15:03:37 <slangasek> #startmeeting
15:03:37 <meetingology`> Meeting started Wed May  2 15:03:37 2012 UTC.  The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
15:03:37 <meetingology`> 
15:03:37 <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
15:04:20 <slangasek> [TOPIC] Lightning round
15:04:55 <slangasek> $ echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek ogra infinity cjwatson xnox)
15:04:58 <slangasek> doko stgraber cjwatson slangasek barry xnox jodh ogra ev infinity bdmurray
15:05:01 <slangasek> $
15:05:07 <ogra_> €
15:05:13 <slangasek> £
15:05:38 <slangasek> doko: you first :)
15:06:05 <doko> - gcc updates for quantal
15:06:05 <doko> - investigate and fix arm biarch builds
15:06:05 <doko> - may 1st was bank holiday
15:06:08 <doko> (done)
15:06:28 <stgraber> Was off Friday and Tuesday:
15:06:28 <stgraber> - Blueprints
15:06:28 <stgraber> - Registered community-q-edubuntu, foundations-q-ltsp, foundations-q-iso-tracker and foundations-q-networking
15:06:32 <stgraber> - Installer
15:06:34 <stgraber> - Merged some dconf improvements and turned on compositing in the installer.
15:06:37 <stgraber> - Fixed some netboot bugs and turned on cifs support in casper.
15:06:40 <stgraber> - Merged os-prober from Debian
15:06:42 <stgraber> - Release
15:06:45 <stgraber> - ISO testing
15:06:47 <stgraber> - Ubuntu/Edubuntu 12.04 release.
15:06:50 <stgraber> - TODO this week
15:06:52 <stgraber> - Look for any missing blueprint and register them
15:06:55 <stgraber> - Continue with merges (currently working on ifupdown)
15:06:57 <stgraber> - Push some SRUs (nagios-nrpe and ldm are currently on my list)
15:07:00 <stgraber> (DONE)
15:07:05 <cjwatson> Release sprint.  Nothing went too badly wrong. :-)
15:07:05 <cjwatson> Opened quantal and got auto-syncs running.  Some degree of fighting with Launchpad, but less than usual.  Piles of merges.
15:07:08 <cjwatson> Removed the data.tar.xz dpkg Pre-Depends requirement from Launchpad.
15:07:11 <cjwatson> Sorted out a few specs.
15:07:13 <cjwatson> Got a bit carried away with update-manager and ubiquity Python 3 porting.  Lots of dependencies still, and I haven't got much into bytes/unicode data modelling, but making good progress on tests.
15:07:17 <cjwatson> .
15:08:30 <slangasek> * upstart in Debian:
15:08:30 <slangasek> * prereq patches for lsb, sysvinit, and debhelper are done or pending
15:08:30 * infinity nudges Mr Manager.
15:08:30 <slangasek> * hacking on invoke-rc.d to implement the Debian policy proposal (doing what /lib/init/upstart-job does currently)
15:08:33 <slangasek> * plymouth patches wanted for mountall, looking at it this week
15:08:35 <slangasek> * SRUs for multiarch libraries, hdparm, other things
15:08:38 <slangasek> * fix upgrade issue in libnfsidmap (bug #933870)
15:08:39 <ubottu> Launchpad bug 933870 in libnfsidmap (Ubuntu) "libnfsidmap2 fails to install" [High,Fix released] https://launchpad.net/bugs/933870
15:08:40 <slangasek> * don't restart kdm on eglibc upgrade (bug #985735)
15:08:41 <ubottu> Launchpad bug 985735 in eglibc (Ubuntu) "update-manager restarts KDM and interrupts update process" [High,Fix released] https://launchpad.net/bugs/985735
15:08:43 <slangasek> * help fix upgrade issues with python (bug #986374)
15:08:44 <ubottu> Launchpad bug 986374 in python2.7 (Ubuntu) "oneiric->precise upgrade failed: E:Internal Error, Could not early remove python-minimal" [Critical,Fix released] https://launchpad.net/bugs/986374
15:08:45 <slangasek> * sponsor upstart fixes from James
15:08:48 <slangasek> * merges merges merges
15:08:50 <slangasek> * python3 porting: bluez ported (thanks Laney for the upload), gdb breaking my brain
15:08:53 <slangasek> * UDS planning - get those blueprints in, everybody!
