15:03 #startmeeting Weekly Ubuntu Foundations team 15:03 Meeting started Thu Mar 21 15:03:41 2019 UTC. The chair is xnox. Information about MeetBot at http://wiki.ubuntu.com/meetingology. 15:03 15:03 Available commands: action commands idea info link nick 15:03 #topic Lightning rounds 15:04 $ echo $(shuf -e vorlon bdmurray xnox tdaitx doko sil2100 rbalint infinity cyphermox mwhudson juliank waveform) 15:04 mwhudson doko xnox bdmurray cyphermox sil2100 juliank infinity rbalint vorlon tdaitx waveform 15:04 doko, ? 15:04 * xnox goes to fix the script to not have mwhudson it in. 15:04 - more openjdk-11-transition work 15:04 - new OpenJDK upstreams: 8u202, 8u212, Openjdk 12 ga, OpenJDK 11.0.3+4, OpenJDK 13 15:04 - Help tracking down https://bugs.openjdk.java.net/browse/JDK-8221083 15:04 - Fix PR jit/87808, backport fix for PR tree-optimization/89505 15:04 - GCC 9 update 15:04 - 360s 15:04 (done) 15:04 * made fuse-zip fail on all arches, instead of just s390x! 15:04 - possibly a way out here is more patches to fuse 15:04 - why did we merge fuse, past DebianImportFreeze?! 15:04 * openssl backports discussions with security/pat 15:04 * working on unbreaking subiquity on power/serial 15:04 * waiting on ibm to continue optimized atlas builds for z13/z14 15:04 * vorlon, still awaiting openssl 1.1.1 accept into 15:04 bionic-proposed. All feedback responded to.... 15:05 https://bugs.launchpad.net/ubuntu/+source/openssl/+bug/1797386 15:05 Launchpad bug 1797386 in openssl (Ubuntu) "[SRU] OpenSSL 1.1.1 to 18.04 LTS" [Undecided,In progress] 15:05 * on holidays next week ⛷️ 15:05 (done) 15:05 bdmurray, 15:05 converted two ET cronjobs from pycassa to python cassandra 15:05 submitted RT, mojo spec MP regarding update of daisy to r824 in prod ET 15:05 modified the above RT to also include updating apport to r3227 see below 15:05 discovered an issue with how apport calls gdb fixed in r3227 15:05 submitted RT, mojo spec MP regarding update of daisy to r826 in prod ET 15:05 review of rls-dd-incoming bug tasks 15:05 submitted PR for metrics and jobs w/ a metric for retracers-results 15:05 submitted PR for metrics and jobs w/ a metric for retracer-avg-process time 15:05 created some test graphs of the above 15:05 reviewed, merged, uploaded xnox's u-r-u lintian fixes for disco 15:05 review of netplan SRU exception 15:05 special SRU review of keystone, python-ldappool 15:06 cyphermox 15:06 Oh, or did bdmurray not finish? 15:06 I did finish 15:07 xnox: yeah, haven't had any SRU day time since last we discussed 15:07 vorlon, ack. 15:08 sil2100, go 15:08 No cyphermox ? 15:08 - SRU reviews and releases 15:08 - A lot of new kernel reviews (new cycle) 15:08 - Some package approvals for the OpenJDK-11 transition 15:08 - Review of the pi spec 15:08 - Reviews and infrastructure help for ubuntu-canary 15:08 - Helping out with an iproute2 backports upload 15:08 - Few FFe reviews 15:08 - Shepherding disco first set of language-pack delta updates 15:08 - Fixing kernel-sru-review to handle kernel backports of flavor kernels without crashing 15:08 - 360 15:08 - core18: 15:08 gah, sorry, I'm here 15:08 * Finalized gadget snap code branch location and permissions 15:08 * Moar work on core18 promotion - scripts almost there, only the jenkins job needed 15:08 * Backporting missing change from core for support of out-of-core18 wpa-supplicant (PR) 15:08 * Modified version number to include the date (and thinking about that a bit) 15:08 * Checked status of snap prepare-image for classic seeding 15:09 - Tomorrow: out-of-office o/ 15:09 (done) 15:09 cyphermox, go 15:09 shim review: