/srv/irclogs.ubuntu.com/2014/11/25/#juju-dev.txt

alexisbnatefinch, you still around?00:05
=== kadams54 is now known as kadams54-away
thumperfwereade_: OMG...how big is that?00:31
anastasiamacthumper: size envy?01:09
anastasiamacthumper: there seem to have been an intent change for 85K lines yesterday?.. no? ;-)01:10
thumperanastasiamac: just wait...01:34
thumperpatience...01:34
* anastasiamac waits for thumper?01:34
* thumper runs the tests one more time01:38
thumper518 files changed, 9228 insertions(+), 8964 deletions(-)01:39
thumper$ git diff master | wc -l01:39
thumper6361801:39
thumperWINNING01:39
thumper\o/01:39
* anastasiamac rolls eyes01:42
anastasiamacthumper: yes, urs is bigger01:42
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
thumperhttps://github.com/juju/juju/pull/121901:49
* thumper wonders if review board will shit itself...01:49
anastasiamacthumper: do i need to change all jc.IsNil that i have put in the last couple of days too?01:50
thumperanastasiamac: afraid so01:50
thumperanastasiamac: once this lands01:50
thumperanastasiamac: my branch changes those that have landed01:50
thumperanastasiamac: simple global replace though :)01:51
anastasiamacthumper: yes, global replaces are always simple...01:51
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
thumpergithub says: Sorry, we could not display the entire diff because too many files (517) changed.01:53
thumperhttp://reviews.vapour.ws/r/528/01:53
thumperhaha01:53
thumper26 pages of diff01:53
thumperglobal search and replace sounds trivial to me...01:54
thumperpitty we don't have the "trivial" review tag on this project01:55
thumperwallyworld: you there?01:56
thumperaxw: there?01:57
axwthumper: I am01:57
thumperaxw: hey there01:57
axwwallyworld went to get lunch01:57
thumperaxw: trade you a bug fix (for the thing I broke on Friday) for a +1 on http://reviews.vapour.ws/r/528/01:58
thumperaxw: everything you need is in the first of the 26 pages of diff01:58
thumpervery mechanical change01:58
axwheh01:58
axwok01:58
thumperthe hardest bit was adding the import to the files that missed it01:58
axwthumper: done02:00
thumperaxw: ta02:01
* thumper goes to fix jujud02:01
wallyworld_thumper_: hi02:38
=== _thumper_ is now known as thumper
thumperoh hai wallyworld02:39
wallyworldoh, looks like i missed out on your mega diff02:39
thumperaxw: here is the fix for bug 1395564  http://reviews.vapour.ws/r/529/02:40
mupBug #1395564: jujud constantly spews `juju.worker runner.go:219 exited "identity-file-writer": StateServingInfo not available and we need it` <juju-core:In Progress by thumper> <https://launchpad.net/bugs/1395564>02:40
thumperwallyworld: you did02:40
wallyworld\o/02:40
thumperwallyworld: I have to land it asap as I feel it will conflict with everyone :)02:41
wallyworldindeed02:41
thumperwallyworld: also, I think I have found out why I get no all-machines.log02:42
thumperthe rsyslog config is fubared02:42
wallyworldwouldn't surprise me02:42
wallyworldi get all-machines.log02:42
thumperwallyworld: really?02:42
wallyworldi think so02:42
thumperon local with master?02:42
thumpercan I get you to check plz?02:43
wallyworldi tought so, but would need to check02:43
wallyworldsure02:43
wallyworldthumper:02:45
wallyworld$ sudo ls /var/log/juju-ian-local/02:45
wallyworldall-machines.log  ca-cert.pem  logrotate.conf  logrotate.run  machine-0.log  rsyslog-cert.pem  rsyslog-key.pem02:45
thumperah phooey02:45
thumperI wonder why I don't get it02:46
thumper...02:46
wallyworldnfi02:46
wallyworldrsyslog is kinda cryptic02:46
menn0thumper: have a look in /etc/rsyslog.d and see if you have a config file left behind that shouldn't be there02:54
menn0thumper: that's what happened to me02:55
thumpermenn0: I have an old running environment02:55
menn0thumper: that's the problem02:55
thumperwhich I probably don't really need...02:55
menn0thumper: the way we set up rsyslog doesn't work with multiple local envs02:55
menn0thumper: i've reported this under 138738802:56
