16:01 #startmeeting 16:01 Meeting started Thu Dec 3 16:01:09 2015 UTC. The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology. 16:01 16:01 Available commands: action commands idea info link nick 16:01 [TOPIC] Lightning round 16:01 o/ 16:03 $ echo $(shuf -e barry doko bdmurray slangasek caribou infinity sil2100 robru cyphermox pitti tdaitx xnox chiluk) 16:03 barry infinity sil2100 doko cyphermox tdaitx pitti xnox robru slangasek caribou chiluk bdmurray 16:03 \o/ 16:03 short week due to usa holiday 16:03 LP: #1519890; LP: #1520568; python issue 25698 (mysterious cpickle test regressions); debian bug #806824; debian bug #806826 and LP: #1518150; pyyaml 3.11-3; pip 7.1.2 (unsuccessfully unfutzing git repo & gbp issues); returning to dirtbike (rewheel) 16:03 Launchpad bug 1519890 in xapian1.3-core (Ubuntu) "[MIR] xapian1.3-core" [Undecided,Fix committed] https://launchpad.net/bugs/1519890 16:03 Launchpad bug 1520568 in nss (Ubuntu) "All queries fails when 'google' is used: ERR_SSL_PROTOCOL_ERROR" [Critical,Confirmed] https://launchpad.net/bugs/1520568 16:03 Debian bug 806824 in src:libpeas "libpeas: Split Python 2 and 3 support" [Wishlist,Open] http://bugs.debian.org/806824 16:03 Debian bug 806826 in src:pyyaml "pyyaml does not support literals in unicode over codepoint 0xffff" [Normal,Fixed] http://bugs.debian.org/806826 16:03 --done-- 16:04 No infinity today 16:04 - Landing team work, silo coordination, preparing landing e-mails 16:04 - RTM Status meetings 16:04 - system-image: 16:04 * Work on the testability branches, making tag-image and copy-image testable 16:04 * Initial work on default-phased-percentage per channel 16:04 - +1-maintenance: 16:04 * Publishing the network-manager-applet 1.0.6 merge (with a tiny whoops) 16:04 * Publishing the network-manager-pptp 1.0.6 merge 16:04 * Preparing merge for n-m-a 1.0.8 16:04 - Review of OTA-8.5 hotfix bugs 16:04 - Discussions about auto-download of puritine in ubuntu-pd 16:04 - Improving the safe-copy-image wrapper to do more extensive checks 16:04 - Promoting a devel-proposed image to devel 16:04 (done) 16:05 doko: hi there 16:05 - updated GCC 5 fsf and linaro packages 16:05 - evaluated linaro test builds 16:05 - build cross compiler defaults from gcc-defaults, get rid off all gcc-defaults-cross-* sources 16:05 - bug forwarding for binutils and GCC 16:05 - still investigating python2.7 test failures 16:05 - llvm-toolchain-snapshot build now on all archs 16:05 - some merges 16:05 (done) 16:05 - ubiquity still waiting for review: 16:05 - https://pad.lv/c/~mathieu-tl/ubiquity/disable-verification/+merge/278668 16:05 - network-manager (1.0.4-0ubuntu5.2): fixes for IPv6 MTU flap 16:06 - many d-i merges: 16:06 - netcfg, user-setup, apt-setup, installation-report, preseed, rootskel 16:06 - partman-auto-raid, partman-iscsi, partman-btrfs 16:06 - ... and more still on the list. 16:06 - in progress merges: 16:06 - tasksel, kbd 16:06 - lots of d-i installs for smoke testing 16:06 - grub2: more TFTP timeout cherry-picks for bug LP: #1521612 16:06 Launchpad bug 1521612 in grub2 (Ubuntu) "Continued TFTP timeouts when PXE booting via grub" [High,New] https://launchpad.net/bugs/1521612 16:06 (done) 16:06 \o/ amazing 16:06 xnox: can you help out with reviewing that ubiquity MP cyphermox has waiting? 16:07 slangasek, yes. 16:07 let me go last, still writing down the stuff from last week 16:07 to go with the d-i component merges there may be some conversion from bzr to git in my future. 