Is there something I should get worried about? In such a case, how to diagnose and fix?
I didn't notice that behavior prior upgrading to rc7. I'm using openwrt since 23.05.5.
I have not touched to the hardware/cables since openwrt was installed last september. Same cables that used to be perfectly OK.
Router is connected to a Netgear switch with a 18 inches ethernet cable. No wifi whatsoever. Router is ERL3.
root@OpenWrt:~# logread|grep link|grep connecti
Tue Jan 28 21:52:45 2025 daemon.notice netifd: Interface 'loopback' has link connectivity
Tue Jan 28 21:52:46 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Tue Jan 28 21:52:49 2025 daemon.notice netifd: Interface 'wan' has link connectivity
Tue Jan 28 21:52:51 2025 daemon.notice netifd: Interface 'wan6' has link connectivity
Wed Jan 29 17:11:20 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Wed Jan 29 17:11:22 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Wed Jan 29 19:14:17 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Wed Jan 29 19:14:18 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Thu Jan 30 10:10:50 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Thu Jan 30 10:10:52 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Jan 31 01:07:49 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Jan 31 01:07:51 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Jan 31 06:09:34 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Jan 31 06:09:36 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Jan 31 08:19:50 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Jan 31 08:19:51 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Jan 31 08:53:20 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Jan 31 08:53:21 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Jan 31 09:18:32 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Jan 31 09:18:34 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Jan 31 14:32:29 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Jan 31 14:32:31 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Jan 31 21:10:37 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Jan 31 21:10:38 2025 daemon.notice netifd: Interface 'lan' has link connectivity
(...)
okay, so we have a baseline, 23.05.5 worked.
so we can opkg install owut and run:
owut check --version-to
This should list every available 24.10.0 version you could install to test, so you could go owut download --version-to 24.10.0-rc6 and then owut install to install rc6 and see if the issue is still there.
If still there, you can go to rc5, rc4, rc2 and rc1 etc.
Looks like they might, @efahl any ideas if your USG commit has affected the Edgerouter Lite, as the makefile suggests both devices reference each other when being built.
Oops, might have talked too fast, sorry. Downgrading to rc5
OpenWrt 24.10.0-rc6, r28388-58d0057481
-----------------------------------------------------
root@OpenWrt:~# logread|grep link|grep connecti
Wed Feb 5 15:55:37 2025 daemon.notice netifd: Interface 'loopback' has link connectivity
Wed Feb 5 15:55:39 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Wed Feb 5 15:55:41 2025 daemon.notice netifd: Interface 'wan' has link connectivity
Wed Feb 5 15:55:42 2025 daemon.notice netifd: Interface 'wan6' has link connectivity
Fri Feb 7 15:01:16 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Feb 7 15:01:18 2025 daemon.notice netifd: Interface 'lan' has link connectivity
-----------------------------------------------------
OpenWrt 24.10.0-rc5, r28304-6dacba30a7
-----------------------------------------------------
root@OpenWrt:~# logread|grep "link connecti"
Fri Feb 7 15:58:40 2025 daemon.notice netifd: Interface 'loopback' has link connectivity
Fri Feb 7 15:58:42 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Feb 7 15:58:43 2025 daemon.notice netifd: Interface 'wan' has link connectivity
Fri Feb 7 15:58:54 2025 daemon.notice netifd: Interface 'wan6' has link connectivity
OpenWrt 24.10.0-rc5, r28304-6dacba30a7
-----------------------------------------------------
root@OpenWrt:~# logread|grep "link connecti"
Fri Feb 7 15:58:40 2025 daemon.notice netifd: Interface 'loopback' has link connectivity
Fri Feb 7 15:58:42 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Feb 7 15:58:43 2025 daemon.notice netifd: Interface 'wan' has link connectivity
Fri Feb 7 15:58:54 2025 daemon.notice netifd: Interface 'wan6' has link connectivity
Fri Feb 7 18:44:52 2025 daemon.notice netifd: Interface 'lan' has link connectivity loss
Fri Feb 7 18:44:53 2025 daemon.notice netifd: Interface 'lan' has link connectivity
-----------------------------------------------------
OpenWrt 24.10.0-rc4, r28211-d55754ce0d
-----------------------------------------------------
root@OpenWrt:~# logread|grep "link connecti"
Fri Feb 7 19:05:03 2025 daemon.notice netifd: Interface 'loopback' has link connectivity
Fri Feb 7 19:05:05 2025 daemon.notice netifd: Interface 'lan' has link connectivity
Fri Feb 7 19:05:06 2025 daemon.notice netifd: Interface 'wan' has link connectivity
Fri Feb 7 19:05:10 2025 daemon.notice netifd: Interface 'wan6' has link connectivity
It should not have changed anything, but for adding a new entry into the meta data in the img and profiles.json files (all the old names remain as-is). As far as I could tell from the code, both the USG and ERL are identical hardware in different cases.
The best way to go about this is to build from source and use 'git bisect'. That should narrow it down to the offending commit. I'd jump to RC1 immediately instead of trying all release candidates, if that's broken as well, then it's bisect. With 23.05 (.5?) as your last working commit.
There's no deep git knowledge required, all you need are commit hashes. Git bisect will ask you to mark a commit as good and another one as bad. Then its algorithm will start picking random commits to narrow the scope, which you again mark as bad (functionality broken) or good (working).
I don't think it's the USB drive. It's only a few months old and has been flashed manually from a dd image to recover from a mistake.
Could be eth0 though. When I was with Bell, eth0 was the WAN and I experienced quite a lot of 'waiting for PADO' timeouts. With my current ISP, WAN was now on eth2 and no timeouts with Ubiquiti firmware.
Now with Openwrt firmware, eth0 is back in business, serving LAN this time.
Coud very well be the config still. It's an old router model, I purchased it second hand on Ebay years ago, there are no real examples of working config available and I did all from scratch using a combination of manual edit and Luci.
Contemplating the possibility it's a hardware issue. Installed 24.10.0 and connected the LAN to eth2. Eth0 disconnected, unused. Required edits to /etc/board.json and /etc/boards.d/* in addition to /etc/config/network in order to get Luci not to display the default eth0 in the page https://router/cgi-bin/luci/admin/status/overview. Currently under test.