cheryljhey wallyworld, looks like there's a windows unit test failure on master from the new xdg home dir work:  http://reports.vapour.ws/releases/3580/job/run-unit-tests-win2012-amd64/attempt/1891#highlight02:23
wallyworldcherylj: ok, will look. we really need to gate on windows unit tests. not sure why we don't yet02:34
cheryljyeah, would be nice02:34
wallyworldwe have been talking about it for months and months02:34
cheryljjust hasn't made it to the top of their list of priorities, I guess02:35
wallyworldyeah. seems like a small change for a big win though02:36
wallyworldaxw: a small tweak to fix windows unit tests http://reviews.vapour.ws/r/3763/02:56
axwwallyworld: are we trying to use XDG things on Windows? :/02:57
wallyworldaxw: no, there's a separate func to get the right location, but the tests have always been linux specific02:58
wallyworldaxw: on windows, it's %APPDATA%/juju02:59
axwwallyworld: done02:59
axwwallyworld: ok, thanks02:59
cheryljoh god, so many test failures.  My eyes!  MY EYES!!04:10
cheryljwallyworld, perrito666, so did the new juju home stuff break out JUJU_HOME into JUJU_DATA and something else?04:11
cheryljbecause it looks like CI wasn't04:11
wallyworldcherylj: JUJU_DATA replaces JUJU_HOME, but it will default to something sensible. if CI uses JUJU_HOME, then i'll need to back out that change04:29
wallyworldor i might set JUJU_DATA to JUJU_HOME till CI is updated04:31
cheryljwallyworld: that sounds like a good compromise.  I'll bring it up to the QA guys in the morning04:31
wallyworldok. i'll do that now04:31
cheryljthanks, wallyworld!04:31
wallyworldsure, sorry about the breakage04:32
wallyworldaxw: another small one http://reviews.vapour.ws/r/3764/04:44
axwwallyworld: done04:45
axwwallyworld: whenever you're ready: http://reviews.vapour.ws/r/3765/04:47
wallyworldaxw: how much smaller is asn encoding vs json?04:57
wallyworlddid it make much of a difference?04:57
axwwallyworld: it was reasonably significant. I can't recall the overhead exactly. I can measure again if you like05:04
wallyworldno need, just curious05:04
wallyworldaxw: a couple of small things05:21
wallyworldaxw: when you are free, another trivial http://reviews.vapour.ws/r/3767/05:33
davecheneywhat's changed now ? https://bugs.launchpad.net/juju-core/+bug/154298506:09
mupBug #1542985: Juju won't switch environments <juju-core:New> <https://launchpad.net/bugs/1542985>06:09
davecheneyaxw: wallyworld what has juju switch been renamed to ?06:13
wallyworlddavecheney: that one stays the same06:13
wallyworldlet me double check06:13
davecheneysee linked bug06:13
wallyworldyup, stays as switch06:14
davecheneywhy can't I change to ap-southeast-2 anymore06:14
wallyworldhmmm, haven't see that bug till now06:14
davecheneyit's there in juju/environemnts.yam06:15
davecheneyit's there in juju/environemnts.yam06:15
davecheneyit's there in juju/environemnts.yaml06:15
wallyworldit's supposed to work06:15
wallyworldsomething broke :-(06:15
davecheneylucky(~/src/github.com/juju/juju) % juju bootstrap -v --debug06:16
davecheney2016-02-08 06:15:41 INFO juju.cmd supercommand.go:59 running juju [2.0-alpha2 gc devel +fa5e547 Sun Feb 7 03:18:28 2016 +0000]06:16
davecheney2016-02-08 06:15:41 ERROR cmd supercommand.go:448 the name of the model must be specified06:16
davecheneylucky(~/src/github.com/juju/juju) % juju bootstrap -v --debug ap-southeast-206:16
davecheneywe really hate our users06:16
davecheneyerror: unrecognized args: ["ap-southeast-2"]06:16
davecheneylucky(~/src/github.com/juju/juju) % juju bootstrap -v --debug -m dev06:16
davecheney2016-02-08 06:16:36 INFO juju.cmd supercommand.go:59 running juju [2.0-alpha2 gc devel +fa5e547 Sun Feb 7 03:18:28 2016 +0000]06:16
