| === junaidali is now known as details | ||
| === details is now known as help | ||
| === help is now known as Guest51665 | ||
| devops | /query junaidali | 01:45 |
|---|---|---|
| === test is now known as Guest70724 | ||
| stokachu | guessing there isa problem with the charmstore? | 02:19 |
| stokachu | http://paste.ubuntu.com/23357102/ | 02:19 |
| === scuttlemonkey is now known as scuttle|afk | ||
| bdx | https://github.com/juju/charmstore/issues/686 | 04:23 |
| kjackal_ | Good morning Juju world | 06:21 |
| cargill | when following the getting started document from docs, are you supposed to end up with ~/.juju/environments.yaml configured? 'charm test' is confused that it doesn't exist | 09:16 |
| cargill | yet juju (2.0) is fine with everything | 09:17 |
| magicaltrout | cargill: where did you read that? | 09:18 |
| magicaltrout | thats 1.2.x stuff | 09:19 |
| magicaltrout | a) the location these days is ~/.local/share/juju | 09:19 |
| magicaltrout | b) environments.yaml doesn't exist any more | 09:19 |
| cargill | $ charm test | 09:20 |
| cargill | juju-test INFO : Starting test run on None using Juju 2.0.0 | 09:20 |
| cargill | juju-test CRITICAL: /home/oku/.juju/environments.yaml file does not exist | 09:20 |
| cargill | this is what I get | 09:20 |
| * magicaltrout defers to kjackal_ | 09:21 | |
| magicaltrout | it may not have been revamped for 2.x I dunno | 09:21 |
| kjackal_ | hello! | 09:21 |
| cargill | I'll check for updates just in case | 09:21 |
| cargill | kjackal_: hi! | 09:21 |
| kjackal_ | hi cargill, magicaltrout | 09:22 |
| kjackal_ | hm... charm test I thought this is not anailable on the recent juju+charm tools | 09:23 |
| magicaltrout | hammer and nailable? | 09:23 |
| kjackal_ | cargill: where did you read about "charm test" | 09:24 |
| cargill | oh, it's meant to be bundletester now? | 09:25 |
| cargill | just tried it, thinking charm build, charm proof, charm test would be the sequence :) | 09:25 |
| kjackal_ | https://jujucharms.com/docs/2.0/tools-charm-tools look at this it is here!! | 09:25 |
| magicaltrout | seems to suggest charm test is indeed valid terminology | 09:26 |
| cargill | so do I use charm test or bundletester? and is bundletester in the repos/ppa? | 09:30 |
| kjackal_ | cargill: I would suggest you use bundletester, at least this is what i am using here | 09:32 |
| kjackal_ | bundle tester should be installable by pip | 09:32 |
| kjackal_ | sudo pip install bundletester | 09:32 |
| cargill | hmm, pip means you don't get updates unless you remember to ask for them... | 09:33 |
| cargill | (the joys of system package managers :)) | 09:33 |
| neiljerram | Morning all! | 10:16 |
| neiljerram | In my tests, I sometimes see 'juju ssh ...' failing with "Permission denied (publickey)." for just one or two units in the deployment; it's fine for all the other machines. | 10:17 |
| neiljerram | There are google hits for this ("juju ssh permission denied (publickey)"), but they're quite old. | 10:18 |
| neiljerram | So it looks like Juju is failing to copy my public key onto those machines... | 10:19 |
| neiljerram | Does anyone else see this, and is there a workaround? | 10:20 |
| junaidali | neiljerram, are you able to manually ssh to those nodes? | 10:23 |
| junaidali | ssh ubuntu@<node-ip> | 10:23 |
| junaidali | from maas | 10:23 |
| neiljerram | junaidali, I'm not sure, because this happens in the middle of an automated test. I can try modifying the script to add that in, if you think it will help. | 10:24 |
| neiljerram | junaidali, I'm trying that now, let's see... | 10:45 |
| neiljerram | junaidali, Unfortunately I am still seeing 'Permission denied (publickey)' with one of the units, even when I do 'ssh ubuntu@<ip>' first. I get 'Permission denied (publickey)' for 'ssh ubuntu@<ip> ls', and then also for 'juju ssh <unit> ls'. | 11:02 |