16:07 xnox: ta 16:07 tdaitx: check 16:07 pitti: 16:07 autopkgtest: 16:07 - Automate daily maintenance for cleaning up behind cloud trouble 16:07 - Disable apt pinning for PPA requests (#1520505) 16:07 - Fix missing restricted/multiverse apt sources in precise images (fallout from cloud-init SRU) 16:07 - Generate http://autopkgtest.ubuntu.com/testlist.shtml 16:07 - Rework charm configs and multiple cloud config handling, so that we are back to fully automatic deployment and keeping all configs in git (this had become too hackish and hand-customized over time) 16:07 - In progress: test LXD in scalingstack on arm64, figure out some config issues, firewall troubles (RT pending), LP #1522026 16:07 Launchpad bug 1522026 in Auto Package Testing "armhf lxd container does not start on arm64 system" [Undecided,New] https://launchpad.net/bugs/1522026 16:07 - make it possible to run britney in a local setup against a PPA (for CI train) 16:07 - rework britney code and data structures to be simpler and more robust, making the code much easier to understand and fixing some bugs 16:07 Merges: grabbed a few old ones from other people: cdebconf, check, debhelper, docbook-dsssl, gegl, gnupg, pptpd 16:07 misc: 16:07 - fix dkms crash in apport hook (#1303275), fix FTBFS of goplay 16:07 - some sponsoring (gnome-desktop3, rsyslog, etc.) 16:07 - systemd: Add autopkgtest for handling cryptsetup LUKS devices 16:07 - in progress: Discuss moving ISO and upgrade testing from s-jenkins to Jenkaas 16:07 EOT 16:07 xenial bootstrap archive was almost completed, got pushed to <<100 failures 16:07 a bunch of people put all of that into launchpad, hence the builds are running there as of today. amazing work all of them. 16:07 resolving in-launchpad FTBFS now 16:07 spent a day in bluefin catching up with people, and CPC team 16:07 Filed MIR requests: https://bugs.launchpad.net/~ubuntu-mir/+subscribedbugs?field.tag=s390x 16:08 next up: d-i & iso priority. 16:08 .. 16:08 robru: next! 16:08 xnox: what do you mean by d-i and iso priority? 16:09 cyphermox, getting them built and working for s390x. neither of those two tasks are done, and there is a 16th of december deadline for those. 16:09 alright. just making sure again that we don't step on each others' toes 16:10 we shouldn't.... 16:11 * short week, Thanksgiving and was off on Monday 16:11 * merges 16:11 * s390x bootstrapping 16:11 * UEFI 2.5 discussions 16:11 * discussions around ZFS support on Ubuntu 16:11 * looking at a race condition in the cryptsetup initramfs script (LP: #1481536) 16:11 Launchpad bug 1481536 in cryptsetup (Ubuntu) ""cryptsetup: unknown fstype, bad password or options?" error unlocking / decrypting LUKS volume at boot" [High,Confirmed] https://launchpad.net/bugs/1481536 16:11 (done) 16:11 caribou: 16:11 * doko sees the first s390x builder online ... 16:11 Bugfix: 16:11 - Icehouse failed migration 16:11 - LP: #1318111 - repetitive crashkernel 16:11 Launchpad bug 1318111 in kexec-tools (Ubuntu) "Adds more and more copies of ‘crashkernel=384M-:128M’ in /etc/default/grub when upgrading or reinstalling grub-pc" [Medium,Confirmed] https://launchpad.net/bugs/1318111 16:11 - LP: #1496317 - kdump OOM killer failure 16:11 - Nut merge 16:11 - Rsyslog merge 16:12 Launchpad bug 1496317 in kexec-tools (Ubuntu Wily) "kexec fails with OOM killer with the current crashkernel=128 value" [High,In progress] https://launchpad.net/bugs/1496317 16:12 Knowledge Brief review & redaction 16:12 ☑ Done 16:12 doko, it did do 4k builds already. they are cloning that builder at the moment to get more going. 16:12 xnox: 4k? https://launchpad.net/ubuntu/xenial/+builds?build_text=&build_state=built&arch_tag=s390x shows only 154 successful builds 16:12 Working on 1432871 coreutils ftbfs - Turns out the testcase needed to be updated. None of the testcases were being run prior because df would fail to read the mount table, and thus were auto-skipped. The root issue appears to still exist upstream. debdiff that fixes the failing testcase is waiting in 1432871 for upload and SRU approval. 