Eickmeyer | vorlon: Figured out the ubiquity issue. Even though it was reported as a permissions problem in the logs, it was really whining about not having a slideshow🤦: https://bugs.launchpad.net/ubuntu/+source/ubiquity-slideshow-ubuntu/+bug/2008731/comments/5 | 00:52 |
---|---|---|
-ubottu:#ubuntu-devel- Launchpad bug 2008731 in ubiquity-slideshow-ubuntu (Ubuntu) "Edubuntu needs slideshow - Ubiquity crashes without it" [Critical, Triaged] | 00:52 | |
Eickmeyer | (logs were less than helpful) | 00:53 |
Eickmeyer | arraybolt3: ^ | 00:53 |
arraybolt3 | Eickmeyer: I don't see how that's possible, Xubuntu Minimal (formerly Xubuntu COre) has been working perfectly fine without a slideshow for a while. | 00:54 |
arraybolt3 | (I was the main tester for Xubuntu Core last cycle, and have never seen a slideshow in it.) | 00:54 |
Eickmeyer | arraybolt3: I don't know. All I know is I installed a slideshow in the live session, and installation worked fine. | 00:54 |
arraybolt3 | Oh weird. | 00:54 |
arraybolt3 | I mean I guess we're going to have a slideshow eventually, so no big deal? | 00:55 |
Eickmeyer | I'm working on resurrecting the old edubuntu one now so there's at least *something*. | 00:55 |
arraybolt3 | jbicha: Thanks for the suggestion about python3-distro, I have confirmed that it should do exactly the job needed and am testing Lubuntu's installer with it now. | 01:00 |
arraybolt3 | Does anyone know of a good alternative to apt-cacher-ng? For some reason it's horribly unstable on my Jammy laptop - it will download about 20 packages if they're missing from the cache and then give 503 errors for everything else. I can run "sudo systemctl restart apt-cacher-ng" and then do another sbuild and another 20 packages or so will download, and then a bunch of 503 errors. | 02:55 |
arraybolt3 | (On Kinetic this happened sometimes, but much more rarely.) | 02:56 |
sarnold | arraybolt3: squid-deb-proxy has been way more reliable than apt-cacher-ng for me | 02:56 |
arraybolt3 | K, I'll look into it. This is almost unusable. | 02:57 |
sarnold | I found it unusuable, yes | 02:57 |
arraybolt3 | sarnold: Is there a good setup guide for squid-deb-proxy? | 02:59 |
arraybolt3 | (I don't want to start installing things haphazardly and end up with broken apt or worse, no Internet :P) | 02:59 |
arraybolt3 | Meh, it looks simple enough. | 03:01 |
sarnold | arraybolt3: heh, good question. I set it up so long ago I can't recall much about the initial setup; the recurring problem is /etc/squid-deb-proxy/mirror-dstdomain.acl.d/* -- I had to set up private-ppa.launchpadcontent.net and esm.ubuntu.com since those (weren't? aren't?) in the default policies, and every time my archive mirror gets a new IP address because my home router is just the worst, I have | 03:02 |
sarnold | to add the new IP address in a file in there.. | 03:02 |
sarnold | One Of These Days [tm] I'm going to get all my 'real' computers on static IPs and there will be puppies and unicorns for all | 03:03 |
arraybolt3 | Hey, this is relatively straightforward so far! | 03:08 |
JackFrost | Unfortunately I find acng to be the best, but I use apt-cacher-ng-3.7.4+really3.6.4 :P | 03:13 |
arraybolt3 | Hah, so something happened in 3.7.4 that ruined it? | 03:13 |
arraybolt3 | Wow, I am making one config file boffo after another here :P | 03:17 |
arraybolt3 | sarnold: Could you share your /etc/squid-deb-proxy/mirror-dstdomain.acl? I cannot figure out how to get it to accept the mirror-dstdomain.acl.d directory. | 03:20 |
arraybolt3 | (I mean I probably could figure it out eventually but you're right here :P) | 03:20 |
sarnold | arraybolt3: with the warning that these have grown over time organically... https://termbin.com/9vqn | 03:22 |
arraybolt3 | No, I wanted the non .d file so I could figure out how to get the directory added. | 03:22 |
