/srv/irclogs.ubuntu.com/2016/02/24/#juju.txt

=== cherylj_ is now known as cherylj
=== aluria` is now known as aluria
gennadiyHi everybody. could you help me with my bundle again? https://code.launchpad.net/~dataart.telco/charms/bundles/dataart-telco-demo/bundle11:43
neiljerramgennadiy, what help do you need?12:09
gennadiyi know that somebody could check error log12:09
gennadiymy bundle is not bublished yet. so i think there are some errors12:10
gennadiy*published12:10
neiljerramHave you posted the error log somewhere?12:11
magicaltroutI think gennadiy means errors to the charm store12:11
gennadiyyes, you are right12:11
gennadiyi reviewed my bundle manually and run bundle proof, seems everything is ok12:12
gennadiyanyway do we have some mechanism to get publish error? maybe by email or web portal?12:13
neiljerramgennadiy, OK - I'm sorry then, because I have not reached that stage myself yet :-)12:13
gennadiybefore pushing i checked bundle on local env. i imported it to juju-gui.12:13
magicaltroutthere's nothing that is publicly published gennadiy12:14
magicaltroutits all magic behind the scenes stuff12:14
magicaltrouti'm not sure how bundles are deployed though gennadiy as I've just done single charms12:15
magicaltroutbut you don't have a a ubuntu version in your charm push path, dunno if that has any impact or not12:15
magicaltroutah i see it in the docs12:16
magicaltroutignore me12:16
magicaltroutalthough looking at the requirement to file a bug12:17
magicaltroutit might be a manual process unlike general charm deployment12:17
magicaltroutso you might just be sat in the queue gennadiy12:17
magicaltrouthttps://jujucharms.com/docs/1.25/charms-bundles12:17
magicaltroutstep 4, 5, 6 and then12:17
magicaltrout"12:17
magicaltroutSomeone will come along and review your bundle for inclusion."12:17
magicaltroutlooks like you just need to wait it out12:17
gennadiythere are 2 different types of bundles: with public namespace(like official version) and user space(with user name prefix in path). so i use user space bundles. the should be deployed automatically12:19
gennadiygenerally it takes 40 min12:20
gennadiyif you have not got result during this time your bundle has error :)12:20
gennadiya few people in this chat has access to error log.12:20
magicaltroutwell I dunno, for charms thats certainly true12:32
magicaltroutthose docs seem to suggest bundles are different12:33
magicaltroutbut I have no idea :)12:33
gennadiyis anybody joined who has access to publish error logs?13:12
jrwrengennadiy: we don't exactly, but i can look at other things.13:19
gennadiyjrwren: do you have possibility to check why bundle is not published yet?13:33
jrwreni don't.13:36
stubStarting from the example at https://github.com/cmars/nrpe-external-master-interface, do I also need an upgrade-charm hook wired up to initialize or reintialize the nrpe-external-master.available handlers?13:42
dweaverHi All, I have a Juju environment on 1.25.0 and I am getting a service constantly reconfigured.  I have 3 cinder units working fine, but then it starts to run cluster-relation-changed, says unit is ready, and then does it again, and again .....  Restarting the services constantly in the process, nagios is detecting it as down.  The cluster relations have not changed, so I am wondering how you tell why Juju is running this constantly in a loop?  Debug-l13:56
dweaverog doesn't seem to tell me what triggers hooks.13:56
dweaverIs there somewhere I can dig further to find out?13:58
jrwrengennadiy: bundle verification failed: ["cannot validate service \"drop-conference\": configuration option \"DID_DOMAIN\" not found in charm \"cs:~dataart.telco/trusty/drop-conference-1\"","cannot validate service \"drop-conference\": configuration option \"PHONE_NUMBER\" not found in charm \"cs:~dataart.telco/trusty/drop-conference-1\""]14:24
magicaltroutI know we touched on some of this stuff in Ghent, are things like that --^ going to become available to developers without having to prod you guys in Juju 2.... and the new review queue etc14:25
magicaltrouta feedback mechanism about what your charm/bundle is upto in the deployment process at least14:27
jrwrenmagicaltrout: yes, there will be no ingestion and you will be able to upload directly to charmstore yourself.14:28
magicaltroutboom14:28
magicaltroutgood to know14:29
=== natefinch is now known as natefinch-afk
lazyPoweraxino o/ heyo14:53
axinohey lazyPower14:54
axinoo/14:54
