16:02 #startmeeting 16:02 Meeting started Thu Feb 1 16:02:42 2018 UTC. The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology. 16:02 16:02 Available commands: action commands idea info link nick 16:02 \ / o 16:02 [TOPIC] Lightning round 16:03 $ echo $(shuf -e slangasek bdmurray xnox tdaitx doko sil2100 rbalint infinity cyphermox mwhudson rcj philroche Odd_Bloke tribaal fginther juliank) 16:03 doko rcj mwhudson tdaitx cyphermox infinity Odd_Bloke tribaal bdmurray philroche juliank xnox slangasek fginther rbalint sil2100 16:03 doko: hello! 16:03 o/ 16:04 ouch 16:04 I go last ... 16:04 rcj 16:04 doko, we're saving the best for last 16:04 * made clouds faster 16:04 * went first in meeting 16:05 * CVE builds for cloud-images 16:05 (done) 16:05 no mwhudson 16:05 tdaitx: 16:06 * OpenJDK 8 security update: 16:06 - still massaging armhf hotspot 16:06 * Ran tests with openjdk 9 and 10 as default jdk on bionic 16:06 - 24 reverse build-depends on default-jdk + default-jdk-headless 16:06 - 8 failures using jdk 9, 10 failures using jdk 10 16:06 * Investigating maven-archiver ftbfs (Debian #886875) 16:06 Debian bug 886875 in src:maven-archiver "maven-archiver FTBFS: test failure" [Serious,Open] http://bugs.debian.org/886875 16:06 * Backporting OpenJDK 8 fixes to OpenJDK 7 16:06 (done) 16:06 - MIR 16:06 - argos2 review 16:06 - discussing mtd-utils MIR with rbalint, triggered by flash-kernel 16:06 - juju-core review for SRU to xenial, from balloons' code branch 16:06 - reviewed subiquity keyboard code 16:06 - netplan: 16:06 - implement 'netplan ip leases' 16:06 - refactoring for private python lib 16:06 - coverage/ci fixes 16:06 - field kernel team questions about netplan 16:06 - debugging grub net timeout/retransmit/repeated RRQs of files it already received 16:06 - shim/grub: 16:06 - validating grub2{,-signed},shim{,-signed} SRU in xenial, artful 16:06 (done) 16:07 * Snapping part of the build infrastructure for cloud images. 16:07 * Partner work 16:07 * Work on https://bugs.launchpad.net/ubuntu/+source/squid-deb-proxy/+bug/1456238 (on my spare time) 16:07 Launchpad bug 1456238 in squid-deb-proxy (Ubuntu) "Rewrite urls to mirror" [Wishlist,In progress] 16:07 (done) 16:07 (out of order \o/ My battery is almost dead) 16:07 * Vacation, travel and catch-up 16:07 (done) 16:07 bdmurray: 16:07 tribaal: any more? 16:07 bdmurray: nope 16:07 oops 16:08 Investigation into non-unique whoopsie ids (LP: #1636954) 16:08 Discussed non-unique whoopsie ids at some length 16:08 Switched rabbitmq charm in mojo spec so nagios check passes 16:08 Updates to mojo spec for nagios changes / recommendations 16:08 Launchpad bug 1636954 in Whoopsie "Whoopsie should use something else if non-unique uuid" [High,Triaged] https://launchpad.net/bugs/1636954 16:08 Documented how to setup whoopsie to send crashes to daisy w/ self-signed cert 16:08 Modified whoopsie to respect an env var to not check ssl cert of daisy server 16:08 Reviewed and merged jibel's apport branch fixing ubiquity package hook (LP: #1582950) 16:08 Launchpad bug 1582950 in apport (Ubuntu Artful) "broken apport hook: TypeError: a bytes-like object is required, not 'str'" [Medium,Fix released] https://launchpad.net/bugs/1582950 16:08 Fixed apport bug re: recommending python-apport (LP: #1729879) 16:08 Launchpad bug 1729879 in apport (Ubuntu Bionic) "apport-collect is still requesting python-apport installation" [Medium,Fix released] https://launchpad.net/bugs/1729879 16:08 Investigation into and resolution of apport-test-crashes creation failure 16:08 Confirmed apport-test-crashes are good - had them retraced in devops ET 16:08 Was on holiday last week 16:08 done 16:08 philroche: 16:09 philroche is out today. 