ericsnow | wallyworld: I think I figured it out (notes logged to the lp issue) | 00:11 |
---|---|---|
wallyworld | ericsnow: sorry, still in meeting | 00:12 |
ericsnow | wallyworld: the worker closes the channel when its loop finishes; so if the runner restarts the worker then it's using a closed channel | 00:12 |
ericsnow | wallyworld: np :) | 00:12 |
ericsnow | wallyworld: and I'm sorry you're stuck in meetings :/ | 00:12 |
wallyworld | ericsnow: ah, good pickup. i now have 40 mins till next meeting :-) | 00:48 |
ericsnow | wallyworld: yeah, I'm not aware of the precedent for dealing with that situation and simple solution isn't coming to mind | 01:13 |
ericsnow | wallyworld: perhaps we should only close the channel when the worker's tomb is killed | 01:13 |
mup | Bug #1494542 opened: unit does not go to error state <juju-core:New> <https://launchpad.net/bugs/1494542> | 01:16 |
wallyworld | ericsnow: maybe, i need to look atthe code in detail, haven't had a chance yet | 01:19 |
ericsnow | wallyworld: np, I've got to EOD so if you don't pick it up I will tomorrow | 01:20 |
=== natefinch-afk is now known as natefinch | ||
natefinch | waigani: hey, merge my PR! https://github.com/waigani/GoOracle/pull/19 | 02:17 |
waigani | natefinch: yikes, I haven't looked at this in a while! thanks :) | 02:18 |
natefinch | waigani: :) Figured as a coworker, I had the right to poke you, especially since it's just docs ;) | 02:18 |
waigani | natefinch: done | 02:20 |
natefinch | waigani: awesome :) | 02:20 |
natefinch | I know how it is. I have a couple outstanding PRs on my projects I need to attend to. | 02:20 |
axw | wallyworld: just managed to bootstrap :) I manually added a public IP to machine-0, but it's not hard to do that in code | 03:27 |
wallyworld | axw: whoohoo | 03:27 |
wallyworld | awesome | 03:27 |
axw | wallyworld: I'll write up my feedback email now. | 03:27 |
wallyworld | tyvm | 03:28 |
wallyworld | be gentle with them :-) | 03:28 |
axw | wallyworld: :) | 03:28 |
natefinch | gah, our inconsistent use of Id vs ID is sooo annoying | 03:38 |
axw | wallyworld: sent a big wall of text. lunch time | 04:23 |
wallyworld | axw: np, ty, was just otp | 04:40 |
thumper | well this week has pretty much sucked | 05:29 |
thumper | glad it's over | 05:29 |
thumper | time for wine | 05:29 |
* thumper hopes for better next week | 05:29 | |
=== urulama__ is now known as urulama | ||
frankban | TheMue: (not sure if message arrived) when you have time, could you please take a look at http://reviews.vapour.ws/r/2628/ ? note that this work is proposed for merging into a feature branch currently. thanks! | 07:36 |
TheMue | frankban: yep, will take a look | 07:38 |
frankban | TheMue: ty | 07:38 |
=== akhavr1 is now known as akhavr | ||
TheMue | frankban: you've got a review | 08:05 |
frankban | TheMue: thanks! | 08:06 |
frobware | dimitern, TheMue: standup time | 09:03 |
dimitern | frobware, we're there | 09:04 |
TheMue | frankban: we're, retrospective and planning | 09:04 |
dimitern | frobware, ah, sorry - it's a different link today | 09:04 |
TheMue | oops, frobware | 09:05 |
TheMue | https://plus.google.com/hangouts/_/canonical.com/sapphire | 09:05 |
frobware | aha | 09:05 |
frankban | TheMue: could you also take a look at the followup branch in http://reviews.vapour.ws/r/2630/ (still proposed against the feature branch)? | 09:16 |
TheMue | frankban: currently in retrospective HO, but then, yes | 09:17 |
frankban | TheMue: ty! | 09:17 |
mup | Bug #1494661 opened: Old logs are not compressed or removed <juju-core:New> <https://launchpad.net/bugs/1494661> | 09:38 |
fwereade | dimitern, reviewed http://reviews.vapour.ws/r/2593/ at last, you might have thoughts re some of my comments | 11:33 |
dimitern | fwereade, sweet! thanks! | 11:35 |
perrito666 | morning all | 11:35 |
perrito666 | fwereade: could you re-look at http://reviews.vapour.ws/r/2592/ | 11:37 |
perrito666 | pretty please? | 11:37 |
fwereade | perrito666, looking | 11:37 |
* perrito666 really wishes he wasn't freezing | 11:46 | |
voidspace | dimitern: frobware: TheMue: dooferlad: morning! | 12:10 |
TheMue | voidspace: heya, you missed a pretty long meeting today. ;) | 12:13 |
voidspace | TheMue: full team meeting? | 12:18 |
TheMue | voidspace: retro/planning | 12:18 |
voidspace | TheMue: ah! shame | 12:18 |
voidspace | TheMue: I should check the board | 12:18 |
voidspace | TheMue: it's just past 8am here, I still need coffee! | 12:18 |