davecheney2016-02-08 06:16:36 ERROR cmd supercommand.go:448 open /home/dfc/.local/share/juju/environments.yaml: no such file or directory06:16
wallyworldjuju home has changed06:17
wallyworldi'm in the process of updating release notes06:17
wallyworldit only landed on the weekend i think06:17
axwdavecheney: FYI, in the near future we will completely ignore environments.yaml06:17
wallyworldjust copy ~/.juju to ~/.local/share/juju06:17
axwmore info later06:18
davecheneylucky(~/.local/share) % ls -al . | grep juju06:18
davecheneylrwxrwxrwx  1 dfc dfc    15 Feb  8 17:18 juju -> /home/dfc/.juju06:18
davecheneyproblem solved06:18
wallyworld~/.juju will be ignored in a day or so06:18
wallyworldas soon as CI is updated06:19
wallyworldthen it will be ~/.local/share/juju06:19
davecheneylooks like it's being ignore now06:19
wallyworldi landed a change maybe 2 hours ago to temporarily use it06:19
wallyworlduntil CI is fixed06:19
wallyworldso just copy stuff to ~/.local/share/juju to be future proof06:20
davecheneycommit fd9cd4bc60d55571ce2db8562462703ed3baa22806:20
davecheneyAuthor: Ian Booth <ian.booth@canonical.com>06:20
davecheneyDate:   Mon Feb 8 14:41:59 2016 +100006:20
davecheneyFall back to JUJU_HOME if JUJU_DATA not set06:20
davecheneyyup, i have that rev06:20
wallyworldyou need $JUJU_HOME set06:20
wallyworldthat's what it falls back to06:20
davecheney2016-02-08 06:20:42 DEBUG juju.provider.common bootstrap.go:318 connection attempt for failed: Warning: Permanently added '' (ECDSA) to the list of known hosts.06:20
davecheney/var/lib/juju/nonce.txt does not exist06:21
davecheney2016-02-08 06:20:48 DEBUG juju.provider.common bootstrap.go:318 connection attempt for failed: /var/lib/juju/nonce.txt does not exist06:21
davecheney^ these won't exist06:21
davecheneymy user has no permission to write to those locations06:21
wallyworldright, but that's not what we're taling about06:21
wallyworldit's the XDG data home06:21
wallyworldas opposed to XDG config home06:22
wallyworldor XDG runtime home etc06:22
mupBug #1542985 opened: Juju won't switch environments <juju-core:New> <https://launchpad.net/bugs/1542985>06:22
mupBug #1542988 opened: bootstrap tries to write to /var/lib/juju on local machine <juju-core:New> <https://launchpad.net/bugs/1542988>06:22
wallyworldoh yuk06:22
wallyworldwhat provider06:22
wallyworldthe nonce.txt is written by cloud init i believe06:24
wallyworldthe bootstrap client has never used that file06:24
axwyes, written by cloud-init. bootstrap looks for it to check that cloud-init has finished06:24
wallyworldhow the fark it is getting written locally i'm not sure06:24
davecheneymaybe that error is leaking through from the remote side06:24
axwwallyworld: PTAL06:27
wallyworldaxw: lgtm. we need to double check that a newly added user on a controller only has access to the models they create or are shared with them. i can't recall whether in the initial poc done with the initial jes work all users had admin access on the controller or not06:30
mupBug #1542990 opened: tools version checker restarts every 3 seconds <juju-core:New> <https://launchpad.net/bugs/1542990>06:52
mupBug #1542992 opened: ec2: destroy-controller causes rate limiting failure <juju-core:New> <https://launchpad.net/bugs/1542992>06:52
axwwallyworld: yep. I did test, but not very rigorously, and it appeared to be checking access07:16
wallyworldaxw: ok, we'll need to ensure going forward it works as we hope07:17
wallyworldaxw: if you get a chance at some point, could you take a another pass at http://reviews.vapour.ws/r/3737/ now that it's been updated?07:37
axwwallyworld: yup, was about to, just fixing my branch07:37
wallyworldawesome ty07:38
axwwallyworld: what would you call the combination of controllers.yaml, accounts.yaml, models.yaml (sic), etc.? currently called "Cache" in that branch, but it doesn't sit well with me08:02