15:09:13 <barry> slangasek, cjwatson: \o/
15:10:28 * barry assumes slangasek is done...
15:10:30 <barry> (reporting for last two weeks)  py3 porting status: debian bug 669301 (py3 packaging for python-openssl, should probably temporarily delta this into quantal).  freedesktop bug 48904 (backward compat problem w/ dbus.gobject_service).  worked on packaging for python-mode.el for sid.  released flufl.* packages with updated cheeseshop py3 trove classifiers.  also spent some time on bug 924079 before release, but punted.  reviewed debian bug
15:10:30 <barry> 625509 (debian-python port).  reviewed slangasek's patch for bluez py3.  bug 990145 (remove gwibber dependency on mx.DateTime).  bug 992195 (remove duplicity dependency on GnuPGInterface).  TODO: will be sprinting w/local pythonistas tomorrow on pep 420 (namespace packages for py3.3) and continue with bug 992195.  done.
15:10:34 <slangasek> er yes, sorry
15:10:35 <ubottu> Debian bug 669301 in pyopenssl "pyopenssl: Build the Python 3 version of the package" [Wishlist,Open] http://bugs.debian.org/669301
15:10:37 <ubottu> Freedesktop bug 48904 in python "switching from pygobject2 to pygi in 1.0 was an incompatible change" [Normal,Reopened: ] http://bugzilla.freedesktop.org/show_bug.cgi?id=48904
15:10:38 <ubottu> Launchpad bug 924079 in apt (Ubuntu Oneiric) "do-release-upgrade fails to upgrade from Oneiric to Precise: Couldn't configure pre-depend libtinfo5 for libncurses5, probably a dependency cycle" [High,Triaged] https://launchpad.net/bugs/924079
15:10:39 <ubottu> Launchpad bug 990145 in Gwibber "Remove dependency on mx.DateTime" [Undecided,New] https://launchpad.net/bugs/990145
15:10:40 <ubottu> Launchpad bug 992195 in Duplicity "Remove dependency on GnuPGInterface" [Undecided,In progress] https://launchpad.net/bugs/992195
15:11:20 <xnox> did hr/new employee procedure tasks, some are still pending
15:11:21 <xnox> - finished testing lvm2 merge, it got updated in debian,
15:11:21 <xnox> remerged, needs further retest
15:11:22 <xnox> - did one sync request (libgpg-error)
15:11:22 <xnox> - finishing up merging sudo, should be ready for sponsorship later
15:11:23 <xnox> - created uds spec, wrt failing hardware notifications
15:11:23 <xnox> - just did a merge proposal for boost1.49 transition tracker
15:11:24 <xnox> NEXT
15:12:03 <infinity> jodh: You're up, since you were disconnected for the randomizer. ;)
15:12:07 <slangasek> xnox: I'd like to upload lvm2 .88 now and worry about .95 later, especially since .88 is the most recent version that I know upstream recommends
15:12:21 <xnox> slangasek: ok.
15:12:26 <slangasek> xnox: but I'll test it on my own system first before doing so
15:13:02 <slangasek> either way, good to have one of our top packages knocked off of http://qa.ubuntuwire.org/oldmerges/ :)
15:13:31 * ogra_ wonders if there is actually a human being behind the jodh nick :)
15:13:59 * ogra_ just moves forward for now
15:14:03 <ogra_> done:
15:14:03 <ogra_> * more compiz upgrades right before/on release
15:14:03 <ogra_> * omap, omap4 and ac100 release image tests
15:14:03 <ogra_> * may 1st was labour day in germany ...
15:14:03 <ogra_> * packaged the new nvidia tegra armhf driver (sadly very buggy on the ac100 atm (suspiciously looks like the GPU hardlocks
15:14:05 <ogra_> after a while, nvidia devs are notified)) (the test package can be found at http://people.canonical.com/~ogra/nvidia-tegra/ atm)
15:14:08 <ogra_> todo:
15:14:10 <ogra_> * merges: i only have procps on the list, feel free to drop stuff on me, else i'll randomly grab bits off the list
15:14:13 <ogra_> * pondering if we shouldnt have a spec for MBR and partition table backup/restore
15:14:15 <ogra_> (i know that came up various times in the past but i couldnt find a spec for it)
15:14:17 <ogra_> UDS:
15:14:19 <ogra_> registered https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-embedded-rootfs
15:14:21 <ogra_> registered https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-hwpack-integration
15:14:25 <ogra_> (done)
15:15:58 <slangasek> ev:
15:15:59 <ev> - http://errors.ubuntu.com is live, though due to a programming error, things
15:16:00 <ev> went down over the weekend. Fortunately, the system is designed to cope with
15:16:01 <ev> this and clients resent crashes when it came back up on Monday. We're still
15:16:03 <ev> getting things hooked into the webops pager system.