TheMue | voidspace: no, too early for your time zone. you're still in the US? | 12:19 |
voidspace | TheMue: I'm back in Europe on Monday | 12:19 |
voidspace | TheMue: yep, still at Wayne's house | 12:20 |
voidspace | TheMue: he's now married and the big celebration is tomorrow | 12:20 |
TheMue | voidspace: relaxing after the wedding party :D | 12:20 |
TheMue | voidspace: oh, tomorrow, ic | 12:20 |
TheMue | frankban: you've got a review | 12:22 |
frankban | TheMue: ty! | 12:22 |
dimitern | voidspace, hey there :) welcome back! | 12:23 |
frankban | TheMue: great suggestion Errorf, I have been looking for something like that but did not find it for some reason | 12:26 |
TheMue | frankban: I would have called it Newf, so it's more easy to find. *bg* | 12:27 |
frankban | TheMue: yeah | 12:27 |
mup | Bug #1494726 opened: TestGetUsesCache different mime-type centos <centos> <ci> <test-failure> <unit-tests> <juju-core:Triaged> <https://launchpad.net/bugs/1494726> | 12:44 |
mup | Bug #1494729 opened: Panic created by net/http.fileTransport.RoundTrip <ci> <panic> <test-failure> <unit-tests> <juju-core:Triaged> <https://launchpad.net/bugs/1494729> | 12:44 |
frobware | voidspace, hiya - let's catch up monday as I have a couple of meetings RSN | 12:47 |
frobware | dimitern, want to sync up? | 12:55 |
dimitern | frobware, I started to summarize my thoughts around the needed changes - how about I finish them and send you a mail with those? | 12:58 |
frobware | dimitern, sure | 12:58 |
voidspace | frobware: ok, no problem | 12:59 |
dimitern | frobware, cheers | 12:59 |
mattyw | fwereade, ping? | 13:01 |
fwereade | mattyw, pong | 13:01 |
mattyw | fwereade, do you have 10 minutes to receiving a whinge? | 13:02 |
fwereade | mattyw, certainly, just between reviews :) | 13:02 |
mup | Bug #1494734 opened: Panic in jujud/agent on ppc64el <juju-core:New> <https://launchpad.net/bugs/1494734> | 13:02 |
mup | Bug #1494734 changed: Panic in jujud/agent on ppc64el <juju-core:New> <https://launchpad.net/bugs/1494734> | 13:08 |
mup | Bug #1494734 opened: Panic in jujud/agent on ppc64el <juju-core:New> <https://launchpad.net/bugs/1494734> | 13:14 |
mup | Bug #1494743 opened: TestMachineAgentRunsCertificateUpdateWorkerForStateServer timeout <ci> <i386> <test-failure> <unit-tests> <juju-core:Triaged> <https://launchpad.net/bugs/1494743> | 13:21 |
=== tvansteenburgh1 is now known as tvansteenburgh | ||
frankban | TheMue: time for a quick third and last one along the same line? http://reviews.vapour.ws/r/2633/ | 13:49 |
Odd_Bloke | Hello all. I have a service comprised of three units (all on different machines). If one of those machines disappears, should I expect to see one or both of cluster-{broken,departed} to be triggered on the other units? | 13:50 |
mup | Bug #1494749 opened: TestWorkerRetriesOnPublishError fails on wily <ci> <intermittent-failure> <lxc> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494749> | 13:51 |
mup | Bug #1494754 opened: TestActionsReceived failed on precise <ci> <intermittent-failure> <precise> <unit-tests> <juju-core:Triaged> <https://launchpad.net/bugs/1494754> | 13:51 |
Odd_Bloke | (tvansteenburgh: ^^^) | 13:51 |
tvansteenburgh | Odd_Bloke: yeah, i'm eagerly awaiting the answer :) | 13:52 |
pmatulis | is this the only place where users can look in order to choose a tools version to upgrade to? https://streams.canonical.com/juju/tools/releases/ | 13:56 |
ericsnow | regarding #1494661, didn't we fix log rotation (including compressing/deleting) already? | 13:58 |
mup | Bug #1494661: Old logs are not compressed or removed <juju-core:New> <https://launchpad.net/bugs/1494661> | 13:58 |
bogdanteleaga | pmatulis: yes, those are the latest stable tools, there's also a devel folder with newer, but usually alpha or beta versions | 14:02 |
ericsnow | fwereade: you have a few minutes (in a little while) to talk about how to fix #1493123? | 14:13 |
mup | Bug #1493123: Upgrade in progress reported, but panic happening behind scenes <landscape> <landscape-release-29> <upgrade-juju> <juju-core:In Progress by ericsnowcurrently> | 14:13 |
mup | <juju-core 1.24:In Progress by ericsnowcurrently> <juju-core 1.25:In Progress by ericsnowcurrently> <https://launchpad.net/bugs/1493123> | 14:13 |
fwereade | ericsnow, sure | 14:14 |
ericsnow | fwereade: OTP, but I'll ping you soon :) | 14:14 |
ericsnow | fwereade: you free? | 14:22 |
fwereade | ericsnow, yeah | 14:23 |