axwwallyworld: only the model info is really a cache08:02
wallyworldaxw: agreed08:02
wallyworldLocalControllerState  ?08:04
wallyworldit's local to that machine, it's related to controllers (and accounts for those), and it's state cached in memory or on disk08:05
wallyworldnaming is hard :-)08:05
axwwallyworld: a bit better, but I'm not completely sold on that either. I'll think about it for a bit.08:06
wallyworldyeah, i wasn't that happy with that suggestion08:06
axwwallyworld: I'm thinking ControllerStore, NewFileControllerStore, NewMemControllerStore, DefaultControllerStore08:09
wallyworldhmmm, that works08:09
axwwallyworld: along the lines of existing configstore08:09
wallyworldit's similar to what's there08:09
anastasiamac_wallyworld: axw: my problem with "store" is that its prime meaning is not "storehouse, warehouse" but actually is closer related to "retail"08:28
anastasiamac_apart from that - i'm phased at all.. whatever08:28
wallyworlddisagree :-)08:28
wallyworldin computer terms, store ahas a well defined meaning08:28
anastasiamac_m not*08:30
dimiternwallyworld, hey there08:31
dimiternwallyworld, re that juju home 2.0 change08:31
wallyworldshould i brace?08:31
dimiternwallyworld, how is going to work with non-public clouds - i.e. maas ?08:31
dimiternwallyworld, juju bootstrap mymaas maas/?08:31
dimiternassuming the lack of environments.yaml08:32
axwdimitern: there's a new clouds.yaml for private clouds08:32
wallyworlddimitern: hop onto #juju08:32
frobwaredimitern, morning!08:52
frobwaredimitern, are merging master -> maas-spaces?  I started but need to decide whether I should continue.08:53
dimiternfrobware, morning08:54
dimiternfrobware, yeah, I think it's good to do it often until we can merge08:54
frobwaredimitern, I didn't finish resolving all the conflicts08:55
dimiternfrobware, the bigger merge landed on friday08:55
dimiternfrobware, so this morning it was much easier - only 1 test to fix08:55
frobwaredimitern, ah - that's the one I'm talking about. I started on Friday, but didn't complete.08:55
dimiternfrobware, I've run make check after the merge on friday and did a quick live test to be sure08:56
frobwaredimitern, and that's now landed in maas-spaces?08:57
dimiternfrobware, this morning's merge is landing now08:57
dimiternfrobware, and the friday's one is already in08:57
dimiternfrobware, I have a few tasks planned to unblock maas-spaces merging into master08:58
dimiternfrobware, so we can then concentrate on the remaining tasks08:58
voidspacefrobware: we had a CI run on Friday, on maas-spaces, that never completed09:32
voidspacefrobware: http://reports.vapour.ws/releases/357809:32
voidspacefrobware: but there was a xenial unit test failure which isn't good - featuretests09:32
voidspacefrobware: that maybe looks space discovery related, which is unexpected09:34
dimiternvoidspace, 2 of those attempts where due to issues with the cloud archive getting the wrong package09:34
voidspacefrobware: however they pass locally09:34
voidspacedimitern: ah right, infrastructure problem then?09:35
voidspacedimitern: the featuretests pass locally for me (wily), but I get a cmd/status failure09:35
dimiternvoidspace, and the last one is flaky due to the controller space not getting handled09:35
voidspaceyeah, we're still expecting that to fail09:35
voidspaceI believe09:35
voidspacecmd/juju/status fails consistently for me09:41
voidspacechecking master09:41
dimiternvoidspace, how does it fail?09:41
voidspacedimitern: six failures, unexpected number of machines in some of them09:42
voidspacedimitern: same failure for me on master09:43
voidspacedimitern: huuuuge amounts of logging so hard to summarise the failures, I'll pick one of the tests and investigate09:43