| SDB_Stefano | hi Hjackal | 12:14 |
| icey | is it possible for a juju action to take a large JSON blob as a value? | 12:58 |
| magicaltrout | icey: its a pain in the balls | 13:00 |
| magicaltrout | you can, but you need to escape it all | 13:00 |
| icey | wtf -_- | 13:01 |
| icey | escape it even though it's wrapped in single quotes ? | 13:02 |
| magicaltrout | yeah, I found for a charm where the save format is json, I couldn't get it to ingest the stuff unless it was escaped | 13:02 |
| magicaltrout | this was a while ago, but I doubt its changed | 13:02 |
| icey | magicaltrout: yeah, giving errors like: ERROR json: unsupported type: map[interface {}]interface {} | 13:02 |
| magicaltrout | yeah it tries to parse it..... | 13:02 |
| * icey cries | 13:04 | |
| SDB_Stefano | Hi, I'm creating a charm for our product ScaleDB UDE, I have some questions, could I propose them ? | 13:04 |
| magicaltrout | go wild SDB_Stefano | 13:05 |
| SDB_Stefano | ok, I have to create a plain/easy charm that the only thing we need is to execute a install/hook to install our software, so : | 13:06 |
| SDB_Stefano | I started using https://github.com/juju-solutions/layer-basic as template | 13:06 |
| SDB_Stefano | about the layer.yaml : | 13:07 |
| SDB_Stefano | should it contains ' includes: ['layer:basic']' ? or it's not mandatory, I don't need it | 13:07 |
| icey | SDB_Stefano: it should | 13:08 |
| magicaltrout | when you said "use layer basic as a template" you mean you git cloned it? | 13:08 |
| magicaltrout | or did `charm build` ? | 13:08 |
| SDB_Stefano | yes git clone | 13:08 |
| magicaltrout | wrong answer | 13:08 |
| SDB_Stefano | I : | 13:08 |
| marcoceppi | SDB_Stefano: the better way to start, is to run `charm create` | 13:08 |
| magicaltrout | sorry yeah create | 13:08 |
| marcoceppi | SDB_Stefano: charms incur a compilation process, which means things like layer basic will be included in the final, compiled product | 13:09 |
| magicaltrout | https://jujucharms.com/docs/stable/authors-charm-writing <- SDB_Stefano | 13:09 |
| SDB_Stefano | the current charm is working, I think that I have only to refine, it's need that I restart using ' charm create' | 13:12 |
| marcoceppi | SDB_Stefano: I'd behappy to help shape your charm into a layer if it's available somewhere | 13:13 |
| SDB_Stefano | ok, great , I'm going to deploy the current version in github, just a moment | 13:14 |
| magicaltrout | marcoceppi is like a charm rewriting ninja | 13:14 |
| kjackal_ | Hello SDB_Stefano, what's up? | 13:14 |
| SDB_Stefano | Hi kjackal, I have some questione about the charm | 13:15 |
| SDB_Stefano | could we have a short skype call, maybe it could be easier ? | 13:16 |
| kjackal_ | Sure | 13:16 |
| junaidali | neiljerram, looks like your node is not up | 13:16 |
| cargill | hmm, great, using bundletester complains about python linting issues in the pgsql layer and that makes it fail | 13:17 |
| cargill | not that I have control over that layer in the first place... | 13:18 |
| neiljerram | junaidali, I'm sure that the node was up; juju status had given a successful report for it. | 13:18 |
| neiljerram | junaidali, you can see this in the transcript of the bug that I just reported, at https://bugs.launchpad.net/juju-core/+bug/1635622 | 13:20 |
| mup | Bug #1635622: 'juju ssh <unit> ...' fails with Permission denied (publickey), for only one or two machines in a deployment <juju-core:New> <https://launchpad.net/bugs/1635622> | 13:20 |
| kjackal_ | cargill: if you run bundletester with -F it will run all the tests and not stop on the first error | 13:33 |
| cargill | how is that handled when I submit the charm for review? | 13:39 |
| kjackal_ | cargill: we usualy ping the author of the other layer to fix the problems, but that might block you as well since we might not be able to verify the correctness of your charm either. | 13:42 |