mup | Bug #1494765 opened: TestWorkerRemovesDeadAddress fails on ppc64el <ci> <intermittent-failure> <ppc64el> <unit-tests> <juju-core:Triaged> <https://launchpad.net/bugs/1494765> | 14:24 |
mup | Bug #1494774 opened: MetricSenderSuite fails on ppc64el panic <ci> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1494774> | 14:24 |
rogpeppe | my juju is in a state where | 14:25 |
rogpeppe | "juju bootstrap -e local" hangs forever | 14:25 |
rogpeppe | after "Bootstrapping Juju machine agent" | 14:25 |
rogpeppe | anyone have an idea what might be going on? | 14:25 |
bogdanteleaga | tried --debug? | 14:26 |
rogpeppe | i can't even interrupt the bootstrap | 14:26 |
rogpeppe | it's hung at "Interrupt signalled: waiting for bootstrap to exit" | 14:26 |
rogpeppe | bogdanteleaga: so i *would* try --debug if i could get it to stop | 14:27 |
bogdanteleaga | C^Z and kill process? :) | 14:27 |
rogpeppe | yeah, i think i'll sigquit it | 14:27 |
bogdanteleaga | might have to clean up after it if you do that tho | 14:27 |
rogpeppe | bogdanteleaga: any idea what i might need to clean up? | 14:27 |
bogdanteleaga | any machines/networks it created I'm thinking | 14:28 |
bogdanteleaga | I usually just power off maas machines | 14:28 |
bogdanteleaga | not sure about local | 14:28 |
rogpeppe | bogdanteleaga: lxc-ls shows one entry | 14:28 |
rogpeppe | juju-trusty-lxc-template | 14:28 |
* rogpeppe runs lxc-destroy | 14:29 | |
rogpeppe | bogdanteleaga: i've got an lxcbr0 network interface | 14:30 |
rogpeppe | bogdanteleaga: do you think i should remove that too? | 14:30 |
bogdanteleaga | rogpeppe, hmm, not sure, I'm really not familiar with lxc | 14:31 |
bogdanteleaga | rogpeppe, the only lxc machine I have seems to not create that | 14:31 |
pmatulis | bogdanteleaga: alright thanks | 14:32 |
rogpeppe | ha, "destroy-environment -y local" still leaves me with an entry in environments/cache.yaml | 14:33 |
rogpeppe | with an empty key | 14:33 |
bogdanteleaga | try --force | 14:33 |
rogpeppe | bogdanteleaga: i did | 14:33 |
rogpeppe | bogdanteleaga: (sorry, i didn't mention that) | 14:33 |
* rogpeppe removes cache.yaml | 14:33 | |
bogdanteleaga | rogpeppe, I never actually saw a cache.yaml before | 14:33 |
rogpeppe | bogdanteleaga: you only get it with the jes feature flag | 14:34 |
bogdanteleaga | rogpeppe, that makes sense :) | 14:34 |
katco | ericsnow: natefinch: ready? | 14:36 |
natefinch | yep | 14:36 |
rogpeppe | bogdanteleaga: for future reference: looking in ~/.juju/local/log/machine-0.log is dead useful :) | 14:40 |
bogdanteleaga | rogpeppe, I'll cherish the day I get to use local and not worry about boringly long deployment times | 14:41 |
rogpeppe | bogdanteleaga: :) | 14:42 |
rogpeppe | bogdanteleaga: filed https://bugs.launchpad.net/juju-core/+bug/1494787 | 14:46 |
mup | Bug #1494787: bootstrap cannot be interrupted if machine agent fails to start <juju-core:New> <https://launchpad.net/bugs/1494787> | 14:46 |
mup | Bug #1494782 opened: should *-broken *-departed hooks run when a unit goes AWOL? <juju-core:New> <https://launchpad.net/bugs/1494782> | 14:57 |
mup | Bug #1494787 opened: bootstrap cannot be interrupted if machine agent fails to start <juju-core:New> <https://launchpad.net/bugs/1494787> | 14:57 |
mup | Bug #1494774 changed: MetricSenderSuite fails on ppc64el panic <ci> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1494774> | 15:15 |
mup | Bug #1494798 opened: Juju fails to report it cannot create buckets <ci> <ec2-provider> <storage> <juju-core:Triaged> <https://launchpad.net/bugs/1494798> | 15:15 |
mup | Bug #1494798 changed: Juju fails to report it cannot create buckets <ci> <ec2-provider> <storage> <juju-core:Triaged> <https://launchpad.net/bugs/1494798> | 15:18 |
mup | Bug #1494774 opened: MetricSenderSuite fails on ppc64el panic <ci> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1494774> | 15:18 |
mup | Bug #1494774 changed: MetricSenderSuite fails on ppc64el panic <ci> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1494774> | 15:21 |
mup | Bug #1494798 opened: Juju fails to report it cannot create buckets <ci> <ec2-provider> <storage> <juju-core:Triaged> <https://launchpad.net/bugs/1494798> | 15:21 |
mup | Bug #1455625 opened: TestStateWatcherTwoEnvironments fails <ci> <test-failure> <wily> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1455625> | 15:27 |
