lemmi

Thank you for the fast response. I'd love to contribute, but I'm not that fluent with rust and the earliest I can take a closer look is in about a...

So now that I can finally come back to this, I noticed that the original issue with enabling `autoconf` was for IPv4-only setups. Wouldn't it make more sense to couple...

https://bugzilla.redhat.com/show_bug.cgi?id=2126243 concerns IPv4-only networks. Personally I see little harm in extra IPv6 addresses, but again, if a network is not supposed to be configured via SLAAC, maybe the better solution...

No problem. Then I'm going to first give this a spin myself for the moment.

I already mentioned it in the cleanup #27, but to clarify a couple bits here: The log above still shows what happens, but it kind of hides the actual issue....

I'm guessing the error was introduced during https://github.com/pyinfra-dev/pyinfra/commit/331865ae62e25aaf5faf3475b481160962e60a59?diff=unified&w=0#diff-706bc934a0c72456c2c88f6435770ac3fe3e13f1232f7b5464e7fcc44a9fd664L84-R88 The changes remove the `host.connection` attribute in favor of some other attribute including `host.connected`. Yet, [`server.reboot`](https://github.com/pyinfra-dev/pyinfra/blob/a7be6892254e32271aca92bca47fbf6edd18e743/pyinfra/operations/server.py#L89-L95) still makes use of `host.connection`: https://github.com/pyinfra-dev/pyinfra/blob/a7be6892254e32271aca92bca47fbf6edd18e743/pyinfra/operations/server.py#L89-L95...

So, just ran the checks locally and they fail with the same error. I'm not yet sure what to make of it, tbh.

@afbjorklund that was it, thank you.

@CameronNemo ping

I think it's a good idea. Since a few hosters started to charge extra for IPv4 addresses, I have a few machines that are happily running IPv6-only. They also happen...