I’m currently running Selenium in Debian 12 docker containers. There are 5 servers: 1 manager and 4 workers using docker. Quite often the first set of selenium tests will fail because the container doesn’t have internet access, so it can’t connect to the test site. Some or all of the servers can have this issue at the same time, but they do not always/consistently have the issue based on any discernable pattern. They will also magically resolve the issue. There will also be times where one automated test run is having problems connecting and a separate test run is working fine. Network manager says there are no error messages on the two DNS servers. I was previously using Centos 9 and didn’t have this problem.
The server error message is:
msg=“[resolver] failed to query external DNS server” client-addr=“udp:172.18.x.xx:45568” dns-server=“udp:10.yy.yy.yyy:53” error="read udp 172.18.x.xx:45568->10.yy.yy.yyy:53
/etc/resolve.conf file (verified not a link):
Generated by NetworkManager
search company.domain
nameserver 10.yy.yy.yyy
nameserver 10.yy.yy.zzz
Things I’ve tried with no improvements:
Hardcoding DNS addresses in /etc/docker/daemon.json
Downgrading docker
Upgrading docker
Restarting docker
Restarting servers