=== lazypower_ is now known as lazyPower
=== urulama|__ is now known as urulama
kjackalmicvog: is that you Michael?08:49
kjackalGood morning juju world!08:49
admcleodjhobbs: hi - i ... can have a look at that but not really sure if you've got the right person09:12
magicaltroutfix it admcleod !09:13
admcleodmagicaltrout: *sends box of white chocolate blue cheese and anchovy balls*09:14
* magicaltrout hurls them over the fence09:14
admcleodhaha.. chocolate anchovies are a thing. thanks google.09:18
magicaltroutbad times09:18
kjackalhey admcleod, we setup a VM for micvog and we are ready to move on with juju09:28
micvogindeed thanks kjackal!09:29
kjackalmicvog: I would also like you to meet magicaltrout. magicaltrout is an awesome community member and he basicaly gives us a reason to work on Juju09:30
kjackalmicvog: magicaltrout is also London based (at least for now)09:31
micvoghey all !09:31
kjackalok, enough with the introductions09:31
kjackalthe first thing to do is setup a monitoring09:32
kjackalso micvog, to get the status of services/applications/machines you just do a "juju status"09:32
micvogso, should we bootstrap first ?09:33
kjackalahh yes you are absolutely right09:33
kjackal"juju bootstrap" it is then!09:34
admcleodkjackal: how bout you set up a shared tmux/screen session so we can watch? :}09:34
kjackaladmcleod: that sounds good, never done this but should be easy, right?09:35
admcleodkjackal: sure! computers!09:35
micvogso, bootstrap has installed some packages on remote host (which host?) and started the agent09:37
micvogI assume everything is local since environments.yaml is not configured09:37
kjackalyes, everything is local09:37
kjackaladmcleod: what is your public ssh key?09:37
admcleodkjackal is a fan of lxd, so its probably an lxd instance known as 'machine-0' which is the juju controller09:38
admcleodkjackal: https://launchpad.net/~admcleod/+sshkeys (top one)09:38
micvogso the .juju/environments/local.jenv was created with bootstrap and contains all info about the current state ?09:40
kjackalmicvog: yes you are correct09:41
kjackalbut the file with the credentials you should be conserned with is this .juju/environments.yaml09:41
kjackalmicvog: .juju/environments.yaml has placeholders for configuring all supported providers09:41
kjackalmicvog: can you do a "tmux attach-session -t shared"09:43
micvog ~$ Michael? Can you read this?09:44
micvogI think we're good09:44
kjackalmicvog: when you do a "juju deploy apache-spark " juju will forst provision a vm09:54
kjackalin this case juju will provision an lxc container09:54
micvogah ok09:54
kjackalfirst time takes a long time because an ubuntu image has to be downloaded09:55
kjackalafter that lxc container provisioning is very fast09:55
kjackalmicvog: after getting the container juju will "bootstrap" the vm/unit and will deploy apache spark on it09:56
kjackalmicvog: we are not installing the containers software "apt-get update & upgrade etc"09:56
kjackal"Fetching resources" means we are downloading apache-spark09:58
admcleodkjackal: micvog sorry guys afk for 10 min09:58
kjackalah we can see what is happening by looking at the logs09:59
kjackalconsole 3 shows the logs "juju debug-log"09:59
kjackalbtw micvog we are deploying https://jujucharms.com/apache-spark/1010:01
micvogis this the remote machine logs  or the juju "controller" logs10:01
kjackalmicvog: spark deployment finished10:02
kjackalwe have spark on our machine 110:02
kjackalit is in cluster mode but has only one worker10:02
kjackalshould we login to that cluster or should we add some extra nodes?10:03
kjackalmicvog: ^10:03
admcleodkjackal: do both :}10:03
kjackalack lets add some units10:03
micvoghave you started it in HA mode ?10:03
micvogusing quickstart ?10:04
kjackalfor HA we need zookeeper10:04
kjackalso here is what we will do10:04
kjackallets add 3 more units: juju add-unit -n 3 apache-spark10:04
micvoglet me try10:05
kjackalin the mean time lets also deploy and apache-zookeeper: "juju deploy apache-zookeeper"10:05
kjackalmicvog: go ahead and try the above10:06
micvoghow do you know the zookeeper and spark are compatible? or how do you enforce versions?10:06