arraybolt3 | I have plenty in the directory, I just can't figure out how to get the directory to be recognized. | 03:22 |
arraybolt3 | Oh, wait, there it is. | 03:23 |
arraybolt3 | Wait... you still have the line commented out? | 03:23 |
sarnold | hmm, I wonder if it requires a restart to regenerate the configs? | 03:23 |
arraybolt3 | I restart it every time I change something. | 03:23 |
arraybolt3 | There was a line I uncommended that you have commented, that's probably the problem. | 03:24 |
arraybolt3 | Bah. It's ignoring my entire include directory. | 03:24 |
sarnold | heh, I think that's 100% a documentation line.. | 03:24 |
arraybolt3 | Yeah, it just is ignoring mirror-dstdomain.acl.d. | 03:27 |
arraybolt3 | It just gives me 403 Forbidden errors. | 03:27 |
arraybolt3 | For everything that isn't in the main file, but in the directory instead. | 03:27 |
arraybolt3 | Found it. | 03:29 |
arraybolt3 | I had the wrong proxy port set. | 03:29 |
arraybolt3 | I was trying to access port 8000, I needed to access 3128. | 03:29 |
sarnold | aha! :) | 03:31 |
arraybolt3 | Though now it's even working for sites that I *don't* have set as allowed??? | 03:31 |
sarnold | hmmm. I wonder if then you've got a 'usual' squid running? o_O | 03:31 |
arraybolt3 | deb.tuxedocomputers.com should not be allowed yet since I forgot to add it, yet there it is. | 03:32 |
arraybolt3 | sarnold: Maybe, lets find out. | 03:32 |
arraybolt3 | Hah, I did in fact have two copies of squid running! | 03:33 |
sarnold | that might be worth a bug report; it seems surprising | 03:33 |
arraybolt3 | Nah, user error. | 03:33 |
sarnold | ah good good | 03:33 |
arraybolt3 | "sudo systemctl restart squid" != "sudo systemctl restart squid-deb-proxy" | 03:33 |
sarnold | ahhhhh yeah. I have to imagine I've done that myself. | 03:34 |
arraybolt3 | Very good. Now it lets through everything I put in and blocks everything I failed to put in :P \o/ | 03:35 |
sarnold | JackFrost: what I don't understand is that you're not the only one I've talked with who's had better success with acng than with squid-deb-proxy, but I've seen acng give 503s and hash-sum mismatches and file-size mismatches dozens of times over the years. Why does it work okay for some people and not at all for others? | 03:36 |
sarnold | JackFrost: and squid has been around for decades as a solid reliable tool that folks forget even exists .. so why does it *not* work for those same people? :) | 03:37 |
JackFrost | Rule of "least terrible", not "actually works great" :P | 03:40 |
sarnold | ahhhhhhhh | 03:40 |
arraybolt3 | I'm thankful there's just some other solution that acng, it was getting to the point where I was actually starting to consider doing `while [ true ]; do sudo systemctl restart apt-cacher-ng; sbuild -d lunar-amd64-shm; done` | 03:41 |
arraybolt3 | Which I'm only now realizing might have left me in a hopeless loop :P | 03:42 |
arraybolt3 | **DO NOT RUN THE ABOVE COMMAND** | 03:42 |
arraybolt3 | Probably should have added a counter or something before doing that. | 03:42 |
sarnold | arraybolt3: I'm 80% sure I did something similar and it kept corrupting files that it had already downlaoded | 03:43 |
arraybolt3 | Oh fine. | 03:43 |
arraybolt3 | Alright, I finally have everything set up right. Now to get all of my sbuild containers to use it. | 03:44 |
Eickmeyer | Your first mistake was "while true; do ... done". | 04:11 |
arraybolt3 | :P | 04:13 |
=== JanC_ is now known as JanC | ||
cpaelzer | dannf: now that I understood the case better we have several ways to overcome the momentary test issue, I have filed bug 2008865 and assigned you to pick one | 07:38 |
-ubottu:#ubuntu-devel- Bug 2008865 in qemu (Ubuntu) "edk2 autopkgtest spotted a real issue, still we need to mitigate for now" [Undecided, Confirmed] https://launchpad.net/bugs/2008865 | 07:38 | |
