/srv/irclogs.ubuntu.com/2014/06/13/#juju.txt

=== vladk|offline is now known as vladk
=== CyberJacob|Away is now known as CyberJacob
=== CyberJacob is now known as CyberJacob|Away
vilajam, mgz: Hi there ! We are severely affected by bug #1200267 to the point where people feel discouraged to write integration tests. The bug description perfectly capture our use case, yet, the importance is 'Low'. What's your pov on that ?07:51
_mup_Bug #1200267: Expose when stable state is reached <canonical-is> <canonical-webops> <papercut> <juju-core:Triaged> <https://launchpad.net/bugs/1200267>07:51
vilajam, mgz: Don't get me wrong, I realize you may have more urgent issues but I'd like to understand what can be expected from juju or amulet in the short term07:52
=== isviridov|away is now known as isviridpv
=== isviridpv is now known as isviridov
=== psivaa-afk is now known as psivaa
jam1vila: https://docs.google.com/a/canonical.com/document/d/1XZN2Wnqlag9je73mGqk-Qs9tx1gvOH0-wxMwrlDrQU4/edit#heading=h.mqzqdofq7hgv is a spec that we worked on for having charms be able to report when they've finish what they're working on to report "ready".10:08
jam1also some discussion on it here: https://docs.google.com/a/canonical.com/document/d/1XZN2Wnqlag9je73mGqk-Qs9tx1gvOH0-wxMwrlDrQU4/edit#heading=h.9euay17obra710:08
jam1I believe it falls under "Observability" which is tasked to Tanzanite (wallyworld's team)10:08
jam1but later in this cycle10:08
jam1say 3 months out or so10:08
* wallyworld_ takes note10:08
vilawallyworld_: you're on my radar now ;-D10:09
wallyworld_\o/10:09
vilajam1: thanks ! /me reads10:09
jam1wallyworld_: is it accurate to say that "charm status" is on your work items?10:09
jam1I was trying to sort through all the stuff that we discussed and if it actually got scheduled10:10
wallyworld_it's all a bit open ended, but it can be if it's considered important10:10
jam1wallyworld_: I certainly know people who want it, the main question is relative importance10:11
jam1"observability" got scheduled to you, but I think it covers a lot of items10:11
wallyworld_yep. i've reached out to stakeholders (dan, antonio, kapil etc) to ask for input10:11
jam1wallyworld_: there is also https://bugs.launchpad.net/juju-core/+bug/1254766 sounds dupe-ish10:12
_mup_Bug #1254766: unit "started" status is unhelpful <cloud-installer> <hooks> <landscape> <relations> <status> <usability> <juju-core:Triaged> <https://launchpad.net/bugs/1254766>10:12
vilajam1: so that would be the 'ready/unready' hooks in 'State, status, charm reporting' ?... Hmm, right, may be more even in that section10:12
jam1there was also a separate discussion about an "idle" indicator, separate from charm health10:12
wallyworld_yes, we did discuss that10:12
wallyworld_our todo lists is not yet fully formed10:13
vilajam1: anyway, you've perfectly answered, thanks10:14
vilawallyworld_: ETA please !10:14
vila:-D10:14
vilawallyworld_: kidding ;)10:14
wallyworld_lol10:14
gnuoyI'm trying to debug an issue with juju not bootstrapping on openstack. Running in debug mode I see it retrieving the json with the image list but then "index has no matching records". What I don't see in the debug is a statement of what key is being used to do the lookup. I'm asking for a trusty amd64 image, I see trusty amd64 images listed so I'm not sure where to go from here10:15
wallyworld_vila: we're finishing up on the github migration and some other feature work eg availability zones. next we'll be looking at what customer facing issues / bugs we should target10:15
wallyworld_gnuoy: what version of juju?10:16