mup | Bug #1455625 changed: TestStateWatcherTwoEnvironments fails <ci> <test-failure> <wily> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1455625> | 15:30 |
mup | Bug #1455625 opened: TestStateWatcherTwoEnvironments fails <ci> <test-failure> <wily> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1455625> | 15:39 |
katco | natefinch: hey can you update bug 1486553 with the latest status? | 15:56 |
mup | Bug #1486553: i/o timeout errors can cause non-atomic service deploys <cisco> <landscape> <juju-core:Triaged> <juju-core 1.25:In Progress by natefinch> <https://launchpad.net/bugs/1486553> | 15:56 |
natefinch | review anyone for a critical fix? katco, ericsnow, etc? http://reviews.vapour.ws/r/2636/ | 16:09 |
natefinch | breaking for lunch | 16:09 |
=== natefinch is now known as natefinch-afk | ||
ericsnow | natefinch-afk: I'll take a look | 16:10 |
bdx | https://bugs.launchpad.net/charms/+source/openstack-dashboard/+bug/1494829 | 16:35 |
mup | Bug #1494829: Support for custom panels <designate> <designate-dashboard> <django> <horizon> <murano> <murano-dashboard> <openstack> <openstack-dashboard (Juju Charms Collection):New> <https://launchpad.net/bugs/1494829> | 16:35 |
mup | Bug #1494831 opened: Windows instances on GCE will have the same hostname <juju-core:Incomplete> <https://launchpad.net/bugs/1494831> | 16:42 |
mup | Bug #1494831 changed: Windows instances on GCE will have the same hostname <juju-core:Incomplete> <https://launchpad.net/bugs/1494831> | 16:45 |
mup | Bug #1494831 opened: Windows instances on GCE will have the same hostname <juju-core:Incomplete> <https://launchpad.net/bugs/1494831> | 16:48 |
mup | Bug #1494848 opened: 1.24+ cannot upgrade in canonistack <canonistack> <openstack-provider> <streams> <upgrade-juju> <juju-core:Triaged> <https://launchpad.net/bugs/1494848> | 17:12 |
mup | Bug #1494848 changed: 1.24+ cannot upgrade in canonistack <canonistack> <openstack-provider> <streams> <upgrade-juju> <juju-core:Triaged> <https://launchpad.net/bugs/1494848> | 17:24 |
mup | Bug #1494848 opened: 1.24+ cannot upgrade in canonistack <canonistack> <openstack-provider> <streams> <upgrade-juju> <juju-core:Triaged> <https://launchpad.net/bugs/1494848> | 17:27 |
=== natefinch-afk is now known as natefinch | ||
natefinch | ericsnow: lol @ comment about using a struct for the params... i have code doing exactly that on my machine, but getting this fix in first. | 17:39 |
ericsnow | natefinch: np :) | 17:39 |
ericsnow | natefinch: it's just less churn overall if you do it in this patch | 17:40 |
natefinch | ericsnow: yeah, but this patch needs to be in ASAP. I was going to do it all together, but it turned out this fix was needed faster than we expected. | 17:40 |
ericsnow | natefinch: k | 17:40 |
katco | natefinch: +1 | 17:41 |
natefinch | ericsnow: making a test for it right now | 17:41 |
ericsnow | natefinch: thanks | 17:41 |
natefinch | katco: gah, 1.25 is blocked. | 18:04 |
katco | natefinch: hmm | 18:05 |
katco | mgz: ping? | 18:07 |
mgz | katco: hey | 18:09 |
katco | mgz: we'd like to land a fix for bug 1486553 | 18:09 |
mup | Bug #1486553: i/o timeout errors can cause non-atomic service deploys <cisco> <landscape> <juju-core:In Progress by natefinch> <juju-core 1.25:In Progress by natefinch> <https://launchpad.net/bugs/1486553> | 18:09 |
katco | mgz: but master is blocked | 18:10 |
katco | mgz: thoughts on jfdi? | 18:10 |
mgz | yeah, please don't | 18:10 |
natefinch | katco: 1.25 | 18:10 |
mgz | look at the history of OS-deployer job | 18:10 |
natefinch | katco: (I mean, master is also blocked, but the bug is for 1.25) | 18:10 |
katco | natefinch: mgz: oops sorry, for 1.25 | 18:10 |
mgz | fine on 1.24 and feature branches | 18:10 |
mgz | screwed on 1.25 and master | 18:10 |
mgz | it's already too hard to pin down cause because too much stuff was piled on | 18:11 |
katco | mgz: this is a fairly hot bug | 18:11 |
natefinch | katco: if 1.25 is broken currently, getting my fix in won't help get it to customers | 18:11 |
mgz | sure, but we can't release in the current state | 18:11 |
mgz | so it's not like that fix on 1.25 will get anywhere | 18:11 |
natefinch | mgz: I guess we're past the point of being able to revert whatever it is? | 18:12 |
mgz | natefinch: master can be reverted back to some kind of sanity | 18:12 |
mup | Bug #1494864 opened: TestBlockChangeServiceUpdate fails on windows <blocker> <ci> <regression> <test-failure> <unit-tests> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1494864> | 18:12 |
