Flint dhcp issue

Technical Support for Routers
1

I’m getting the same issue that described here

system log on flint:
Sat Dec 4 22:33:40 2021 daemon.notice netifd: dhcp (3460): udhcpc: lease lost, entering init state
Sat Dec 4 22:33:40 2021 daemon.notice netifd: Interface ‘dhcp’ has lost the connection
Sat Dec 4 22:33:46 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending select for X
Sat Dec 4 22:33:46 2021 daemon.notice netifd: dhcp (3460): udhcpc: lease of X obtained, lease time 600
Sat Dec 4 22:33:46 2021 daemon.notice netifd: Interface ‘dhcp’ is now up
Sat Dec 4 22:33:46 2021 user.info mwan3rtmon[3306]: Detect rtchange event.
Sat Dec 4 22:33:46 2021 user.notice firewall: Reloading firewall due to ifup of dhcp (eth0)
Sat Dec 4 22:38:46 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to X
Sat Dec 4 22:38:46 2021 daemon.notice netifd: dhcp (3460): udhcpc: send: Network unreachable
Sat Dec 4 22:38:46 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:41:16 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:42:31 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:43:08 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:43:26 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:43:36 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:43:40 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:43:42 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:43:43 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending renew to 0.0.0.0
Sat Dec 4 22:43:43 2021 daemon.notice netifd: dhcp (3460): udhcpc: lease lost, entering init state
Sat Dec 4 22:43:43 2021 daemon.notice netifd: Interface ‘dhcp’ has lost the connection
Sat Dec 4 22:43:49 2021 daemon.notice netifd: dhcp (3460): udhcpc: sending select for X
Sat Dec 4 22:43:49 2021 daemon.notice netifd: dhcp (3460): udhcpc: lease of x obtained, lease time 600
Sat Dec 4 22:43:49 2021 daemon.notice netifd: Interface ‘dhcp’ is now up

I’m using last release version 3.207
thanks

2 Likes
2

This bug provokes udp connection “disconnect” like for example i’m getting constant hangs (like about for 3 sec each 6-10 min?) while using google meet and bc of this i can’t really use this router and have to fallback gl-mifi with last openwrt instead. So i guess i’ll need to try other dhcp client with flint or some new version of udhcpc? In the thread (url above) you could find link to github issue (udhcpc: no MSG_DONTROUTE when sending packet by adiov · Pull Request #1017 · openwrt/openwrt · GitHub) which supposedly solves the problem.

1 Like
3

Thank you very much for your feedback.

  • You can update busyBox below.
    rm /usr/bin/flock && opkg install busybox_1.28.3-4_ipq.ipkbusybox_1.28.3-4_ipq.zip (279.7 KB)
4

I flashed last snapshot qsdk-ax1800-3.208-1213.img then installed this package and it works!
Thanks!

5

Can you confirm that the problem of Google meet hang is fixed using this patch?

6

Sure, i can confirm tomorrow. But now it’s able to renew dhcp without issues

1 Like
7

I had no issues yesterday with google meet so i think it’s fixed now

8

This is what I had been reported before as the internet offline and online within few seconds. Could you please guide me how to flash the busybox? Or the snapshot 3.208 can resolved the issue?

9

You can install the package openssh-sftp I believe and then connect via filezilla to your router via sftp.

Or maybe you can use curl <url> --OUTPUT package.ipk and then follow the commands.

–

Also running 3.208 here and I can confirm this firmware and the newer busybox update works very stable, no more dhcp dropping here.

The only issue I found was mwan3 causing sometimes problems with dropping aswell especially on different interfaces, I was searching as for why in the issue tracker on openwrt.

As from what I understand is that mwan3 sometimes chooses the wrong wan interface especially when the configuration is not properly set.

What I mean with not properly is that each interface such as vlan interface for example or a isolated routed AP as interface, shares metric 0 or 10, this could cause unpredictfull behaviour, if I would fill in the gateway in one of the interfaces luci shows under stats the vlan as wan, the issue got solved when I made the metrics invidual different for each, I noticed far more less dropping from both stubby and vlans by doing this change.

Im not sure if theres any upstream patches for mwan3 for this, but it could be great.

10

I tried 3.208 and NextDNS integration seems to broken for me.
In the 3.208 firmware, if I select NextDNS and give in my id, dns request are not using NextDNS.

11

You could use scp
scp -oHostKeyAlgorithms=+ssh-rsa busybox_1.28.3-4_ipq.ipk root@192.168.8.1:/tmp
this will upload it to /tmp of the router
then ssh into router
ssh -oHostKeyAlgorithms=+ssh-rsa root@192.168.8.1
and
cd /tmp
rm /usr/bin/flock && opkg install busybox_1.28.3-4_ipq.ipk

12

I did steps below, but the problem remains, however I am not sure if there is any improvement or not.

13

I’m login to ssh then command rm /usr/bin/flock && opkg install busybox_1.28.3-4_ipq.ipk but failed. Or how to attached the * unzip pnSH0upWHkYChbZtETT91jKzWwN.zip