16:09 juliank, 16:09 merged lvm2 16:09 fixed some ubuntu-release-upgrader stuff (bug #1744722) 16:09 cryptsetup 2.0 transition 16:09 rsyslog merge 16:09 bug 1744722 in ubuntu-release-upgrader (Ubuntu Artful) "Unknown bad source brings up during 'zesty' to 'artful' upgrade and It break the process" [Critical,Triaged] https://launchpad.net/bugs/1744722 16:09 apt test suite fix 16:09 filed argos2 (for cryptsetup) MIR 16:09 fastjson (for rsyslog) MIR 16:09 working on lscpu xenial fix now (bug 1732865) (done) 16:09 bug 1732865 in util-linux (Ubuntu Xenial) "[LTCTest][OPAL][FW860.20] lscpu failed to list cpu max and min frequencies" [Undecided,In progress] https://launchpad.net/bugs/1732865 16:09 ok, the (done) should have been its own line 16:09 xnox: 16:09 Work on migrating 1.0.2n (upgrade from g release), openssl migrated 16:09 Working on completing migration of my uploads (ppc/s390x) 16:09 Fix a couple of easy 1.1.0 openssl compatibilities 16:09 Merged systemd v237 16:09 Back from japan on sunday, off to brussels for fosdem tomorrow. 16:09 done 16:10 * working through maas+grub netboot regression possibly triggered my meltdown mitigations (LP: #1743249) 16:10 Launchpad bug 1743249 in MAAS "Failed Deployment after timeout trying to retrieve grub cfg" [Undecided,Incomplete] https://launchpad.net/bugs/1743249 16:10 * sort out livecd-rootfs regression in bionic (LP: #1746631) 16:10 Launchpad bug 1746631 in livecd-rootfs (Ubuntu) "livecd-rootfs 2.496 needs a PROPOSED=1 build test before release to bionic" [Undecided,In progress] https://launchpad.net/bugs/1746631 16:10 * proposed-migration work 16:10 * digging into the arm64 runner slowdown over the past week (but not getting very far) 16:10 * retries, overrides, etc. of autopkgtest failures 16:10 * discussions w/ server team about default behavior of cloud-init on unconfigured NICs 16:10 * discussions around fixing MAAS ability to install ESM-supported precise 16:10 * discussions around clean shutdown of complex storage configs (finalrd) 16:10 (done) 16:10 fginther: 16:11 rbalint: 16:11 * finish flash-kernel merge 16:12 * mtd-utils MIR for flash-kernel (still needs fixing tests) LP: #1746126 16:12 Launchpad bug 1746126 in mtd-utils (Ubuntu) "[MIR] mtd-utils (WIP)" [Undecided,New] https://launchpad.net/bugs/1746126 16:12 * merge shadow 16:12 * fix cdbs dependency on scour which broke shadow's build 16:12 * fix FTBFS of libnfs 16:12 * internal testbed setup 16:12 * gce-compute-image-packages update, SRU-s pending LP: #1746588 16:12 Launchpad bug 1746588 in gce-compute-image-packages (Ubuntu) "Update google compute-image-packages to 20180129" [Undecided,Fix released] https://launchpad.net/bugs/1746588 16:12 (done) 16:12 - Off on Tuesday, very fragmented week in overall 16:12 - SRU reviews and releases 16:12 - Kernel SRU reviews and releases 16:12 - ubiquity: 16:12 * Testing the ubiquity early keyboard selection branch 16:12 * Figuring out how to run the autopilot-tests locally on a VM 16:12 * Working on fixing autopilot tests 16:12 - Checking edge cases of ubuntu-server installs (making sure quiet and splash aren't there) 16:12 - DMB meeting 16:12 - 16.04.4 discussions and slip announcements 16:12 (done) 16:12 doko: 16:13 - came back from LinuxConf.AU, short week 16:13 - updated mpfr/mpclib, and all GCC packages, including cross packages 16:13 - GCC 7.3 and binutils 2.30 releases 16:13 - glibc fix for binutils 2.30 16:13 - gdb 8.1 release 16:13 - and unfortunately handling a lot of unfinished transitions ... 16:13 - build failures ... 16:13 (done) 16:13 any questions over status? 16:13 Oh, I forgot: I've been hacking around a bit on Launchpad to add Valid-Until to release files (WIP) 16:14 juliank, sounds interesting and scary. e.g. in the context of ESM and the fact that release pocket is frozen and not republished. 16:14 juliank, and esm is kind of open-eneded in end dates. 16:14 xnox: for -updates, -security, -proposed, and -backports I think. We'll see how it works out :) 16:15 juliank: for ESM, the -updates and -security pockets need to be usable for years after they stop being updated 16:15 but that's not a reason for LP to not support it in principle 16:15 slangasek: I'd imagine we can come up with something for that 16:15 and ppas.... like fips, esm, cloud archive, partner archive, dbgsysms, 16:16 [TOPIC] Bugs 16:16 [LINK] http://reqorts.qa.ubuntu.com/reports/rls-mgr/rls-bb-incoming-bug-tasks.html#foundations-bugs 16:17 anything here we should be taking? 