mup | Bug #1494868 opened: TestAzureWindows fails on wily and centos <centos> <ci> <test-failure> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494868> | 18:12 |
katco | natefinch: how hard would it be to get into 1.24? | 18:12 |
mup | Bug #1494870 opened: TestMeterStatusEvents fails on wily and vivid <ci> <race-condition> <test-failure> <unit-tests> <vivid> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494870> | 18:12 |
mgz | but yeah, as no one is taking my suggestion on backing out broken changes we now have regressions without clear blame | 18:13 |
katco | mgz: ty for the input... we won't land in 1.25 right now. natefinch, just learned this fix is needed in 1.24.6, so let's land there | 18:15 |
natefinch | katco: even better. | 18:15 |
natefinch | mgz: http://cdn.meme.am/instances2/500x/1820980.jpg | 18:15 |
katco | very small review for a critical bug on 1.24.6: http://reviews.vapour.ws/r/2638/ | 18:23 |
bogdanteleaga | mgz: are wily and centos tests running with another go version? | 18:23 |
mgz | bogdanteleaga: wily uses go 1.5 | 18:24 |
mgz | sinzui: what version of go is on centos? | 18:25 |
bogdanteleaga | bug 1494868 above is really interesting | 18:25 |
mup | Bug #1494868: TestAzureWindows fails on wily and centos <centos> <ci> <test-failure> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494868> | 18:25 |
bogdanteleaga | it should fail everywhere | 18:25 |
bogdanteleaga | but it fails only on wily and centos | 18:25 |
bogdanteleaga | because the test doesn't get ran on other platforms | 18:25 |
mgz | bogdanteleaga: that is interesting. | 18:27 |
sinzui | mgz: bogdanteleaga looking. I thought the test log printed that for us | 18:27 |
mgz | sinzui: I see the juju version but not the go version | 18:28 |
sinzui | mgz: our python test runner doesn't do a go version like our bash version :( | 18:28 |
* sinzui visits the machine | 18:28 | |
sinzui | mgz: bogdanteleaga go version | 18:29 |
sinzui | go version go1.4.2 linux/amd64 | 18:29 |
mgz | bogdanteleaga: thae windows-public-clouds branch was retested and passed, which further persuades me it's not the cause of recent breakage | 18:29 |
mgz | sinzui: ta! | 18:30 |
bogdanteleaga | http://reviews.vapour.ws/r/2640/ | 18:31 |
katco | natefinch: http://reviews.vapour.ws/r/2639/diff/# is the same as the one that was targetted to 1.25? | 18:32 |
bogdanteleaga | mgz: well, it can't be since you got the same failures on master | 18:32 |
bogdanteleaga | mgz: and, afaik that didn't get in master yet | 18:32 |
mgz | bogdanteleaga: we got so many issues on master I was unpersuaded of anything | 18:32 |
bogdanteleaga | :) | 18:32 |
voidspace | fwereade: thanks for the review | 18:33 |
voidspace | fwereade: I tried to make the changes without any changes to the external api - e.g. the allwatcher | 18:33 |
voidspace | fwereade: where empty address was used to signal no address set yet | 18:33 |
voidspace | fwereade: I think that's specifically been changed on master already though | 18:33 |
voidspace | fwereade: so the same work on master looks better | 18:34 |
natefinch | katco: yep | 18:34 |
voidspace | fwereade: a bunch of good suggestions though (especially the transaction testing) | 18:34 |
voidspace | fwereade: I have left a couple of replies, working through the others | 18:34 |
katco | natefinch: go for it | 18:34 |
katco | natefinch: and i could use a review for: http://reviews.vapour.ws/r/2638/ | 18:34 |
natefinch | katco: ok looking | 18:35 |
katco | that goes for anyone ^^ 22 line change | 18:35 |
katco | critical bug for 1.24.6 | 18:35 |
mgz | bogdanteleaga: renderers.ToBase64 returns a string? | 18:36 |
bogdanteleaga | fwiw, you ought to get https://bugs.launchpad.net/juju-core/+bug/1493887 fixed on windows first | 18:36 |
mup | Bug #1493887: statusHistoryTestSuite teardown fails on windows <blocker> <ci> <regression> <test-failure> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1493887> | 18:36 |
bogdanteleaga | since it might break other tests | 18:36 |
natefinch | katco: lgtm | 18:36 |
bogdanteleaga | I had a mail about this kind of failure on the ml before | 18:36 |
katco | cherylj: hey what's the status on bug 1494356 ? | 18:36 |
mgz | bogdanteleaga: yeah, there's a bunch of things on master I'm afraid | 18:36 |
mup | Bug #1494356: OS-deployer job fails to complete <blocker> <ci> <regression> <juju-core:Triaged by cherylj> <juju-core 1.25:Triaged by cherylj> <https://launchpad.net/bugs/1494356> | 18:36 |