voidspacedimitern: it could be an oddity with my machine09:43
voidspacehah, latest run of master on CI failed to build09:44
voidspaceWily unit tests passed on Friday on CI though09:45
dimiternvoidspace, I see both master and maas-spaces-merge-master-7* pass here, although I did see the LoginsDuringUpgrade test failing once09:48
voidspacedimitern: how did it fail?09:48
dimiternvoidspace, something like can login as machine expected IsTrue, got false09:54
voidspacewith status test, I'm getting five machines when only one was expected09:56
voidspacedigging into it09:56
voidspacedimitern: http://pastebin.ubuntu.com/14992568/10:50
wallyworlddimitern: here's a trivial one line fix for a failing windows unit test (i hope), can you take a quick look? http://reviews.vapour.ws/r/3771/12:19
dimiternwallyworld, looking12:20
dimiternwallyworld, LGTM12:21
dimiternwallyworld, and I assume a better fix will be done there once possible, right?12:21
wallyworlddimitern: that whole line will be deleted12:21
wallyworldsince JUJU_HOME won't be supported and the fallback in the code will be gome12:22
wallyworldbut CI scripts need to be updated12:22
dimiternwallyworld, right12:22
wallyworldthe test fails because of a temporary fallback in the code12:22
voidspacewallyworld: hey, you know anything about status tests?12:29
wallyworlda littlw12:29
wallyworldi do know i hate them12:30
voidspacewallyworld: there's a few that fail consistently on my machine (but not anywhere else)12:30
voidspacewallyworld: here's one http://pastebin.ubuntu.com/14992568/12:30
voidspacewallyworld: the test is trying to limit status to an exposed service but gets five machines instead of one12:30
voidspacewallyworld: that's running just that test with the preceding tests in the loop removed (so line numbers won't match yours) to eradicate state leakage between tests as a cause12:31
wallyworldand it's intermittent?12:31
wallyworldnot on your machine12:31
voidspacewallyworld: consistent on my machine, not seen elsewhere yet I believe12:32
voidspacewallyworld: but in means I can't do a full unit test run without failures on my machine - so I'd like to find out the cause12:32
wallyworldi get mongo failures quite regularly :-(12:33
wallyworldvoidspace: i don't see anything obvious, how long has it been doing this?12:36
voidspacewallyworld: new today12:39
voidspacewallyworld: thanks for looking anyway12:39
wallyworldvoidspace: i do notice what appears to be old agent-state values in the yaml output. those are gone from master. you could try rebasing12:40
voidspacewallyworld: I've just had a thought - I copied environments.yaml from CI12:41
voidspaceI wonder if that is screwing with things12:41
voidspacewallyworld: I'm on current master head12:41
wallyworldprobably not, but yoy could try without, i'd have to see what's in it i guess12:41
wallyworldshouldn't affect the tests12:42
voidspacenah, doesn't affect it - still 6 failures12:44
wallyworldhmmm, there must be something unique to your setup, but i can't think of what12:56
voidspaceupdated version of go as well, no difference13:12
dimiternvoidspace, check if you have cache.yaml in ~/.juju/environments/13:12
dimiternvoidspace, and delete it if it's there before retrying13:12
voidspacedimitern: I blew away .juju and did a fresh init13:12
dimiternvoidspace, hmm - well, then check ~/.local/share/juju ?13:13
voidspacedimitern: I blew that away before the init too :-)13:13
voidspaceI'll check though13:13
voidspacejust reinstalling lxc to fix (hopefuly) another test failure on my machine13:13
voidspacedimitern: no cache.yaml there13:14
dimiternvoidspace, what do you have in ~/.local/share/juju ? I only have ssh/ with juju_id_rsa + juju_id_rsa.pub13:17
voidspacedimitern: I have an environments.yaml too13:17
