Dnsmasq-script exception after upgrade

I had auto updates off for a while and enabled just when krack was announced of course. :slight_smile:

The update was made successfully according to the gui homepage.

But I think after that the forwarding to my other LAN DNS stopped working. DNS works in general but none of my LAN names are resolved because that happens at the LAN DNS that turris routes through when going outside.

I’ve worked around it by setting the upstream DNS manually.

Tried rebooting the turris. Did not want to try resetting yet because I did that a few days ago. What do you think?

2017-10-17T19:18:09+02:00 info dhcp_host_domain_ng.py[]: DHCPv4 new lease
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]: DHCPv4 new lease
2017-10-17T19:18:09+02:00 warning dhcp_host_domain_ng.py[]: Add_lease, hostname check failed
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]: Add_lease, hostname check failed
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]: Traceback (most recent call last):
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]:   File "/etc/resolver/dhcp_host_domain_ng.py", line 304, in <module>
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]:     dd.update_dhcp(sys_op, hostname, ipv4)
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]:   File "/etc/resolver/dhcp_host_domain_ng.py", line 233, in update_dhcp
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]:     log("DHCP remove old hostname [%s,%s]" % (op, hostname, ipv4), LOG_INFO)
2017-10-17T21:18:09+02:00 debug dnsmasq-script[7464]: TypeError: not all arguments converted during string formatting
2017-10-17T21:18:09+02:00 warning dnsmasq[7464]: script process exited with status 1

Hi,
This should be fixed in 3.8.5. :slight_smile:

Do you mean a coming release?

Yes, that’s a “crash” in the script that dynamically updates the LAN names.

Information a week ago was that it’s supposed to be fixed in 3.8.4, but it’s possible the KRACK created a shift in the releases. EDIT: better link: https://github.com/CZ-NIC/turris-os-packages/pull/39

1 Like

Until 3.8.5 is out, is there a short term workaround available for this?

Several times a day, randomly, for a few minutes, my ‘internet’ stops working meaning I can’t access stuff like google.com but I also can’t access 192.168.1.1 etc.

If I can’t access internal addresses, it’s not a WAN fault obviously but rather, a fault with name resolution or something I guess. And everytime this happens, I look at my Luci System Logs and see a tonne of dnsmasq-script related issues so I’m guessing my problem is similar to OP?

That sounds like a different problem. See Kresd writing to disk every 10 minutes

1 Like