15:16:04 <ev> - Fixed a few bugs that have arisen in the web side of the crash database
15:16:06 <ev> since release.
15:16:07 <ev> - Fixed a critical bug in apport as version 2.0.1-0ubuntu7. We were directing
15:16:09 <ev> people to launchpad.net post-release.
15:16:10 <ev> - We now have (the start of) a web API for the crash reporting statistics.
15:16:12 <ev> This was initially written to provide webops with a means on monitoring the
15:16:13 <ev> health of the retracers.
15:16:13 <ev> - Created the 12.04 USB disks for the store.
15:16:20 <ev> - As it turns out, 4 i386 and 4 amd64 retracer instances wasn't enough
15:16:20 <ev> post-release. Initial testing suggests we could get by with about 16 of
15:16:22 <ev> each. We'll also need a lot more disk space, as between the backlog of core
15:16:23 <ev> dumps and the per-retracer per-release caches we're quickly eating up the
15:16:25 <ev> existing disk. Worked with webops to rememedy the immediate problem. New
15:16:26 <ev> hardware is on order.
15:16:27 <ev> - Registered blueprints for UDS:
15:16:29 <ev> - https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-crash-database
15:16:30 <ev> - https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-metrics
15:16:32 <ev> - https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-updates-from-crash-reports
15:16:33 <ev> - https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-fix-ddebs
15:16:35 <ev> - https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-bucketing-improvements
15:16:36 <ev> - https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-phased-updates
15:16:39 <ev> - Worked with Matthew to try to write a better mean time between failures
15:16:41 <ev> algorithm. Mathematicians wanted.
15:16:42 <ev> - Built out support to collect statistics on the release, package, version,
15:16:44 <ev> year, month, and day level (and combinations therein).
15:16:45 <ev> - Started moving the web UI to consuming its own API to generate the charts,
15:16:47 <ev> in support of being able to switch chart views via fancy-pants AJAX.
15:16:48 <ev> - Discussed the http://errors.ubuntu.com design with Matthew, Kate, and Colin.
15:16:50 <ev> If you get a chance, I'd love your thoughts in reply to the email I sent
15:16:51 <ev> about it on April 13th.
15:16:52 <ev> (done)
15:16:53 <ev> I'll post some fun stats during AOB
15:16:57 <infinity> ev: When will errors.ubuntu.com be usable by !canonical people (or will it ever?)
15:17:27 <infinity> - Release sprint last week:
15:17:28 <infinity> - 12.04 got released, rather uneventfully
15:17:28 <infinity> - 12.10 got opened while no one was looking
15:17:28 <infinity> - This week:
15:17:28 <infinity> - Fixed d-shlibs to deal with the armhf linker
15:17:30 <infinity> - Fixed a few bugs in the fpc/armhf bootstrap
15:17:33 <infinity> - Worked on fixing some llvm/clang-on-arm issues
15:17:35 <infinity> - Rebootstrapped the tex* stack to sort out an
15:17:38 <infinity> accidental out-of-order sync that blew up the world
15:17:40 <infinity> - Registered and fiddled with spec for UDS.
15:17:43 <infinity> ...
15:17:46 <ev> infinity: well, I'm not shouting about it from the rooftops yet as the code could use some performance improvements
15:17:56 <ev> and I really, really, really don't want Hacker News hitting it
15:17:58 <ev> just yet
15:18:02 <slangasek> ev: speaking of retracers, I noticed the backtraces shown in the errors reports don't appear to be a 'bt full' - is there any plan to make this available?
15:18:09 <infinity> ev: Sure, but I mean it's currently behind SSO demanding that you're in the canonical group. :)
15:18:19 <slangasek> it's relevant often enough that I think we should have it by default
15:18:45 <ev> infinity: but the game plan is canonical group + bug control being able to access the actual reports
15:18:48 <ev> just like on launchpad
15:18:53 * infinity nods.