kjackalmicvog: spark and zookeeper share an interface10:07
kjackalif we are to create versions that do not relate we will have to change the interface10:07
kjackalyou can see which interface is used in the web pages and readme files of each charm10:08
kjackallets look at the states10:08
Guest_94447Allah is doing10:08
Guest_94447sun is not doing allah is doing10:09
micvogfetching resources10:09
micvogthats to download spark on the 3 new nodes10:09
Guest_94447moon is not doing allah is doing10:10
Guest_94447stars are not doing allah is doing10:10
Odd_BlokeWhat's that command?10:10
kjackalmicvog: isn't it nice? 4 node spark cluster on your machine EXACTLY how it would look on AWS10:11
Guest_94447planets are not doing allah is doing10:11
Guest_94447galaxies are not doing Allah is doing10:12
kjackalmicvog: and you basicaly you didn;t do much "juju deploy, juju add-unit"10:12
Guest_94447oceans are not doing Allah is doing10:12
magicaltroutits like slow polite trolling10:12
admcleodmagicaltrout: i think i see where this is going10:12
magicaltrouti like a bit of morning poetry.....10:13
Guest_94447mountains are not doing Allah is doing10:13
admcleodmagicaltrout: at some point we're going to encounter the eclesiastical rift concerning whether god is doing white chocolate and blue cheese.10:13
magicaltroutGuest_94447: I don't believe in God, last time I checked a good muslim friend of mine told me I was going to burn in hell10:13
magicaltroutso we had a good chat about that10:13
magicaltroutbut I'm okay with it10:13
Guest_94447trees are not doing Allah is doing10:14
magicaltroutbloody trees10:14
kjackalmicvog: you see now we have a spark cluster with 4 nodes and 1 master10:14
micvogall ready10:14
kjackalmicvog: we also have a zookeeper10:14
kjackallets relate zookeepr and spark10:15
kjackaljuju add-relation apache-spark apache-zookeeper10:15
kjackaltry this ^ micvog10:15
kjackaland look at the state10:15
Guest_94447mom is not doing Allah is doing10:15
Guest_94447dad is not doing Allah is doing10:15
Guest_94447boss is not doing Allah is doing10:16
micvogwhere can I see they're now related apart from debug logs?10:16
Guest_94447job is not doing Allah is doing10:16
Guest_94447dollar is not doing Allah is doing10:16
Guest_94447degree is not doing Allah is doing10:16
micvogdamn he's posting more frequently now10:16
Guest_94447medicine is not doing Allah is doing10:17
magicaltroutit is making me chuckle10:17
Guest_94447customers are not doing Allah is doing10:17
Guest_94447you can not get a job without the permission of allah10:17
kjackalwho is the moderator here?10:17
* magicaltrout asks the permission of no one......10:17
micvogah by looking at juju status you can see the relation ?10:18
kjackalyes, the full status10:18
Guest_94447you can not get married without the permission of allah10:18
kjackalbut you see micvog that spark is now in HA mode10:18
kjackalcool eh?10:19
kjackalautomagicaly :)10:19
Guest_94447nobody can get angry at you without the permission of allah10:19
kjackallets ssh to one of these machines10:19
kjackalconsole 410:19
kjackalmicvog: you do not need to remember the ips or the strange strings each cloud provider will give you10:21
Guest_94447light is not doing Allah is doing10:21
Guest_94447fan is not doing Allah is doing10:21
micvogyes that's really handy10:21
Guest_94447businessess are not doing Allah is doing10:21
kjackalmicvog: you just do "juju ssh <unit>"10:21
magicaltroutmicvog fyi: /ignore Guest_9444710:21
Guest_94447america is not doing Allah is doing10:21
magicaltroutshould sort you out10:21
Guest_94447fire can not burn without the permission of allah10:22
kjackalmicvog: admcleod: I am being called for lunch10:22
kjackallets see bundles and micvog work when I get back ok?10:22
micvogcatch up later10:23
micvogI'll try to run smth10:23
Guest_94447knife can not cut without the permission of allah10:23
Guest_94447rulers are not doing Allah is doing10:23
Guest_94447governments are not doing Allah is doing10:23
Guest_94447sleep is not doing Allah is doing10:24
Guest_94447hunger is not doing Allah is doing10:24
Guest_94447food does not take away the hunger Allah takes away the hunger10:25