| cargill | if it's just python linting issues? | 13:43 |
| kjackal_ | cargill: linting errors will most likely not block you | 13:43 |
| BlackDex | If i want to set config-flags for nova-compute i can't set nfs_mount_options=nfsvers=3 | 14:05 |
| BlackDex | because there are 2 = signs | 14:05 |
| marcoceppi | BlackDex: quote the value | 14:12 |
| marcoceppi | BlackDex: furthermore, I don't see nfs_mount_options as config on the charm https://jujucharms.com/nova-compute/ | 14:13 |
| BlackDex | marcoceppi: tried that not working | 14:17 |
| marcoceppi | BlackDex: lwhat's the error you're getting, exactly? | 14:17 |
| BlackDex | its juju set nova-compute config-flags='nfs_mount_options="nfsvers=3"' | 14:17 |
| BlackDex | that is what needs to be done | 14:18 |
| BlackDex | that the config-flag value is not correct | 14:18 |
| marcoceppi | BlackDex: taking a look | 14:35 |
| marcoceppi | BlackDex: you may want to ask in #openstack-charms this might be a bug with how commands are being parsed | 14:36 |
| cargill | how do I propose a charm to the store again? | 15:01 |
| BlackDex | marcoceppi: thx i will go there :) | 15:02 |
| marcoceppi | cargill: http://review.jujucharms.com/ | 15:02 |
| cargill | marcoceppi: thanks | 15:49 |
| === freyes__ is now known as freyes | ||
| cory_fu | mbruzek: Your wish is my command: https://github.com/juju-solutions/charms.reactive/pull/86 | 16:25 |
| cory_fu | (And thank you for keeping me honest) | 16:25 |
| mbruzek | Sorry to be so pedantic | 16:25 |
| mbruzek | cory_fu: merged | 16:27 |
| mbruzek | Cynerva: ping | 16:35 |
| Cynerva | mbruzek: pong | 16:35 |
| mbruzek | Cynerva: or ryebot: I am getting this error with a defined action ERROR invalid params schema for action schema restore-snapshot: bool is not a valid type | 16:35 |
| mbruzek | type: bool | 16:35 |
| Cynerva | mbruzek: i believe the correct type is boolean | 16:36 |
| ryebot | mbruzek: hmm I think the type needs to be 'boolean' | 16:36 |
| mbruzek | I knew you guys would know. | 16:36 |
| mbruzek | Thank you | 16:36 |
| Cynerva | mbruzek: cool, glad to help! | 16:37 |
| mbruzek | I should have tried that before asking but I gave up after I searched the documentation and found no boolean options | 16:37 |
| Cynerva | mbruzek: I've found it useful to refer to json-schema docs for action stuff. The primitive types (including boolean) are here for example: http://json-schema.org/latest/json-schema-core.html#anchor8 | 16:39 |
| mbruzek | Thanks Cynerva | 16:40 |
| === ubuntu is now known as Guest33397 | ||
| suchvenu | Hi | 17:38 |
| suchvenu | I am trying to deploy Openstack charm from the store : https://jujucharms.com/openstack-base/ | 17:38 |
| suchvenu | I have a KVM and MAAS is configured in one guest from that box. Juju is installed and could deploy a sample charm as well. | 17:38 |
| suchvenu | However when i deploy Openstack charm from store I get "Failed deployment" as the status | 17:38 |
| suchvenu | The MAAS log ahows as : | 17:38 |
| suchvenu | Oct 20 15:38:32 maascontroller maas.node: [INFO] vm3: Status transition from DEPLOYING to FAILED_DEPLOYMENT Oct 20 15:38:32 maascontroller maas.node: [ERROR] vm3: Marking node failed: Node operation 'Deploying' timed out after 0:40:00. Oct 20 15:38:37 maascontroller maas.node: [INFO] vm4: Status transition from DEPLOYING to FAILED_DEPLOYMENT | 17:38 |
| suchvenu | Any idea on this error ? Will increasing the timeout fix the issue ? | 17:39 |
| gaurangt | Hi, for the storage support in MAAS for juju version 1.25.. the documentation says - This storage provider is static-only; it is currently only possible to deploy charms requiring block storage to a new machine in MAAS, and not to an existing machine. | 17:47 |
| gaurangt | What exactly new machines mean here? | 17:49 |