15:19:03 <Laney> use ubuntu-dev instead for now, or bugcontrol?
15:19:14 <ev> slangasek: can you provide me with an example?
15:19:22 <ev> it's definitely not intentional
15:20:43 <slangasek> ev: looking
15:20:50 * bdmurray starts then
15:20:53 <bdmurray> bug triage of ubiquity, iso-testing and update-manager bug reports
15:20:53 <bdmurray> investigation into reporters submitting dist-upgrade bugs multiple times
15:20:53 <bdmurray> closed apport-package reports regarding package install failures due to not a debian package
15:20:56 <bdmurray> wrote a bugbot routine for 'not a debian format archive bug reports'
15:21:03 <bdmurray> closed texinfo bug reports due to a syntax error in /etc/environment
15:21:04 <bdmurray> whoopsie apport-package duplicate bug consolidation into bug 988995
15:21:04 <bdmurray> tried to recreate bug 987956 regarding ubiquity
15:21:04 <bdmurray> wrote a bug pattern for bug 937546
15:21:06 <ubottu> Launchpad bug 988995 in whoopsie-daisy (Ubuntu) "package whoopsie 0.1.32 failed to install/upgrade: el subproceso instalado el script post-installation devolvió el código de salida de error 1" [Undecided,Confirmed] https://launchpad.net/bugs/988995
15:21:07 <ubottu> Launchpad bug 987956 in ubiquity (Ubuntu) "Installer Deletes Contents from Separate HOME partition without WARNING!" [High,Incomplete] https://launchpad.net/bugs/987956
15:21:08 <ubottu> Launchpad bug 937546 in nautilus-dropbox (Ubuntu) "package nautilus-dropbox (not installed) failed to install/upgrade: subprocess installed post-removal script returned error exit status 2" [Undecided,Invalid] https://launchpad.net/bugs/937546
15:21:15 <bdmurray> reported apport bug 989929 regarding checking if a package is an ubuntu package
15:21:16 <ubottu> Launchpad bug 989929 in apport (Ubuntu) "does not check ubuntuness of packages being removed or upgraded" [High,Confirmed] https://launchpad.net/bugs/989929
15:21:18 <bdmurray> conversion of harvest opportunities from direct launchpad database queries to using the api
15:21:21 <bdmurray> blogged regarding duplicate bug title consolidation
15:21:23 <bdmurray> created a list of people who improved bug reports during the precise dev cycle for skaet
15:21:32 <bdmurray> investigation into remastersys and emailed contact person regarding bug 984276
15:21:33 <ubottu> Launchpad bug 984276 in casper (Ubuntu Precise) "installing casper on a non live system causes update-initramfs to fail" [Medium,Confirmed] https://launchpad.net/bugs/984276
15:21:35 <bdmurray> created a launchpad development virtual machine
15:21:37 <bdmurray> created a couple of blueprints for Q
15:21:48 <bdmurray> done
15:21:58 <slangasek> ev: "no symbol table info available", so maybe this was a missing debug symbol problem. https://errors.ubuntu.com/oops/008aa98e-92ea-11e1-b399-2c768aafd08c
15:22:18 <doko> msg infinity did you give back gcc-4.7? these mysterious give backs become annoying ...
15:22:22 <doko> oops
15:22:45 <slangasek> bdmurray: texinfo> can you give us a bug about fixing texinfo to not muck around in /etc/environment where it doesn't belong?
15:22:46 <infinity> doko: I did a mass-give-back of the world yesterday, nothing today...
15:23:05 <infinity> doko: But that wouldn't have hit PPAs, if your PPA upload is what got given-back.
15:23:09 <doko> right
15:23:10 <bdmurray> slangasek: yes, will do
15:23:13 <slangasek> ta
15:24:21 <slangasek> jodh: your turn, if you're around :)
15:24:28 <jodh> - boot/upstart: - Raised MP on quantal upstart to fix compile error.
15:24:28 <jodh> - planning/blueprints:
15:24:28 <jodh> - discussions around multiseat.
15:24:28 <jodh> - service readiness:
15:24:31 <jodh> https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-upstart-service-readiness
15:24:35 <jodh> - ptrace limitations:
15:24:38 <jodh> https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-upstart-overcome-ptrace-limitations
15:24:41 <jodh> - libs/nih: bug 776532 (required for Upstart quantal work). Wrote tests and
15:24:44 <jodh> tested. Approved by Keybuk.