admcleodif someone is in the blufin building, or logged in over the vpn: /oper ${your_ud-ldap_uid} ${irc password without the leading NICK}10:26
Guest_94447water does not take away the thirst Allah takes away the thirst10:26
Guest_94447seeing is not doing Allah is doing10:26
Guest_94447hearing is not doing Allah is doing10:26
magicaltroutaww admcleod10:26
magicaltroutyou're so mean10:26
Guest_94447seasons are not doing Allah is doing10:27
Guest_94447weather is not doing Allah is doing10:27
Guest_94447humans are not doing Allah is doing10:27
Guest_94447animals are not doing Allah is doing10:27
Guest_94447the best amongst you are those who learn and teach quran10:28
Guest_94447one letter read from book of Allah amounts to one good deed and Allah multiplies one good deed ten times10:29
Guest_94447hearts get rusted as does iron with water to remove rust from heart recitation of Quran and rememberance of death10:30
Guest_94447heart is likened to a mirror10:30
Guest_94447when a person commits one sin a black dot sustains the heart10:31
Spadssounds healthy10:31
* Spads sustains his heart10:33
magicaltroutbloody black dots10:33
Guest_94447to accept Islam say that i bear witness that there is no deity worthy of worship except Allah and Muhammad peace be upon him is his slave and messenger10:33
admcleodlunch is not doing, All.. oh wait yes, im going to lunch10:34
magicaltroutallah made your lunch?10:34
magicaltroutI need to have a word10:34
magicaltroutmines still in the cupboard10:34
admcleodi dunno, i suspect i may have some pork based product, since spain. anyway, bbl10:35
magicaltroutallah will cast you down with increased risk of bowel cancer then admcleod10:36
magicaltroutso says......10:36
magicaltroutthe doctors10:36
kjackal /ignore Guest_9444710:40
Guest_94447read book http://www.fazaileamaal.com10:40
Guest_94447read book http://www.muntakhabahadith.com10:40
Guest_94447need spiritual teacher visit http://www.alhaadi.org.za10:41
Guest_94447allah created the sky without any pillars10:41
magicaltroutaww kjackal10:41
magicaltrouta) remove the space10:41
magicaltroutb) where's your sense of adventure, you should join up10:41
Guest_94447allah makes the sun rise from the east and Allah makes the sun set in the west10:41
Guest_94447allah gives life and Allah gives death10:42
kjackalmagicaltrout: my irc skills are non existent!10:42
Guest_94447all creation are useless,worthless,hopeless10:42
Guest_94447can not do10:42
magicaltroutso it seems kjackal10:42
Guest_94447can not benefit10:42
Guest_94447can not harm10:42
Guest_94447allah is the doer of each and everything10:42
magicaltroutthat said its better than some of the ibm'ers who like to address kwmonroe as kevin... or <kwmonre> with some weird copy paste thing10:43
magicaltrouttab completion IBMers! tab completetion!10:43
kjackalmagicaltrout: lol10:43
Guest_94447when Allah wants us to stand we stand10:43
Guest_94447when Allah wants us to sit we sit10:43
Guest_94447i am not doing Allah is doing10:43
Guest_94447you are not doing Allah is doing10:44
Guest_94447atom bomb is not doing Allah is doing10:44
Guest_94447rice is not doing Allah is doing10:44
Guest_94447all creation get together can not create one grain of rice10:44
Guest_94447all humans get together can not stop rain10:45
Guest_94447all humans get together can not make anybody hungry10:45
magicaltroutGuest_94447: can you impart your wisdom on those in ##saiku they all need your help10:46
Guest_94447all humans get together can not move sun one second up or down10:46
magicaltroutor ##pentaho10:46
magicaltroutthey also need som guidance10:46
Guest_94447we can not count the hair on our head10:46
Guest_94447we can not count the rain drops10:46
Guest_94447we can not count the particles of sand10:46
Guest_94447medicine has no power to cure10:47
Guest_94447two people take same medicine one passes away and one does not10:47
Guest_94447degree has no power to give job10:47
kjackalmicvog: admcleod: I am back10:47
Guest_94447many people have degrees but do not have jobs10:47
Guest_94447sustenance does not depend on effort10:48
