I've been trying out OpenWRT recently, and really appreciate all the features it grants, though I am suffering from an incredibly frustrating issue.
Every hour or so (it's a bit random, sometimes shorter, sometimes longer), I lose internet connection on all my devices for about 10 seconds, then it resumes like nothing happened.
I have tried many things to try resolve it:
Hardcoded DNS on client side
Changed DNS on luci
Disabled software offloaing
Downgraded to v22
The system logs show nothing when it occurs.
Nothing seems to resolve it. I don't think its a hardware issue as I don't have any problems on stock firmware.
Anyone suffering from same issue? / Any suggestions?
no error in the openwrt logs, would indicate the problem's elsewhere.
what if you run a continuous ping from a client to the router, and another
one from the router to say 8.8.8.8 ?
it'd be interesting to see what happens during one of the outages.
Not sure how I’d check this exactly, but definitely over a day. Also IP same before/after this issue so don’t think it’s this. (I’m with Virgin Media UK btw)
Thu Oct 13 19:17:22 2022 daemon.notice netifd: wan (2547): udhcpc: started, v1.33.2
Thu Oct 13 19:17:22 2022 daemon.notice netifd: wan (2547): udhcpc: sending discover
Thu Oct 13 19:17:22 2022 daemon.notice netifd: wan (2547): udhcpc: sending select for xx
Thu Oct 13 19:17:22 2022 daemon.notice netifd: wan (2547): udhcpc: lease of xx obtained, lease time 476933
You have a 132 1/2 hour lease. By design udhcpc should request a renewal at lease half-life (66 1/4 hours), and repeat the process if not acknowledged by the ISP again at the remaining half-life until the renewal request is honoured, or the timer runs out which can cause a momentary pause until the ISP obliges by renewing your lease.
The timestamps in the logread output should ideally indicate that your ISP is renewing the lease at the first half-life request. i.e my ISP provides a 4 hr lease and it’s renewed every 2 hrs.
Your router booted with a time stamp of the last configured /etc/config file stamped time and the udhcpc started before you ntp time server corrected the time.
A good clean reboot would clear up your log.
and if you issue this command from cli it will reset the lease from your ISP kill -USR1 $(pidof udhcpc)
Then a good clean reboot would clear up your log and refresh the lease fully.
Tue Feb 7 16:06:08 2023 daemon.notice netifd: wan (2210): udhcpc: started, v1.33.2
Tue Feb 7 16:06:09 2023 daemon.notice netifd: wan (2210): udhcpc: sending discover
Tue Feb 7 16:06:09 2023 daemon.notice netifd: wan (2210): udhcpc: sending select for xx
Tue Feb 7 16:06:09 2023 daemon.notice netifd: wan (2210): udhcpc: lease of xx obtained, lease time 469505