bogdanteleaga | mgz: it very well could return a string | 18:36 |
bogdanteleaga | mgz: I just wanted to use deepequals everywhere for consistency | 18:37 |
mgz | I mean, does it? it's a function with a call value | 18:37 |
bogdanteleaga | mgz: now the tests runs on all platforms and it's good | 18:37 |
bogdanteleaga | mgz: yeah, think of it like a transformer | 18:37 |
mgz | just want to avoid the ugliness of a mismatch on []uint8 | 18:37 |
bogdanteleaga | mgz: that needs everything to move from jc.DeepEquals to str(x), gc.Equals, str(y) | 18:38 |
bogdanteleaga | mgz: because some functions in those suites give back strings and some give back byte arrays | 18:38 |
bogdanteleaga | mgz: can't really remember which do which right now | 18:38 |
mgz | ;_; | 18:39 |
mgz | bogdanteleaga: lgtmed | 18:39 |
bogdanteleaga | mgz: wasn't the best idea in hindsight :) | 18:39 |
bogdanteleaga | mgz: 1.25 is blocked though, right? | 18:39 |
natefinch | bogdanteleaga: what you need to do is write a jc.StringEquals that'll do the string conversion for you :) | 18:40 |
mgz | katco: I think she's hoping for more data on lxc specifics still, I'm going to rerun against 1.25 with that shortly | 18:40 |
katco | mgz: ah ok | 18:40 |
mgz | bogdanteleaga: yep | 18:40 |
bogdanteleaga | natefinch: that sounds useful, yeah | 18:41 |
natefinch | bogdanteleaga: it's 50% useful and 50% horrible.... since it produces a very squishy test that can pass when it really shouldn't. | 18:42 |
cherylj | katco: yeah, what mgz said | 18:42 |
cherylj | :) | 18:42 |
katco | cherylj: k ty :) | 18:42 |
bogdanteleaga | natefinch: I'm actually wondering if contains would work if they're equals | 18:43 |
mgz | cherylj: note that the test just passed on the windows-pulbic-clouds branch which is 1.25 as-of-a-few-revisinos-back | 18:43 |
* bogdanteleaga goes to the playground | 18:43 | |
cherylj | mgz: I'm wondering if the test is just taking a long time, as we've seen some of the containers come up in some of the test runs. | 18:44 |
mgz | cherylj: that is certainly part of it. | 18:44 |
cherylj | mgz: I'm running some manual testing in canonistack and it took over an hour to download the image to create the template container | 18:45 |
cherylj | don't know if it could be related or not | 18:45 |
bogdanteleaga | natefinch: deepequals still doesn't test that they have the same type though | 18:45 |
mup | Bug #1494876 opened: TestFatalErrors fails on wily <ci> <test-failure> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494876> | 18:46 |
mgz | but compare the 1.24 runs. they're either done in 35-40 mins, or hit the *internal* deployer timeout | 18:46 |
mup | Bug #1494887 opened: uniterV1Suite.SetUpTest <ci> <test-failure> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494887> | 18:46 |
cherylj | mgz: hmm, good point | 18:46 |
mgz | the master runs are *still deploying units* an hour in when our external job timeout hits | 18:46 |
cherylj | ouch | 18:47 |
mgz | so like, at minimum it's a 50% perf regression due to either our box, the network, (but not as of today), or the code in master | 18:47 |
cherylj | mgz: Is there any way to access the machines once the test completes? I'm wondering if we need to examine the containers to see what's going on | 18:49 |
cherylj | or are they immediately released? | 18:50 |
mgz | cherylj: let me check the job, most of them have a flag we can set to not destroy-environment after | 18:50 |
mgz | cherylj, sinzui: I am setting --keep-env on OS-deployer job | 18:53 |
sinzui | mgz: I hope you remember to undo it before you sleep :) | 18:54 |
mgz | also setting failure-threshhold to 1 | 18:55 |
sinzui | mgz: I've dont that. Once the job starts, you can restore the script., then wai 40 minutes for the job to get to your point of investifation | 18:55 |
mgz | as this one ties up most of our maas we really can't be doing other things at the same time | 18:55 |
voidspace | fwereade: hmmm... and yes, setting the preferred addresses on set rather than on read is better | 18:59 |
voidspace | fwereade: only nuisance is we have two setters for addresses. Ah well, still better | 18:59 |
cherylj | mgz: hopefully it won't take too long to see what's going on with the containers. | 19:04 |
mup | Bug #1494894 opened: TestWatchInterfaces fails on wily <ci> <test-failure> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494894> | 19:19 |