Guest_94447one person is working very much but is earning very less10:48
micvogI uploaded the jar to one of the spark nodes, seems that the input files are missing - working on it10:48
Guest_94447other person is working very less but is earning very much10:48
magicaltroutwhere's he gone10:50
micvog"Found 57 outliers" - looks like we're onto something10:52
kjackalmicvog: nice!11:00
kjackalmicvog: give me some time to go over the blog and the program11:01
micvogfyi channel, we're trying to deploy this using juju - https://micvog.com/2016/05/21/using-spark-for-anomaly-fraud-detection/11:04
=== micvog is now known as micvog__
=== micvog_ is now known as micvog
kjackalmicvog: ok I think I got the big picture11:16
kjackalI am thinking...11:17
kjackalwe have a good start11:17
kjackaland a relatively good idea what we want to get out of this...11:18
kjackalSo... we will need to complete the stack of charms11:18
kjackalWe will need a way to feed realtime data to a kafka queue as shown in the chart11:19
kjackalthen have spark consume the queue and persist everything (eg to HDFS)11:20
kjackaland then visualise the results. it would be great to do that through zeppelin11:21
micvogsorry dc11:22
micvoglast message was that the list of charms need to be completed11:22
magicaltrout 12:12  kjackal| We will need a way to feed realtime data to a kafka queue as shown in the chart                              │ blr11:25
magicaltrout 12:13  kjackal| then have spark consume the queue and persist everything (eg to HDFS)                                        │ blr11:25
magicaltrout 12:14  kjackal| and then visualise the results. it would be great to do that through zeppelin11:26
kjackalAny ideas magicaltrout?11:26
magicaltroutI wasn't paying a great deal of attention whilst mulling over life with the quran11:29
magicaltroutwhat part are you mulling over kjackal ?11:29
magicaltrouttry that again11:30
magicaltroutI wasn't paying a great deal of attention whilst mulling over life with the quran11:30
magicaltroutwhat part are you mulling over kjackal ?11:30
kjackal_magicaltrout: do you recommend any tool for awesome visualisations on (streaming) data that can interface with spark?11:32
magicaltroutzeppelin seems like a sound choice on the Juju platform11:36
magicaltroutone consideration I would ponder is how do you present it to the user.11:36
magicaltroutIf you're streaming data, it would be good to hook up websockets or long polling and have it update the zeppelin charts on the fly without a user refreshing the page etc11:36
magicaltrouthttps://gist.github.com/granturing/a09aed4a302a7367be92 this chap did some streaming spark map example11:38
magicaltroutor just write some websocketed thin client with your own choice of vis toolkit11:38
magicaltroutand use streaming spark or write it back to kafka and have kafka write it to the sockets11:39
kjackal_sounds cool! Thanks magicaltrout11:40
micvogthis looks interesting reg visualization - https://www.youtube.com/watch?v=BhCXL8AHiB0&index=9&list=PL-x35fyliRwiy50Ud2ltPx8_yA4H34ppJ11:40
micvogif you can send data through websockets then D3.js is a very good option (and realtie)11:41
magicaltroutyeah its very lightweight as well11:41
magicaltroutso your browser shouldn't drown11:41
magicaltroutwhich used to be an issue with streaming data.. when to flush it11:41
=== dpm is now known as dpm-bbiab
rcjcoming late to the 'multiseries' charm game.  Had a question on updating an existing charm...  Was told to push to somewhere new and launchpad.net/ubuntu-repository-cache was suggested.  Is it really necessary to create an entire LP project for the charm?  What are the requirements?13:32
rcj ^ re: https://code.launchpad.net/~rcj/charms/trusty/ubuntu-repository-cache/multiseries/+merge/29947213:33
rick_h_rcj: so you can push using the charm push command from anywhere. I think the key thing is that you're pushing a charm that works on multiple series to a series-i-fied url which seems a bit untrue13:34
rcjrick_h_, yeah I wanted a diff that showed the 4 line change against the current code so no one is looking at this as entirely new.13:35
rick_h_rcj: understand. The new review queue will show the diff of actual files which will help13:35