gnuoywallyworld_, 1.18.4-trusty-amd6410:17
wallyworld_private cloud? or do you have access to http://streams.canonical.com/?10:17
gnuoywallyworld_, private cloud and I  believe the list in our cloud is updated daily so something could have gone wrong in the simplestream update process. But from what I can see it all looks good10:18
vilawallyworld_: I'm subscribed to bug #1254766 and bug #1200267 and officially nominate them to be my 3 wishes ;-D10:19
_mup_Bug #1254766: unit "started" status is unhelpful <cloud-installer> <hooks> <landscape> <relations> <status> <usability> <juju-core:Triaged> <https://launchpad.net/bugs/1254766>10:19
_mup_Bug #1200267: Expose when stable state is reached <canonical-is> <canonical-webops> <papercut> <juju-core:Triaged> <https://launchpad.net/bugs/1200267>10:19
wallyworld_sure, np, thanks for the input10:19
gnuoywallyworld_, http://paste.ubuntu.com/7638201/10:19
wallyworld_gnuoy: what does "juju metadata validate-images" say?10:20
gnuoywallyworld_, http://paste.ubuntu.com/7638212/10:21
gnuoyline 54 of http://paste.ubuntu.com/7638201/ shows a "trusty 14.04 amd64" entry10:22
wallyworld_gnuoy: if you look inside your index.json file, does it have data for region "serverstack" and endpoint " http://10.98.191.25:5000/v2.0" ?10:22
gnuoywallyworld_, yes, http://paste.ubuntu.com/7638231/10:23
wallyworld_can you paste the whole file?10:24
gnuoywallyworld_, sure10:24
gnuoywallyworld_, http://paste.ubuntu.com/7638235/10:25
wallyworld_gnuoy: and you have daily streams configured?10:26
gnuoywallyworld_, sorry, I don't understand. The data is updated daily, is that what you're asking ?10:27
jam1gnuoy: there are often 2 different streams10:28
jam1"released"10:28
jam1and "daily"10:28
wallyworld_gnuoy: simestreams uses a stream type eg daily or released to distinguish images10:28
wallyworld_com.ubuntu.cloud.daily:server:14.04:amd6410:28
jam1you have to configure "image-stream: daily" in order to get daily streams10:28
wallyworld_is the key in the index file10:28
wallyworld_what jam1 said10:28
wallyworld_your index file contains daily images10:28
wallyworld_so juju needs to be told thst10:29
wallyworld_or else it looks for released images by default10:29
gnuoyah, so in environments.yaml  ?10:29
gnuoyYes, I have that set10:29
wallyworld_yup10:29
wallyworld_hmmm10:31
wallyworld_gnuoy: looks like it can only find 1.18.3 tools10:33
wallyworld_2014-06-13 10:05:40 DEBUG juju.environs.simplestreams simplestreams.go:883 metadata: &{map[com.ubuntu.juju:12.04:i386:{ 1.15.0 i386   map[20140603:0xc2100f1d20]} com.ubuntu.juju:13.04:amd64:{ 1.14.1 amd64   map[20140603:0xc210106120]} com.ubuntu.juju:13.10:amd64:{ 1.14.1 amd64   map[20140603:0xc210106360]} com.ubuntu.juju:14.04:amd64:{ 1.16.2 amd64   map[20140603:0xc2101065a0]} com.ubuntu.juju:14.10:amd64:{ 1.18.2 amd6410:33
wallyworld_map[20140603:0xc210106a80]} com.ubuntu.juju:14.10:arm64:{ 1.19.2 arm64   map[20140603:0xc210106ba0]} com.ubuntu.juju:14.10:i386:{ 1.18.2 i386   map[20140603:0xc210106cc0]} com.ubuntu.juju:14.10:ppc64:{ 1.19.2 ppc64   map[20140603:0xc210106de0]} com.ubuntu.juju:12.04:amd64:{ 1.14.1 amd64   map[20140603:0xc2100f1c00]} com.ubuntu.juju:13.10:i386:{ 1.15.0 i386   map[20140603:0xc210106480]} com.ubuntu.juju:14.04:i386:{ 1.16.2 i38610:33
wallyworld_map[20140603:0xc210106840]} com.ubuntu.juju:12.10:amd64:{ 1.14.1 amd64   map[20140603:0xc2100f1ea0]} com.ubuntu.juju:12.10:i386:{ 1.15.0 i386   map[20140603:0xc210106000]} com.ubuntu.juju:13.04:i386:{ 1.15.0 i386   map[20140603:0xc210106240]} com.ubuntu.juju:14.04:arm64:{ 1.17.2 arm64   map[20140603:0xc210106720]} com.ubuntu.juju:14.04:ppc64:{ 1.19.2 ppc64   map[20140603:0xc210106960]}] map[] Tue, 03 Jun 2014 08:02:36 +0000 products:1.010:33
