15:06 #startmeeting 15:06 Meeting started Thu Oct 3 15:06:42 2013 UTC. The chair is slangasek. Information about MeetBot at http://wiki.ubuntu.com/meetingology. 15:06 15:06 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 15:08 [TOPIC] Lightning round 15:08 $ echo $(shuf -e barry doko stgraber jodh ev bdmurray slangasek cjwatson xnox stokachu) 15:08 stgraber xnox cjwatson doko barry jodh ev slangasek bdmurray stokachu 15:09 stgraber: you're first :) 15:10 oops, one sec 15:10 Blueprint-related work: 15:10 - Image based updates (BLUEPRINT: foundations-1305-image-based-updates) 15:10 - Updated the specification on the wiki 15:10 - Added phased-percentage support on the server side 15:10 - Extended test coverage (still need tests covering the file generators) 15:10 - Prioritized the work for the system-image clien in 13.10 15:10 - Implemented the boot time hooks for touch (somewhere in the landing queue) 15:10 - LXC 15:10 - Moved all the bugs from sourceforge to github (well, fixed most of them rather than move them) 15:10 - Code reviews 15:10 - Some discussions with Google on cgroup management 15:10 - Rewrote lxc-info to use the LXC API and show IP addresses by default. 15:10 - Updated all the API based tools to check the user has control access over the container instead of failing in mysterious ways. 15:10 15:10 Other work: 15:11 - Other 15:11 - Quite a lot of queue reviews 15:11 - Implemented a script to auto-accept unseeded packages into the archive 15:11 15:11 TODO (hopefully this week): 15:11 - Get to 100% code coverage in the system-image server side tests 15:11 - Write a slightly more user friendly tool to release an image from saucy-proposed to saucy 15:11 - Spend some more time debugging the filesystem unmount sequence, fixing anything that needs fixing 15:11 - LXC upstream work (some more bugs to move to github, website work, ...) 15:11 (DONE) 15:12 ah. 15:12 skip me please. 15:14 cjwatson: your turn then 15:14 Launchpad: 15:14 * Fixed resetting of non-virtual builders (bug 1232131). 15:14 * Coordinated rollout of buildd slave fixes (bug 993642, bug 1227086). Much cleanup of broken builders following upgrade. 15:14 bug 1232131 in Launchpad itself ""'BuilderVitals' object has no attribute 'failBuilder'" when trying to auto-reset builder" [Critical,Fix released] https://launchpad.net/bugs/1232131 15:14 * Fixed recipe handling regression (bug 1234621) and helping to coordinate another rollout to virtual builders now. 15:14 bug 993642 in launchpad-buildd "Breaks if the build doesn't produce the expected changes file" [Critical,Fix released] https://launchpad.net/bugs/993642 15:14 bug 1227086 in launchpad-buildd "Slaves on 32-bit architectures crash on >2GiB build logs" [Critical,Fix released] https://launchpad.net/bugs/1227086 15:14 * Uploaded apt SRUs (bug 1234691, bug 1234705), to support upgrading Launchpad production to precise. 15:14 bug 1234621 in launchpad-buildd "lp-buildd crashes on recipe failures: too many values to unpack" [Critical,Fix committed] https://launchpad.net/bugs/1234621 15:14 Archive: 15:14 bug 1234691 in apt (Ubuntu Precise) "apt-utils: Ignore APT::FTPArchive::Packages::SHA512 option" [High,Fix committed] https://launchpad.net/bugs/1234691 15:14 bug 1234705 in apt (Ubuntu Raring) "apt-ftparchive writes SHA256 checksums in place of SHA512 in Sources" [High,Triaged] https://launchpad.net/bugs/1234705 15:14 * Arranged for generation of http://people.canonical.com/~ubuntu-archive/proposed-migration/saucy_uninst.txt and started cleaning it up. i386 uninstallables have gone from ~80 to ~30 this week. 15:14 Click: 15:14 * Fixed crash when unregistering a preinstalled package (bug 1232066), and a regression caused by that fix (bug 1233280). 15:14 bug 1232066 in click (Ubuntu) "click unregister on preinstalled app causes exception" [High,Fix released] https://launchpad.net/bugs/1232066 15:14 bug 1233280 in click (Ubuntu) "Permission error while removing preinstalled click before installing any other click" [High,Fix released] https://launchpad.net/bugs/1233280 15:14 * Implemented fat package support. 15:14 * Extended "click info" interface to allow printing the manifest for a single package (bug 1232118). 15:14 bug 1232118 in URL Dispatcher "Don't read manifest files directly" [High,Confirmed] https://launchpad.net/bugs/1232118 15:14 * Forced unpacked files to be owner-writeable (bug 1232128). 