lazyPoweraxino i have it on good authority you guys have a bomb-diggity logstash charm over in IS14:54
* axino gives lazyPower a Chouffe14:54
lazyPowerWoo14:54
lazyPowerthis early?14:54
rick_h_lazyPower: maybe hit up jrwren as well. I know we were playing with one15:00
lazyPowerrick_h_ i'm already all over that one :)15:00
lazyPowerrick_h_ - its claimed as abandoned, but i think its about to get some TLC15:00
rick_h_lazyPower: gotcha15:01
neiljerramAnyone know why 'juju deployer -b ...' would fail to download code from a launchpad branch?15:01
neiljerramHere's the transcript:15:02
neiljerramroot@calico14:~/md4# rm -rf trusty/15:02
neiljerramroot@calico14:~/md4# juju deployer -b -c liberty.yaml15:02
neiljerram2016-02-24 14:59:49 Using deployment envExport15:02
neiljerram2016-02-24 14:59:49 Starting deployment of envExport15:02
neiljerram2016-02-24 15:00:50 Could not branch lp:~project-calico/charms/trusty/neutron-api/liberty to trusty/neutron-api15:02
neiljerram bzr: ERROR: [Errno 2] No such file or directory15:02
neiljerram2016-02-24 15:00:50 Deployment stopped. run time: 61.5415:02
neiljerramAnd in .bzr.log I have:15:03
neiljerram0.097  bazaar version: 2.7.0dev115:03
neiljerram0.097  bzr arguments: [u'co', u'--lightweight', u'lp:~project-calico/charms/trusty/neutron-api/liberty', u'trusty/neutron-api']15:03
neiljerram0.104  looking for plugins in /home/ubuntu/.bazaar/plugins15:03
neiljerram0.104  looking for plugins in /usr/lib/python2.7/dist-packages/bzrlib/plugins15:03
neiljerram0.129  encoding stdout as sys.stdin encoding 'UTF-8'15:03
neiljerram0.242  ssh implementation is OpenSSH15:03
neiljerram15.043  creating branch reference in file:///home/ubuntu/md4/trusty/neutron-api/15:03
neiljerram17.685  trying to create missing lock '/home/ubuntu/md4/trusty/neutron-api/.bzr/checkout/dirstate'15:03
neiljerram17.719  opening working tree '/home/ubuntu/md4/trusty/neutron-api'15:03
neiljerram18.766  Traceback (most recent call last):15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/commands.py", line 930, in exception_to_return_code15:03
neiljerram    return the_callable(*args, **kwargs)15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/commands.py", line 1121, in run_bzr15:03
neiljerram    ret = run(*run_argv)15:03
mgzneiljerram: please, pastebin15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/commands.py", line 673, in run_argv_aliases15:03
neiljerram    return self.run(**all_cmd_args)15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/commands.py", line 697, in run15:03
neiljerram    return self._operation.run_simple(*args, **kwargs)15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/cleanup.py", line 136, in run_simple15:03
neiljerram    self.cleanups, self.func, *args, **kwargs)15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/cleanup.py", line 166, in _do_with_cleanups15:03
neiljerram    result = func(*args, **kwargs)15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/builtins.py", line 1667, in run15:03
neiljerram    accelerator_tree, hardlink)15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/branch.py", line 1469, in create_checkout15:03
neiljerram    hardlink=hardlink)15:03
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/bzrdir.py", line 907, in create_workingtree15:04
neiljerram    accelerator_tree=accelerator_tree, hardlink=hardlink)15:04
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/workingtree_4.py", line 1565, in initialize15:04
neiljerram    delta_from_tree=delta_from_tree)15:04
neiljerram  File "/usr/lib/python2.7/dist-packages/bzrlib/transform.py", line 2543, in build_tree15:04
neiljerram    del15:04
magicaltroutneiljerram: let me introduce you to pastebin or gist! ;)15:04
neiljerramSure, sorry!15:04
neiljerramSo here's the .bzr.log: http://pastebin.com/rMuqrLsL15:04
neiljerram(Have I done that right?)15:05
mgzneiljerram: yes. and the branch just has a borked symlink15:07
mgzneiljerram: and you're the last commit on it15:07
mgzso... just fix that?15:07
neiljerrammgz, You mean there's a broken symlink somewhere in the working tree?15:08
neiljerrammgz, I'm sorry, I'm not seeing the problem.  So if you are, please do say more!15:12
neiljerramHowever, it occurs to me that I should be able to use strace to see the failing symlink call...15:13
mgzthat's not super useful15:14
mgzbecause it's just a limbo name15:14
mgzneiljerram: so, just get the branch with `bzr branch --no-tree lp:...`15:14
mgzthen you can do `BZR_PDB=1 bzr reconfigure --tree` and drop in at the error site15:14