rick_h_rcj: not sure where that's at atm, but will check on it13:35
rick_h_rcj: but appreciate where you're coming from there13:35
rick_h_rcj: maybe recommend making the change to the new path first, getting that "just migrate existing code" and then do the patch for the 4 lines on top of that?13:36
rcjrick_h_, otherwise, it has already been pushed to https://jujucharms.com/u/rcj/ubuntu-repository-cache and I need to know how to get it queued for review.13:36
rcjrick_h_, either way, how do I get this on the review queue (or view the queue)?  http://manage.jujucharms.com/tools/review-queue linked from https://jujucharms.com/community/charmers just gives me 'Internal Server Error'13:42
rick_h_rcj: ah, bad link: http://review.juju.solutions/ is the current queue.13:44
rick_h_rcj: from the docs "If something needs review, subscribe ~charmers."13:44
rick_h_rcj: looking at https://jujucharms.com/docs/master/reference-reviewers13:44
rcjthx.  Also for bad links from http://review.juju.solutions/ it's suggests reading "Charm Review Process" @ https://juju.ubuntu.com/docs/charm-review-process.html -> DNS_PROBE_FINISHED_NXDOMAIN13:46
rick_h_rcj: yes, filing a bug on that bad link now13:46
rick_h_heh well filing on jujucharms.com/community, guess there's another one from the actual review site13:47
rick_h_tvansteenburgh: what's the latest on new queue timing?13:47
tvansteenburghrick_h_: marcoceppi ^13:49
jhobbsadmcleod: yeah ok probably wrong person haha thanks14:07
=== dpm-bbiab is now known as dpm
admcleodjhobbs: no worries :)14:33
lutostagstub: hoping you could push the built/latest published charm to the branch "built" like you had been previously for git.launchpad.net/postgresql-charm if possible14:53
xiletusing 2.0-beta7-xenial-amd64, any idea where to start looking for why after a full system reboot 'juju status' just hangs?15:05
cory_fumarcoceppi, petevg, and I are doing a live session discussing testing in charms and layers: https://plus.google.com/u/1/events/col2a1aertqj329rjb8tgg314u015:06
cory_fuWe're currently discussing this PR: https://github.com/juju-solutions/layer-apache-bigtop-base/pull/2815:06
cory_fuAnd this PR: https://github.com/juju-solutions/jujubigdata/pull/5915:06
cory_fuxilet: We're currently up to beta11, so one option might be to upgrade to that.  What provider are you using, though?  lxd, aws, etc?15:08
iceyjrwren: unfortunately, ceph_ips = [x['public-address'] for x in self.d.sentry['ceph-mon']] doesn't work: http://pastebin.ubuntu.com/18792425/15:09
xiletIt had deployed an openstack fine, but after the reboot it has been several hours and still nothing,  strace shows something is timing out but isn't giving me any really good indicators. I see quite a few of the services have launched according to ps.15:10
xiletlet me try upgrading juju15:13
magicaltrout... said no one on a Friday15:13
cory_fuxilet: I think that after a reboot the lxc containers might not be set to start back up automatically.  I'm not familiar enough with lxd to know the command to check for the running state of the containers15:19
xiletcory_fu: it has in the past, I ran into a bug  with adding a second nova-compute unit, so Iw as bouncing the whole system to see what might changed, I had rebooted several times prior,  curious if there is some kind of central juju log that I could check15:24
cory_fuxilet: There is a log on the bootstrap container, but I'm not sure if it would have the info you're looking for or not15:25
xiletso other question, I had (I think, I am new to this) upgraded to juju-2 using the xenial-proposed branch,  what is the official method these days for updating it?15:27
stublutostag: It is already pushed15:27
lutostagstub: perfect, thanks15:28
mbruzekmarcoceppi: Do you know where the tab completion code is for juju?  Re: https://bugs.launchpad.net/juju-core/+bug/160025716:05
mupBug #1600257: The tab completion on juju yeilds KeyError: 'services' <juju-core:New> <https://launchpad.net/bugs/1600257>16:05
marcoceppimup: it's probably in the packaging branch16:06
mupmarcoceppi: I apologize, but I'm pretty strict about only responding to known commands.16:06