15:14 bug 1232128 in click (Ubuntu) "click package cannot be installed on desktop" [Medium,Fix released] https://launchpad.net/bugs/1232128 15:14 Installer: 15:14 * Fixed partition recipe size calculation when partitions are reused (bug 1197766). 15:14 bug 1197766 in partman-auto (Ubuntu Precise) "Different partition layout after recovery with keep home partition" [High,In progress] https://launchpad.net/bugs/1197766 15:14 * More work on putting the pieces together for bug 1065281 in precise. 15:14 bug 1065281 in OEM Priority Project quantal "Installer crashed when trying to partition 4k/4k sector hard disks" [High,In progress] https://launchpad.net/bugs/1065281 15:14 * Improved support for various mount options (bug 978032). 15:14 bug 978032 in partman-ext3 (Ubuntu Precise) "partitioning step doesn't have "discard" as a selectable mount option" [Wishlist,In progress] https://launchpad.net/bugs/978032 15:14 Touch: 15:14 * Rearranged build-initrd.sh in initramfs-tools-ubuntu-touch to stop doing documented-unreliable things with fakeroot. 15:15 To do: 15:15 * Review Brian's chroot management branch. 15:15 * Hunt through GRUB code to try to chase down bug 1229458. 15:15 bug 1229458 in grub2 (Ubuntu) "grubnetx64.efi tftp client does not work over ipv6" [High,New] https://launchpad.net/bugs/1229458 15:15 * Continue chasing down archive inconsistencies. 15:15 .. 15:15 doko: your turn? 15:19 * worked on emulator: 15:19 - it boots into TTY now 15:19 - published a prebuild copy at http://people.canonical.com/~xnox/emulator.tar.xz 15:19 - build fully integrated into phablet branches, one simply does a build for cm_goldfish-eng device. 15:19 - NEXT: fix & make android LXC container boot and correctly initialise 15:19 * patch piloting today, mostly trying to get interesting things for saucy 15:19 * upstart reviews, etc. 15:19 .. 15:21 system-image: lp: #1196991, lp: #1233094, lp: #1215943, lp: #1221844, lp: #1233379, lp: #1231628, lp: #1229807, lp: #1204618, lp: #1215946, lp: #1229710, lp: #1234703. s-i 1.6, 1.7, 1.8 15:21 Launchpad bug 1196991 in Ubuntu system image "Support the new download dbus service" [Critical,Fix released] https://launchpad.net/bugs/1196991 15:21 Launchpad bug 1233094 in system-image (Ubuntu) "Upgrade fail" [Undecided,Invalid] https://launchpad.net/bugs/1233094 15:21 Launchpad bug 1215943 in Ubuntu system image "should have a consistent "last upgrade date"" [Critical,Fix released] https://launchpad.net/bugs/1215943 15:21 Launchpad bug 1221844 in Ubuntu system image "Support channel aliases tracking" [Critical,Fix released] https://launchpad.net/bugs/1221844 15:21 Launchpad bug 1233379 in Ubuntu system image "Reset reactor timeout back to 10m on each non-finishing signal" [Critical,Fix released] https://launchpad.net/bugs/1233379 15:21 done 15:22 * upstart: 15:22 - Fixed bug 1227212 and bug 1089159 (uploaded as 1.10-0ubuntu3). 15:22 - Investigated ssh.override issue on Touch. 15:22 - Investigated upstart-local-bridge/upstart-property-watcher issue For 15:22 bug 1227212 in upstart (Ubuntu) "Session logout takes too long" [High,Fix committed] https://launchpad.net/bugs/1227212 15:22 sergiusens. Identified a problem with an lxc pre-start but seems 15:22 bug 1089159 in upstart (Ubuntu) "ADT test-suite failure" [Medium,Confirmed] https://launchpad.net/bugs/1089159 15:22 there is more work to do there. 15:22 * TODO: 15:22 - Add extra debug to upstart-local-bridge. 15:22 - work on overcoming ptrace limitations issue. 15:22 lp: #1231628, lp: #1229807, lp: #1204618, lp: #1215946, lp: #1229710 15:22 š¤’ 15:22 Launchpad bug 1231628 in Ubuntu system image "Support for phased updates" [Critical,Fix released] https://launchpad.net/bugs/1231628 15:22 lp: #1234703 15:22 Launchpad bug 1229807 in Ubuntu system image "auto_download setting not set to '1' by default" [Critical,Fix released] https://launchpad.net/bugs/1229807 15:22 Launchpad bug 1204618 in Ubuntu system image "Plumb progress through DBus signals" [Critical,Fix released] https://launchpad.net/bugs/1204618 15:22 Launchpad bug 1215946 in Ubuntu system image "_NoUpdate sends buggy UpdateFailed" [High,Fix released] https://launchpad.net/bugs/1215946 15:22 Launchpad bug 1229710 in Ubuntu system image "Add autopkgtests" [Medium,Fix released] https://launchpad.net/bugs/1229710 15:22 Launchpad bug 1234703 in ubuntu-system-settings (Ubuntu) "Checking for new updates puts the service out of order" [High,In progress] https://launchpad.net/bugs/1234703 15:23 Short week. Was on holiday in ā€˜merica until Monday. 