voidspacedimitern: which was created by juju init13:17
dimiternvoidspace, in ~/.local/share/juju/ ?13:18
dimiternvoidspace, I don't have this fwiw13:18
voidspacethis is a fresh build with go 1.5, no preexisting juju home directory, then juju init13:19
voidspacewhich created environments.yaml13:19
voidspaceit's just the boilerplate one13:19
dimiternvoidspace, juju init shouldn't be necessary anymore AIUI13:21
voidspacedimitern: it certainly shouldn't cause tests to fail though!13:21
voidspaceI'll get rid of the environments.yaml and see13:22
dimiternvoidspace, yeah13:22
voidspacego rid of canonistack and ec2 credentials from my environment - that didn't fix it either13:22
voidspaceremoving environments.yaml didn't help13:23
dimiternvoidspace, try rm -fr $GOPATH/pkg/ followed by go install github.com/juju/juju/... ?13:25
voidspacedimitern: heh, ok13:25
voidspaceI just updated go version which I think did the same thing13:27
voidspacenope, same problem13:28
voidspaceback to digging into the code13:28
dimiternvoidspace, hmm ok, I hope you find it13:31
bdxhey whats going on everyone??? Can someone give an example of provisioning storage using the openstack provider? Thanks!14:37
natefinchericsnow: ready when you are.15:16
mupBug #1543178 opened: TestBlankJujuXDGDataHomeEnvVar fails because widows isn't looking at xdg location <ci> <regression> <test-failure> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1543178>15:50
mupBug #1543189 opened: testing_test.go: fakeHome declared and not used <ci> <go1.5> <go1.6> <regression> <test-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1543189>16:20
mupBug #1543193 opened: TestConfig fails on windows because of wrong dir <ci> <regression> <test-failure> <windows> <juju-core:Triaged> <https://launchpad.net/bugs/1543193>16:20
mupBug #1542992 changed: ec2: destroy-controller causes rate limiting failure <juju-core:New> <https://launchpad.net/bugs/1542992>16:32
dimiternvoidspace, any luck with those test failures btw?16:56
frobwarevoidspace, dooferlad: sync with rick17:02
mupBug #1542990 changed: tools version checker restarts every 3 seconds <juju-core:New> <https://launchpad.net/bugs/1542990>17:03
cheryljfwereade: did you see waigani's email about moving the state workers to use the dependency engine?17:04
voidspacefrobware: omw17:07
voidspacedimitern: buried in predicate matchers in the apiserver at the moment :-)17:07
voidspacefrobware: can't join17:08
dimiternvoidspace, right, np17:08
mupBug #1543216 opened: EOF uploading charm <charm> <ci> <deploy> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1543216>17:12
mupBug #1543216 changed: EOF uploading charm <charm> <ci> <deploy> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1543216>17:24
fwereadecherylj, dammit, I did, it looked broadly sensible/correct but I haven't done it properly yet17:25
mupBug #1543216 opened: EOF uploading charm <charm> <ci> <deploy> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1543216>17:33
mupBug #1543223 opened: kill-controller fails on missing volume <ci> <kill-controller> <juju-core:Triaged> <https://launchpad.net/bugs/1543223>17:33
mupBug #1543223 changed: kill-controller fails on missing volume <ci> <juju-release-support> <kill-controller> <juju-core:Triaged> <https://launchpad.net/bugs/1543223>17:39
mupBug #1543223 opened: kill-controller fails on missing volume <ci> <juju-release-support> <kill-controller> <juju-core:Triaged> <https://launchpad.net/bugs/1543223>17:42
voidspaceunitMatchSubnet returns true for match when it gets an error!17:51
voidspacewhich seems very odd17:51
voidspacehowever, if there's an error - match shouldn't be checked17:52
voidspaceI think we've got a double-nil error here - the error is != nil but errors.Trace is still returning nil17:52
voidspacesomething like this17:52
