EdgeRouter X hanging occasionally

[dbrb2]

Hi,
I have openWRT running on an EdgeRouter X
It seems to be working fine - with one large caveat:

About once every ~12 hours it hangs, won't respond to pings, and needs a power cycle.
Any ideas what could be causing this - or how I could try to find out?

|Model|UBNT-ERX|
|Architecture|MediaTek MT7621 ver:1 eco:3|
|Firmware Version|OpenWrt 19.07.4 r11208-ce6496d796 / LuCI openwrt-19.07 branch git-20.247.75781-0d0ab01|
|Kernel Version|4.14.195|
|Local Time|2020-11-12 14:17:51|
|Uptime|0h 7m 54s|
|Load Average|0.14, 0.25, 0.13|

[AshkaliN]

Hello everyone, im new to this forum
On topic, im ALSO using ER-X flashed with OpenWrt 19.07.4 r11208-ce6496d796 / LuCI openwrt-19.07 branch git-20.247.75781-0d0ab01, Kernel Version 4.14.195 and after playing around with it for about a month, flashing different version, with this one i loose connection to the router ( no ping to my router from PC with LAN cable) every now and then. This happens only if i have HW OFFLOADING Enabled. IF i use SOFTWARE Offloading, everything is working with no disconnects. Any suggestions?

[dbrb2]

Interesting. Sounds like the same symptoms as me - I am also using HW offloading - but hadn't nailed that down to being the cause.

I used the upgrade method provided by @stman

Operational OpenWrt 19.07.2 Factory Tar file for EdgeRouter X

Here is the link to my repository where you can find my custom made "short receipt" "OpenWRT 19.07 factory tar file" called "openwrt-ramips-mt7621-ubnt-erx-initramfs-factory.tar" You can flash it directly over any EdgeOS, and then replacing it by sysupgrading the router again with a usual "OpenWRT official stock image", like the lastest 19.07.2 sysupgrade stock bin file found here : http://downloads.openwrt.org/releases/19.07.2/targets/ramips/mt7621/openwrt-19.07.2-ramips-mt7621-ubnt-erx-squ…

Which seems to have worked fine. However, HW offloading is my primary reason for wanting to use the ERX - so if anyone has any ideas that would be much appreciated!

For now I'm using a pi4 instead - no hardware offloading, but comporable speeds, presumably due to its CPU grunt

[pwned]

There is a "main" thread about this topic already.

Netfilter "Flow offload" / HW NAT

Looks like Linux is finally getting hardware NAT support: http://patchwork.ozlabs.org/cover/852523/ More information here: https://lwn.net/Articles/738214/ The code is new so no drivers will support hardware offloading off the bat. Still, the code is 2.5 times faster than the regular software path. The patch applies cleanly to 4.14 which seems to be the next LEDE/OpenWrt kernel version.

Beside that there is also:

Flow_offloading=1 is broken on latest snapshot (4.19 issue)

Archer C7 v2, OpenWrt SNAPSHOT, r10451-653e05d27f Flow offload seems to be broken on 4.19. root@router:~# lsmod | grep flow nf_conntrack 71817 18 ipt_MASQUERADE,xt_state,xt_nat,xt_conntrack,xt_REDIRECT,xt_CT,nft_redir_ipv4,nft_redir,nft_nat,nft_masq_ipv4,nft_masq,nft_flow_offload,nft_ct,nf_nat_ipv4,nf_nat,nf_flow_table,nf_conntrack_rtcache,nf_conntrack_netlink nf_flow_table 14399 6 xt_FLOWOFFLOAD,nft_flow_offload,nf_flow_table_ipv6,nf_flow_table_ipv4,nf_flow_table_inet,nf_f…

As far as I remember there where issues in the past IPv6/HW offloading releated. I don't know it this has been fixed already. Maybe it is worth to try a snapshot build to verify. But be aware you will lose LuCI-VLAN configuration. You have to work out a DSA setup manually.

[dbrb2]

Cheers. I tried a snapshot on the ERX last week which seemed to brick it - I had to do a TFTP recovery - but I might try again this weekend if I have time

Luci warns me when I do that the name has changed between release and snapshots, from erx:

http://downloads.openwrt.org/releases/19.07.4/targets/ramips/mt7621/openwrt-19.07.4-ramips-mt7621-ubnt-erx-squashfs-sysupgrade.bin

to edgerouter-x:

https://downloads.openwrt.org/snapshots/targets/ramips/mt7621/openwrt-ramips-mt7621-ubnt_edgerouter-x-squashfs-sysupgrade.bin

Not sure how significant this is. I can force the upgrade anyway....but when I did this, I bricked the unit

[LGA1150]

MT7621 snapshot has switched to DSA, your previous Ethernet configuration is incompatible.

[dbrb2]

Ah - yes OK. So I have to do the upgrade but /not/ keep the settings - then it should hopefully come back up?

The forums seem to suggest HW-NAT may not be supported under DSA yet...but I could be wrong. Will give it a go

[pwned]

@dbrb2

Ah - yes OK. So I have to do the upgrade but /not/ keep the settings - then it should hopefully come back up?

Yes, you cannot keep settings.

@dbrb2

The forums seem to suggest HW-NAT may not be supported under DSA yet...but I could be wrong. Will give it a go

Actually it seems to be it is working on my EdgeRouter (running snapshot). But I did only "test" it by enabling HW offload and watching cpu usage during full speed downloads to verify HW offload is working (100% vs. 0% usage). I cannot "proof"/verify this by other precede (because I don't know other). Beside that I've disabled IPv6 on my device.

[dbrb2]

Ah OK. I'll try it later - I flashed sysupgrade. First flash failed. Second seems fine.
However, WAN<>LAN traffic (PPPoE on WAN) drops from ~930Mbps to 300Mbps - suggesting a problem with hardware offloading.

No worries - the pi4 works fine