wallyworld_com.ubuntu.juju:released:tools}10:33
wallyworld_2014-06-13 10:05:40 INFO juju.environs.bootstrap bootstrap.go:58 picked newest version: 1.18.310:33
gnuoywallyworld_, and that would effect its ability to find the right disk image ?10:33
wallyworld_yes, if you are bootstrapping with 1.18.410:34
wallyworld_i think it needs an exact match10:34
wallyworld_1.18.4 client i mean10:34
gnuoywallyworld_, I took " picked newest version: 1.18.3" to mean, "shrug, I'd have liked 1.18.4 but 1.18.3 will do"10:34
gnuoywallyworld_, ok, let me have a go at fixing that and I'll report back on my result10:35
wallyworld_i think it is saying that it looked in the metadata and found all these versions and it picked the latest10:35
gnuoywallyworld_, I've added the latest tools and I can see bootstrap picking them up. The image lookup is still failing though, http://paste.ubuntu.com/7638333/10:52
wallyworld_gnuoy: can you paste the products file at streams/v1/com.canonical.serverstack.serverstack:ubuntu:daily.json10:58
gnuoywallyworld_, that was the one I pasted before, http://paste.ubuntu.com/7638235/10:59
wallyworld_gnuoy: oh sorry, the other one then, the index file?11:01
gnuoysure, np11:01
gnuoywallyworld_, http://paste.ubuntu.com/7638362/11:02
gnuoywallyworld_, I need to nip out for about an hour11:16
wallyworld_gnuoy: ok, i'm looking at thedata and see no reason for it not to match11:16
wallyworld_:-(11:16
gnuoywallyworld_, it'd be really useful if juju said what key it was looking for, or maybe the logic is more complicated than that.12:10
wallyworld_gnuoy: i can't recall ottomh exactly what it logs, but yeah12:11
wallyworld_i'll have to use your metadata to set up something locally and try and see why it's failing12:12
wallyworld_but it's quite late here and i'm falling asleep sadly so i can only do it tomorrow or something12:12
gnuoywallyworld_, np, I shall keep digging.12:14
wallyworld_ok, email me if you want and i'll see what i can do over the weekend12:14
gnuoywallyworld_, I can pick it up with you on monday12:14
wallyworld_ok12:15
=== vladk is now known as vladk|offline
=== psivaa is now known as psivaa-afk
hazmatmarcoceppi, i haven't tried on osx.. i've resurrected my osx machine though, i can give it a whirl14:22
hazmatmarcoceppi, part of the issue on osx is that the brew setup is out of date, 1.16 style.. and that's not supported14:22
hazmatmarcoceppi, i've got an extant bug that we should be distributing/compiling osx binaries14:22
nookyHello, one question, exist some chance that a Hook do a callback to the machine where was executed juju set for example?14:31
james_whi lazyPower, does your name in the channel title mean you are on the hook for charm reviews?14:40
james_wif so I'd appreciate a look at https://code.launchpad.net/~james-w/charms/precise/haproxy/metrics/+merge/22140814:40
lazyPowerjames_w: ack, i'll take a look shortly. i'm in a meeting14:40
james_wthanks14:40
lazyPowerjames_w: looking now15:00
gnuoywallyworld_, I added some debug to my client http://paste.ubuntu.com/7639281/ which showed that the items were lacking the endpoint and region fields and hence the match was failing. I've spoken to smoser and he was doing some work on the format of the simple streams data published. It looks like this has exposed a bug in juju in that higher level config options should apply to lower level ones but juju when looking for an image match is not using the high15:10
