daemon.warn dnsmasq[15385]: Maximum number of concurrent DNS queries reached (max: 150)
This is the only log that reports a problem, which was solved by restarting the DNS. This has been happening for days, even the sites I'm visiting are freezing, but I've had the same configuration for 6 months and haven't changed it because it works well. My provider is NextDNS, which I've had for 6 months, and in any case, the solution of restarting the cache seems to rule that out. Does anyone know how to fix this? I don't really want to reset...
[edit: there is another frequent error log, which I did not report because I have seen other instances of curl version conflicts with dnsmasq: “user.err ddns-scripts[16367]: glddnsv6: cURL Error: ‘7’”. I would like to add that I do not use VPN on the router, so I do not use dnsmasq for splitting or other rewrites]
To be clear, the problem is this: a few times a day, I browse the open tabs, but when I open new ones it doesn't find the DNS.
Resetting the cache (turning off and on again an option from the GUI) works.
It's been like this for days, without having touched anything for months, the only error logs regarding DNS are the ones above, I don't use DNSmasq for PBR.
Please, does anyone have an idea what could have caused this, when fully operational?
It is difficult to say, but what I would do is first make a full backup via luci, you can bring this back unless you have some special config from other packages things like avahi are not backedup, only contents in /etc/config
Then factory reset, and then only setup nextdns nothing else, if it then happens it should be a bug.
But to me it sound like a misconfiguration.
It can be a issue in combination with rebind protection, you access a top level domain with a local ip, and for some reason causes a loop filling the max concurrent count, maybe you inserted dns settings double somewhere rather than the router ip.
This seem also a little odd to me, did you install luci app for ddns?, it is odd because I cannot remember GL-iNet ddns using ddns-scripts package, if this true it is a third party and may clash with internal ddns scripts edit: not odd I just verified myself on my Brume3.
There is also mentoin of ipv6, but do you have ipv6 and also a working ipv6 dns?
Maybe it is worth to turn ipv6 off, and restart devices in your network to ensure no-one has still a gua set and it is expired, maybe that is the issue that ipv6 expires and this mismatched signals break dnsmasq concurent queries for some reason.
And the reason dns fails is because it tries to resolve on ipv6 which is non existant/expired, this often happens if the prefix is not correctly assigned or there is no block space available anymore, sometimes isp give really small ip blocks and sometimes you may need to use NAT6 even.
I have a backup of the entire overlay, last restored 6 months ago, with the current firmware, current configuration, and no problems. So it wouldn't be a problem, aside from some subsequent customizations, like changing passwords, etc. For this reason, I'd prefer not to use it,
and thank goodness, because nothing would have changed. You're right, something has changed with IPv6, because I recently reset the main router that handles only that (ISP IPv6 MAP). Your reasoning is clear: when a DNS6 query occurs, it retries in a loop and crashes, so I'm marking it as a solution and will proceed with resetting the NAT6.
Thanks, in any case, the DDNS on Beryl AX is integrated, even in IPv6, provided by the GL servers.