!MthpOIxqJhTgrMNxDS:nixos.org

NixOS ACME / LetsEncrypt

91 Members
Another day, another cert renewal42 Servers

Load older messages


SenderMessageTime
4 Sep 2023
@raitobezarius:matrix.orgraitobezariusBut I don't think it's reasonable to block indefinitely something on the hope of seeing it merged13:56:38
@m1cr0man:m1cr0man.comm1cr0man
In reply to @raitobezarius:matrix.org
I do have the time and skills to bring it to systemd ;)
Oh neat okay, I didn't realize 😅
13:56:48
@raitobezarius:matrix.orgraitobezariusBut I cannot grasp the maintenance overhead merging this would create13:56:52
@m1cr0man:m1cr0man.comm1cr0manIt's added complexity to the Acme units. I've been pretty adverse to feature additions because it creates new failure scenarios and it's already got crazy feature creep but in this instance it's pretty important to have a rate limit and I've seen the effects of it first hand.13:58:14
@os:matrix.flyingcircus.ioosnyx (he/him)
In reply to @m1cr0man:m1cr0man.com
I think your arguments are solid. I'm not on board for waiting for systemd to add features (and your hammer saying is the same reason why). Like I said if you're willing to just be around to take questions or PR fixes into that portion of the module, I'm happy to see your one merged. I do think it is more complicated but I can live with that if I'm not the only one that understands how it works.

I would like you to copy over the test case I made though, to prevent future regressions
That's why I wanted to get another opinion of the team regarding not the only one that understands how it works.
I try to get the tests in this week.
13:59:31
@raitobezarius:matrix.orgraitobezariusI think it's fair that we set the "direction of the ACME module" to: we can welcome this feature and urge/usher into an era where systemd will provide it and we can decrease the complexity in the future14:05:46
@m1cr0man:m1cr0man.comm1cr0manI also want to upstream some stuff to Lego, so between the two hopefully complexity will fall over the next while.14:11:31
@os:matrix.flyingcircus.ioosnyx (he/him) * My take on the "let's solve it with systemd unit options alone" approach is just the idea that we must be careful to not fall into the when all you want to use is a systemd-253 hammer, everything looks like a unit option nail.
It might be a hammer you know, but that hammer bight also just be adding things to the evergrowing list of interwoven systemd unit relationships…
14:24:55
@os:matrix.flyingcircus.ioosnyx (he/him) changed their display name from Oliver Schmidt to osnyx (he/him).20:10:48
9 Sep 2023
@os:matrix.flyingcircus.ioosnyx (he/him) m1cr0man: Can you take care of the concurrency PR or shall I use the opportunity at NixCon to ask some other ACME team maintainers as well? 17:13:59
@m1cr0man:m1cr0man.comm1cr0manYo merge it? You need to ask like anyone with merge permissions. I approved it right?17:15:21
@os:matrix.flyingcircus.ioosnyx (he/him)
In reply to @m1cr0man:m1cr0man.com
Yo merge it? You need to ask like anyone with merge permissions. I approved it right?
I cannot see any approval, sorry.
17:40:41
@os:matrix.flyingcircus.ioosnyx (he/him)There's a merge party right now. So I guess I should still squash the commits I guess?17:41:07
@os:matrix.flyingcircus.ioosnyx (he/him)https://github.com/NixOS/nixpkgs/pull/24451117:41:33
@os:matrix.flyingcircus.ioosnyx (he/him)So in case I get your approval, I can certainly get it merged.17:42:26
@m1cr0man:m1cr0man.comm1cr0manAh well I can do that 17:43:06
@os:matrix.flyingcircus.ioosnyx (he/him)I'll squash the commits, give me a sec17:43:21
@os:matrix.flyingcircus.ioosnyx (he/him) m1cr0man: squashed an pushed 17:45:28
@m1cr0man:m1cr0man.comm1cr0manCool ok17:45:42
@os:matrix.flyingcircus.ioosnyx (he/him)Thanks a lot17:54:08
21 Sep 2023
@dedmunwalk:matrix.orgdedmunwalk joined the room.23:08:48
22 Sep 2023
@k900:0upti.meK900 changed their profile picture.09:53:38
25 Sep 2023
@os:matrix.flyingcircus.ioosnyx (he/him) m1cr0man: Hey, you probably want to close your alternative PR https://github.com/NixOS/nixpkgs/pull/246665 21:38:44
@m1cr0man:m1cr0man.comm1cr0manty for the reminder21:52:45
2 Oct 2023
@atra1n:matrix.orgTrain joined the room.01:42:45
@os:matrix.flyingcircus.ioosnyx (he/him)Hey, I might have run into a general issue with acme and nginx again, but wonder whether I am holding it wrong because nobody else has complained about it so far: Initial acme certificate generations fail for newly added vhosts when nginx has already been running12:33:31
@raitobezarius:matrix.orgraitobezariusI don't think I encountered this issue personally12:40:47
@raitobezarius:matrix.orgraitobezariusHow is it failing?12:40:59
@os:matrix.flyingcircus.ioosnyx (he/him) Background:
When changing the config file by e.g. adding new vhosts at switch time, nginx is not immediately reloaded and being made aware of the config changes. One of the reasons is that new vhosts might rely on certificate files yet to be generated by the acme subsystem. Reloading is thus triggered by nginx-reload-config.service.
Its dependencies are configured as such that it runs before the respective acme-domain-finished.target, but after the acme-domain.service renew service. That service though communicates with an acme registry and makes the registry fetch the validation response from nginx. With the config not being reloaded yet, nginx does not know the respective vhost and cannot serve a valid response.
12:45:48
@os:matrix.flyingcircus.ioosnyx (he/him) acme: error: 403 :: urn:ietf:params:acme:error:unauthorized 12:46:49

Show newer messages


Back to Room ListRoom Version: 6