attempting to review OpenSuSE / tumbleweed / SLES shims; hard to even get a build env working to begin with 15:09 shim review: partial review of Debian's shims for ia32, x64, aa64. We're gaining an additional contributor to reviewing shims too yay! 15:09 uploaded grub2 SRUs for http module + debconf prompts (LP: #1787630) (LP: #564853) 15:09 Launchpad bug 1787630 in grub2 (Ubuntu) "[FFe] Include HTTP support in pre-build GRUB module" [High,Fix released] https://launchpad.net/bugs/1787630 15:09 Launchpad bug 564853 in grub2 (Ubuntu Cosmic) "Spurious conffile prompts for /etc/default/grub" [Undecided,Incomplete] https://launchpad.net/bugs/564853 15:09 added mmx64.efi to d-i images (and will end up on dailies) to better address Secure Boot configuration for unfinished installs (LP: #1798171) 15:09 Launchpad bug 1798171 in debian-installer (Ubuntu Disco) "System fails to boot with \EFI\BOOT\mmx64.efi - Not Found" [High,Fix released] https://launchpad.net/bugs/1798171 15:09 initial review of netplan wireguard support PR 15:09 revived netplan PR#34: "openvswitch support" - allow networkd to bring up devices without addresses / unconfigured but IFF_UP. 15:09 netplan roadmap: prioritizing next features to implement 15:09 paperwork for design / user stories for new netplan features 15:09 I still need help to get the netplan.io SRUs reviewed in the bionic, cosmic queues 15:10 fwiw I'm looking now at why the dailies don't yet have mmx64.efi as I thought they would; which is why I was unresponsive, deep hack mode looking at mark-pending-current and such 15:10 (done) 15:10 cyphermox, mark-pending-current.... interesting, i wonder if ci jobs are fialing. 15:10 they def are 15:10 juliank, ! 15:10 (short week, OoO on monday) 15:10 * (last Fri) Re-verified APT SRUs 15:10 * ESM UX review 15:10 * Some PackageKit merge review 15:10 * Improve PackageKit error handling (https://github.com/hughsie/PackageKit/pull/321) 15:11 * Tighten APT dependencies on its own libs (https://salsa.debian.org/apt-team/apt/merge_requests/56) 15:11 * Rebase APT proposed update branches against master and do some more stringview cleanup 15:11 * More ESM UX bikeshedding 15:11 * Analysis of aptdaemon use in software-properties 15:11 * First stab at PackageKit driver installs in software-properties-gtk - (updates works already) 15:11 - Not sure about QApt vs PackageKit in the Qt backend 15:11 (also: apt-side of ESM stuff landed everywhere today, woohoo!) 15:11 (done) 15:11 infinity is not in the channel... 15:11 rbalint: 15:11 (short week) 15:11 * fixed script bzr -> git conversion script to handle software-properties' repo 15:11 * converted and switched software-properties project to git 15:11 * prepared sru patches for systemd-detect-virt wsl detection 15:11 * partner work 15:11 * fixing unattended-upgrades bugs, I'm finalizing tests and push the changes later today 15:11 (done) 15:12 * short week, was off last Friday 15:12 * legwork for fips vs secureboot 15:12 * discussions around UC20 roadmap 15:12 * discussing netplan roadmap for 19.10 15:12 * looked at snap seeding and core18 coming into the disco desktop images 15:12 vorlon, 15:12 * worked with xnox to fix modules.squashfs the right way around (in debian-cd) 15:12 * next week: off (spring break) 15:12 (done) 15:12 tdaitx, 15:12 * openjdk-11 bionic security transition 15:12 - gradle test and fixes (LP: #1820389, LP: #1797761, Debian: #925225) 15:12 - geogebra update for Bionic 15:12 - investigated LP: #1803855 in gradle/groovy but that requires a new groovy update and that is not happening any time soon 15:12 - following