| deanman | Trying to learn Juju 2 by setting up a local cloud (lxd). Controller gets created just fine and i can see it with `lxc list` but when i try to deploy a charm it just hangs. Any hints? | 17:57 |
| rick_h_ | deanman: check out juju status --format=yaml and see if there's anything fishy | 17:58 |
| rick_h_ | deanman: or juju debug-log | 17:58 |
| deanman | ok it seems its downloading a trusty image to deploy the charm by default where i would have expected that it would use a xenial image, same as with controller. | 18:00 |
| deanman | thanks rick_h_ | 18:00 |
| rick_h_ | deanman: hmm, something in the config set trusty as the default? | 18:01 |
| deanman | just following getting started guide from stable docs | 18:01 |
| rick_h_ | deanman: you on xenial? | 18:01 |
| deanman | xenial 64 VM on MacOS host | 18:02 |
| deanman | the deployed charm was mysql | 18:02 |
| rick_h_ | deanman: oic, ok so that's probably because mysql charm isn't xenial ready it looks like | 18:03 |
| * rick_h_ notes it's less crazy then | 18:03 | |
| deanman | ok rick_h_ fair enough, so in general having a xenial64 VM and then using local as a cloud controller to deploy consequent charms in LXD is OK for dev. workflow? | 18:04 |
| rick_h_ | yep | 18:04 |
| skay | just when I want to create a charm, github is down. charm-create fails due to trying to get templates from there | 18:08 |
| skay | :( | 18:08 |
| skay | is that mirrored anywhere else? | 18:09 |
| deanman | rick_h_: is juju gui deployed with controller by default? | 18:10 |
| kwmonroe | cory_fu: this world wide dns problem just got real! "unit-namenode-0: 18:05:54 INFO unit.namenode/0.install Error: Could not connect to https://forgeapi.puppetlabs.com". i didn't care before, but now i'm inclined to write a letter to congress. | 18:13 |
| kwmonroe | (and hope al gore sees it) | 18:13 |
| rick_h_ | deanman: yes | 18:15 |
| rick_h_ | deanman: just run 'juju gui' | 18:15 |
| cory_fu | kwmonroe: lol | 18:17 |
| cory_fu | But yeah, it's a pretty significant attack. I think python.org is being affected as well | 18:18 |
| kwmonroe | good news is that the ddos people failed to remember never to release on a friday. we can all just go home. if they really wanted havoc, they would have waiting till monday. | 18:20 |
| suchvenu | Can anyone please tell me on how to get charm-tools 2.1.4 ? Mine is in 2.1.3 | 18:34 |
| suchvenu | I already did sudo apt update && sudo apt upgrade | 18:34 |
| brad_nokia | anyone can help with charm storage? using a cinder volume and the charm hangs in a pending mode awaiting attachment | 19:05 |
| mbruzek | suchvenu: you can sudo pip install charm-tools | 20:03 |
| mbruzek | to get the latest verison | 20:03 |
| cory_fu | suchvenu: Best I can tell, from https://github.com/juju/charm-tools/blob/master/setup.py#L11, a version newer than 2.1.2 doesn't exist | 20:23 |
| cory_fu | pypi would disagree with me https://pypi.python.org/pypi/charm-tools | 20:25 |
| lutostag | question... so with reactive how many reactive functions can trigger in a file for a single hook? | 20:52 |
| cory_fu | lutostag: As many as match | 20:52 |
| cory_fu | As long as their preconditions match, they will be triggered. There's no limit otherwise. | 20:52 |
| lutostag | cory_fu: ok, that's what I thought, if I'm in debug-hook I assume I can list the states, so I'll try that, must not have a state I should | 20:52 |
| cory_fu | lutostag: Easy way to see the states from the CLI: charms.reactive -y get_states | 20:53 |
| lutostag | cory_fu: that was easy, thanks | 20:55 |
| lutostag | charm publish -> charm release... that was recent I'm not crazy right? | 21:24 |
| kwmonroe | yeah lutostag, you're not crazy. it was right around the beta->rc timeframe | 21:27 |
| === rcj is now known as Guest57444 | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!