mbruzekmarcoceppi: I searched for python in juju/juju could not find it16:06
marcoceppimbruzek: it's probably in the packaging branch16:06
marcoceppimup: when did you get so saucy16:06
mupmarcoceppi: In-com-pre-hen-si-ble-ness.16:06
marcoceppiomg. who taught you to talk back.16:06
mbruzekmup you steal all my best lines16:06
mbruzek*the best lines16:07
marcoceppiwow, now mup is all quiet16:07
* marcoceppi feels accomplished.16:07
mbruzekYeah mup must know who is in charge16:07
lazyPowermarcoceppi - its been doing that for quite some time now16:11
lazyPowermup what is the air speed velocity of an unladen swallow?16:11
lazyPowerwelp, ya broke it16:11
babbageclunkmbruzek: it's in juju/juju/etc/bash_completion.d (in case you haven't found it)16:33
mbruzekYes I just found it.16:33
babbageclunkmup: anything exciting planned for the weekend16:33
mupbabbageclunk: Roses are red, violets are blue, and I don't understand what you just said.16:33
babbageclunkthat's fun16:33
mbruzekThe error looked like a python traceback, I didn't expect the file to be bash, with embedded python.16:33
mbruzekthanks babbageclunk16:33
babbageclunkmbruzek: yeah, it's a bit of an obscure location16:34
babbageclunkmbruzek: Are you seeing that error, or do you just need it for something else?16:36
mbruzekbabbageclunk: I am seeing the error, and looking to try to fix it16:36
babbageclunkmbruzek: It's already been fixed, I think - you can reinstall the completions by running `make install-etc`.16:37
babbageclunkmbruzek: (I went through this same series of steps a while back.)16:37
mbruzekbabbageclunk: OK I see that the bash completion file does not contain "services" in juju (master) but I see "services" on the completion file on my system in /etc/bash_completion.d/juju2 on my system16:39
babbageclunkmbruzek: Right - that is what I found when I had this error.16:40
mbruzekbabbageclunk: OK thanks, just trying to help16:41
petevgmarcoceppi: not production ready, but here's an invocation that fixes the import in the docker-layer unit tests we discussed in the hangout: http://paste.ubuntu.com/18798385/16:44
petevg(Basically, you make sure that the "charms" module exists, and then you patch a mock object into it.)16:45
neiljerramlazyPower, mbruzek - Hi there; just wondering what the plan/schedule is for publishing a new etcd charm with all recent fixes.  Currently I'm using cs:~lazypower/etcd-19 - which is fine - but I guess there should be a new publication to cs:~containers ?16:48
lazyPowerneiljerram - cs:~containers/etcd-4 exists16:48
lazyPowerits currently development channel, i want to have an ironclad migration from etcd-2 to etcd-4+ before it goes stable16:48
neiljerramlazyPower, Oh, that's easy then! :-)16:49
lazyPowerso its fine to do new deploymets, but not recommended for upgrades yet16:49
neiljerramlazyPower, perfect.  Thanks!16:49
petevgmarcoceppi: or you can do this, which is much simpler :-)  http://paste.ubuntu.com/18798975/16:51
petevg(I always forget about the create=True flag)16:52
lazyPower#TIL with mock.patch('charms.layer', create=True):16:58
lazyPowerpetevg - i'm going to have to try and osmose some of that testing knowledge you have been dropping the last couple weeks16:58
petevglazyPower: feel free to ping me with questions, or pull me into a hangout. I'm always happy to spread the testing love :-)17:00
lazyPowerdo we have an idea on what the upper boundary limits are for filesizes  and resources?17:26
rick_h_lazyPower: I think it's in the couple hunderd of MB atm. I was just talking with the team on upping that while we work on some quota-like safety measures17:28
rick_h_lazyPower: but I think it's a timing things vs a size thing so it's how many bytes you can push in the timeout window17:28
lazyPowerThats what i was looking for, ta rick_h_17:29
* lazyPower decides its a bad idea ot publish over a mifi with 1 bar of service, and a 2gb payload.17:29
jrwrenlazyPower: yup, we can't currently process 2GB :[ sorry.17:30
jrwrenlazyPower: Incidently, what is the 2GB payload?17:30
lazyPowerjrwren - i was actually being silly, i believe the payload is quite large though. This was in response to a versioning issue/question from Nexenta.17:30
