Interested in joined effort + armv32/armv64 builds about docker-pdns HOT 6 CLOSED

Hey there! Thank you very much for your willingness to contribute to this repo. All of what you have proposed makes sense to me and I'll gladly accept your contributions. However, your PR is just too big to do a meaningful review - could you please split it up logically into multiple PRs?

What is really important to me - the updates cannot break anything for existing users. I hate it when it happens to me, so let's try to avoid it as possible for users of these images.

Regarding the 0x97 vs ngoduykhanh - we can move completely to the ngoduykhanh's repo as it seems to be more active with more features. Just again, let's try to make the update for existing users easy.

Thanks again and I'm looking forward to our cooperation!

from docker-pdns.

Hey!

I think regarding your users, they use your docker-images and not the repo layout so i would suppose any changes to that are uncritical. I cannot split the effort here since things are related. I cannot split the movement to template-generated Dockerfile and then do the arm based builds - one is a dependency of the other.

Even though i introduced a new build system, its not replacing anything you have and it does not touch your core Docker images you actually build.

The only changes which we could discuss which could potentially break anything, but i build them to do so are:

a) I did not change any configuration settings or variable names or defaults for the docker images, only introduced 2 new (missing) variables which do default to the old default.

b) i changed your Core Dockerfile to no longer use CMD only, but use a entrypoint based setup, this said, this has been done to exactly match the old behaviour and should behave the same entirely - its just encapsulation

c) i removed your alpine, based uwsgci builds - but it was broken anyway - nobody could have ever used it

d) i removed the pdns-admin folder / image - again, it was broken, it could not even have been build, so nobody could have used it.

I fully understand why you do not want to break things and i designed anything like that since i hate it myself too.

I tried to do the necessary house keeping and keeping anything sane for users - i hope so.

That said, splitting all that into smaller junks is nothing i would invest time into. This was a project i invested way too much time already and also move on to using unbound due to the "heaviness" of this setup anyway. So i am still interested in participating, but i guess this bigger junk needs to be "eaten alive" first.

I am fine if you do not consider this and its too big - its big - no doubt. It is too big, but splitting it would be a horrible amount of work i am not willing to put into - sorry. Not for such simple changes in it :)

If you cannot go by that, no worries, close that, ditch the PR, i am fine. I will / can build my own images there anytime, so no worries. Thanks!

from docker-pdns.

Thanks for your honest answer. I'll see what I can do :-)

from docker-pdns.

Is there any updates on this? I'm interested in having a arm64 build, and I can send a PR on this.

from docker-pdns.

We now have linux/arm64 image variants available in docker hub for pdns-mysql and pdns-recursor images 🎉

pdns-admin is not available as linux/arm64 for now, as the flask asset build timeouts the job in Github CI on arm64 VM. For now I don't see another reasonable way to go around this, but the pdns-admin project is now "under new management" and the development seems to be active again, so maybe the situation will be improved in the near future.

from docker-pdns.

There's now a linux/arm64 version of docker image for pdns-admin as well. I'm considering this now done, even though it was originally also about armv32, but I don't think that's important right now.

from docker-pdns.