voidspaceno, not the problem17:53
tych0cherylj: looks like i finally got #4314 to work :)18:01
voidspacedimitern: frobware: ah, it's matching on the subnet - now to work out why18:02
mupBug #1543235 opened: UniterSuite.TestUniterRelations again fails <ci> <i386> <intermittent-failure> <ppc64el> <regression> <juju-core:Triaged> <juju-core 1.25:Triaged> <https://launchpad.net/bugs/1543235>18:03
voidspacedimitern: frobware: it's fucking BT dns18:04
frobwarevoidspace, been there.18:04
voidspacedimitern: frobware: BT causes net.ResolveIPAddr("started") to work...18:04
voidspacefsking bollox18:04
voidspaceok, at least I found it18:04
frobwarevoidspace, this is the problem I had with "testing.invalid"18:04
voidspacenow to see if I can fix it18:04
voidspacefrobware: yeah, I will have that problem too...18:05
voidspaceso matchSubnet returns true for the "started" status filter18:05
dimiternvoidspace, wow! that really sucks for BT :)18:11
cheryljtych0: yay!!18:14
cheryljtych0: CI should pick up your branch to run now since you've updated it.18:15
cheryljtych0: I see unit test failures, though, now that it's gotten through the build.18:17
cheryljtych0: http://juju-ci.vapour.ws:8080/job/github-merge-juju/6271/console18:18
tych0cherylj: i think that's the last build18:18
tych0the n-1th build18:18
tych0i fixed those and the jujubot actually merged it18:19
cheryljok :)18:19
tych0cherylj: what i don't remember is how to go to the fancy web page you sent me to18:20
tych0cherylj: and if jujubot merged that, are the additonal tests that are going to be run now that i need to check that fancy web page for, or not?18:20
cheryljtych0: http://reports.vapour.ws/releases#lxd-container-type18:22
cheryljtych0: there are also reports emailed out from CI with summaries.18:22
tych0ah, reports.vapour.ws :)18:22
cheryljtych0: I can keep an eye out for yours and let you know if there are any failures seen there that are unique to your branch.18:23
tych0cherylj: oh, that would be swell, thanks!18:23
tych0cherylj: is there anything else you need from me right now? (modulo any fixes, of course)18:23
cheryljtych0: nope, not until your branch runs through CI18:24
tych0cherylj: ok, cool. thanks!18:24
mupBug #1542985 changed: Juju won't switch environments <juju-core:Invalid> <https://launchpad.net/bugs/1542985>18:33
natefinchericsnow: trivial review?18:36
dimiternvoidspace, dooferlad, frobware, http://reviews.vapour.ws/r/3773/ please have a look when you can18:38
natefinchericsnow: should I just remove the existing show-charm-resources command (currently called "resources")?  It conflicts with the card I'm working on that wants list-resources to have an alias of "resources"18:40
ericsnownatefinch: sure18:40
ericsnownatefinch: I'll be removing it in my patch as well18:40
natefinchericsnow: figured.18:40
natefinchericsnow: cool18:40
natefinchrick_h__ or rick_h___: we have a task to change push-resource to only accept a single resource at a time, in anticipation of adding a --comment flag... in that case, does it make sense to keep the format push-resource service name=file, or should we just make it push-resource service name file  (i.e. drop the equals and make the name and file separate arguments)?18:50
mupBug #1538241 opened: 2.0-alpha2 stabilization <blocker> <juju-core:Triaged> <https://launchpad.net/bugs/1538241>19:06
mupBug #1538241 changed: 2.0-alpha2 stabilization <blocker> <juju-core:Triaged> <https://launchpad.net/bugs/1538241>19:12
mupBug #1538241 opened: 2.0-alpha2 stabilization <blocker> <juju-core:Triaged> <https://launchpad.net/bugs/1538241>19:15
natefinchericsnow: any idea whay this PR didn't get a review on reviewboard?19:42
natefinchericsnow: https://github.com/juju/juju/pull/433219:42
* ericsnow takes a look19:42