menn0bug 1387388 even02:56
mupBug #1387388: rsyslogd configuration does not support multiple local environments <tech-debt> <juju-core:Triaged> <https://launchpad.net/bugs/1387388>02:56
menn0thumper: different configs get written out for different envs but using the same rsyslog port02:57
menn0thumper: rsyslogd picks one02:57
menn0thumper: and the other config(s) don't work02:58
thumperah...02:58
thumperyes...02:58
thumperthat makes sense02:58
thumperanother thing that fixing the local provider will fix02:58
menn0thumper: exactly02:58
menn0thumper: this also causes the rsyslog workers on each machine and unit to continually restart, in the env who's config didn't win02:59
menn0thumper: because the certs used by rsyslogd are for the other env and don't match what the rsyslog workers are trying to use03:00
thumperhaha03:00
thumperoops03:00
thumpermenn0: there is an rsyslog port that can be configured03:01
thumperbut I think this is my problem of two different users using the came default config03:01
menn0thumper: cool, well that's the workaround I guess03:01
anastasiamacaxw: u have ben so instrumental to block/unblock cmd so far03:04
anastasiamacaxw: and everyone is asking u to review stuff today03:04
axwheh03:04
* axw braces himself03:04
anastasiamacaxw: and u r so attentive during reviews03:04
anastasiamacaxw: ;-)03:04
axwmaybe I'll just swap my OCR day...03:05
anastasiamacaxw: could u plz, pretty plz, have a look at http://reviews.vapour.ws/r/523/03:05
anastasiamacaxw: swap is not my call but it seems fair :-p03:05
thumperanastasiamac surely knows how to butter people up...03:06
thumperheh...03:06
* anastasiamac puzzled as to what thumper implies...03:06
thumperoh lord, you are so big, we're all really impressed down here I can tell you03:06
axwanastasiamac: I'll take a look after I'm done with wallyworld's03:06
anastasiamacaxw: thank you sooooo much :-003:07
thumperanastasiamac: it is the 'u r so attentive', and 'pzr, pretty plz'...03:07
wallyworldaxw: anastasiamac: i'm already reviewing03:07
axwok03:08
anastasiamacwallyworld: thnx ;p03:08
anastasiamacthumper: butter is an important cooking ingredient used for food03:10
* anastasiamac would never dream of puuting food ingredients on ppl03:10
* thumper does a little dance03:19
thumpermega branch landed03:19
axwthumper: why do you have to wait till after upgrade to write the identity file?03:25
wallyworldthumper wins the award for the biggest diff EVER03:25
thumperaxw: because the upgrade steps were potentially changing it03:27
thumperwallyworld: I'll think about one that would be bigger still!!!03:27
axwahh03:27
wallyworldthumper: it's not the size that counts, right?03:28
thumperwallyworld: you keep thinking that...03:28
axwthumper: wouldn't it make sense just to write it in the upgrade step then?03:28
axwoh... it *is* written in there... I'm lost03:28
thumperaxw: this is run every startup03:28
thumperaxw: on the off chance that a machine goes from not a state machine -> a state machine03:29
thumperaxw: I didn't write that bit03:29
thumperaxw: just had to make sure it worked03:29
axwagent.WriteSystemIdentityFile is called on startup already, in StateWorker03:29
axwso between that and the upgrade step which also writes the file, I don't understand the point of the additional worker...03:30
thumperaxw: um... I thought it was removed from state worker03:55
thumperit should have been03:55
thumperIIRC03:55
axwthumper: sorry, was looking at my out of date branch04:03
axwthumper: reviewed04:11
thumperta04:11
* thumper goes to make dinner04:30
thumperlater folks04:30
menn0RB is flaking out on me04:42
menn0I've hit save after editing a comment and it's stuck "Loading..."04:42
menn0I've tried twice now04:42
=== kadams54 is now known as kadams54-away
TheMuemorning08:19
anastasiamacTheMue: morning :-)08:21
TheMueanastasiamac: oh, still online? should be quite late in your TZ, or am I wrong?08:22
anastasiamacTheMue: 6pm... was going to induction but m usually online later (around 9pm+) - u r rite08:23