up on android-tools (LP: #1820513) 15:12 - quick check on an octave update for bionic for better openjdk-11 support, but that would require a soversion update 15:12 Launchpad bug 1820389 in gradle (Ubuntu Cosmic) "Gradle on OpenJDK 8: java.lang.NoSuchMethodError: java.lang.invoke.MethodHandles.privateLookupIn()" [Undecided,Fix committed] https://launchpad.net/bugs/1820389 15:12 * openjdk-7 packaged and uploaded for review by security team 15:12 Launchpad bug 1797761 in gradle (Ubuntu) "Gradle 4.4.1 package missing alternative dep to Java8" [Undecided,Fix released] https://launchpad.net/bugs/1797761 15:12 Debian bug 925225 in gradle "gradle: Fix gradle compatibility with OpenJDK 8" [Normal,Open] http://bugs.debian.org/925225 15:12 Other: 15:12 Launchpad bug 1803855 in gradle (Ubuntu) "Illegal reflective access operation" [Undecided,Confirmed] https://launchpad.net/bugs/1803855 15:12 - 360 fun 15:12 Launchpad bug 1820513 in android-platform-tools-base (Ubuntu) "android-* packages not working well with OpenJDK 11" [Undecided,Incomplete] https://launchpad.net/bugs/1820513 15:12 - working some odd hours this and the next weeks (gym on Mon/Fri afternoon + apartment renovation^Wproblems going on) 15:12 - happy 321 holidays! 15:12 (done) 15:13 waveform: go! 15:13 * More work on pi spec (answering reviews) 15:13 * SRU for RPi.GPIO to bionic 15:13 * RPi.GPIO fixes accepted upstream; tested updated package on ubuntu under aarch64 15:13 * ITP colorzero to Debian (new dep of gpiozero) 15:13 * Reading up on live-build (overlays for pi configuration hacks) 15:13 * Worked on first boot design (3 scenarios to support) 15:13 * Experimented with seeding cloud-init from boot partition on the pi 15:13 * piwheels configuration for ubuntu (ongoing) 15:13 (done) 15:14 Any status questions? 15:14 #topic Release incoming bugs (disco) 15:14 #link http://reqorts.qa.ubuntu.com/reports/rls-mgr/rls-dd-incoming-bug-tasks.html#foundations-bugs 15:15 I went ahead and set the importance for all the rls-dd-incoming 15:15 bug 1790205 15:15 bug 1790205 in systemd (Ubuntu) "systemd journals take up too much space, aren't vacuumed automatically" [High,Confirmed] https://launchpad.net/bugs/1790205 15:15 bdmurray, so... depending on how big ones instalation is, one will have bigger or smaller journal usage. 15:15 I was surprised to find 4G of logs on my system 15:16 bdmurray, so if people have large hard drive, you will have gigs of logs. 15:16 bdmurray, is it more than what journal declares to be the limit? one sec: 15:16 xnox: and if my hard drive becomes small because of the logs then what happens? 15:16 bdmurray, $ journalctl -b -u systemd-journald.service -n 1 15:16 -- Logs begin at Mon 2018-09-17 14:21:50 BST, end at Thu 2019-03-21 15:16:25 GMT. -- 15:16 Mar 21 11:27:37 ottawa systemd-journald[571]: System journal (/var/log/journal/1b8df0fa27039f0163586c6756a6d401) is 3.7G, max 4.0G, 231.6M free. 15:16 for me, the max is 4G, and it's below that at the moment, so that is working correct. 15:17 if logs are more, than max, that would be a bug, so far i have not seen anybody, prove that limits are neither in place, nor observed. 15:17 bdmurray, what does above say for you? 15:17 Mar 19 07:50:32 impulse systemd-journald[608]: System journal (/var/log/journal/5daa40091e7e4b6fa27e6439f0a48bdc) is 3.9G, max 4.0G, 23.8M free 15:17 bdmurray, or is it the case that 4GB is "omg logs grow uncontrollable" 15:17 bdmurray, so you are definately capped already. 15:19 $ journalctl --disk-usage 15:19 Archived and active journals take up 3.7G in the file system. 