only first name server attempted in finding the puppet server

[rtprio]

Describe the Bug

I recently noticed all my puppet agents were failing when the primary DNS resolver was out. This was on FreeBSD.

Expected Behavior

I would expect that puppet would use the next nameserver in the list to look up the address of the puppet server.

Steps to Reproduce

Steps to reproduce the behavior:

1. Configure a host with two resolvers, the first one nonfunctional
2. Run the agent, I used puppet agent -t -v --logdest console
3. observe an error similar to the following

puppet-agent[29201]: Could not retrieve catalog from remote server: Host is down - sendto(2) for "10.0.17.2" port 53

Environment

• Version 8.7.0
• Platform FreeBSD 14.1

Additional Context

Add any other context about the problem here.

[joshcooper]

@rtprio can you provide more details about how you configured a "host with two resolvers, the first one nonfunctional"? Also does the hostname you're connecting to have multiple IPv4 addresses, or does it have both IPv4 and v6? Could you also add --trace and include the backtrace in the description.

[rtprio]

sure thing;

the resolv.conf of the host was like this:

nameserver 10.0.17.2
nameserver 10.0.17.4

In this case, .2 crashed and was no longer answering DNS. These hosts only have a single ipv4 at the moment, I have not tested the ipv6 side of this condition.

[rtprio]

I suppose I should have investigated deeper when it was happening; It seems to be properly using the next nameserver now. 🤷

[joshcooper]

Thanks for letting us know. I'm going to close this, but if you can reproduce feel free to reopen.