gnuoyer level options15:10
=== psivaa-afk is now known as psivaa
marcoceppihazmat: brew has 1.18.315:32
hazmatmarcoceppi, cool good to know it got updated.15:37
smoserhey15:37
hazmatmarcoceppi, doesn't really change my perspective we should be distributing osx binaries the same way we do for windows15:37
smosercan someone read juju core code for me (as i'm to lazy)15:38
marcoceppihazmat: fwiw, juju-docean didn't quite work on osx15:38
hazmatsmoser, what do you need?15:38
smoserand tell me if15:38
smoser https://bugs.launchpad.net/simplestreams/+bug/132980515:38
_mup_Bug #1329805: juju search for image does not find item if endpoint and region are inherited <juju-core:Confirmed> <simplestreams:Fix Released> <https://launchpad.net/bugs/1329805>15:38
smoserif juju will look "up" at all for endpoint/region, or if the tags explicitly have to be on the item level.15:38
hazmatmarcoceppi, hmm.. i'll take a look, there's some extant patches for client that might resolve15:38
smoseri know it will not look up to sibling-of-products.15:38
marcoceppihazmat: cool, thanks15:38
lazyPowerjames_w: excellent tests! sorry its taken me a moment i got sidetracked, back on the review now15:48
lazyPowerjames_w: merged. Thanks for the contribution!16:57
james_wthanks lazyPower17:06
nuclearbobis it possible to run juju in openstack from an instance in that same openstack region?17:33
marcoceppinuclearbob: I'm not sure I understand your question17:36
marcoceppibut I think the answer is yet17:36
marcoceppiyes*17:36
nuclearbobmarcoceppi, I'm got an instance setup in canonistack lcy 02, and I want to run juju in that instance, but whenver I run canonistack-sshuttle, I lose my ability to connect to the instance.  I'm trying to set it up now without using sshuttle, but now I'm getting bootstrap failures, and I'm not sure how to reset things17:47
marcoceppinuclearbob: are you running canonistack-sshuttle from the instance?17:47
nuclearbobmarcoceppi, yes17:47
marcoceppinuclearbob: don't you're already in canonistack you don't need to run sshuttle17:48
marcoceppiit should just work without any shuttleing17:48
nuclearbobmarcoceppi, okay, that makes sense17:48
nuclearbobmarcoceppi, now when I try to bootstrap, I get this: http://pastebin.ubuntu.com/7640009/17:48
marcoceppinuclearbob: run with --debug see if it gives you more information17:49
nuclearbobmarcoceppi, it does: http://pastebin.ubuntu.com/7640012/17:50
marcoceppinuclearbob: run juju destroy-environment --force17:51
marcoceppithen try to bootstrap again17:51
mattrae_hi, i have units with failed, hooks.. resolve --rtry tells me "cannot set resolved mode for unit".. "already resolved" http://pastebin.com/msfTZSA417:51
mattrae_any way to get past that?17:51
nuclearbobmarcoceppi, http://pastebin.ubuntu.com/7640024/17:51
marcoceppimattrae_: how long has it been since you first ran resolved? and did it first work or has it been suck always? are you in debug hooks?17:52
mattrae_marcoceppi: i tried to do debug-hooks to debug this after trying --retry.. now i'm thinking maybe the hook is still running.. just very slowly17:53
mattrae_marcoceppi: i didn't see any output for awhile from the unit log.. but looks like it moved forward17:54
mattrae_yeah the hook finally finished running17:55
=== roadmr is now known as roadmr_afk
josemarcoceppi: will you need me to host that call?18:18
lazyPowerjose: for TSII?18:25
josemhm18:25
joselazyPower: ^18:29
lazyPowerjose: Wouldn't hurt. Castro is out and marco's in a meeting.18:30
joseok18:30
lazyPowerhe's going to be up to the wire18:30