mgz | katco: can I persuade one of you to revert the last change on master before the weekend? | 19:23 |
mgz | the dep change broke like four things, I don't see a good reason to leave it in place | 19:23 |
katco | mgz: sec meeting... was it cmars patch? | 19:27 |
cmars | mgz, commit hash please? | 19:28 |
natefinch | katco: I can pick that up ^^ if you want. Pretty trivial... from Rogpeppe & frankban | 19:28 |
katco | natefinch: yep +1... go for it | 19:28 |
mgz | katco, cmars: http://reviews.vapour.ws/r/2606/ | 19:29 |
cmars | mgz, what tests fail? can I see a log? | 19:29 |
mgz | bug 1494441 bug 1494864 two more unrelated windows test failures and who knows what else after the build succeeds | 19:30 |
mup | Bug #1494441: ppc64el: cannot find package "encoding" <blocker> <ci> <ppc64el> <regression> <unit-tests> <juju-core:Triaged> <gccgo-go (Ubuntu):Invalid> <gccgo-go (Ubuntu Trusty):In Progress> <https://launchpad.net/bugs/1494441> | 19:30 |
mup | Bug #1494864: TestBlockChangeServiceUpdate fails on windows <blocker> <ci> <regression> <test-failure> <unit-tests> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1494864> | 19:30 |
cmars | mgz, ty | 19:30 |
natefinch | cmars: can you rubber stamp the revert? http://reviews.vapour.ws/r/2641/ | 19:31 |
natefinch | it's just a PR created with the revert button in github | 19:31 |
cmars | natefinch, i'm going to look at the test failures first | 19:31 |
mgz | cmars: so, the ppc64 build failure is 'not your fault' but prevents us doing proper testing | 19:31 |
natefinch | cmars: I've looked at the 1494441: ppc64el: cannot find package "encoding" one | 19:32 |
mgz | we had a fix to ggcgo that for whatever reasons didn't get back into trusty-updates | 19:32 |
natefinch | cmars: what he said ^ also, the windows one looks like a pretty standard "Someone doing file moves/deletes before they close the file" in windows | 19:34 |
perrito666 | mgz: still here? | 19:34 |
cmars | natefinch, mgz done | 19:35 |
mgz | perrito666: I am | 19:35 |
perrito666 | mgz: how much do you know about mong ca files and stuff? | 19:35 |
mgz | cmars: I will file other bugs from the results we got, but can address and reland easily enough next week | 19:36 |
mgz | perrito666: not much of the detail | 19:37 |
perrito666 | mmmpf, bad luck | 19:37 |
* perrito666 tries to setup mongo3 and it seems to want to authenticate with ssl certs | 19:37 | |
perrito666 | aghh finally | 19:38 |
perrito666 | --sslWeakCertificateValidation | 19:38 |
perrito666 | and by weak they mean nil | 19:39 |
mgz | perrito666: well, that sounds better than what I think we had before which is just a password-style key file | 19:40 |
mgz | of some number of random bytes | 19:40 |
perrito666 | mgz: I am trying to get juju to run on a minimal mongo 3 | 19:40 |
perrito666 | we can implement certs later on | 19:40 |
perrito666 | but with that option it should be the same as we have now | 19:40 |
natefinch | I really hate that JFDI requires __ JFDI__ because the underscores get hidden by markdown | 19:47 |
natefinch | I screw it up every time | 19:47 |
natefinch | WTF | 19:48 |
natefinch | I guess having __JFDI__ in a previous comment is not good enough. sigh | 19:49 |
mup | Bug #1494912 opened: TestHookContextEnv fails on windows <blocker> <ci> <test-failure> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1494912> | 19:52 |
mgz | natefinch: this isn't a jfdi, it's a fixes-1494441 | 19:52 |
natefinch | mgz: it doesn't fix anything, it just backs out the change... I mean, I guess it sorta does, but I figured it might be better to leave the bug around for when the code actually gets fixed. | 19:55 |
perrito666 | mgz: since we are in it, can we have fixes be a bit more permissive? | 19:55 |
perrito666 | mgz: fixes does not remove the block, it just lets your merge pass | 19:56 |
perrito666 | I meant natefinch | 19:56 |
perrito666 | mgz: could we have it also take fix- fix_ and fixes_ ? | 19:56 |
mgz | we can still use the bug | 19:57 |
natefinch | mgz: I guess I find the wording of "fixes" to be inaccurate in the case of just backing out a change... but maybe I'm being too pedantic. | 19:58 |
mgz | that bug is really about the ftb with gccgo, which backing out the change does address. I agree that that thinks can get complicated when using all the features of version control. | 19:59 |
mgz | well, this is fun. | 20:07 |
fwereade | voidspace, yeah, the two setters are a bit of a hassle | 20:13 |