16:17 bug 1744318 - should this move to targetted? 16:17 bug 1744318 in update-manager (Ubuntu) "changelogs.ubuntu.com should be using HTTPS" [Undecided,New] https://launchpad.net/bugs/1744318 16:18 xnox: I think we had discussed LP: #1707898 again while you were out... what should we do with this? 16:18 Launchpad bug 1707898 in systemd (Ubuntu) "systemd translations are not synced with upstream" [Critical,New] https://launchpad.net/bugs/1707898 16:18 bdmurray: 1744318 - yes I think so 16:20 Laney, do you think we can trollo-fy the above systemd translation task, on like desktop team trello backlog? it needs a spike investigation to come up with a plan of action. 16:20 slangasek, I think xnox is going to reach out to laney 16:20 xnox: I think we should drop the rls-bb-incoming tag, since regardless of criticality it isn't for us to accept 16:20 sounds reasonable 16:20 if it's another team that needs to fix it 16:21 and then i am happy to tackle it, or to delegate, as long as there is agreement on how to "make translation in langpacks, and be usable by systemd, et.al and all working fine as needed" 16:21 slangasek: So will you modify 1744318? 16:21 bdmurray: yes 16:21 Laney, possibly need changes to langpacks, and pkgmangler, to mangle systemd in a special snowflake way. 16:21 I have some others but will wait for the systemd discussion to tend. 16:21 * xnox is done. 16:21 tend = totally end 16:22 xnox: I've not really touched langpacks, so you probably want to start with seb128 16:22 definitely not a typo 16:22 I see I was @-ed in that bug but not subscribed to it ;-) 16:22 oh no I am, wtf 16:23 Laney, re:touching langpacks, me neither.... but some sensible plan / guidance, would be nice before shooting in the foot. Like choose which foot, and whoose =) 16:23 bdmurray, xnox: takes 1707898 out of our incoming queue, will trust that it gets followed up w/ desktop team 16:23 X-Launchpad-Message-Rational: subscribed-by @xnox: file in /dev/null :-) 16:23 +e 16:23 twat 16:23 =) 16:23 Is bug 1729491 worth targetting to the release? 16:23 bug 1729491 in apport (Ubuntu) "ubuntu-bug says that snap packages are not installed" [Medium,Triaged] https://launchpad.net/bugs/1729491 16:23 I actually took those langpacks from seb, but I guess seb might know more on the actual mechanics of how they work 16:23 I just know and deal with their generation 16:23 It should just be a string change somewhere. 16:24 hm, which reminds me, I forgot to re-enable bionic update packs, eh 16:24 bdmurray: hmm yes, let's take that one 16:24 bdmurray: what would be the proper way to report a snap bug? 16:24 (done) 16:24 Laney, could we start a trello card for this on dekstop board and/or subscribe/add me to that card? it will need some discussion and coordination. 16:24 juliank: that's a separate question from having the bug-reporting tools giving confusing error messages 16:24 xnox, Laney: could you also include me in that card? 16:25 juliank: I don't know if there is one. 16:25 alright. 16:25 yes, unfortunately there's not a consistent story there currently 16:25 bug 1736072 seems important - I haven't dug into it at all 16:25 bug 1736072 in ubiquity (Ubuntu) "Encrypted swap does not work" [Undecided,New] https://launchpad.net/bugs/1736072 16:26 1736072 is the reason xnox doesn't get a straight backport of swapfiles into 16.04 as an SRU 16:26 xnox: sil2100: Chat to Seb in #ubuntu-desktop in the first instance I think 16:26 slangasek, .... but i thought it was fixed by cyphermox , no? 16:26 juliank: I'm still doubting about another json implementation in main ... 16:26 I don't know, has it been? 16:26 cyphermox: does encrypted swap work in bionic? should LP: #1736072 be closed? 16:26 Launchpad bug 1736072 in ubiquity (Ubuntu) "Encrypted swap does not work" [Undecided,New] https://launchpad.net/bugs/1736072 16:27 doko: Well, yeah, what can we do? I don't know. 16:27 slangasek: should be working, but we should give it another check 16:27 juliank: can it still be built with json-c? 16:27 juliank: yell at upstream? block at the current version of rsyslog for this cycle? 