15:19 so maybe --disk-usage, should state the max, and free, like it does on startup. 15:19 sounds sensible 15:19 vorlon, ^ 15:20 ok. 15:20 I do think 4GB sounds like a surprisingly large amount for logs however 15:20 is there so much more data in journals vs. traditional syslog that we need to allocate more space? 15:21 "Archived and active journals take up 4.0G in the file system." for me 15:21 mar 18 21:44:34 tdaitx-P65 systemd-journald[32600]: System journal (/var/log/journal/e0e3d1032c874b5ba18c951f8ae3cbfa) is 1.0G, max 1.0G, 0B free. 15:21 Archived and active journals take up 1.0G in the file system. 15:22 but I set it to 1G in journald.conf 15:22 SystemMaxUse=1G 15:23 $ journalctl | grep gnome-shell 15:25 ^ gnome-shell sucks 15:26 my thinking here is that the journal should be rotating by date. Because if something's going on /currently/ that makes the logs balloon, we should keep those (up to the limit), but I don't think we should let systemd consistently fill up a fixed amount of space on disk 15:27 MaxRetentionSec=1m 15:27 and MaxFileSec control that 15:28 my desktop has a 1.3G journal on a 12G rootfs. is that going to grow up to 4G? (How do I check on cosmic the limit?) 15:28 ah found it 15:28 Mar 12 11:01:49 virgil systemd-journald[6237]: System journal (/var/log/journal/c856e8839ee179de1933c6264c9c4260) is 1.2G, max 1.1G, 0B free. 15:28 vorlon, it used to be the case but time just correlates with log size, while log size is the cost paid for the logs on systems thus size base limit fits the cost better and now it can proberly be enforced 15:29 vorlon, that looks buggy. because 1.2 is more than 1.1 15:29 xnox: I notice! 15:29 vorlon, that is worth filing a bug. 15:29 before journald size based limits could not have been enforced thus time-based was the best heuristics 15:29 10% max used, 15% min free 15:29 vorlon, and it's not weird variable size fs, right ? i.e. compressed btrfs? 15:29 each capped to 4 GB 15:29 xnox: haha no 15:29 ok 15:30 The --disk-usage switch gave me a different number than the system journal did too 15:30 it likely grew at boot, and then vacuumed later and it got down? 15:30 System journal (/var/log/journal/5daa40091e7e4b6fa27e6439f0a48bdc) is 3.9G 15:30 Archived and active journals take up 4.0G in the file system. 15:30 well --disk-usage shoes the current size, i think, which might be both persistent (/var) and runtime (/run) 15:30 time to let it go for now 15:30 don't make me sing 15:31 Okay, lets move on. xnox you'll follow up with the bug yes? 15:31 bdmurray, xnox is on vacation after tomorrow 15:31 no Hasselhoff? 15:31 bdmurray, yes. 15:31 bdmurray, doing typy typy now. 15:32 bug 1807479 - there's a patch but why was this assinged to the team xnox? 15:32 bug 1807479 in system-config-kickstart (Ubuntu) "Hashed passwords stored as MD5 hashes in /etc/shadow" [Medium,Triaged] https://launchpad.net/bugs/1807479 15:32 err 15:32 system-config-kickstart? 15:32 yo! 15:35 We've decided that it doesn't need to be assigned to the team 15:35 ando somebody'll have a look at the patch 15:36 the fix seems correct anyway 15:38 bug 1811694 15:38 bug 1811694 in aptdaemon (Ubuntu) "/usr/sbin/aptd:RuntimeError:_inline_callbacks:get_uid_from_dbus_name:return_value:_inline_callbacks:_check_simulated:_inline_callbacks" [High,Confirmed] https://launchpad.net/bugs/1811694 15:38 there are hundreds of dupes of that in the error tracker 15:38 make it go away! 15:38 aah 15:39 But yes, that seems like aptdaemon needs to be adapted to python 3.7 or something 15:39 bdmurray: +1 for accepting it and putting it on the backlog 15:39 will do 15:40 xnox: is bug 1815599 on your radar? 