katco | natefinch: did your patch merge? | 20:19 |
mgz | katco: both seem to have. | 20:21 |
mup | Bug #1494913 opened: TestNoSpoolDirectory fails on windows <blocker> <ci> <test-failure> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1494913> | 20:22 |
mup | Bug #1494917 opened: TestEnvSetsPath fails on windows <blocker> <ci> <test-failure> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1494917> | 20:22 |
katco | mgz: ty | 20:22 |
natefinch | katco: yep | 20:26 |
natefinch | katco: what do you think about making the rest of my changes for the add service stuff to 1.25 or master? Or should I be trying to get the rest of it into 1.24 as well? | 20:32 |
natefinch | gotta run for a while, will check in later | 20:40 |
=== natefinch is now known as natefinch-afk | ||
katco | natefinch: i'd say no | 20:40 |
natefinch-afk | yeah, the other half is trickier, I'd prefer 1.25 or later | 20:40 |
mup | Bug #1494936 opened: imageSuite.TestDownloadEnvironmentPath <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494936> | 21:16 |
mup | Bug #1494938 opened: Panic DeploySuite.TestConfig on wily <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494938> | 21:16 |
mup | Bug #1494939 opened: Panic backupsSuite.TestAuthRequiresClientNotMachine on wily <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494939> | 21:16 |
mup | Bug #1494936 changed: imageSuite.TestDownloadEnvironmentPath <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494936> | 21:19 |
mup | Bug #1494938 changed: Panic DeploySuite.TestConfig on wily <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494938> | 21:19 |
mup | Bug #1494939 changed: Panic backupsSuite.TestAuthRequiresClientNotMachine on wily <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494939> | 21:19 |
mup | Bug #1494936 opened: imageSuite.TestDownloadEnvironmentPath <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494936> | 21:22 |
mup | Bug #1494938 opened: Panic DeploySuite.TestConfig on wily <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494938> | 21:22 |
mup | Bug #1494939 opened: Panic backupsSuite.TestAuthRequiresClientNotMachine on wily <ci> <intermittent-failure> <panic> <unit-tests> <wily> <juju-core:Triaged> <https://launchpad.net/bugs/1494939> | 21:22 |
mup | Bug # opened: 1494947, 1494948, 1494949, 1494951 | 21:31 |
mup | Bug # changed: 1494947, 1494948, 1494949, 1494951 | 21:34 |
mup | Bug # opened: 1494947, 1494948, 1494949, 1494951 | 21:37 |
mup | Bug # changed: 1494947, 1494948, 1494949, 1494951 | 21:43 |
mup | Bug # opened: 1494947, 1494948, 1494949, 1494951 | 21:46 |
benji | f/quit | 21:54 |
ericsnow | anyone still around that could give me a quick review? http://reviews.vapour.ws/r/2644/ | 22:06 |
katco | ericsnow: tal | 22:16 |
katco | ericsnow: so for now the channel never gets closed, but it doesn't matter b/c its lifecycle is the same as the process? | 22:18 |
katco | *agent process | 22:18 |
ericsnow | katco: pretty much | 22:29 |
katco | ericsnow: what is the harm in having a defer close? | 22:29 |
ericsnow | katco: also, william indicated that not closing the channel would be fine | 22:29 |
katco | ericsnow: ah wait, that would close it as the function exited | 22:29 |
katco | ericsnow: not as the workers died | 22:30 |
ericsnow | katco: yep :) | 22:30 |
katco | ericsnow: +2, ship it. good work | 22:30 |
ericsnow | katco: thanks | 22:30 |
ericsnow | blech, currently 7 blockers on master :( | 22:32 |
perrito666 | wtf happened | 22:33 |
perrito666 | hi btw | 22:33 |
perrito666 | :) | 22:33 |
katco | perrito666: o/ | 22:35 |
katco | ericsnow: wow | 22:35 |
katco | ericsnow: i thought there was only 1 and nate had removed it | 22:35 |
mgz | katco: you got us back to our previous blockers :) | 22:36 |
katco | mgz: face palm | 22:36 |
mgz | which was why I was keen on the revert | 22:36 |
katco | mgz: understood... ty for pushing us in the rigth direction | 22:36 |
mgz | trying to unpick which were due to the dep update and which were due to maltese-falcon was causing me headaches | 22:37 |
katco | fun fact: the rigth direction is an early Gaelic belief that an undiscovered direction existed. later, it was recoined as the "right" direction | 22:37 |
katco | or, alternatively, i just can't type. | 22:37 |
mgz | now you've removed the dep update from the confusion | 22:38 |
katco | good to hear | 22:38 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!