neiljerramI guess these are the relevant calls: http://pastebin.com/8Bmt9axy15:15
mgzbut getting the borked filename out is mildly annoying15:15
magicaltroutdon't ya just love bzr weirdness! ;)15:15
neiljerram'limbo'?  I see that in the strace too, but I have no idea what it means here.15:15
mgzbzr stages files when creating trees15:16
mgzdoesn't overwrite in place15:16
magicaltroutreminds me of how sad perforce used to make me ;)15:17
neiljerramI'm still not understanding a lot of the terminology here.15:17
neiljerramBut I think you're saying that the bzr history includes a commit somewhere that added a broken symlink.15:17
neiljerramAnd because of that, bzr is always going to fail for me with this branch.  Is that right?15:18
neiljerramAnyway, I will try the commands you recommended!15:18
magicaltroutthats basically it, and you need to extract a copy without a working tree to unbork it15:19
magicaltroutnot that i'm any more clued up on bzr either15:19
mgzneiljerram: hooks/install15:19
mgzneiljerram: the target is "" - which is bogus15:19
neiljerrammagicaltrout, thanks15:20
neiljerrammgz, thanks also!15:20
mgzI can just fix this, might even have push rights to the branch15:21
mgzcommit 142 "unknown <james.page@ubuntu.com>" dunno what he did to screw this, but goodjob15:21
mgzhistory rewriting gone wrong? something else?15:23
=== natefinch-afk is now known as natefnich
=== natefnich is now known as natefinch
mgzjamespage: your git switch breaking things? ^15:23
neiljerramHere's what I see with the bzr reconfigure session: http://pastebin.com/XWG7LFsG15:24
neiljerramDoes 'new-15' mean that the problem was introduced in revision 15 ?15:24
jamespagemgz, not yet15:24
mgzneiljerram: nope, it's just a placeholder name in the stagin area15:25
mgzneiljerram: it's actually more useful to do `bzr diff -r0..-1 and grep the diff for the symlink lines15:25
neiljerramOK...15:26
mgzjamespage: I am confused as to how a change from september that made it impossible to checkout a tree wasn't seen till today15:26
mgzwithout history rewriting15:26
jamespageI don't even know how todo that :-)15:27
mgzjamespage: well, I guess the branches don't have push overwrite protection on lp? any one or any bot could have changed it?15:28
jamespagemgz, we've not done any push --overwrites for a while now15:28
beisnermgz, jamespage - afaik, 15.04 was the last push --overwrite on ~openstack-charmers space15:31
neiljerramDo you already know whether the upstream openstack-charmers branch has the same problem?  I wonder if something went wrong when I pushed my fork of that to lp:~project-calico/charms/trusty/neutron-api/liberty15:36
neiljerramHmm; I don't see the same problem if I use lp:~openstack-charmers/charms/trusty/neutron-api/next instead of my fork.  So I guess there was a problem in how I pushed my fork.  I'll try killing and recreating my fork.15:45
jamespageneiljerram, you might be please to know that as of next week, openstack charm development should be under the openstack project so git/gerrit rather than bzr/launchpad....15:45
neiljerramjamespage, Yes, I will be very pleased about that!  Thanks for your work on this.15:46
gennadiyjrwren: thanks. i have to disable copy/paste functionality15:53
=== natefinch is now known as natefinch-lunch
josembruzek: so, Ryan's in? :D17:34
neiljerramHi all, now seeing a problem with bootstrap, I wonder if anyone might help?17:36
joseneiljerram: sure!17:36
joseneiljerram: what's going on?17:36
neiljerramAppears problem is that machine-0 doesn't manage to bring up juju-br017:36
neiljerramAnd when that happens, I can't ping the machine; the logs that I can see in the vSphere console indicate that it has lost its IP address.17:37
neiljerramIf I deploy the same machine using just MAAS, it's fine.17:38
neiljerram(Using Trusty)17:38
josewhat substrate are you deploying on?17:38
neiljerramNot sure what 'substrate' is - MAAS?17:39
joseyep, what are you deploying on (aws, maas)17:39
neiljerramYes, MAAS then.17:39
jose^those are just examples17:39
neiljerramAnd if I just use MAAS to deploy the same machine, it's fine.17:40
josehmmk. I'm not too much into maas, but maybe there's something missing in Juju, so let's take a look!17:40
neiljerramI.e. it has an IP and I can log in via ssh.17:40
josecan you please execute `juju bootstrap --debug` and paste the results on paste.ubuntu.com?17:40
neiljerramSure...17:40
josethanks :D17:40
neiljerramBy the way, what is meant by the juju "tools" ?17:42