15:24:45 <ubottu> Launchpad bug 776532 in libnih "nih_dir_walk_scan passes incorrect value to file filter" [Undecided,Triaged] https://launchpad.net/bugs/776532
15:24:48 <jodh> - TODO:
15:24:51 <jodh> - ConsoleKit+logind investigations + multiseat blueprint with robert_ancell.
15:24:54 <jodh> - upstart job disabling blueprint.
15:24:57 <jodh> - upstart state passing blueprint.
15:25:00 <jodh> ⨌
15:25:58 * slangasek feels all integrated
15:26:16 * infinity feels dirty
15:26:29 <stgraber> jodh: can you make sure I'm subscribed to the multiseat blueprint? (I have an interest because of past experience with these and because of LTSP)
15:27:09 <jodh> stgraber: will do.
15:27:13 <slangasek> jodh: please also register a general "upstart roadmap" blueprint - we should have that session again just to make sure we're still aligned with what people need
15:27:14 <cjwatson> I didn't even know there *was* a quadruple integral operator.
15:27:25 <jodh> slangasek: ack.
15:28:16 <stgraber> jodh: thanks
15:28:22 <slangasek> anyone have questions over the above?
15:30:29 <slangasek> [TOPIC] UDS blueprints
15:31:04 <slangasek> so with UDS half a week away, any remaining blueprints you mean to have on the schedule need to get in the system ASAP
15:31:07 <slangasek> tomorrow at the latest
15:31:18 <cjwatson> I've got "btrfs sync-up with server team" and "automated bootloader testing" to do from our call yesterday
15:31:24 <cjwatson> will register those today
15:31:46 <slangasek> and by Friday at the latest, please have a look at what's on the schedule for UDS and get yourself subscribed to the sessions you need to be at
15:31:53 <slangasek> so that the autoscheduler can do its work over the weekend
15:32:04 <slangasek> instead of having it throw conflicts at us Monday :)
15:32:09 <slangasek> cjwatson: thanks
15:32:29 <infinity> (For the record, if people have sessions they think I should be at, feel free to subscribe me)
15:33:23 <ogra_> infinity, do we have an arm v8 spec btw ?
15:33:41 <infinity> Yep.
15:33:43 <slangasek> I expect that as usual none of you should have any trouble finding things to occupy your time at UDS; but if anyone should happen to feel they don't have enough that they're participating in, let me know
15:33:44 <ogra_> (assuming it would be yours if it exists)
15:33:51 <infinity> ogra_: https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-aarch64-porting
15:33:54 <ogra_> thx
15:34:05 * ogra_ subscribes
15:34:12 <infinity> Cleverly hidden under the arch name that no one knows.
15:34:20 <ogra_> hehe
15:34:50 <cjwatson> Anyone know if wookey's already planning to set up a quantal sbuild-ma buildd in time for UDS?
15:34:52 <slangasek> ev, bdmurray: incidentally, do you think one session about the crashdb is sufficient? (https://blueprints.launchpad.net/ubuntu/+spec/foundations-q-crash-database)
15:34:58 <cjwatson> I'd ask him directly but he's not on IRC right now
15:35:12 <slangasek> he's on #multiarch@OFTC :)
15:35:19 <slangasek> (I don't know)
15:35:24 <ev> slangasek: as in should we have foundations-q-crash-database-2?
15:35:28 <infinity> I was just about to mention he was on oftc. :P
15:35:36 * ogra_ sees him in #linaro
15:35:36 <ev> or are you asking if we should split it out into more specs?
15:35:38 <slangasek> ev: as in, are there subtopics that warrant an hour of their own :)
15:35:49 <cjwatson> ah, I'll ask him there
15:36:01 <ev> slangasek: I'll give that a think tonight
15:36:05 <ev> it does link to some subtopics
15:36:26 <ev> but yeah, maybe the mean time between failures is an hour talk itself
15:37:20 <slangasek> ev: if it's simpler we can just block out multiple hours for the umbrella blueprint
15:37:28 <ev> yeah, lets do that
15:37:42 <slangasek> though sometimes that just means the first topic takes two hours and you never get to the end of the list ;)
15:37:47 <ev> heh
15:37:56 <ev> I'll still think about how we can further break this down tonight
15:38:03 <ev> but lets tentatively plan for two hours
15:38:05 <slangasek> any other concerns about blueprints at the moment?