16:28 ah, the bug says it is now racy 16:28 cyphermox: shall I assign to you to follow up? 16:28 xnox: you're right swap should have been fixed last cycle 16:28 yes 16:28 (and claim it for the cycle, it's clearly important, just not clearly a bug) 16:28 slangasek: rsyslog upstream is the one who forked the json library 16:28 i wonder if there are dependencies missing, and/or not specified x-systemd-after things in fstab 16:28 and made it incompatible and dropped their json-c compat layer 16:28 sweet 16:28 juliank, did they embed a partial tree of systemd too? like NM? 16:28 I don't think so 16:29 * xnox says bad very very bad deal 16:29 juliank: this does not preclude yelling ;) 16:29 but everything is possible :) 16:29 That's it for rls-bb-incoming as far as I'm concerned 16:29 bdmurray: there seemed to be a lot of high incoming bugs that we skipped over 16:30 I think juliank should claim LP: #1725861 :) 16:30 Launchpad bug 1725861 in apt (Ubuntu) "APT::AutoRemove::SuggestsImportant "false" should be the default" [High,New] https://launchpad.net/bugs/1725861 16:30 slangasek: ah, indeed 16:30 slangasek: Oh well 16:31 juliank: do you think it's fair to say this is something we should fix before the next LTS, even if not implemented in the way the bug title suggests? 16:32 slangasek: I think it might make sense with my other plan to only autoremove "new garbage" 16:33 I'm not convinced it's the right thing to do 16:33 I'm not convinced it's the wrong thing to do either. 16:33 juliank: ok, let's claim it for now and we can discuss further w/o blocking the team meeting :) 16:34 slangasek: to be clear "claim it" also means create a card for it? 16:34 bdmurray: yes, doing 16:35 slangasek: and so every rls-bb-tracking tagged bug should have a card too correct? 16:35 bdmurray: should, yes 16:35 bdmurray: I have been doing those inline here 16:36 xnox: your last comment on LP: #1739672 says you weren't uploading while the build farm was blocked. Is this now fix-committed? 16:36 slangasek: Is there a way to audit the bugs for cards? Do they get tagged during the import? 16:36 Launchpad bug 1739672 in systemd (Ubuntu) "Regression in getaddrinfo(): calls block for much longer on Bionic (compared to Xenial), please disable LLMNR" [High,Triaged] https://launchpad.net/bugs/1739672 16:37 bdmurray: they have not been getting tagged during the import; I thought they were meant to be but maybe that's a separate job; I've noticed a lot of my imported cards not being tagged. fginther? 16:37 * juliank was a bit confused too about the systemd bug 16:37 I didn't realize the tracking bugs you were getting automagically imported 16:37 slangasek, well, but systemd is not building atm on arm64 because binutils / relocation / EFI 16:37 slangasek, but i should stage that in git at least, yes. 16:38 slangasek, yeah that should be happening. Let me look at logs 16:38 xnox: oh. who has the baton for fixing that, and is there a bug? 16:38 * juliank is checking if gnu-efi is causing the systemd FTBFS 16:39 slangasek, doko / juliank / and I are on the hook for it. 16:39 It might be either toolchain, gnu-efi, or systemd 16:39 systemd uses gnu-efi. 16:39 because of course it does. 16:39 slangasek, there are multiple things affected, e.g. kernel, gnu-efi, systemd.... which i guess is gnu-efi. 16:39 ok, so juliank is on this? 16:39 and I have a gnu-efi transition 16:39 slangasek, re:LLMNR should it be a default for resolved.conf; .network setting for a network; or a netplan provided default? 16:39 slangasek: yeah, I'm on the gnu-efi part 16:40 slangasek, i'm guessing we want it as default off, but not sure if that means in resolved, or on every network.... 16:40 juliank: can you (create a bug and) give me a bug number so I can follow without pestering? 