TheMueanastasiamac: ah, ic08:23
* TheMue looks out of the window, 1° and a bit foggy. *brrr*08:24
anastasiamacTheMue: it's 6pm, very very very hot outside. we r in the aircon set for 21C and it barely copes :-( visit Australia08:26
TheMueanastasiamac: oh, you're inviting me? :D sure I would like to, never have been so far away from home and especially in the southern hemisphere08:27
anastasiamacTheMue: the best bits r inverted seasons = Xmas on z beach :p08:29
TheMueanastasiamac: trying to imagine this, would feel strange. :)08:30
* fwereade_ bbiab08:47
jamdimitern: I might be a couple minutes late to the interview, my son is coming home right now on the bus08:53
dimiternjam, not to worry, I've prepared questions, etc.08:53
jamestunnicliffejam, dimitern: I can wait until you are both ready if you like.08:54
dimiternjamestunnicliffe, hey! welcome :)08:55
jamestunnicliffedimitern: hi :-)08:55
dimiternjamestunnicliffe, no, it's ok - we can start on time and jam will join us08:55
dimiternjamestunnicliffe, we do have a standup meeting right after :)08:56
jamestunnicliffedimitern: great. See you in a moment.08:56
TheMuerogpeppe: ping09:22
rogpeppeTheMue: pong09:22
mattywmorning all09:22
TheMuerogpeppe: I migrated your codec checkers to juju testing. do we have tests for the checkers? meta-testing :D09:22
TheMuerogpeppe: otherwise could you take a look at https://github.com/juju/testing/pull/39. they are simply copied09:23
rogpeppeTheMue: no, we don't have tests, but they've been used a lot09:24
rogpeppeTheMue: i would suggest one change when copying them, that i've been meaning to make for a while.09:25
TheMuerogpeppe: so kinda verification by usage :)09:25
=== psivaa-holiday is now known as psivaa
TheMuerogpeppe: I'm listening09:25
rogpeppeTheMue: i'd suggest that you make the first argument to the checker a string rather than a []byte09:25
rogpeppeTheMue: that way if the test fails, it prints the JSON in string form rather than a load of hex byte values09:26
TheMuerogpeppe: would make the usage more convenient, yes09:26
rogpeppeTheMue: feel free to write some tests if you like09:26
TheMuerogpeppe: ok, will change and then add tests. I'll ping you later again for a review.09:27
mattywdimitern, ping?09:41
jamdimitern: firefox just crashed on me09:53
jamgive my apologies09:53
dimiternjam, it's ok I think - my chrome almost crashed as well :)09:54
dimiternmattyw, hey, sorry was in a meeting09:54
voidspacejam: dimitern: TheMue: got logged out10:07
voidspaceback in a minute10:07
voidspacedimitern: are you around most of today - or are there times you're in meetings?10:17
voidspacedimitern: I have some little "detaily" questions I need to ask at some point10:17
voidspacedimitern: I might collect a few of them together...10:17
voidspacequestions that is10:17
dimiternvoidspace, I have a 1:1 with alexisb in 4.5h, apart from that I'm available10:20
voidspacedimitern: thanks - appreciated10:20
voidspacedimitern: for AddSubnet - rather than taking 8 parameters, I've created a SubnetInfo struct to use as the arg passed in10:20
voidspacedimitern: this mirrors what AddNetwork does10:21
dimiternvoidspace, +1 - we can probably use the same struct for updating existing subnets10:21
voidspacedimitern: cool10:21
voidspacedimitern: "subnets" for the collection name?10:27
voidspacedimitern: and will we need an upgrade step to create the collection - or will it be created on first use?10:27
=== ashipika is now known as ashipika_afk
dimiternvoidspace, "subnets" sgtm10:30
dimiternvoidspace, no need for upgrade step - the collection will be created along with all the others when we try to insert for the first time10:31
voidspacedimitern: cool, thanks10:34
voidspacerebooting11:34
jamwallyworld_: are we still on for storage discussion in 5 min ?11:54
wallyworld_yep11:54
wallyworld_jam: fwereade_: meeting?12:02
mattywfwereade_, when you have 2 minutes12:30
rogpeppedoes anyone know if debug-log is supposed to work correctly on local environments?13:36
