[11:17] <Laney> cking: hey, happy new year - here's a present for you: can you look into the autopkgtests of stress-ng please? tlb-shootdown is somehow murdering the test runner
[11:18] <Laney> I think it's killing the parent or something
[11:18] <Laney> reproduces in the qemu runner fwiw
[11:18] <cking> Laney, how urgent is it?
[11:19] <Laney> I can blacklist for now
[11:19] <Laney> so depends if you want results
[11:19] <cking> blacklist it for now, I'll look at it sometime next week, thanks for the heads-up on this, it is a torture test that probably is too hard on the test runner for sure
[11:20] <Laney> I run that stressor over SSH and it kills my SSH connection :-)
[11:21] <cking> oops, that's a bit harsh of it
[11:23] <cking> Laney,  OK, I'll add some oom'ing magic to that one to make sure it gets oom'd before anything else. I think that's the issue
[11:25] <Laney> cking: I don't see any OOM stuff in dmesg from these runs
[11:25] <Laney> anyway, should reproduce for you I'd imagine
[11:25] <cking> Laney, how many CPUs and how much memory on the box?
[11:25] <cking> and which kernel?
[11:27] <Laney> 1 / 1024 / 4.13.0-17
[11:27]  * Laney did this: http://packaging.ubuntu.com/html/auto-pkg-test.html#executing-the-test
[11:28] <cking> ok, that's a good starting point, I'll try and get around to that next week
[11:28] <Laney> righto
[12:40] <juliank> apt's happy eyeballs in progress for security.ubuntu.com: https://paste.ubuntu.com/26312732/ - I artificially lowered the connection attempt delay to like 10 seconds to get quite a few attempts
[12:40] <juliank> It's a host that prefers IPv4, but you also note that the second v4 attempt completed before the first, so that one is picked.
[12:41] <juliank> This also means that if one of the addresses fails, APT now falls back to another within that delay (250 ms by default), rather than waiting for 2 minutes
[12:42] <juliank> I'm considering adding a linear back off factor, but currently it's a constant as the RFC specifies for stupid clients
[12:44] <juliank> It seems a natural back off the RFC 8305 does not specify would be to just always wait clamp(250 ms, time we spent so far, 2000 ms). Opinions welcome :D
[16:04] <wxl> ypwong: saw the recent changes to The Bug™. does this signal that we'll be respinning and releasing 17.10 again?
[16:36] <juliank> Huh. I just heard that loading rtsx_pci enables support for 01:00.0 Unassigned class [ff00]: Realtek Semiconductor Co., Ltd. RTS5229 PCI Express Card Reader (rev 01)
[16:36] <juliank> how odd
[16:56] <juliank> -> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1741084
[17:44] <letterman> does the default install pam log user passwd history
[17:46] <nacc> letterman: you're in the wrong channel.
[18:33] <wxl> oh ho! interesting fact: 4.15rc6 is being suggested to fix broken BIOS due to Intel SPI on bug 1734147 .... and that theoretically includes all of the x86 page table isolation patches https://lwn.net/Articles/742449/
[18:33] <wxl> so maybe we could kill two bugs with one stone? :)
[18:35] <dannf> slangasek: looks like some fixes i had queued for d-i/bionic got dropped from the repo before the last upload - was there a problem with those?
[18:36] <dannf> or maybe i failed to push them...
[18:38] <slangasek> dannf: sorry, I didn't knowingly drop anything
[18:38] <dannf> slangasek: ok - probably my fault then :)
[23:31] <JackFrost> Laney: https://sigma.unit193.net/source/debhelper_11ubuntu0.1~ubuntu16.04.dsc - https://sigma.unit193.net/source/debhelper_11ubuntu0.1~ubuntu17.10.dsc btw.
[23:31] <JackFrost> Xenial needs dh-autoreconf too, but that's super easy.
[23:33] <tsimonq2> It seems that all Launchpad builders are currently disabled. See the topic in #launchpad for more details.
[23:47] <JackFrost> My bad, dh-autoreconf already *has* a backport.  Not very recent, but should do the trick.
[23:48] <juliank> Yes, 12 is the minimum for debhelper
[23:48] <tsimonq2> juliank: We're on to debhelper 12 already? O_O
[23:48] <JackFrost> I'm aware, I didn't see xenial had it.
[23:48] <juliank> The other ones are mostly speed ups :D
[23:49] <juliank> tsimonq2: dh-autoreconf 12
[23:49] <juliank> :D
[23:49] <tsimonq2> juliank: hehehehe I was gonna say... :D
[23:49] <juliank> but debhelper 12 compat level exists in development :D
[23:49] <JackFrost> juliank: There was a discussion in #debian-derivatives, I may have accidentally volunteered to help backport debhelper.
[23:50] <juliank> I'm obviously highlighted by dh-autoreconf and happy to do any backports of that :D
[23:50] <juliank> but none needed so
[23:52] <JackFrost> I'd say "sorry for the ping", but your fault putting that on highlight. ;)
[23:52] <juliank> Re backports: I still gotta build an apt snap
[23:52] <JackFrost> ...Unless you want me to poke you again about xdt-autorun. :>
[23:52] <juliank> Latest apt everywhere :D
[23:53] <JackFrost> I'm presuming https://anonscm.debian.org/git/apt/apt.git/commit/?id=8d23827be3043daf7fed1b86da1d41578889eaeb won't get backported?
[23:56] <juliank> JackFrost: Oh, that should be backported to 1.4.y and 1.2y.
[23:57] <JackFrost> Ooh nice, won't have to wait for Bionic then.
[23:58] <juliank> I gotta cleanup 1.5.2 and release that
[23:58] <juliank> and then I can add that to the 1.4.y and 1.2.y queues
[23:58] <JackFrost> I still have to wait for Bionic for the new apt-file though, of course.