15:23 Lots of time spent getting caught up after three weeks away. 15:23 Mostly still helping Alex bootstrap the CI team. 15:23 Got up to speed on the Cassandra failures. We had a massive Ceph outage which knocked over Cassandra, which obviously knocked over Daisy. Reports resend every two hours, so I donā€™t believe we missed out on anything, but Iā€™m still waiting on access to the production database to confirm these sorts of things: 15:23 https://rt.admin.canonical.com//Ticket/Display.html?id=63860 15:23 We are *way* behind on the existing retracer architectures and we still donā€™t have an armhf retracer for Touch. There appears to have been some work done on this while I was away, but it looks like they dropped it again. I tried following up in #webops, but we are without a vanguard. Iā€™ll touch base with James or Tom later. 15:23 https://rt.admin.canonical.com//Ticket/Display.html?id=58019 15:23 In anticipation of webops requesting it, Iā€™m migrating the error tracker deployment over to juju-deployer. 15:23 TODO: 15:23 I need to find a replacement for gnetworkmonitor, as it waking up frequently on busy networks continues to piss people off: 15:23 https://bugs.launchpad.net/whoopsie/+bug/991481 15:23 I now have enough details from Acunu to finish implementing the improvements to the average errors per calendar day calculation. 15:23 (done) 15:23 Ubuntu bug 991481 in Whoopsie "Constant dns traffic for daisy.ubuntu.com" [High,Confirmed] 15:24 https://wiki.canonical.com/IncidentReports/2013-09-30-Ceph-Rebalance-Outage is the Ceph outage 15:25 ev: bah, it's only causing around 20% of the requests to my DNS servers ;) 15:25 hah 15:25 1 request every 15s per machine on the network and I apparently have a lot of those ;) 15:25 stgraber: And your own substation? 15:26 ev: anti-social =) 15:26 ev: retracers> thanks - I think we need to get this sorted out ASAP, it's alarming that we have no errors data at all about the phone at this stage. 15:26 slangasek: I couldn't agree more. Any help leaning on them would be greatly appreciated 15:27 * slangasek nods 15:27 ev: what events do you actually want from gnetworkmonitor? you can have whoopsie e.g. start on net-device-up, then each time a connection is established, you can retry uploading reports. 15:27 cjwatson: ;) I do occasionaly run LXC load testing with around a thousand container on the same machine, all running whoppsie, then those DNS queries are pretty much the only thing I see going through the network ;) 15:27 Or telling the other Ubuntu Engineering teams to stop taking up all of webops' time 15:27 ev: I would've yelled at last week's sync call about this had I realize it still hadn't moved :/ 15:27 ev: *cough* 15:28 xnox: from memory, I'm using gnetworkmonitor to know when we're really connected 15:28 in my testing CONNECTED_ALL or whatever the NetworkManager state was didn't convey that 15:28 ev: Hopefully I've finished getting launchpad-buildd upgraded on things for, oh, at least a few weks 15:28 *weeks 15:28 (dear GOD is that process painfully manual) 15:28 * helped with getting final beta out last week 15:28 * discussions about how to do freeze management for phone packages so that we're not putting up additional roadblocks to phone development for 13.10 15:28 * worked with Barry on testing system-image / ubuntu-download-manager so this could land in the phone images \o/ 15:28 * fighting with my network, which has decided to stop passing dhcp requests reliably following a power outage over the weekend 15:28 * picked at the edges of the proposed-migration queue (insighttoolkit4) 15:28 * followed through on SRUs for shim update in 12.04.4 15:28 * got /etc/adjtime out of our live images, since it's not used in Ubuntu and just causes confusion 15:28 (done) 15:29 bug triage of update-manager, ubuntu-release-upgrader, ubiquity bugs 15:29 recreated update-manager bug 1202754 15:29 reviewed and merged apt-check (update-notifier) merge proposal speeding it up 15:29 bug 1202754 in update-manager (Ubuntu Saucy) "update-manager crashed with SystemExit in exit(): 