rogpeppefwereade_, dimitern: ^13:37
dimiternrogpeppe, it should work if you have all-machines.log13:39
rogpeppedimitern: i'm only seeing results for machine-013:39
dimiternrogpeppe, in all-machines.log ?13:40
rogpeppedimitern: yup13:40
rogpeppesw13:41
rogpeppe% sudo grep -v '^machine-0:' all-machines.log | wc13:41
rogpeppe      0       0       013:41
rogpeppe% sudo wc all-machines.log13:41
rogpeppe  29962  324235 4314099 all-machines.log13:41
fwereade_rogpeppe, that would be a bug then :/13:41
dimiternrogpeppe, hmm it looks like a bug :)13:41
rogpeppeok, i'll report it13:41
rogpeppei'm guessing not many people use debug-log in the local env13:42
dimiternrogpeppe, I haven't even tried - sudo less ~/.juju/localenv/logs/machine-0.log is easier13:42
rogpeppedimitern: i don't have a localenv dir13:43
rogpeppedimitern: i was looking in ~/.juju/local/log/all-machines.log13:43
rogpeppedimitern: i was writing a document and wanted to give instructions that would work for both a local or a remote env13:44
dimiternrogpeppe, localenv is the name of my local env13:44
rogpeppedimitern: ah13:44
rogpeppedimitern: i hadn't realised that dir was named after the local env13:44
alexisbjam, you still around?13:45
dimiternrogpeppe, yeah13:45
rogpeppefwereade_, dimitern: https://bugs.launchpad.net/juju-core/+bug/139615913:49
mupBug #1396159: debug-log only reports machine 0 entries in local environment <juju-core:New> <https://launchpad.net/bugs/1396159>13:49
dimiternrogpeppe, thanks!13:57
rogpeppedimitern: np13:57
rogpeppedimitern: well, only minor probs :)13:57
=== ChanServ changed the topic of #juju-dev to: https://juju.ubuntu.com | On-call reviewer: see calendar | Open critical bugs: None
* rogpeppe lunches14:18
alexisbnatefinch, perrito666 cloudbase call14:30
ericsnownatefinch: ping15:22
natefinchericsnow: howdy, I'm in moonstone15:28
ericsnowk15:28
aznashwantime-appropriate greetings everyone! :D16:33
aznashwancould I please get a quick second opinion on this PR please: https://github.com/juju/juju/pull/74816:33
=== kadams54 is now known as kadams54-away
hazmatfwereade_, how's leader el coming?16:44
marcoceppiHey guys, I'm not 100% sure of the answer to the question in #juju, is there a migration path for 0.7 => 1.X ?16:46
hazmatmarcoceppi, no16:49
marcoceppihazmat: thanks16:49
hazmatmarcoceppi, i think thats the first user request i've seen for pyjuju upgrades16:50
hazmatto core16:50
marcoceppihazmat: the mythical beast exists!16:50
=== kadams54-away is now known as kadams54
=== kadams54 is now known as kadams54-away
alexisbhazmat, katco can update you on what she has been doing recently for leader elections17:55
alexisbhazmat, any reason for your question?17:56
hazmatalexisb, desparate need to avoid hacks in charms17:58
hazmatalexisb, just caught up date via priv message18:02
voidspaceg'night all18:41
katcoso i need to move MachineAgent out of main and into an actual package19:06
katcoany opinions on where it might best be placed?19:06
katcoright now it's in cmd/jujud/ which is in "main"19:07
natefinchkatco: maybe /agent?  no idea if that might cause circular references, though19:41
katconatefinch: root to juju/juju? (i'll take care of any dependency graphs, just looking for logical placement)19:44
katcoit does seem incorrect to keep it resident as a child of cmd/jujud19:46
katcosince that is just that -- building an executable19:46
natefinchkatco: yeah, the stuff in package main should be about *the executable* not any other logic.  to me, that means stuff to parse the CLI and that's about it19:51
katconatefinch: ok, so we're talking about github.com/juju/juju/agent?19:52
natefinchkatco: seems reasonable19:52
katcocool... i'll ping dave tonight if i get a chance since i know he's been looking at this stuf19:52
katcof19:52
katco(dependencies etc)19:53
natefinchcool.  He'll probably have an idea too19:53
katcoi thought he might19:53
dpb1Hey -- what is the current state of juju networking on 1.21?  is there any docs about it?20:02