joseI'll go and have lunch, then, bbiab18:30
=== CyberJacob|Away is now known as CyberJacob
marcoceppijose: ?18:51
marcoceppiI can run it, if I find the password in time18:51
josemarcoceppi: troubleshoothing one18:51
joseI already set it up18:51
joses/one/two/18:51
marcoceppiyay18:51
lazyPowerTroubleshooting II is starting everyone!19:01
lazyPowerIf yo have any questions, feel free  to target me and i'll make sure we get them covered19:01
ali1234what is Troubleshooting II?19:02
mbruzekThe second edition of Juju Troubleshooting19:02
lazyPowerali1234: http://www.ubuntuonair.com19:02
mbruzekhttp://ubuntuonair.com/19:02
lazyPowerwait, i dont see it. i see the plenary on UOA19:03
lazyPower@josedid we get UOA updated?19:03
ali1234yeah me too19:03
lazyPowerali1234: refresh :)19:03
ali1234so..... any idea why i can't do a local bootstrap?19:04
mbruzekali1234, We need the error, put the error on pastebin and we could help19:04
ali1234mbruzek: https://bugs.launchpad.net/ubuntu/+source/juju-core/+bug/132942919:04
_mup_Bug #1329429: local bootstrap fails <amd64> <apport-bug> <trusty> <juju-core (Ubuntu):New> <https://launchpad.net/bugs/1329429>19:04
lazyPowerali1234: i've pinged in #juju-dev with your issue pending response from someone taking a look19:07
lazyPowerthey may have additional questions - so you might want to preemptively join #juju-dev to field those as they come in19:08
lazyPowerwith it being a null pointer dereference, its obv. a bug in the code somewhere and requires more intervention than I can provide as a non-developer.19:08
joseany questions?19:21
mattrae_if i'm doing debug-hooks and I want to pause execution leaving a hook in an error state, by using 'exit 1'.. if i see juju move on to execute the next hook, is this normal? i though exit1 should pause execution of hooks so I can exit from tmux and run juju resolved --retry to restart execution19:33
lazyPowermattrae_: not sure if you're on the troubleshooting hangout following along19:34
lazyPowerhowever, its a bug and known behavior - debug-hooks doesn't seem to respect the exit code you provide the shell.19:35
mattrae_lazyPower: nope i'm not on the hangout. is there info about this troubleshooting hangout?19:35
lazyPowermattrae_: ubuntuonair.com19:35
mattrae_lazyPower: oh cool, thanks.. so its a bug19:35
nuclearboblazyPower, or whoever, I've got a different problem with juju local than I did last time.  it seems to just hang when it gets to "Bootstraping Juju machine agent"  debug output is here: http://pastebin.ubuntu.com/7640395/19:36
mattrae_nuclearbob: not sure if it will help.. but maybe try adding --upload-tools?19:38
nuclearbobmattrae_: I'll give it a try19:38
mattrae_nuclearbob: also if it was working previously.. maybe something is corrupt in .juju.. after destroy-environment --force it should be pretty empty19:39
nuclearbobmattrae_, okay, I'll give it a chance to run again and then try moving that out if I need to19:40
nuclearbobmattrae_, if I move .juju somewhere else, do juju init, and then restore my .juju/environments.yaml, is that a good way to restart things?19:40
mattrae_nuclearbob: yeah that should give you a fresh start.. i've seen sometimes where containers get left around as well.. i'm not sure if you're running into that in this case19:43
nuclearbobmattrae_: what container names should I look for to delete?  ones starting with my username?19:43
mattrae_nuclearbob: they'll be something like 'juju-machine-1-lxc-0' if when you run sudo lxc-ls19:44
nuclearbobmattrae_: okay, I don't see any containers with juju in the name19:45
