[11:35] jugmac00: launchpad seems to be in trouble, I seem to get occasional errors from chrome: [11:35] The connection to launchpad.net is not secure [11:35] You are seeing this warning because this site does not support HTTPS. Learn more [11:35] Can't say for sure what's going on [11:35] Did not have developer tools open [11:36] but it at least happened twice [11:42] What this means is that it's not responding on https port, 3 times now, so single-digit percentage [11:44] mtr --tcp --port 443 looks clean hmm [11:51] I don't see anything very alarming in our graphs [11:53] Not seeing any connection problems from home either [11:57] hmm [11:57] So I guess mtr is not super helpful as it will always mtr the same endpoint? [11:57] 600 consecutive requests so far, no errors [11:58] You might have broken IPv6 or something [11:58] it might be a backbone issue [11:59] but it's hard to diagnose [11:59] There are two haproxy frontends, each with an IPv4 and an IPv6 address [12:00] But they're in the same DC so should be similar from a backbone perspective [12:00] one on nutmeg, one banana something [12:01] yes, banana and nutmeg [12:01] lo.annegrit.canonical.com loses packets [12:01] with mtr --tcp --port 443 [12:02] 65% on the route to nutmeg [12:02] reply times up to 7s [12:06] copied mtr to IS [12:08] looks plausible - my mtr doesn't go that way so I didn't see it [13:28] @cjwatson Thanks for jumping in! [19:32] I'm trying to make a functionally "empty" ppa --- an apt repository that might hold some package to be used by already-configured clients, but normally holds nothing. The only way I can think to do this is to publish a "placeholder" package for each of the distributions I want to support. Is there a better way to do this? [19:33] I think I've heard that it doesn't "really" exist until you publish a package there, yeah [19:46] you can publish a placeholder package, and delete it right away [19:47] oh handy :) [19:47] maybe there's a better way though, I'll let the experts answer :) [20:03] you can usse the 'mark-suite-dirty' script in ubuntu-archive-tools [20:03] Yeah, what RikMills said is what I was just going to say [20:04] It takes an --archive argument - see its --help [20:05] :D [21:09] nice :) [21:10] RikMills: neat! how do I query which ones are currently "dirty"? [21:16] realtime-neil: Why do you need to? [21:16] to avoid doing something unnecessary [21:17] Doesn't matter, the operation is idempotent [21:17] cjwatson: okay, understood [21:17] The information isn't exposed at the moment [21:17] cjwatson: what kind of auth is this thing doing? and how do I know if it's using the wrong creds? [21:18] https://help.launchpad.net/API/launchpadlib [21:18] If it's using the wrong creds you'll get an error ... [21:19] cjwatson: is it the same error for a nonexistent archive. Because I'm seeing "No such archive". [21:19] You probably typoed somehow [21:19] What was your invocation? [21:20] cjwatson: `$ ./mark-suite-dirty --archive=ppa:realtime-robotics/ros --suite=bionic` [21:20] That doesn't match the --help [21:21] needs to be ppa:realtime-robotics/ubuntu/ros [21:22] But also that archive already has actual packages published to bionic [21:22] I mean you can do it and it's harmless, but are you sure that's what you meant? [21:22] cjwatson: yep, I'm just trying to do a no-op for now [21:22] OK [21:24] cjwatson: I'm seeing a slightly different help message: https://paste.debian.net/1219778/ [21:24] Or, more likely, I don't understand the usage message conventions of this repo