16:12 http://pad.lv/1347788 - Fix exists on unstable branch. May consider moving xenial tun unstable find-utils like fedora, centos, and suse. 16:12 --done-- 16:13 Launchpad bug 1347788 in findutils (Ubuntu Xenial) "find crashed when current working directory is not readable and -exec or -execdir used" [Low,Confirmed] 16:13 working on a mojo spec for the error tracker 16:13 researched and gave stub info about how to rebuild bucketversionscount 16:13 uploaded apport to xenial resolving issue parsing dpkgterminallog with python2 16:13 verified the fixes for LP: #1463680 and LP: #1511783 are good 16:13 Launchpad bug 1463680 in ubuntu-release-upgrader (Ubuntu) "mirrors only added to mirrors.cfg, never removed" [Medium,Fix released] https://launchpad.net/bugs/1463680 16:13 for phased updates overrode some nautilus, xorg-server regressions 16:13 Launchpad bug 1511783 in ubuntu-release-upgrader (Ubuntu) "release upgrader can create a too minimal sources.list" [High,Fix released] https://launchpad.net/bugs/1511783 16:13 T, W SRU of bug fix for python-apt bug 1503979 16:13 emailed jibel about how to get data out of errors.u.c using API 16:13 updated merge-o-matic to check if there is a new package version in -proposed (LP: #1518113) 16:13 bug 1503979 in python-apt (Ubuntu Wily) "ubuntu-support-status throws exeption No date tag found (regression)" [Medium,In progress] https://launchpad.net/bugs/1503979 16:13 Launchpad bug 1518113 in Merge-o-Matic "Please show -proposed package version information on merges.u.c" [Medium,Fix released] https://launchpad.net/bugs/1518113 16:13 merged cyphermox, slangasek changes to merges.ubuntu.com 16:13 slangasek, yes. launchpad is weird it was giving me out of 4k builds until i force refresh things. oh well =) wishful thinking. and probably stale yui js. 16:13 added subscribe to package code to ubuntu-qa-tools 16:13 subscribed foundations-bugs to a bunch of s390 packages 16:13 SRU review of glib2.0 for seb128 16:13 created google code in task for launchpad-gm-scripts 16:13 ✔ done 16:14 yeah, 4k would be too many, that would be main done. 16:14 Short week: on vacation from Nov 26 to Dec 02 16:14 Updates from Nov 19 to Nov 25: 16:14 - Provided OpenJDK 6b37 1.13.9 packages for the security team (new update for Wily and backports for Vivid, Trusty, and Precise) 16:14 - Follow up on OpenJDK 7u91 2.6.3 updates to the security team 16:14 - OpenJDK bug triage and fixes 16:14 - Upstreamed LP: #1518741 16:14 Launchpad bug 1518741 in openjdk-7 (Ubuntu) "A pkg-config file is need for OpenJDK distributions" [Undecided,New] https://launchpad.net/bugs/1518741 16:14 - Need help from someone who knows pkgconfig better than I do on LP: #1518741 16:14 - Gone through lots of emails since I was back yesterday afternoon/evening 16:14 EOF 16:15 robru: around? 16:15 ops, back copy&paste: upstreamed LP: #932274 16:15 Launchpad bug 932274 in Iced Tea "Regression: Unreadable menu bar with Ambiance theme in Java/Swing GTK L&F" [Medium,In progress] https://launchpad.net/bugs/932274 16:16 any questions over status? 