mattrae_nuclearbob: cool, just something to watch out for if one ever gets left around after destroy-environment19:45
lazyPowerThanks everyone for watching. Juju Troubleshooting II charmschool is officially closed19:48
lazyPowermake sure you join us next week for working with / troubleshooting the local provider19:49
=== roadmr_afk is now known as roadmr
nuclearboblazyPower, when is the session on troubleshooting the local provider?19:54
lazyPowerlet me check the calendar, i'm pretty sure its same time next friday but it may be in 2 weeks19:55
lazyPowerInterestingly enough I dont have it on the calendar. I'll track this down for you an ping you back with the info nuclearbob19:56
lazyPowernuclearbob: We're going to schedule it for 3-4PM EDT next friday.19:57
nuclearboblazyPower, cool, thanks19:58
lazyPowerNP. We'll send a reminder email to the list next week19:59
cjohnstonis a units public IP address something that I can do a relation-get for or does it have to have a relation-set first?20:05
=== Ursinha is now known as Ursinha-afk
=== Ursinha-afk is now known as Ursinha
ali1234if i reboot my computer, do all my local environment get started up automatically?20:42
lazyPowerali1234: yep. there are links left behind to autostart the containers20:50
=== Ursinha is now known as Ursinha-afk
sebas5384hey lazyPower o/20:57
lazyPowerhey sebas i'm im nearly home. I'm going to be about 5 minutes late20:57
sebas5384what do you think if we start a hangout live20:57
sebas5384of course, no rush :)20:57
=== CyberJacob is now known as CyberJacob|Away
lazyPowersebas5384: you want to hangouts on air?21:03
sebas5384just for recording21:03
sebas5384there are some people here that can't join us right now21:04
lazyPowersure thats fine21:04
sebas5384but want to watch it later21:04
lazyPoweri need a minute or two to get settled and i'm good 2 go21:04
sebas5384awesome :)21:04
lazyPowerhttps://plus.google.com/hangouts/_/hoaevent/AP36tYfF6VnfsnVX5hl_jMiONm-ybp7bfiniyN2j9tR3c48Z2Vsjhg?authuser=0&hl=en   sebas538421:09
=== isviridov is now known as isviridov|away
sebas5384lazyPower: https://github.com/sebas5384/juju-vagrant-plugin22:19
JoshStroblHey, I noticed that there is documentation on juju debug-log (https://juju.ubuntu.com/docs/config-LXC.html) regarding a bug that prevents it from working as expected. However the bug (https://bugs.launchpad.net/juju-core/+bug/1202682) shows that the fix has been released, so the bug at this point no longer "exists" and I think the docs should be updated to reflect that.22:43
_mup_Bug #1202682: debug-log doesn't work with lxc provider <cts-cloud-review> <debug-log> <local-provider> <papercut> <ssh> <ui> <juju-core:Fix Released by thumper> <https://launchpad.net/bugs/1202682>22:43
alexisbJoshStrobl, thanks for the catch, can you send a note to dev list?22:58
JoshStroblalexisb: #juju-dev ? Gotta add myself back on, but yea I'll send off the mail.23:08
=== Ursinha-afk is now known as Ursinha
sebas5384lazyPower: here's the link of the session I told you https://amsterdam2014.drupal.org/session/devops-where-start23:34
wallyworld_gnuoy: that's great detective work. when i looked at your data i saw that the region/endpoint were defined higher up and assumed they were inherited. i'll investigate and may need to fix juju if there's been a regression23:46
SIGILLWhen I start a new charm, juju will choose the *first* matching/free machine and deploy the charm without trying to be clever about chosing "the best" machine, right?23:47

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