15:40 bug 1815599 in multipath-tools (Ubuntu) "multipath shows '#:#:#:#' for iscsi device after error injection" [High,New] https://launchpad.net/bugs/1815599 15:40 no 15:40 bdmurray, i hope that like cpaelzer and/or server team can take it 15:41 xnox: have you done more than hope? 15:42 gaughen: you'll bring it up with the server team yes? 15:42 doing so right now 15:42 so we're not taking it for disco but we "hope" server team does? :) 15:42 vorlon, yes 15:43 the last hope 15:43 no, a new hope 15:43 the final hope 15:43 the meow hope 15:44 bug 1798369 15:44 bug 1798369 in ubiquity (Ubuntu) "Reinstall Ubuntu (with preserving existing data) shows error message due to "Could not get lock /target/var/cache/apt/archives/lock"" [High,Triaged] https://launchpad.net/bugs/1798369 15:44 xnox: was not on my radar at all so far 15:44 the last best hope for mankind 15:44 A unicorn's hope 15:44 cpaelzer, can it be? i think it's "normal" to not recover the multipaths, but i might be wrong. they claim "it never recovers", or i think it does but "takes a while" 15:44 YES! 15:45 xnox: they didn't really answer to Frank, but I'll take a look 15:46 cpaelzer, yeah, try to move it along, by sounding to have multipath knowledge 15:46 the great pretender ... 15:49 I'll test the reinstall option and see if it still exists 15:49 cool 15:50 Okay, I think that's good for today. 15:50 bdmurray, you might need to trick the iso to think it is disco final.... 15:50 bdmurray, by like bind-mounting final sounding things into .disk 15:50 and /etc/os-release et.al. 15:50 that sounds involved 15:50 bdmurray, cause disco-on-disco will currently thinkg it is an upgrade from disco-to-19.04 15:51 instead of 19.04-on-19.04 reinstall 15:51 #topic Team proposed-migration report 15:51 #link http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses_by_team.html#foundations-bugs 15:52 u-u fix is on the way to unlock python-apt 15:53 doko, do we need the new fuse? 15:55 there was a reason ... 15:55 but the regression in fuse-zip is real 15:55 so we shouldn't let new fuse in, as is at the moment. 15:55 (as seen in the rebuild of fuse-zip, with new fuse support) 15:56 nest up gdb/python3.7 kind of related 15:56 python3.7 are blocked by openssl 1.1.1b compat (there is a bug report aobut that) 15:57 there's a new comment on the python issue 15:57 but should we let gdb through then? 15:57 (cause python3.X->gdb is openssl issues, and shouldn't block new gdb) 15:57 rbalint, python-apt blocked by unattended-upgrades regressions. Did you look at that? 15:58 u-u fix is on the way to unlock python-apt 15:58 https://bugs.python.org/issue35998 15:58 doko, fuse has uint/int abi bugs in the interface layear, which cuases hangs in fuse-zip testsuite, that hangs the VM. 15:58 rbalint, great! 15:59 re:perl -> there is progress in debian, the gdn/ndb regression compat is real, but we don't know if we care about acient dbs to be usable on new releases. 15:59 i guess i should upload that into ubuntu direct, instead of waiting for the fixed up upload in debian. 15:59 why? I don't think it's urgent 15:59 I think we should let the Debian maintainer fix it 16:00 unless we think that's not happening soon 16:00 vorlon, well it is urgent as IBM observe it, in their testing of disco. 16:00 vorlon, it was raised as a partner hwe issue. 16:01 it's buggy and they've identified the bug and it will be fixed by release 16:01 ok. 16:01 I don't think IBM have said it's blocking them from testing other stuff, have they? 16:07 #topic AOB 16:07 no 16:07 none here 16:07 #endmeeting