15:38:43 <ogra_> any suggestions about hving a blueprint for a partition table and MBR backup restore ?
15:38:47 <ogra_> *having
15:38:58 <ogra_> (mentioned in my report above)
15:39:16 <slangasek> ogra_: I wasn't sure I understood the use case for this
15:39:41 <ogra_> having backups of your partition table in case you falsely change it ?
15:39:49 <slangasek> who does that though? :)
15:39:57 <ogra_> being able to restore a messed up MBR a user fiddled to death
15:40:22 <ogra_> dunno, i kno win the past having grub do backups on grub-update came up
15:40:30 <cjwatson> update-grub doesn't touch the partition table or the MBR
15:40:35 <ogra_> (and we had such discussions for u-boot as well)
15:40:37 <cjwatson> You may be thinking of grub-install
15:40:47 <ogra_> oh, indeed, wrong term, sorry
15:41:02 <slangasek> right, and grub-install doesn't get run except at install time
15:41:10 <cjwatson> I dunno, it seems like a special case of having good system backups
15:41:11 <ogra_> k, so its moot
15:41:21 * xnox .oO(UEFI)
15:41:35 <slangasek> ogra_: I suspect we can't justify a session for it - but thanks for bringing it up
15:41:38 <ogra_> right, might make more sense in context of an advanced backup tool
15:41:40 <cjwatson> ... which we still don't do well
15:42:00 <slangasek> do we want system backups on the agenda?
15:42:11 <slangasek> or would that need to be desktop / server instead of foundations?
15:42:13 <ev> don't we have deja dup?
15:42:19 <ev> it sometimes manages to back up to u1
15:42:26 <cjwatson> ev: Have you ever tried to do system restores based on duplicity backups?
15:42:31 <cjwatson> I have.  It doesn't work.
15:42:37 <ev> cjwatson: yikes
15:42:37 <ogra_> heh ... sometimes
15:43:23 <cjwatson> (Among its faults, it stores file ownership by user/group name rather than number, so you can't reliably restore from a live CD without having to spend ages fiddling with permissions afterwards.  It's obviously meant as a single-user backup tool, not a system backup tool.)
15:44:06 <cjwatson> slangasek: I suspect nobody has time for it this cycle :-)
15:44:15 <slangasek> right-o
15:44:18 <slangasek> moving on then
15:44:22 <slangasek> [TOPIC] Bugs
15:44:24 <cjwatson> Maybe I should talk to mterry and see if duplicity can be bludgeoned into the right shape, since it's clearly desktop's direction
15:44:32 <ev> cjwatson: noted - my experience with it has been fairly light
15:44:37 * xnox bacula - is ok for system backups
15:44:56 <slangasek> bdmurray: are there any bugs showing up post-release that we should be worrying about SRUing?
15:44:57 <barry> duplicity is one of those py3 application ports, so there's even more to bludgeon mterry with :)
15:45:00 <slangasek> (ones that we aren't already)
15:45:13 * stgraber uses a central backuppc server with rsync for all the Linux systems, works great but not exactly something we can fit in the default install
15:45:51 <bdmurray> slangasek: well, I ran across bug 988995 which reminded me of another bug
15:45:52 <ubottu> Launchpad bug 988995 in whoopsie-daisy (Ubuntu) "package whoopsie 0.1.32 failed to install/upgrade: el subproceso instalado el script post-installation devolvió el código de salida de error 1" [Undecided,Confirmed] https://launchpad.net/bugs/988995
15:46:00 <bdmurray> bug 523896
15:46:01 <ubottu> Launchpad bug 523896 in shadow (Ubuntu) "useradd: cannot lock /etc/passwd; try again later." [Medium,Confirmed] https://launchpad.net/bugs/523896
15:47:12 <cjwatson> stgraber: Right, roughly me too, but it's all home-grown
15:47:18 <ev> yeah, I did ask about that in #ubuntu-devel
15:47:30 <ev> I'm not sure what else we can do if /etc/passwd is locked
15:47:47 <slangasek> 523896> the blocker there was that we haven't reached consensus with the upstream yet, so I was reluctant to diverge :/
15:48:37 <slangasek> ev, bdmurray: what's surprising is the large number of reports of it *being* locked, all of a sudden
15:49:48 <slangasek> Why should users not be seeing locked /etc/passwd until installing whoopsie?  Is it just that the stale lock has been there for years, but nothing has tried to add a new user on upgrade?