16:20 [TOPIC] AOB 16:20 pitti, xnox: could we sync quickly regarding autopkgtest on s390x? 16:21 we did in a PM, but sure, let's repeat here 16:21 (here, or after meeting) 16:21 so my "minimal must have" requirements are: 16:21 - deboostrap works 16:21 pitti: we haven't all three of us synced in a PM ;) 16:21 - LXC works 16:21 - I get at least 15 CPUs and 30 GB of RAM to have enough capacity to keep up 16:21 (preferably distributed over two VM instances, or perhaps 4 16:22 that. 16:22 there's a longer list of things which make this nicer (boot loader for vmdeboostrap and QEMU, cloud images, scalingstack etc.), but that's not critical 16:22 slangasek, once debootstrap works and lxc works, and like i can run one autopkgtest successfully the request will be to get machine(s) access to pitti to hook the rest of things up. 16:23 pitti: so one thing I heard said earlier was that we maybe didn't want to turn on autopkgtest gating for s390x until more is built because otherwise things will fail due to dependencies 16:23 and I disagree with that 16:23 why wouldn't we do that? 16:23 we only hold back on regressions, not on failures 16:23 and we now track regression vs. alwaysfail per-arch 16:23 why wouldn't we enable normal gating, I mean 16:23 slangasek, i never said anything about disabling gating... 16:23 I think we should turn on gating sooner rather than later 16:24 xnox: ok, that's what I heard, so I wanted to be clear :) 16:24 so even if all tests fail initially, that just won't catch regressions, but not hold anything back 16:24 slangasek, in that case - when can we have autopkgtest machines provisioned? =) 16:24 and shall those just be clones of buildds ?! =) 16:24 xnox: I don't know; we need to talk to kip/tinoco/cpaezler 16:25 (well, autopkgtest need different network and firewall, so assume it's not the same as lp-buildds) 16:25 the buildds aren't done with the same VM config we requested for autopkgtest 16:25 slangasek, and by we you mean I, or You? =) 16:25 xnox: yes 16:25 anyway 16:25 lovely. 16:26 assuming there are no hang-ups on the VM provisioning side, I think we should be aiming to have autopkgtest enabled for s390x by early next week 16:26 ack. 16:26 since a lot of packages are getting built right now, and this may be our one chance to *get* a baseline for their test results in xenial 16:27 unless pitti has another idea about how to get baseline results en masse, after the packages have already built+migrated to xenial? 16:27 i assume we can retrigger testing of all -release pocket on s390x. 16:27 slangasek: we can trigger the tests against -release too 16:28 and if they pass, that will become the baseline. 16:28 ok 16:28 britney won't do it for us of course, but it's easy enough to come up with a list of run-autopkgtest commands to do it 16:28 then that's probably a good idea to do, and then we don't have to worry so much about the timing of the autopkgtest bring-up 16:28 I still want us to get that ASAP however :) 16:28 i. e. we grab s390x Packages.gz, and issue a request for all packages that we built 16:29 [ACTION] slangasek and xnox to pursue s390x VMs for autopkgtesting before next week 16:29 * meetingology slangasek and xnox to pursue s390x VMs for autopkgtesting before next week 16:30 alrighty 16:30 xnox: and nice work getting s390x going 16:31 thank you, there is a lot of stuff to fix still. 16:31 I'd congratulate infinity too but I'm glad he's elsewhere :) 16:31 anything else? 16:32 well him, colin and william did the black magic of bring up first buildd and that was dandy. 16:32 heh, its disabled now 16:33 the builder? it's torn down for cloning 16:33 but I like the build times, some under 1min 16:33 they'll be making 28 of them shortly 16:34 probably the easiest new arch bring-up ever, at least once the hardware is in place ;) 16:35 and it seems we've got plenty this time 16:35 I still remember the scarcity of arm back then 16:35 doesn't take much mainframe to split up into chunks that are useful 16:35 yes, there's generally no shortage of server-grade zSeries hardware 16:36 obviously what we should do is move all of prodstack onto the z13 16:37 #endmeeting