lazyPowerjrwren - its packing in chef and some other build time deps like ruby libs17:30
jrwrenlazyPower: ah. cool.17:31
lazyPowerwhat theyd o today is they host a bintray style repo and fetch over the wire. They haven't versioned the bins, just kept a running-tip publication.17:31
lazyPowersooo, either they need to version, split up payloads and do multiple payloads, or should be actively participating in planning sessions for the feature as they would be a big consumer of it.17:31
lazyPowerat least thats my take on it17:32
jrwrenlazyPower: to expand further, my current working theory is that the timeout happens because of server side processing time. The upload completes. The charmstore processes the charm and this can take some time. Currently we have a 50s timeout(I'm about to up it) so if the server takes longer than 50s to process a large charm it doesn't send a response in that 50s and it times out.17:32
lazyPowerah yeah, i can see that being problematic17:32
jrwrenlazyPower: We've also, AFAIK, not observed the limit on resources. Its likely we don't process resources as heavily and so they would be done more quickly.17:33
jrwrenlazyPower: As a general rule, I'd say many resources for each item is better than bundling things into a single resource.17:33
lazyPowerjrwren - i've only used 2 myself, and they were small golang bin attachments.  I as well would +1 multiple smaller deliveries rather than one large payload.17:35
marcoceppipetevg: nice, thanks!17:35
lazyPowerit seems like the bigger they get, the more trouble they cause17:35
jrwrenlazyPower: i wonder if we can document this guidance?17:36
lazyPowersure can17:38
jrwrenlazyPower: thanks.17:41
lazyPowerjrwren - ping me if you submit and i'll be happy to review17:41
cory_fugeetha: You were asking about http://pastebin.ubuntu.com/18774917/ and saying that you were not seeing that handler being called even after you changed one of the config values.  That snippet looks fine to me.  Is there any chance you can provide me with some of the juju logs from that unit?17:47
cory_fuOr perhaps someone else here can spot something in that code that I'm missing.17:48
lazyPowercory_fu is @when_not_all an alias for when_not?17:48
lazyPoweror is when_not more like when_not_any17:49
geethaSame handler function being called in juju 1.25 when I change config value17:49
geethausing 'juju set' command17:50
cory_fulazyPower: It is not the same as when_none (there is no when_not_any).  when_not == when_none, when_not_all is its own thing (i.e., trigger if one of the given states is not set, vs when_none only triggers if not a single one state is set)17:52
cory_fugeetha: Hrm.  So it's specific to Juju 2.0?  That's very strange17:52
rcjstub, Why do I need to create a new home for ubuntu-repository-cache prior to promulgation?  I don't want to create a new top-level project in LP and my understanding would be that once promulgated it will end up in charms/ubuntu-repository-cache, right?  If it ends up there I don't want a second place to confuse people.17:54
rcj ^ re: https://bugs.launchpad.net/charms/+bug/160024317:55
mupBug #1600243: review ubuntu-repository-cache charm now with multiseries support <Juju Charms Collection:New> <https://launchpad.net/bugs/1600243>17:55
stubrcj: When it is promulgated it ends up in cs:ubuntu-repository-cache (jujucharms.com/ubuntu-repository-cache).17:58
=== scuttle|afk is now known as scuttlemonkey
stubrcj: Where do you want the main branch to live? Currently it is in bzr at lp:~charmers/charms/trusty/ubuntu-repository-cache/trunk, with the alias lp:charms/trusty/ubuntu-repository-cache.18:01
stubrcj: I can merge it in there, but it seems silly having a multi-series charm in a trusty specific namespace. And it still needs to be promulgated by the ecosystem team.18:02
stubrcj: It also means you still need ~charmers to land stuff in the future, since it is owned by that team.18:03
geethacory_fu: you can see the log when I change the config option http://pastebin.ubuntu.com/18805681/18:12
