What are some of the key points to calling a poudriere server wtih pkg over HTTPS?
I set up a poudriere server that I'd like to call through a dynamic dns on a site that uses https with a cert from Let's Encrypt. I notice that a lot of the pkg examples seem to be geared toward http. When I try to use https as the protocol in the url property of the config file for the pkg repos, I end up with an authentication error. This is alerting on the web server's cert; it's not alerting on the package signing key for poudriere. When I set up for HTTP, the call times out in pkg. When I look at fetch (in general), it seems like we could ask to turn https off; but, I'd like to get the authentication done. Many of the examples have used certs generated by other means. Since the Let's Encrypt cert works with a browser, I think I'd like to get pkg to use that one to verify the site.
I consulted references like these:
In cases like this, would we be expected to import a copy of that cert to the calling client? That seems a little much. Is this what the "srv" and "fingerprint" options are for? Is there a file I should derive from the PEM on the website cert to share with the calling client? Do you have advice or a tutorial recommendation that could show how to set up pkg on the calling client? Thanks.
I set up a poudriere server that I'd like to call through a dynamic dns on a site that uses https with a cert from Let's Encrypt. I notice that a lot of the pkg examples seem to be geared toward http. When I try to use https as the protocol in the url property of the config file for the pkg repos, I end up with an authentication error. This is alerting on the web server's cert; it's not alerting on the package signing key for poudriere. When I set up for HTTP, the call times out in pkg. When I look at fetch (in general), it seems like we could ask to turn https off; but, I'd like to get the authentication done. Many of the examples have used certs generated by other means. Since the Let's Encrypt cert works with a browser, I think I'd like to get pkg to use that one to verify the site.
I consulted references like these:
- https://www.digitalocean.com/commun...r-freebsd-servers#configuring-package-clients
- https://www.freebsd.org/doc/handbook/pkgng-intro.html
- https://www.freebsd.org/doc/en/books/porters-handbook/testing-poudriere.html
- https://www.freebsd.org/cgi/man.cgi...opos=0&manpath=FreeBSD+12.1-RELEASE+and+Ports
- https://www.freebsd.org/cgi/man.cgi...opos=0&manpath=FreeBSD+12.1-RELEASE+and+Ports
- https://www.freebsd.org/cgi/man.cgi...opos=0&manpath=FreeBSD+12.1-RELEASE+and+Ports
- https://www.freebsd.org/cgi/man.cgi...opos=0&manpath=FreeBSD+12.1-RELEASE+and+Ports
- https://www.freebsd.org/cgi/man.cgi...opos=0&manpath=FreeBSD+12.1-RELEASE+and+Ports
- https://github.com/freebsd/poudriere/wiki/pkg_repos
In cases like this, would we be expected to import a copy of that cert to the calling client? That seems a little much. Is this what the "srv" and "fingerprint" options are for? Is there a file I should derive from the PEM on the website cert to share with the calling client? Do you have advice or a tutorial recommendation that could show how to set up pkg on the calling client? Thanks.