsch