15:02 <slangasek> #startmeeting
15:02 <meetingology> Meeting started Wed Jun 12 15:02:24 2013 UTC.  The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
15:02 <meetingology> 
15:02 <meetingology> Available commands: #accept #accepted #action #agree #agreed #chair #commands #endmeeting #endvote #halp #help #idea #info #link #lurk #meetingname #meetingtopic #nick #progress #rejected #replay #restrictlogs #save #startmeeting #subtopic #topic #unchair #undo #unlurk #vote #voters #votesrequired
15:02 <slangasek> [TOPIC] lightning round
15:02 <slangasek> $ echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek cjwatson xnox stokachu)
15:02 <slangasek> stokachu slangasek bdmurray doko stgraber ev barry jodh cjwatson xnox
15:02 <ev> wooo, time
15:03 <bdmurray> I will type slowly to bide you time
15:03 <stokachu> i just got one bug on the list bug 962046
15:03 <ubottu> bug 962046 in python-boto (Ubuntu Quantal) "EC2 metadata retrieval fails with spaces in a resource name" [Medium,In progress] https://launchpad.net/bugs/962046
15:03 <stokachu> the sru template is in the comment but ill move it to the description in a min
15:03 * xnox quickly runs to generate a cup of coffee
15:03 <stokachu> done.
15:05 <slangasek> stokachu: is the bug still applicable to saucy?  It's marked 'triaged'
15:05 <stokachu> dosaboy: ^
15:05 <slangasek> stokachu: and this is at the "Needs upload sponsorship" stage for the SRUs?
15:05 <ev> bdmurray: :D
15:06 <stokachu> slangasek: as far as im told they are waiting on upload as well
15:06 <stokachu> if i dont hear from dosaboy in the next hour ill double check saucy status
15:06 <slangasek> ok
15:07 <slangasek> anyone here want to pick those up for SRU sponsorship?
15:08 <xnox> i can
15:08 <stgraber> slangasek: I'll do it once we know that it's indeed fixed in saucy
15:08 <slangasek> xnox got there first :)
15:08 <slangasek> xnox: thanks :)
15:08 <slangasek> * seem to have fixed my kernel swap death issue now by disabling /tmp on tmpfs - thanks, xnox!
15:08 <slangasek> * working with Phonedations team on the container flip - udev vs. ueventd, partitioning questions
15:08 <slangasek> * established a regular biweekly sync with the QA team to keep aligned on priorities
15:08 <slangasek> * prep for Plumbers and for the upcoming client sprint
15:08 <slangasek> (done)
15:09 <stokachu> slangasek: is that ods in nov?
15:09 <bdmurray> uploaded update-manager to S after making it no longer check for auto-launch
15:09 <bdmurray> uploaded update-notifier to S removing auto-launch key again
15:09 <bdmurray> searched ubuntu archive using ubuntu-security-tools for update-notifier auto-launch usage
15:09 <bdmurray> tested bug 981461 regarding WoL
15:09 <ubottu> bug 981461 in ifupdown (Ubuntu Precise) "Network interfaces are not correctly brought down on halt, disrupting Wake-on-LAN" [High,Triaged] https://launchpad.net/bugs/981461
15:09 <bdmurray> research into ubuntu-drivers-common / nvidia problem seen on errors
15:09 <bdmurray> errors - redesign of bucketsystems column family
15:09 <bdmurray> submitted merge proposals for daisy, oops-repository, errors
15:09 <bdmurray> fixed 'All versions' being missing from the versions table in errors
15:09 <bdmurray> fixed an OOPS with errors
15:09 <slangasek> stokachu: Plumbers is in September; the client sprint is a tech-lead-y, manager-y meeting at the end of July
15:10 <stokachu> ah
15:10 <bdmurray> fixed a unicode issue with errors
15:10 <bdmurray> email to ubuntu bug control regarding error tracker
15:10 <bdmurray> merge proposal for change-override to set the phased_update_percentage
15:10 <bdmurray> tested Phased-Update-Percentage=0 and Always-Install-Phased-Updates apt option
15:10 <bdmurray> improvements to my phased-updater code
15:10 <bdmurray> irc discussion and investigation into bug 982082 with barry
15:10 <ubottu> bug 982082 in ubuntu-release-upgrader (Ubuntu) "do-release-upgrade crashed with IOError in init_proxy(): [Errno 5] Input/output error" [Medium,Incomplete] https://launchpad.net/bugs/982082
15:10 <bdmurray> doneā€¼
15:11 <doko> - updated cross compilers
15:11 <doko> - fixed cross build of native compiler
15:11 <doko> - binutils update to the trunk, updated cross binutils
15:11 <doko> - preparing gcc upload using new binutils
15:11 <doko> - preparing first test rebuild of saucy
15:11 <doko> - arm64 stage0 bringup, together with cjwatson
15:11 <doko> (done)
15:11 <stgraber> slangasek: I hope to find some time this week to look at the container flip again, last week I was mostly stuck on issues related to read-only / and the Android container behind a bit unhappy about this (ueventd wouldn't start, so SF would fail to start and the container would die taking the system with it)
15:11 <xnox> doko: when will that rebuild start? weekend?
15:12 <doko> I hope so. need to upload that gcc
15:12 <dosaboy> stokachu: bug just says precise and quantal but is quite old so maybe we should apply to saucy too?
15:13 <stgraber> Blueprint-related work:
15:13 <stgraber> - Image based updates (BLUEPRINT: foundations-1305-image-based-updates)
15:13 <stgraber> - Waiting on review of the GPG implementation plan, need to get the keys generated and signed after that.
15:13 <stgraber> - Waiting for IS to setup the image based update server.
15:13 <stgraber> - Implemented xz compression/decompression functions in the server code
15:13 <stgraber> - Implemented GPG keys generation, keyring generation and file signature functions in the server code
15:13 <stgraber> - Good progress on the publishing code (generation of channels.json, index.json and individual update publication)
15:13 <stgraber> - Registered a new LP project (ubuntu-system-image) and team to store all the code related to image based upgrades
15:13 <stgraber> - Clarified some details in the specs
15:13 <stgraber> - Discussed OEM/carrier customizations and how that'll work with image based updates
15:13 <xnox> dosaboy: how does one exactly test this? should I start an instance in ec2? cause just running the testcase script here locally, simply hangs indefinately.
15:13 <stgraber> Other work:
15:13 <stgraber> - release/cdimage
15:13 <stgraber> - Some follow-up e-mails on ~ubuntu-sru expiration of some current members
15:13 <stgraber> - Ubuntu touch
15:13 <stgraber> - Some more discussions on partitioning and containers
15:13 <stgraber> - Discussed OMA with asac and lool
15:13 <stgraber> - LXC
15:13 <stgraber> - Usual code reviews
15:13 <stgraber> - Got bored of my laptop wasting a lot of CPU keeping thunderbird happy, so switched back to offlineimap+mutt, still tweaking the configs so poke me if you see something that doesn't look right (encoding, gpg signature, source address, ...)
15:13 <stgraber> 
15:14 <stgraber> TODO:
15:14 <stgraber> - Try and get the rest of the image updates server tools implemented and tested so they are ready to be used once we have the server online and the keys generated
15:14 <stgraber> - Follow-up with the phonedation team on getting the container fliped images to use a read-only rootfs and use the planned partition layout
15:14 <stgraber> - Finish self-rebuilds feature implementation on nusakan
15:14 <stgraber> - Process some pending merges (ifupdown and resolvconf)
15:14 <stgraber> (DONE)
15:14 <ev> - Small bug fixes to our back population jobs (that are still falling over -
15:14 <ev> working with gnuoy to find out what is sending them SIGKILL).
15:14 <ev> - Investigating the retracers all falling over because all the Cassandra
15:14 <ev> connections went away. Discussed with Tom and decided to increase the timeout
15:14 <ev> for individual connections (it was silly-low) and provide better metrics on
15:14 <ev> connection failures.
15:14 <ev> - Fixed our usage on txstatsd reporting into Graphite, with help from sidnei.
15:14 <ev> We now have graphs of pycassa connection creations, failures, and disposals.
15:14 <ev> This should really help debug some of the "OMG every node is telling us to
15:14 <ev> go away" problems we see from time to time, especially on the retracers:
15:14 <ev> https://graphite.engineering.canonical.com/render/?width=1129&height=614&_salt=1371049545&target=statsd.whoopsie-daisy.daisy.cassandra_connection_failures.rate&from=-48hours
15:14 <ev> - More discussion with the server team on error reporting. Making the case for
15:14 <ev> a default option in the d-i module of, "Yes, I'd like to help make Ubuntu
15:14 <ev> better by turning on automatic error reporting."
15:14 <ev> - Lots of investigation around why Keystone is *still* throwing 403s at us
15:14 <ev> throughout the day when writing core dumps into Swift. We've got myself, the
15:14 <ev> GSAs, webops, and U1 looking into it now, so hopefully we'll have a solution
15:14 <ev> soon.
15:14 <ev> - Made my biggest changes to lp:canonical-memento yet, removing HAProxy from in
15:14 <ev> front of the Cassandra nodes:
15:14 <ev> https://rt.admin.canonical.com/Ticket/Display.html?id=62027
15:14 <ev> - Discussions with legal on giving Kees access to the error reporting data for
15:14 <ev> the purposes of hunting security vulnerabilities. Now have an amended
15:14 <ev> document written, just waiting for the signatories to return from events.
15:14 <ev> - Code review for Brian.
15:14 <ev> - Our usual review process failed quite spectacularly for some changes in
15:14 <ev> support of phased updates. Shepherded some quick fixes through with the help
15:14 <ev> of webops, but there was a deeper problem.
15:14 <ev> - In investigating what was going wrong with the above, finally hooked the unit
15:14 <ev> tests for oops-repository up to Tarmac (so we have a safety net for human
15:14 <ev> error in the merge review process). I also fixed the bug.
15:14 <ev> (done!)
15:15 <barry> image based upgrades - the gpg regime is almost complete (one last pathological case to handle).  should be merged into the main trunk today or tomorrow.
15:15 <barry> LP: #1058884 - all SRUs uploaded, awaiting approval
15:15 <ubottu> Launchpad bug 1058884 in python3.3 (Ubuntu Raring) "Race condition in py_compile corrupts pyc files" [High,In progress] https://launchpad.net/bugs/1058884
15:15 <barry> LP: #982082, LP: #1094218 - thought i had sru'd the fix, but maybe i didn't?
15:15 <ubottu> Launchpad bug 982082 in ubuntu-release-upgrader (Ubuntu) "do-release-upgrade crashed with IOError in init_proxy(): [Errno 5] Input/output error" [Medium,Incomplete] https://launchpad.net/bugs/982082
15:15 <ubottu> Launchpad bug 1094218 in lsb (Ubuntu) "lsb_release crashed with IOError in getstatusoutput(): [Errno 10] No child processes (called by teamviewerd)" [Medium,In progress] https://launchpad.net/bugs/1094218
15:15 <barry> done
15:15 <stokachu> stgraber: are you a fan of teh docker.io?
15:16 <stgraber> stokachu: I haven't used it myself, but we are in touch with the guys behind it and try to improve the upstream LXC library to better fit their need and reduce the code duplication (they currently wrap our binaries which isn't really ideal)
15:17 <stokachu> nice!
15:17 <xnox> stgraber: if you have any offlineimap patches, feel free to poke me and include them upstream/debian/ubuntu.
15:17 <dosaboy> xnox: i did not personally do the test, let me find out
15:17 * xnox really should resurrect my offlineimap setup.
15:17 <slangasek> stgraber: resolvconf> so I know Thomas wanted us to drop some of our delta around the maintainer scripts... have you worked through whether that's the right thing to do / safe?
15:18 <stgraber> slangasek: no. I haven't actually looked at the new resolvconf yet but I remember Thomas' e-mail. My guess is that I'll start with a simple merge, then we can look at reducing the delta after that (should reduce the number of regressions vs doing all of that at once)
15:19 <xnox> dosaboy: yeah, at the moment I feel like setting the bug to incomplete because it's not possible to follow the test case steps to reproduce the problem. I am guessing it assumes some kind of special environment where that python snippet should be executed.
15:19 <slangasek> stgraber: ok.  I think the "simple merge" is going to have to deal with this issue, because of how much the code has changed on the Debian side.
15:20 <stgraber> slangasek: ah, is it that bad? as I said, I haven't looked at the new version yet, but if he did some major changes to the maintainer scripts, then I guess we won't have much choice indeed...
15:20 <slangasek> cjwatson: no jodh today, you're up
15:20 <cjwatson> Usual round of syncs, merges, transition management, and pushing NBS up the hill.
15:20 <cjwatson> Fixed udev-udeb startup script regression that broke d-i (bug 1188864).
15:20 <ubottu> bug 1188864 in systemd (Ubuntu) "/dev/pts not getting mounted before install" [Critical,Fix released] https://launchpad.net/bugs/1188864
15:20 <cjwatson> foundations-1305-arm64-bringup: Assembled stage0 for new arm64 port, with doko.  Lots of cross-builds, and some fixes pushed to saucy and forwarded to Debian.  debootstrap --variant=buildd now works under emulation, plus apt and debhelper.
15:20 <cjwatson> foundations-1305-click-package: Added buildsource subcommand to click-package.
15:20 <cjwatson> Worked on autopkgtest integration in proposed-migration.  I have the general shape of this now and am debugging it into existence.
15:20 <cjwatson> Debugged a crash in proposed-migration, caused by incorrect undo handling for overlapping virtual packages in different sources involved in the same hint.  Several hours for a one-line fix.
15:21 <cjwatson> ..
15:21 <xnox> * Upstart: Uploaded full serialisation cherrypick into saucy, with
15:21 <xnox> additional postinst checks as per slangasek's comments. Waiting on
15:21 <xnox> smoser to test. And will upload into raring-proposed this week.
15:21 <xnox> * Helped out hunting down why ecryptfs-tools moved into minimal, and
15:21 <xnox> moved it back by uploading adduser.
15:21 <xnox> * gcc-bionic toolchain: gitified using repo tool, dropped java
15:21 <xnox> build-dep, enabled parallel build, pushed to github, send email to
15:21 <xnox> doko & phonedations.
15:21 <xnox> * cryptsetup: changed to conditionally include cryptsetup tools and
15:21 <xnox> modules only when required. Tested all cross-combinations of common
15:22 <xnox> setups and uploaded. Wrote a blog post to ubuntu planet about the
15:22 <xnox> changes.
15:22 <xnox> * Generated some more overall copyright documents.
15:22 <xnox> * Upstart Jobs: util-linux: forwarded hwclock upstart jobs to
15:22 <xnox> debian. (adjusted to suit debian)
15:22 <xnox> * Installer: have a patch to drop pkexec (testing remaining flavours),
15:22 <xnox> todo fix up logind integration, todo investigate why U1 page hangs.
15:22 <xnox> ..
15:22 <slangasek> xnox: cryptsetup++
15:23 <slangasek> now we get to wait for all the bug reports to trickle in against plymouth from the users who were using ecryptfs and now only see the Ubuntu splash screen for a second at boot ;)
15:23 <xnox> slangasek: still needs a forward to debian, as it does a lot to determine whether it's needed or not, only to say "setup="yes"" and include it unconditionally anyway.
