The DNS server stops working every couple of hours

[ElliottDCooper]

The DNS server stops working every couple of hours (does not respond to DNS queries) and it continues to work after a restart. Everything works properly after the restart.

I am using docker image ghcr.io/mosajjal/sniproxy:v2.0.3 with default configuration + domain acl. The log level is info and there is nothing in the logs (like no one sends any DNS queries).

[mosajjal]

been running sniproxy on my machine for a few hours with no issues. what are you using as upstream proxy? maybe change that to something like this as a test?

upstream_dns: udp://8.8.8.8:53

[ElliottDCooper]

Thanks for replying.
Yes, I'm using udp://8.8.8.8:53. Actually, I haven't seen the issue in the last three days, but it was happening before. No config changed. Maybe because there was less load on the system in the last three days.

[ElliottDCooper]

I used a script that performs a lot of DNS queries (like a query per 5ms for 5 minutes), and it happened again.
The CPU and memory usage are normal.

CONTAINER ID   NAME                          CPU %     MEM USAGE / LIMIT     MEM %     NET I/O           BLOCK I/O        PIDS
xyz   sni-proxy-1       0.00%     5.438MiB / 1.883GiB   0.28%     328kB / 258kB     246kB / 8.19kB   5

Output of a sample dig:

$ dig @serverip google.com

; <<>> DiG 9.16.1-Ubuntu <<>> @serverip google.com
; (1 server found)
;; global options: +cmd
;; connection timed out; no servers could be reached

[mosajjal]

I've made some changes in DNS (still work in progress) in #63, feel free to install it and see how that works. doesn't support DNS over SOCKS5 proxy though (for now).

go install github.com/mosajjal/sniproxy@f07c7f317a3634797ad6c5768e3a77fb905f6cb1

[ElliottDCooper]

I switched to this version. I haven't tested yet.

[ElliottDCooper]

Hi,
I used version f07c7f317a3634797ad6c5768e3a77fb905f6cb1, and it happened again. The DNS server stopped working and it continued to work after a docker restart.

dig's result when it's not working:

$ dig @serverip google.com
; <<>> DiG 9.16.1-Ubuntu <<>> @serverip google.com
; (1 server found)
;; global options: +cmd
;; connection timed out; no servers could be reached

The server's log is kinda frozen too.

[almdandi]

I would like to add to this problem. I experienced today the same problem. I first tried to ask the dns server for google.com über nslookup from my windows pc. I did saw the dns traffic from my windows pc to the sniproxy but no outbound dns traffic ware made. I then checked if the docker host system can lookup domains, and dns worked. After that a spawned a bash shell in the container uns tried to make a http request with curl and this also worked. At last i just restarted the container and that solved the dns problem. I ran the sniproxy in DEBUG mode, so i can provide the log but there ware no errors in to logs. The only thing i noticed was the a missing INFO line in the logs INF dns.go:91 > returned origin address for fqdn google.com. and rtt 38.296349ms service=dns was missing. After the restart it showed up again.

[mosajjal]

I'll consider this issue closed due to the changes I've made to the proxy. if the problem still exists, please re-open the issue