| DarkTrick | sarnold, Well.. I never did package or stuff... I expected to deb files to have some instructions what to do at installation time | 02:17 |
|---|---|---|
| alkisg | Hi, is there a problem going on with the mirrors? From one ISP, http://us.archive.ubuntu.com/ shows a directory "ubuntu", and within it, the archive folders like dists, pool | 07:51 |
| alkisg | From another ISP, the same URL shows the archive folders dists and pool directly, with no ubuntu directory | 07:52 |
| alkisg | The result is that `apt update` fails at least in half of Greece, is this a global problem? | 07:52 |
| blahdeblah | alkisg: Can you run `host us.archive.ubuntu.com` (or nslookup or dig if you don't have host installed) from both places? | 08:00 |
| alkisg | In one: 91.189.91.38, 91.189.91.39, 2001:67c:1562::18, 2001:67c:1562::15 | 08:01 |
| alkisg | The other is the same, but reversed, 39 first | 08:02 |
| alkisg | Trying to force .39 to one... | 08:02 |
| alkisg | That works | 08:02 |
| alkisg | So .38 appears broken | 08:02 |
| blahdeblah | alkisg: They both look identical from here: https://pastebin.ubuntu.com/p/mRYd2gKfTg/ | 08:07 |
| alkisg | Hrm, it was fixed here too now. Maybe the problem was fixed a few minutes ago... testing more... thanks btw | 08:09 |
| blahdeblah | Maybe someone is messing with your http streams? Or your DNS, selectively? | 08:12 |
| alkisg | Eh, it had the complete archive, but outside the ubuntu/ directory, so I doubt it. Maybe it was some mirror synchronization glitch. It appears to be fine now. | 08:32 |
| RAOF | Hm. Is there an easier way of testing old versions of the toolchain than wandering over to https://launchpad.net/ubuntu/hirsute/+source/binutils, selecting the relevant release, and then manually downloading stuff? | 09:09 |
| laney | RAOF: there's pull-lp-debs, might be marginally easier | 09:13 |
| RAOF | Oh, that's nicer, yeah | 09:14 |
| RAOF | Pity launchpad isn't exposed as snapshots.archive.ubuntu.com 😀 | 09:14 |
| laney | now I'm imagining an ubuntu bisection tool | 09:14 |
| laney | creates a new ephemeral container for each iteration | 09:15 |
| RAOF | I wish to subscribe to your newsletter. | 09:35 |
| RAOF | Boo. pull-lp-debs apparently times out in canonistack. | 09:35 |
| alkisg | blahdeblah: could you visit http://ports.ubuntu.com/ just now? It doesn't have the "ubuntu" directory | 09:47 |
| alkisg | I think some new "server syncing" script was introduced | 09:48 |
| alkisg | That, while syncing, breaks the servers, by not having an /ubuntu/ dir | 09:48 |
| alkisg | After the sync then it works again, and /ubuntu appears again | 09:48 |
| alkisg | For later reviewing, here's what I get with `wget http://ports.ubuntu.com/ -O - | nc termbin.com 9999`: https://termbin.com/44vr | 09:52 |
| blahdeblah | alkisg: For ports, the directory is called /ubuntu-ports rather than /ubuntu :-) | 09:54 |
| alkisg | blahdeblah: ah, sorry, ignore this then :) A user in #ubuntu reports the ports issue, and I saw the dists folders etc, and I thought it was the same problem | 09:55 |
| alkisg | Although the [ ]Archive-Update-in-Progress-ports-banjo2021-07-09 09:53 file does hint at my "server syncing" theory | 09:55 |
| alkisg | Are servers supposed to be broken while syncing, or is it supposed to be atomic from the user point of view? | 09:56 |
| blahdeblah | The fact that an archive update is in progress should be transparent to people using the archive, since the pools are updated before the indexes. But I'm far from an expert on that and will leave it to others here who are better qualified to answer. | 09:58 |
| alkisg | Thank you. I will wait until that "update-in-progress" file disappears, then check the contents of /; if the "dists" folders etc don't appear there anymore, then it will be a strong indication that something is indeed broken | 09:59 |
| blahdeblah | I've just tried an apt update on my system which uses ports and everything seems fine. | 10:02 |
| alkisg | The update-in-progress file has disappeared, so I guess it should be working now anyway. But the /dists folders etc are still there (unlike in archive.ubuntu.com) | 10:07 |
| === sem2peie- is now known as sem2peie | ||
| === sem2peie- is now known as sem2peie | ||
| pitti | hello everyone, long time no see! | 12:50 |
| pitti | my ubuntu-core-dev privs timed out during my holidays, and now I can't even upload cockpit to universe any more | 12:50 |
| pitti | (I'd actually be fine with just getting universe privs back, I don't really need core-dev) | 12:51 |
| pitti | or even just PPU for cockpit | 12:53 |
| cjwatson | pitti: Hi! If it hasn't been expired for too long then I think you should be able to drop a note to devel-permissions@lists.ubuntu.com and a DMB member should sort it out (see https://wiki.ubuntu.com/DeveloperMembershipBoard/KnowledgeBase#Accidental_Expiry) | 12:57 |
| pitti | cjwatson: thanks! No, just a few weeks ago | 12:58 |
| pitti | https://launchpad.net/~ubuntu-core-dev/+members doesn't actually show "former members" for me, but that might just be because of my particular login powers | 12:59 |
| cjwatson | Only admins of the team can see that | 12:59 |
| pitti | cheers! | 12:59 |
| cjwatson | The bullet-point list on that wiki page is directed towards DMB members | 13:00 |
| pitti | (I might still go for PPU -- feels more honest) | 13:00 |
| seb128 | pitti, I think you are still trusted for Ubuntu uploads, doesn't feel dishonest to keep your upload rights active in case you feel like fixing something at some point | 13:06 |
| rbasak | pitti: I agree with all of the above. If you want to email devel-permissions@ with your request now, I can JFDI for you. | 13:11 |
| pitti | (in meeting) will do, thanks seb128 and rbasak! | 13:14 |
| pitti | mailed | 13:29 |
| rbasak | done | 13:31 |
| pitti | rbasak: wow, that was super-fast, thank you! | 13:42 |
| rbasak | https://launchpad.net/~package-import is dead now, right? | 13:49 |
| rbasak | Should I kick it out of ~ubuntu-core-dev on the "least privilege" principle? | 13:49 |
| === didrocks999 is now known as didrocks | ||
| cjwatson | rbasak: As far as I know it is very dead, yes, so that's probably a good idea | 14:33 |
| cjwatson | rbasak: I assume you would know if any of your stuff relied on it :-) | 14:33 |
| sarnold | DarkTrick: ah, so the package maintainer scripts. Good Packages won't be doing anything inside user home directories in those scripts, but it's entirely possible that someone who is unfamiliar with debian packaging would prepare one that does do things in user home directories | 18:10 |
| === genii is now known as genii-core | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!