15:50:14 <infinity> Could be that.
15:50:20 <infinity> We don't add new users all that often.
15:50:31 <infinity> Unless you install something that does so, of course.
15:51:34 <cjwatson> Quite a few packages add users; odd that whoopsie would be a particularly common victim
15:51:40 <slangasek> so unless we think whoopsie is somehow misbehaving and *causing* the lock, the right thing to do is to fix the underlying lock
15:51:51 <slangasek> cjwatson: quite a few packages in the default desktop install that will add them on upgrade?
15:51:53 <cjwatson> I suppose it could be like man-db being victimised by trigger bugs
15:52:05 <cjwatson> Maybe not upgrade, no, though whoopsie's surely not the first
15:52:14 <cjwatson> Well, first in this case :)
15:52:38 <cjwatson> Anyway, I see no relevant misbehaviour in whoopsie.postinst
15:52:42 <slangasek> yep
15:53:02 <slangasek> and whoopsie certainly shouldn't be working around the lock if it's not buggily causing it in the first place
15:53:16 <slangasek> so the thing to do is to get the lock cleared (e.g., at boot)
15:53:44 <slangasek> cjwatson: any objections to clearing that lock file at boot?
15:54:04 <cjwatson> Seems fair
15:54:27 <infinity> Seems reasonable (ish) to me, if we've confirmed that shadow/passwd inconsistency is avoided in other ways.
15:54:30 <slangasek> ok
15:54:35 <slangasek> bdmurray: any others?
15:54:39 <ev> https://errors.ubuntu.com/ still lists that python-central bug, which has been floating near or at the top since release. Don't know if it's my parsing of the launchpad crash bugs data being broken or if we really still don't have a bug reported for it.
15:54:43 <bdmurray> Is that thought that there may be some deliquent package leaving the lock?
15:54:44 <ev> apols for jumping in there
15:55:01 <infinity> ev: doko uploaded a fix, we just need to get it through SRU land.
15:55:04 <ev> https://errors.ubuntu.com/oops/00104e3a-9071-11e1-b51d-2c768aafd08c
15:55:05 <ev> ah yay
15:55:14 <infinity> ev: And his fix *did* reference a bug, so maybe your matching isn't working?
15:55:26 <ev> well it will only match apport filed bugs
15:55:33 <infinity> I think it was.
15:55:40 <ev> okay, I'll have a look then
15:55:58 <ev> I'm sure I'm failing to parse a special character properly
15:56:04 <slangasek> I thought doko hand-filed the bug
15:56:17 <infinity> ev: https://bugs.launchpad.net/ubuntu/+source/python-central/+bug/955936
15:56:19 <ubottu> Launchpad bug 955936 in python-central (Ubuntu) "pycentral crashed with NameError in run(): global name 'arch' is not defined" [High,In progress]
15:56:28 <ev> cheers
15:56:28 <infinity> slangasek: Nope.
15:56:38 <slangasek> bdmurray: well, one of the scenarios that caused the lock was if there's a crash while using the guest session
15:56:57 <slangasek> bdmurray: so there's a bug there somewhere, but I wasn't able to spot it when I looked at the time
15:57:11 <infinity> slangasek: Yeah, one might wonder why the guest session is locking passwd?
15:57:14 <ev> infinity: added a note to investigate - thanks
15:57:21 <slangasek> infinity: yep
15:57:47 <bdmurray> ev: is it because the bug has a python traceback and the error doesn't?
15:57:59 <slangasek> infinity: where exactly is the fix for bug #955936?  it's not in the unapproved queue
15:58:00 <ubottu> Launchpad bug 955936 in python-central (Ubuntu) "pycentral crashed with NameError in run(): global name 'arch' is not defined" [High,In progress] https://launchpad.net/bugs/955936
15:58:07 <cjwatson> it's in -proposed
15:58:07 <infinity> slangasek: Accepted.