josethe juju tools are basically what enables juju on the machines17:44
joseall the binaries17:44
neiljerramHere's `juju bootstrap --debug`: http://pastebin.com/cVJP56EN17:45
lazyPowerneiljerram - which version of maas are you using?17:47
neiljerramMAAS Version 1.8.3+bzr4053-0ubuntu1 (trusty1)17:47
neiljerram(that's what it says at the bottom of my MAAS web page)17:48
lazyPoweryep, just making sure it wasnt 1.9, i dont think 1.9 works with 1.25 - but dont hold me to that.17:48
* lazyPower hasn't investigated lately17:48
josehmm, from what I read juju is not creating anything, but assuming the bridge is already there17:49
neiljerramcan you see that in those logs?17:52
neiljerramBTW, I should add, bootstrapping was working for me yesterday, and AFAIK I haven't changed anything relevant.17:54
neiljerramBetween yesterday and today, I've worked on and resolved another problem to do with the exact charm code that I am trying to deploy.  But I would not expect that to be relevant to the bootstrap phase.17:55
neiljerramSo it's quite mysterious why it has stopped working.17:55
neiljerramIs there a way I can see the complete log of what Juju did to the machine after it had MAAS-booted?17:57
neiljerramOr can I configure a password for 'root' or 'ubuntu' on that machine (instead of an SSH key), so that I can log in to its console?18:00
josesorry, I had to run do something but I'm back around18:02
josehmm, let me remember where those logs are stored18:03
neiljerramNo worries - thanks for looking at this!18:05
josehappy to help :)18:05
neiljerramI do have to run in about 5 minutes, though - so no pressure :-)18:05
joseok, so the machine is actually created, but juju cannot connect, right?18:06
neiljerramcorrect18:06
neiljerramI can't ping or connect with SSH, either18:06
=== natefinch-lunch is now known as natefinch
joseok, so can you please ssh into the machine?18:10
neiljerramNo, I can't.18:10
neiljerram(I guess because the Juju post-boot setup on the machine has removed the IP address from eth0.)18:11
joseoh, oh, you can't18:12
josehmm... so the bridge is missing. I'd say we have to manually create the bridge18:13
joseI don't know if someone around knows how to do it, I don't tbh18:13
neiljerramok - well I have to go now, but thanks for looking at this!18:14
joseno prob! let us know when you're back and I'll see what I can do18:15
lazyPowercory_fu - if you have a minute, can you help me wrap my head around why i'm getting a scope error on handling relationship data/states here? http://pastebin.ubuntu.com/15190366/20:00
cory_fulazyPower: It must be that the scope value you're passing in to self.conversation() isn't valid20:01
cory_fuOH wait20:01
cory_fuYou're using RelationBase.get_remote()20:02
lazyPowerits on an interface that is scope: global on provides, scope: service on requires20:02
lazyPowershould i refactor those self.get_remotes() to conv = self.conversation() conv.get_remote('thing')?20:02
cory_fuYou're using self.get_remote() in the requires, which you can't use for anything but global scope20:02
lazyPowerinteresting nuance20:02
cory_fulazyPower: If you're scope is service, it means you can have multiple conversations (one per connected service) so you have to iterate over them20:03
cory_fu*your20:03
cory_fulazyPower: https://github.com/chuckbutler/interface-etcd/blob/master/requires.py is GLOBAL scope.  I assume you're changing it to support multiple services?20:04
lazyPowernope20:04
lazyPoweri clearly misunderstood scopes again :)20:04
lazyPowerand just reverted back to global from service20:04
cory_fulazyPower: Comment added to your PR on the etcd interface layer20:13
cory_fulazyPower: And another comment20:16
jamespagebeisner, hey - are you ok for a quick +1 on https://code.launchpad.net/~james-page/charms/trusty/neutron-gateway/tox8-test-refactor/+merge/28693320:17
jamespageall passing now...20:17
lazyPowercory_fu - added a response20:19
cory_fulazyPower: Responded to your response.  ;)20:19
lazyPowergah sniped me20:20
lazyPoweri mean i'm fine refactoring i'm not married to this, just exhausted after trying to get a working cluster after that refactor :|20:20
cory_fuThe important bit is that the interface layer shouldn't put restrictions on how the charm layer implements things.  Someone might want to create an alternate etcd charm that didn't use a peer relation, or got the port from somewhere other than config, or... something20:21
cory_fuAlso, you tend to end up with much cleaner code when you properly separate your concerns20:22