natefinchdpb1: I don't think we released the network stuff yet.  The 1.21 release notes don't talk about networks.20:15
dpb1natefinch: ok great.  Just noticed it leaked out to juju status command then.  :/20:16
waiganithumper, menn0: the func I was talking about is transformId in state/watcher.go20:17
waigani/ transformId converts a global key for a ports document (e.g.20:17
waigani/ "m#42#n#juju-public") into a colon-separated string with the20:17
waigani/ machine id and network name (e.g. "42:juju-public").20:17
dpb1hrm, I don't see it anymore... must be smoking something.  n/m natefinch20:17
natefinchdpb1: heh, ok good.  Wouldn't be the first time the release notes were lacking something and/or something slipped into the CLI before it was really ready.20:17
waiganiwhy would that be needed?20:18
menn0waigani: it was introduced by dimiter in 14b6409120:26
menn0waigani: the commit message says "openedPortsWatcher no longer exposes global keys"20:26
menn0waigani: I suspect is so that global keys aren't reported over the API20:26
alexisbkatco, dave is out on vacation, just fyi20:30
alexisbback on monday20:30
katcoalexisb: oh doh. ok20:30
katconot a big deal. ty for the info!20:30
waiganimenn0: okay, thanks. I'll focus on working out how the localID is getting pasted into the watcher merge func20:30
* katco pines for her two large screens @ home.20:41
katcothe laptop screen just isn't the same.20:42
perrito666 /query natefinch22:07
perrito666ouch lol22:07
* thumper sighs22:23
* thumper stabs rsyslog in the face22:23
perrito666Is anyone versed enough into deploying your own python soft in windows and want to chat a couple of minutes with me to help me deploy a floss project without going crazy?22:36
perrito666I know many of us pack a windows past22:36
perrito666s/windows/python22:37
thumperI tried to use windows last night22:37
thumperOMG what a PITA22:37
thumperso hard to get anything done22:37
thumperWTF22:37
perrito666I only use windows to create ninja ide installer, but I am having some terribe pains packaging python for windows when my python does complex enough things22:38
thumperkatco, perrito666: do either of you have a local environment running with more than just the bootstrap node?22:38
perrito666thumper: nope22:38
thumpermenn0: just proposing feature flag branch, I'll be taking the discussion to the list, so don't feel like you have to review it now22:42
thumpermenn0: I won't land it until we've had a chat22:42
thumperwhere we is the team22:43
rick_h_thumper: how does that feature flag work out? I was just talking with wayne about those the other week so curious. Did it get handed off your way?22:46
thumperrick_h_: um... just writing an email to the list22:46
thumperrick_h_: been planning on writing it for a while22:46
rick_h_thumper: ok, I'll look forward to it22:46
thumperrick_h_: email sent22:48
* thumper heads to the gym22:48
thumperrick_h_: aren't you supposed to be on holiday this week?22:48
rick_h_thumper: ok, so this looks less like charm flags and more feature flags22:48
rick_h_thumper: I would be if I could get my work deployed22:48
rick_h_thumper: cool, thanks for this PR as we're doing feature flags in python and looking at it in Go so this might be useful as well22:48
thumperrick_h_: yeah, nothing to do with charm flags22:48
rick_h_thumper: have fun at the gym, I'll point my guys at this and might want to chat with you/etc22:49
thumperrick_h_: we could push the generic bits into some utils package22:49
thumperrick_h_: rather than in juju-core22:49
rick_h_thumper: gotcha, yes, please. It's something we're eager to put into place as you can imagine with our use cases22:51
rick_h_thumper: I've sent the team an email to follow up with you and look it over for our own uses tomorrow.22:52
menn0thumper: no problems22:53
=== fuzzy_ is now known as Fuzai
katcothumper: sorry for the delay. i have a manual environment running with lxc hosts on another physical machine.23:08

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