15:58:11 <infinity> ^
15:58:14 <slangasek> oh
15:58:20 <slangasek> sorry, was confused by the 'in progress'
15:58:30 <slangasek> which is from the wrong task
15:58:33 <cjwatson> I guess we'll auto-copy it to quantal once it's verified
15:58:38 <slangasek> ok then
15:58:41 <ev> bdmurray: no, the signatures are not matched somehow. I think it's because I'm incorrectly parsing the signature on our end to what people.canonical.com/~ubuntu-archive/apport... is expecting
15:59:20 <bdmurray> slangasek: that's all I had for bugs
15:59:33 <slangasek> ok, thanks
15:59:41 <slangasek> [TOPIC] AOB
15:59:46 <slangasek> one minute, talk fast ;)
15:59:53 <ev> Some crash database statistics for you:
15:59:58 <ev> - 5,066,166 crashes received since March 20th. That's right, 5 million.
15:59:58 <ev> - 263,280 crashes received on release day and 426,729 the next day.
16:00:00 <ev> - Thanks to having to constantly purge the cache, we're retracing one 12.04
16:00:01 <ev> crash every 47.44 seconds. Our best time was on release day with a full
16:00:03 <ev> cache. We retraced one crash every 25.15 seconds.
16:00:04 <ev> - 8,874 unique problems were seen today (with many instances of those
16:00:06 <ev> seen). Yesterday we saw 12,468 unique problems. Keep in mind that this is
16:00:07 <ev> greatly influenced by how quickly we can retrace crashes (and it currently
16:00:09 <ev> incorrectly places the crashes in the day bucket for today, regardless of
16:00:10 <ev> whether it actually happened days ago and we've only just retraced it). On
16:00:11 <ev> release day we only saw 1,604 unique problems.
16:00:12 <ev> - We've received reports from 364,359 people.
16:00:16 <slangasek> wow
16:00:29 <ogra_> geez, 5mio !
16:00:52 <infinity> Take-home message: (A) we actually have users, (B) software sucks?
16:01:04 <slangasek> what's the distribution of duplicates look like?  https://errors.ubuntu.com/ seems to only give a rolling frequency
16:01:04 <ev> the retracer hardware is causing serious pain. The numbers will shoot up much higher once we sort that.
16:01:21 <xnox> (C) we can now act in more real-time to real problems users are hitting *right now*
16:01:22 <ogra_> still if we say we have 20mio users thats like a quater of them having crashes :)
16:01:38 <slangasek> i.e., I'm pretty sure there've been more than 858 instances of the python-central issue, and that the number shown was higher yesterday :)
16:01:38 <ev> infinity: yes, and yes
16:01:41 <cjwatson> (D) maybe we need a more agile SRU process
16:01:47 <infinity> cjwatson: Indeed.
16:01:52 <ogra_> ++ on that
16:01:56 <infinity> cjwatson: Do we have a bikeshed session for that?
16:02:02 <slangasek> not currently
16:02:06 <ev> slangasek: yeah, the counts look a bit suspicious. I've been investigating but everything checks out so far.
16:02:07 <infinity> We probably should.
16:02:09 <slangasek> someone want to schedule one?
16:02:12 <ev> the code for month and year ranges will land soon
16:02:15 <slangasek> infinity: thanks for volunteering :)
16:02:19 <ev> in the web ui that is, it's already on the backend
16:02:21 <infinity> >:(
16:02:24 * infinity goes to register.
16:02:37 <ev> xnox: yes, this is entirely realtime data and we'll soon have realtime updates on the page itself
16:02:41 <ev> so you can watch a crash climb
16:03:09 <slangasek> ev: well, I'm *sure* the numbers were higher when I loaded the page yesterday - I remember a four-digit count for the top crashers
16:03:20 <ev> it's per day at the moment
16:03:31 <ev> you'll be able to select month and year soon
16:03:40 <ev> and break down by releases, packages, and versions
16:04:16 <slangasek> ok
16:04:53 <ev> and matthew is working on a mockup UI to be able to select date ranges, I believe
16:04:58 <slangasek> would be good to understand things like: what percentage of all crashes does a given crash represent, are users likely to run into this same crash multiple times or just as a one-off, etc
16:05:02 <ev> as that was one of skaet 's requirements
16:05:11 <ev> slangasek: reply to my email :)
16:05:19 <slangasek> heh :)
16:05:23 <slangasek> noted ;)
16:05:25 <ev> IRC is far too lossy for this sort of thing
16:05:32 <ev> cheers!
16:05:44 <slangasek> #endmeeting