ericsnownatefinch: also, be sure to take a look at the tasks for that card (there's more to it than the rename)19:43
natefinchericsnow: ahh, I missed the tasks, thanks19:44
ericsnownatefinch: it looks like you already have a review request up for revision 968fb3719:47
ericsnownatefinch: perhaps you have a draft up? an old PR?19:47
ericsnownatefinch: in either case you have to completely delete (rather than discard) the old review request19:48
mupBug #1356500 changed: Lost relation data after juju killing upgrader/uniter on state connection lost. <sts> <juju-core:Invalid> <https://launchpad.net/bugs/1356500>19:48
mupBug #1433336 changed: juju deploy tags are persistent accross juju deploys <juju-core:Invalid> <https://launchpad.net/bugs/1433336>19:48
mupBug #1543262 opened: keystone V3 support needed <uosci> <Go OpenStack Exchange:New> <juju-core:New> <https://launchpad.net/bugs/1543262>19:48
thumpermorning peeps19:49
natefinchericsnow: weird19:52
mupBug #1543262 changed: keystone V3 support needed <uosci> <Go OpenStack Exchange:New> <juju-core:New> <https://launchpad.net/bugs/1543262>19:54
mupBug #1356500 opened: Lost relation data after juju killing upgrader/uniter on state connection lost. <sts> <juju-core:Invalid> <https://launchpad.net/bugs/1356500>19:54
mupBug #1433336 opened: juju deploy tags are persistent accross juju deploys <juju-core:Invalid> <https://launchpad.net/bugs/1433336>19:54
natefinchericsnow: yeah, there's a draft up.... no idea how that got there.  it has no contents, though.19:54
natefinchericsnow: how do I delete it?19:55
ericsnownatefinch: under the "Close" tab click on "Delete Permanently"19:55
natefinchericsnow: I don't have that, must be admin only?19:55
ericsnownatefinch: weird, I thought you could do that for your own requests19:56
ericsnownatefinch: I'll clean it up19:56
natefinchericsnow: thanks19:56
ericsnownatefinch: what's the URL to the request19:57
mupBug #1356500 changed: Lost relation data after juju killing upgrader/uniter on state connection lost. <sts> <juju-core:Invalid> <https://launchpad.net/bugs/1356500>20:06
mupBug #1433336 changed: juju deploy tags are persistent accross juju deploys <juju-core:Invalid> <https://launchpad.net/bugs/1433336>20:06
mupBug #1543262 opened: keystone V3 support needed <uosci> <Go OpenStack Exchange:New> <juju-core:New> <https://launchpad.net/bugs/1543262>20:06
ericsnownatefinch: kill that PR and and try again20:30
natefinchericsnow: worked, thanks20:33
natefinchericsnow: I have a couple of merges going.  Gotta go snowblow again.  I'll finished up the tasks on that show-service-resources patch when I get back in... I should be able to get that and the change to push-resource done tonight20:34
ericsnownatefinch: cool20:34
=== natefinch is now known as natefinch-afk
mupBug #1543283 opened: [Joyent] 4k ssh key can not be used: "cannot create credentials: An error occurred while parsing the key: asn1: structure error: length too large" <juju-core:New> <https://launchpad.net/bugs/1543283>21:09
rick_h___natefinch-afk: I think name=file is worthwhile as it's explicit.21:28
rick_h___wallyworld: can you bump that until after your standup please?21:51
wallyworldrick_h___: the cli meeting?22:01
rick_h___wallyworld: yes please. I'm flying back in that window but will in a layover after your standup22:02
rick_h___wallyworld: or do you want to chat now?22:02
wallyworldrick_h___: if alexisb is free we could22:02
wallyworldrick_h___: ah, but she has monday off, that's right, so i'll bump22:03
rick_h___wallyworld: ah right, she's on holiday today22:03
wallyworldit tuesday here so i forgot what day it was22:03
wallyworldover there22:03
rick_h___ty wallyworld22:06
wallyworldnp. you sre a busy, busy man22:07
rick_h___wallyworld: wheeeee22:09