16:40 slangasek: I'm checking if sytemd builds with the new gnu-efi or not. 16:40 xnox: my expectation is off in resolved by default 16:40 slangasek: One for gnu-efi transition, systemd FTBFS on arm64, or both? 16:40 juliank: systemd ftbfs 16:41 ack 16:41 slangasek, ok 16:42 slangasek: xnox: https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1746765 16:42 Launchpad bug 1746765 in systemd (Ubuntu) "systemd FTBFS on arm64" [Undecided,New] 16:42 I assigned it to gnu-efi, systemd, and binutils and we can figure out what's doing it 16:42 juliank: thanks 16:43 anything else on bugs? 16:43 bdmurray: please could you have a look at the apport autopkg test failures? 16:43 slangasek: Should we add a card for systemd FTBFS and mark it in progress? 16:44 juliank: done 16:44 thx 16:44 doko: oh sure, I wonder why I haven't gotten an email about that 16:45 autopkg test failures, not build failures 16:46 I thought we were supposed to get emails about both. 16:46 juliank, xnox: my $.02 is that we should disable any use of gnu-efi in our systemd packages as we are certainly not using the EFI artifacts in the distro and have no plans to 16:46 slangasek: I use them :/ 16:46 bdmurray: we email about stuck packages after a grace period 16:46 juliank: but that should not block systemd distro development in 18.04 16:47 right 16:47 [TOPIC] AOB 16:47 anything else? 16:47 we could also just enable them on arm64, though, as that's where it fails, then I'm happy too :) 16:48 juliank: indeed 16:48 slangasek, when i saw it fail to build, i was like "why the hell is _that_ enabled?" 16:50 juliank, you use them - systemd-efi-boot? you wrap apt in an initramfs, with a kernel image, wrapped as a UEFI binary and then you sign it and only boot that? =) 16:50 transition mess again ... 16:50 yes 16:51 I haven't been giving the transitions much attention because I've been focused on autopkgtest 16:51 xnox: I use systemd-boot to boot my system instead of grub, and I do wrap an initramfs with a kernel into one image which I sign 16:51 I'm doing unfinished transitions for the past four days ... 16:51 doko: do you have a pointer to where you think people should start on the transition mess? 16:51 it's now getting autopkg test failures fixed for involved packages 16:52 xnox: I can't really get grub to work since everything except the ESP is in my LUKS->LVM partition 16:52 doko: I see you've been demoting packages to proposed again, which again makes the bottom of update_excuses less useful 16:52 so everything which depends on mpfr4, mpclib, gdbm, perl, php, and more which I don't have yet identified 16:52 doko: we've discussed that these packages should be removed instead of demoted if they are going to require sourceful fixes 16:53 slangasek: ok, I can do that again. that were the packages still depending on python-imaging and not in debian testing 16:53 yes, those are good candidates for removal if they depend on python-imaging directly 16:54 and I'd like to proposed to push perl into the release pocket to disentangle it from other transitions 16:54 juliank, well, maybe we should fix that. 16:54 juliank, imho kernel should drop and load modules off ESP, and or /boot/grub should be on ESP too 16:54 doko: there was at least one autopkgtest regression against perl which looks legitimate, which is why I haven't skiptest'ed yet 16:55 xnox:Well then you have to put it in machine-id specific directories like I do for kernels in sicherboot 16:55 because multiple OS share the ESP 16:55 xnox: /boot/grub on ESP> yes please, though the migration of existing systems would be rough 16:55 juliank: we own /efi/ubuntu on the ESP 16:55 slangasek: What about systems with 2 ubuntus? 16:56 we'll always find corner-cases 16:56 juliank: you still only have one entrypoint via /efi/ubuntu/bootx64.efi, so it doesn't change the character of the problem if we move more stuff into /efi/ubuntu 16:56 juliank, we still own /efi/ubuntu, so we can add machine-id subfolders. plus ubuntu's can boot each other. 16:57 Right 16:57 I just add machine-id directories directly to the root of the ESP 16:57 anyway, interesting discussion, but I think we're done 16:57 #endmeeting