15:24 <slangasek> xnox: hwclock> interesting... considering Debian has (incorrectly) opted to use /etc/adjtime instead of a simple config file listing UTC=yes/no, how did that forwarding go?
15:24 <xnox> slangasek: it's a "port"
15:25 <slangasek> hmmmm, ok. :)
15:25 <slangasek> [TOPIC] Bugs
15:26 <xnox> slangasek: i have a few questions about their "integration" basically the package was moved to use /etc/adjtime because of systemd-timedated blindly ignoring existing hwclock.sh settings or actually changing all other packages that read & use UTC option.
15:26 <slangasek> this is probably a short topic today, stokachu gave his 1 bug earlier and bdmurray seems to have been flagging attention to stuff on the list
15:26 <slangasek> xnox: hmm, I don't think that was the reason for changing the package
15:27 <bdmurray> I agree with slangasek as I don't have anything else
15:27 <slangasek> xnox: mostly it was the sysvinit maintainer wanting to pare down /etc/default/rcS to be able to make it a conffile
15:27 <xnox> slangasek: well it was consolidating two settings into one setting in one location, but chose the wrong file yet the stated claim is that "it's all universal now"
15:28 <slangasek> heh, ok
15:28 <cjwatson> slangasek: did you find time for those clock-setup/ubiquity changes, or do I need to?
15:29 <xnox> slangasek: hmm.... do you have a pointer/URL about this: "wanting to pare down /etc/default/rcS to be able to make it a conffile"
15:29 <xnox> slangasek: ?
15:29 <slangasek> xnox: the historical assertion by Keybuk is that we don't need adjtime /at all/.  I haven't proven this to myself however, so I haven't tried to argue it to the util-linux Debian maintainers... maybe you could dig into this and get it sorted?  It's not too late to reverse the /etc/adjtime usage in Debian
15:29 <slangasek> xnox: um, the pointer is in my brain, it references IRC conversations with rleigh
15:29 <slangasek> cjwatson: haven't done so yet, but I will this week
15:29 <cjwatson> ok
15:29 <xnox> slangasek: systemd uses /etc/adjtime
15:30 <xnox> slangasek: already as integrated in util-linux package in debian.
15:30 <slangasek> xnox: this is no proof that it's correct :P
15:30 <xnox> slangasek: correct, I do not state that.
15:30 <slangasek> xnox: the changes to util-linux were *not* driven by systemd, they were driven by the sysvinit maintainer
15:30 <xnox> slangasek: ok. i shall reconcile all of those things in my ahead at one point.
15:30 <slangasek> [TOPIC] AOB
15:31 <slangasek> anything else?
15:31 <slangasek> rants about bugs?  movie reviews?
15:31 <bdmurray> no spoilers though!
15:31 <barry> kaiser soze
15:31 <cjwatson> the titanic sinks at the end
15:31 <xnox> is the new superman any good / worth going to? (man of steel.... such a pun on 'iron man')
15:32 <bdmurray> Of course it is worth going to
15:32 <bdmurray> Its Superman!
15:32 * xnox is going to see Swan Lake this evening at Royal Albert Hall
15:33 <cjwatson> I'd be surprised if the nickname "Man of Steel" didn't predate the Iron Man comics by a couple of decades :)
15:34 <cjwatson> first Superman comic 1938, first Iron Man comic 1963
15:34 <slangasek> nonsense, all media was invented in 1998
15:34 <cjwatson> films I'm watching lately aren't new - taking the kids through Lord of the Rings and Harry Potter
15:35 <xnox> i'm yet to see the first/old star wars episodes.....
15:36 * bdmurray walks away
15:36 <slangasek> :D
15:36 <slangasek> #endmeeting