cory_fugeetha: Is that all of the log?  I don't see the bit where it actually calls reactive_main to see how it's evaluating the tests.  Also, it might be helpful to add the line `charms.reactive -y get_states` to the top of your handler file to see what the states are when it gets called18:18
rcjstub, I would like charmers to assist with landing changes in the future.  I would hope that it could end up @ lp:charms/ubuntu-repository-cache18:33
rcjAnd my understanding is that this would replace cs:trusty/ubuntu-repository-cache cleanly.  My goal here is a blessed charm with a single entry in the charm store that makes it 100% clear that it is supported with no confusion as to which charm to pick.18:35
stubrcj: Ok. I think you need ecosystem for that branch URL or alias - I'm not sure how to do it. And they can help with the 'charm push', 'charm publish' and promulgation to get that branch promulgated into the charm store.18:44
geethacory_fu: I could not able to paste full log, It has more lines. I have added `charms.reactive -y get_states` and again ran the command to change the config option . It's failing with non zero exit code.http://pastebin.ubuntu.com/18808692/18:45
rcjstub, who to ask?18:51
stubmarcoceppi: Are you or your team available to get rcj promulgated? lp:~rcj/charms/trusty/ubuntu-repository-cache/multiseries / https://bugs.launchpad.net/charms/+bug/160024318:52
mupBug #1600243: review ubuntu-repository-cache charm now with multiseries support <Juju Charms Collection:New> <https://launchpad.net/bugs/1600243>18:52
cory_fugeetha: From that `eval '{config.changed:'` line, it looks like you might have an errant { in your code.  I didn't see it in the pastebin from earlier, so I'd need to see more of the reactive handler file to pinpoint it19:01
marcoceppistub rcj I think we can help19:16
marcoceppircj: will you be the sole person maintaining this charm?19:16
cory_fugeetha: Very strange.  That all looks fine to me19:19
geethacory_fu: Even I have tested in juju 1.25 same code and without resources and terms. It worked fine19:22
cory_fugeetha: I can't imagine what 2.0 would be doing differently that would cause that.19:23
cory_fugeetha: Which beta version of 2.0 are you using?  beta11?19:27
geethacory_fu: No, it's beta719:28
lazyPowermagicaltrout - https://kognitif.bandcamp.com/track/walking-on-sunshine  may the jazz flute compel you.19:29
rcjmarcoceppi, it will be the cpc team <cpc@canonical.com>19:30
rcjwhich includes me19:31
cory_fugeetha: Unfortunately, my best recommendation at this point is to try upgrading to a newer beta.  I can't find anything that points to what's causing it19:38
cory_fugeetha: Is this charm available in Launchpad, GitHub, or jujucharms.com?  I can try to replicate on my side and do some more digging19:39
magicaltroutthanks lazyPower I'll check it out!19:40
geethaThanks cory_fu, I will try to upgrade to newer beta version. This charm is available in jujucharms.com: https://jujucharms.com/u/ibmcharmers/ibm-was-base/trusty/15. But still I'm working on this charm to fix other issues also.19:46
=== scuttlemonkey is now known as scuttle|afk
magicaltroutthats some pretty far out jazz flute lazyPower20:31
lazyPoweri'm sayin20:32
magicaltroutreminds me of woodstock type stuff in some ways20:32
lazyPowerthis entire album was made from samples20:32
lazyPoweri am impress20:32
magicaltroutalso very jazz fusion weather report esque https://www.youtube.com/watch?v=pqashW66D7o20:33
xiletI just upgraded to 2.0-beta11-xenial-amd64, now when I bootstrap it hangs at Running apt-get update,  the instance is up and the apt-get process finishes inside, but nothing returns back20:38
magicaltrouttold you not to upgrade on a friday ;) its against the rules20:39
xiletHah indeed you did, well the entire system was hosed due to an lxc issue, I couldn't even get lxc to return commands, so I figured I would restart from scratch.20:40
lazyPowergroovy, dig this too magicaltrout20:47
magicaltroutyeah we used to play a bunch of weather report and similar tracks in my old jazz band lazyPower20:49
magicaltroutalways preferred them over the traditional stuff20:49

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