adrien | can someone re-trigger https://autopkgtest.ubuntu.com/request.cgi?release=lunar&arch=arm64&package=llvm-toolchain-13&trigger=ocaml%2F4.13.1-4ubuntu1 ? it's llvm-toolchain tests on arm64 for ocaml and it failed due to the testbed not being ready in time; it's probably the (actual) last test blocking ocaml stuff (at least that I can spot) and hopefully, by the time it runs, runners won't be overloaded | 07:54 |
adrien | anymore at all | 07:54 |
tjaalton | what does it need to be able to re-run failed autopkgtests? some team membership besides core-dev/motu? | 08:09 |
adrien | when I attempt to do so, the service tells me "You submitted an invalid request: You are not allowed to upload llvm-toolchain-13 or ocaml to Ubuntu, thus you are not allowed to use this service." | 08:18 |
tjaalton | oh, so it's tied to upload rights | 08:20 |
adrien | I think it makes sense (I also think it's frustrating :D ) | 08:20 |
=== sem2peie- is now known as sem2peie | ||
danilogondolfo | May I ask someone to run this test for me, please? https://autopkgtest.ubuntu.com/request.cgi?release=lunar&arch=s390x&package=pycurl&trigger=curl/7.88.1-1ubuntu1&trigger=pycurl/7.45.2-3 | 11:01 |
ginggs | danilogondolfo: . | 11:05 |
danilogondolfo | thanks ginggs | 11:05 |
bluesabre | vorlon: I've got an easy, low-priority item if you want: https://code.launchpad.net/~xubuntu-website/+git/xubuntu-cdimage-css/+ref/master Some minor tweaks to the Xubuntu cdimage css to improve the minimal image and oversized files | 11:47 |
=== JackFrost is now known as Unit193 | ||
bluesabre | vorlon: This one too: https://code.launchpad.net/~bluesabre/ubuntu-cdimage/+git/ubuntu-cdimage/+merge/438004 (sorry for all the tags and things I'm sending your way) | 11:55 |
=== tjaalton_ is now known as tjaalton | ||
ogayot | Hi, could somebody retry the pycirkuit autopkgtests for me please? Thanks! | 14:03 |
ogayot | https://autopkgtest.ubuntu.com/request.cgi?release=lunar&arch=amd64&package=pycirkuit&trigger=m4%2F1.4.19-3 | 14:04 |
ogayot | https://autopkgtest.ubuntu.com/request.cgi?release=lunar&arch=armhf&package=pycirkuit&trigger=m4%2F1.4.19-3 | 14:04 |
ogayot | https://autopkgtest.ubuntu.com/request.cgi?release=lunar&arch=ppc64el&package=pycirkuit&trigger=m4%2F1.4.19-3 | 14:04 |
ogayot | https://autopkgtest.ubuntu.com/request.cgi?release=lunar&arch=s390x&package=pycirkuit&trigger=m4%2F1.4.19-3 | 14:04 |
jbicha | ogayot: done | 14:31 |
ogayot | jbicha: thanks! | 14:33 |
enr0n | Can a core dev please retry the tests blocking libio-socket-ssl-perl (retry-autopkgtest-regressions -s lunar --blocks libio-socket-ssl-perl)? The libhttp-daemon-ssl-perl test passed for me locally on amd64. The failures from other packages look like test bed failures. | 14:44 |
vorlon | bluesabre: marking as reviewer is preferable to irc pings, as it gets into a persistent queue :) | 16:19 |
bluesabre | vorlon: I added you as a reviewer for ubuntu-cdimage. I don’t think the css has a public branch I can submit a MR for (or does it somewhere?) | 16:22 |
bluesabre | Thanks for the review! | 16:22 |
vorlon | bluesabre: oh, I guess that one doesn't need review, because that's the branch we pull from! | 16:23 |
vorlon | bluesabre: pulled | 16:24 |
bluesabre | vorlon: Thanks a bunch! | 16:25 |
rbasak | vorlon: marking as reviewer> you mean like https://code.launchpad.net/~racb/ubuntu-archive-tools/+git/ubuntu-archive-tools/+merge/429629? :-P | 17:09 |
vorlon | rbasak: I thought we agreed to move that to ubuntu-archive-scripts? :) | 17:54 |
vorlon | rbasak: also I said it went into a persistent queue, I didn't say the queue ever emptied | 17:55 |
vorlon | (because it's actually a horrible cross between a queue and a stack, enjoy) | 17:55 |
rbasak | Ah yes we did, sorry | 19:00 |
adrien | I initially thought llvm-toolchain-13 failures on arm64 were testbed/setup issues but there seems to be a number of dumped code | 19:12 |
adrien | core* | 19:12 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!