0" [High,Confirmed] https://launchpad.net/bugs/1202754 15:29 worked on click chroot code 15:29 ev: should we refactor start.ubuntu.com/network-connectivity check from ubiquity?! =) yeah, supporting states in upstart would help here, as upon network check we could run a quick check against start.ubuntu.com and emit the state of "connected" 15:29 worked on ubuntu-release-upgrader not needing gksu 15:29 uploaded procps bug fix for bug 1150413 15:29 for those jobs that care about htat. 15:29 bug 1150413 in procps (Ubuntu Raring) "Cannot allocate memory if process owned by user with large number of groups" [High,Confirmed] https://launchpad.net/bugs/1150413 15:29 tested kernels for bug 1218004 15:29 bug 1218004 in linux (Ubuntu Saucy) "Apple Wireless Trackpad causes kernel oops" [High,In progress] https://launchpad.net/bugs/1218004 15:29 done 15:29 bug 1211876 needs upload, bug 995719 needs sru approval (done) 15:29 bug 1211876 in keepalived (Ubuntu Raring) "keepalived reload buggy due to improper ipv4 address comparison" [Undecided,Confirmed] https://launchpad.net/bugs/1211876 15:30 bug 995719 in puppet (Ubuntu Precise) "process_name.rb removed in 2.7.11 but still provided by puppet-common" [High,Confirmed] https://launchpad.net/bugs/995719 15:30 xnox: yes - networkmanager has support for something like this, but the tin foil hat crowd nixed it the last time matthieu brought it up 15:30 if memory serves 15:30 xnox: we could reimplement everything in Qt and use its network checking apis, I hear those work a treat 15:30 (by which I mean, bug #1233435) 15:30 xnox: what I'd really like is an API that I could register an address with. "Let me know when we're connected and you get this response from this URL" 15:30 note: 1211876 is a high visibility bug going to mark it as high 15:30 bug 1233435 in ubuntu-download-manager "u-d-m downloads stall if the network configuration changes" [Critical,Triaged] https://launchpad.net/bugs/1233435 15:31 cjwatson: *belated fist shaking* 15:33 anyone want to sponsor stokachu's SRU for bug #1211876? 15:33 bug 1211876 in keepalived (Ubuntu Raring) "keepalived reload buggy due to improper ipv4 address comparison" [High,Confirmed] https://launchpad.net/bugs/1211876 15:33 stgraber, maybe you could have a look? 15:34 bdmurray: would you mind looking over bug 995719 for approval? 15:34 bug 995719 in puppet (Ubuntu Precise) "process_name.rb removed in 2.7.11 but still provided by puppet-common" [High,Confirmed] https://launchpad.net/bugs/995719 15:34 stokachu: I'll have a look today 15:34 bdmurray: thanks man 15:35 stokachu: bug #995719, you say there should be no regression. So it's not possible that something will be relying on this behavior? 15:35 bug 995719 in puppet (Ubuntu Precise) "process_name.rb removed in 2.7.11 but still provided by puppet-common" [High,Confirmed] https://launchpad.net/bugs/995719 15:36 slangasek: nothing should be calling it directly, think of it more like a background job keeping up the process names 15:36 slangasek: yep, can do 15:36 stokachu: so nothing can be relying on those process name changes? 15:36 slangasek: nothing in puppet does, no 15:37 well, but what about things outside of puppet? 15:37 these are the questions we ask for SRUs :) 15:37 its not an external api so nothing would be calling it directly 15:38 but something external to puppet could be relying on the process name updates 15:38 everyone ive interacted with either manually deletes this or has a puppet config that ensures that file doesn't exist 15:38 even IS does it 15:38 sto they all look good to me, will upload those 3 in a minute 15:39 stokachu: ^ 15:39 stgraber: thanks man 15:39 stokachu: that doesn't mean that /everyone/ does it, just the people who are smart enough to be in your friends list :) we have to consider the SRU implications for those not following best practices, too 15:39 slangasek: yea im trying to think of what kind of use case would do this 15:40 slangasek: to help me pinpoint where there might be a problem 15:40 anyway, we can follow up on that out of band 15:40 do we have any puppet experts here? 15:40 ok 15:40 I'd try the server team for puppet expertise 15:41 ok ill ask around and see 15:41 [TOPIC] AOB 15:41 any other topics? 15:41 stokachu: all uploaded 15:42 stgraber: thanks again 15:43 stgraber: want me to unassign sponsors now? 15:43 #endmeeting