beisnerjamespage, definitely good with the tox8 changes.  passing tests are good.  i don't see anything offensive in the other bits. ^20:27
=== scuttle` is now known as scuttle|afk
stokachuhttps://www.irccloud.com/pastebin/81w6oO7Q/20:51
stokachuam i missing something with juju2 and lxd?20:51
rick_h_stokachu: what ubuntu release?20:58
stokachurick_h_: oh, on trusty20:58
stokachuguessing that wont work20:58
rick_h_stokachu: there's a few issues with lxd and beta1 and on the email through with the beta notice20:58
rick_h_stokachu: yes, trusty among them20:58
stokachuok20:58
rick_h_stokachu: also doesn't work with lxd beta3 that just came out due to an api break the team is working on20:59
stokachurick_h_: and also doesn't work with beta4 just fyi20:59
stokachu(just found out)20:59
rick_h_stokachu: yea, jam is working with tose folks to get it working with master21:00
rick_h_stokachu: juju has to play catch up with api changes happening in lxd between beta releases21:00
stokachurick_h_: ok cool, im anxious to mess with it21:00
rick_h_stokachu: +121:00
=== scuttle|afk is now known as scuttlemonkey
lazyPowerrick_h_ - btw days later response to  your statement of cloud credentials - foudn it in the rel notes https://lists.ubuntu.com/archives/juju/2016-February/006618.html22:42
lazyPowertvansteenburgh - got another minute?22:58
lazyPowertvansteenburgh - pushed ~ 20 minutes ago the working charmbox branch i've cooked up. LMK if theres anything blatantly missing in there you need for QA/CI purposes. bugs as required and i'll get it fixed up22:59
tvansteenburghlazyPower: about the charmbox pr?22:59
tvansteenburghlazyPower: roger, will look in a bit, maybe tomorrow am, in the middle of something atm22:59
lazyPowerbut this should get you on 2.0-beta1 faster than cat in a hat :)22:59
tvansteenburghcool22:59
lazyPowerunderstood. thanks mang22:59
tvansteenburghnp22:59
nagyzis there a way to tell juju-deployer to just use the "normal" (upstream?) charms?23:12
nagyzinstead of trying to find them locally?23:12
tvansteenburghnagyz: prefix them with 'cs:'23:15
muxhey - I'm new to this... I'm interested in the 'storage' charm, but I'm getting "ERROR cannot resolve charm URL "cs:trusty/storage": charm not found"23:15
muxI'm sure that's something really simple, like "the charm doesn't exist for this version", but like I said, I'm new23:17
muxwhere can I find details?23:17
nagyztvansteenburgh, ah, stupid me. :) what does cs stand for actually?23:17
tvansteenburghnagyz: charmstore23:17
tvansteenburghmux: looks like the 'storage' charm exists only for precise23:18
muxtvansteenburgh: where do I look to find that?23:18
tvansteenburghmux: there are other options though: https://jujucharms.com/q/storage23:18
nagyztvansteenburgh, thanks23:19
tvansteenburghnagyz: np23:19
tvansteenburghmux: i just went to jujucharms.com and searched for 'storage'23:19
nagyzone more question: when deploying something using juju-deployer and I have the machines listed in the yaml, is there really a need for +1 machine (for the bootstrapping node)?23:19
lazyPowernagyz - juju wont function without the model controller23:19
lazyPowerso yeah, that +1 is pretty mandatory23:20
lazyPoweryou dont need to define it in the yaml though, as that node is special to juju and its implied23:20
nagyzok so that's something that's always assumed and is outside of whatever I have described in my yaml23:20
nagyzright23:20
lazyPoweryep yep23:20
nagyzcheers23:20
lazyPowercory_fu - thanks for the interface pointers23:21
cory_funp23:21
lazyPowerlooks a lot cleaner after refactoring to your suggestion(s).23:21
lazyPowerso, when i get a chance, i really want to pair with you on realtion scopes, and get something put together for a reference23:21
lazyPowereven if its just a video of you re-telling me the same things you've told me five times that i've forgotten23:21
cory_fuYeah23:21
lazyPowerit'll keep me from bugging you regardless23:21
lazyPowerirrespectively23:22
lazyPowerirregardlessly23:22
lazyPowerkwmonroe'ly23:22
cory_fu9_923:22
magicaltroutokay random question23:44
magicaltroutif I wanted to add more useful mysql backup, like automysqlbackup or something that shunted semi-directly to s3, would it make sense to create a subordinate so as not to pollute the existing charm?23:45
magicaltroutI'm thinking about charms like mysql where people have their own preference of backup ideas23:45

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