rick_h___wallyworld: "livin the dream"22:09
wallyworldliving the dream :-D22:09
wallyworldyup :-)22:09
wallyworldme too!22:09
rick_h___wallyworld is my role model22:10
wallyworldfor some value of "role model"22:11
wallyworldthe mind boggles22:11
wallyworldi am good at livin the dream :-)22:11
rick_h___wallyworld: ooh, can I ask a favor? I'm trying to figure out something.22:16
rick_h___wallyworld: if an ubuntu user isn't on a guest ubuntu image, does juju create one?22:16
rick_h___wallyworld: someone thinks this is true but unsure how to check it out22:16
wallyworldafair, we expect an ubuntu user on our public cloud images, i think we create one for manual provider22:17
wallyworldi'd have to check to be sure22:17
wallyworldthumper: can you recall?22:18
rick_h___wallyworld: if you get a sec can you dbl check? We're talking about providing images for someone and they want a non-ubuntu user ootb and I'm worrked about charms not working if there's no ubuntu user22:18
rick_h___wallyworld: or hint me on how to dbl check22:18
wallyworldthumper: do you know for sure or not if we create an ubuntu user if one doesn't exist?22:18
thumperI think we do now22:18
thumpernot for sure, but 85%22:18
wallyworldrick_h___: i'll find out and get back to you22:19
rick_h___wallyworld: ty22:19
rick_h___I'd like to tell them "no" but wanted to make sure I had grounds first :)22:19
wallyworldok, maybe you can say yes :-)22:20
rick_h___no, I'm not good at giving people what they want. I like it my way :P22:21
wallyworld/ SetUbuntuUser creates an "ubuntu" use for unix systems so the juju client22:25
wallyworld/ can access the machine using ssh with the configuration we expect.22:25
wallyworldso we create the ubuntu user in cloud init22:25
rick_h___ah ok coolio ty22:25
davechen1ycan I get a review pls22:29
davechen1yoh dear22:31
davechen1yi have a horrible feeling about the api server22:31
thumperdavechen1y: shipit22:31
thumperwhat is this feeling?22:32
davechen1yevery api server "conection" is a http conneciton22:32
davechen1ywhich is pooled on the client22:32
davechen1ythere is no "session for a client"22:32
davechen1yit's just a bag of http connections22:32
davechen1yso "counting the number of outstanding connections" is complicated22:32
davechen1ydo we want the number of http requests in fly22:33
davechen1ythe number of http connections connectde (but possibly inactive)22:33
davechen1yi think the safest thing is always send Connection: closed after every API request22:33
davechen1ydisabling pooling on the client22:33
davechen1yalthought, this probably isn't the root issue that Cheryl is chasing22:34
davechen1yisnt it web sockets over http ?22:36
davechen1yso actaully there _is_ a long lived connection22:36
davechen1ywhy did it have to be web sockets22:36
davechen1yok, so back in the day we had the gui22:36
davechen1ybut now we have more clients that are _not_ browsers22:36
davechen1y                        modelUUID := req.URL.Query().Get(":modeluuid")23:07
davechen1y                        logger.Tracef("got a request for model %q", modelUUID)23:07
davechen1y^ left over debugging ?23:07
wallyworldanastasiamac_: axw: a minute late23:15
anastasiamac_wallyworld: k. let us know when23:15
wallyworldhere now23:16
davechen1ycherylj: what was the issue # we talked about at standup ?23:23
davechen1ythumper: i think I figured out a way around it23:24
cheryljdavechen1y: https://bugs.launchpad.net/juju-core/+bug/154321623:32
mupBug #1543216: EOF uploading charm <charm> <ci> <deploy> <intermittent-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1543216>23:32
mupBug #1543362 opened: juju debug-log returns error if run too early in bootstrap <juju-core:New> <https://launchpad.net/bugs/1543362>23:37

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!