| *** shakamunyi has joined #kolla | 00:13 | |
| *** barra204 has joined #kolla | 00:13 | |
| *** vinkman has joined #kolla | 00:19 | |
| *** rhallisey has quit IRC | 00:20 | |
| *** zhiwei has joined #kolla | 01:06 | |
| *** barra204 has quit IRC | 01:24 | |
| *** shakamunyi has quit IRC | 01:24 | |
| *** vinkman has quit IRC | 01:27 | |
| openstackgerrit | Martin André proposed stackforge/kolla: Fail image build when build script is missing https://review.openstack.org/187834 | 01:31 |
|---|---|---|
| *** tobe has joined #kolla | 01:33 | |
| *** erkules_ has joined #kolla | 01:39 | |
| *** erkules has quit IRC | 01:41 | |
| *** vinkman has joined #kolla | 01:50 | |
| *** erkules_ has quit IRC | 01:50 | |
| *** erkules_ has joined #kolla | 01:52 | |
| *** sdake has quit IRC | 02:11 | |
| *** vinkman has quit IRC | 02:16 | |
| *** fangfenghua has joined #kolla | 02:56 | |
| *** dims__ has quit IRC | 02:56 | |
| SamYaple | loth: ive talked to a few people outside kolla here and they all want docker volumes. i have no strong opinion so well do what the community wants | 02:57 |
| SamYaple | in the future they may be `docker volumes` comming as a command | 02:57 |
| SamYaple | so that would be cool | 02:57 |
| *** vinkman has joined #kolla | 03:31 | |
| *** vinkman has left #kolla | 03:31 | |
| openstackgerrit | weiyu proposed stackforge/kolla: Check the result of compose cmd https://review.openstack.org/187855 | 03:39 |
| *** weiyu-001 has joined #kolla | 04:20 | |
| *** weiyu-001 has left #kolla | 04:20 | |
| *** weiyu has joined #kolla | 04:20 | |
| *** bmace has quit IRC | 04:21 | |
| *** weiyu has quit IRC | 04:21 | |
| *** weiyu-001 has joined #kolla | 04:21 | |
| *** bmace has joined #kolla | 04:32 | |
| *** fangfenghua has quit IRC | 05:03 | |
| *** shakamunyi has joined #kolla | 06:16 | |
| *** barra204 has joined #kolla | 06:17 | |
| *** dasm|afk is now known as dasm | 06:44 | |
| *** shardy has joined #kolla | 06:55 | |
| *** inc0 has joined #kolla | 07:03 | |
| inc0 | good morning | 07:07 |
| SamYaple | morning inc0 | 07:09 |
| inc0 | SamYaple, do you ever sleep? | 07:09 |
| SamYaple | im asleep right now | 07:09 |
| inc0 | where did you get irc-to-brain interface? | 07:10 |
| SamYaple | china | 07:10 |
| SamYaple | https://code.google.com/p/irssi-connectbot/ | 07:10 |
| inc0 | :) | 07:11 |
| inc0 | I salute your endurance;) | 07:12 |
| SamYaple | its actually my midday right now | 07:12 |
| SamYaple | i work 3rd shift | 07:12 |
| inc0 | ah, I thought yesterday you've said that your midday is about us time? | 07:12 |
| inc0 | did you travel? | 07:12 |
| SamYaple | no. i must not realize what time it is when i tal kto you | 07:13 |
| SamYaple | im normally up 2100-1200 central time | 07:13 |
| inc0 | I'm almost UTC, 9am now | 07:14 |
| *** athomas has joined #kolla | 07:14 | |
| SamYaple | 2am here | 07:14 |
| inc0 | 2am midday?;) not 2pm? or 14? | 07:15 |
| inc0 | sorry, I'm before my first coffee | 07:15 |
| SamYaple | i work 3rd shift "my midday" | 07:15 |
| inc0 | ah.. I see now | 07:16 |
| SamYaple | time is a fluid concept that none of use can fully understand | 07:16 |
| inc0 | yeah, I have particular trouble with it;) | 07:17 |
| *** bradjones has quit IRC | 07:20 | |
| *** bradjones has joined #kolla | 07:22 | |
| *** tobe has quit IRC | 07:52 | |
| diga | SamYaple: Hi | 07:56 |
| diga | can you give me link how to install kolla on VM ? | 07:56 |
| *** dims_ has joined #kolla | 07:58 | |
| SamYaple | diga: its poorly named, but https://github.com/stackforge/kolla/blob/master/docs/dev-quickstart.md | 07:58 |
| SamYaple | that is the "install" guide atm I belive | 07:58 |
| diga | SamYaple: Thanks | 07:58 |
| *** dims_ has quit IRC | 08:03 | |
| *** tobe has joined #kolla | 08:14 | |
| *** shardy_ has joined #kolla | 08:16 | |
| *** shardy has quit IRC | 08:17 | |
| *** shardy_ has quit IRC | 08:22 | |
| *** shardy has joined #kolla | 08:22 | |
| *** gfidente has joined #kolla | 08:24 | |
| *** pdb has quit IRC | 09:42 | |
| *** pdb has joined #kolla | 09:42 | |
| *** zhiwei has quit IRC | 09:48 | |
| inc0 | hmm...guys, I'm thinking how to determine master with multi node deployment | 09:55 |
| inc0 | I'd need to add priority to keepalived | 09:55 |
| inc0 | and master should be slighltly higher | 09:55 |
| SamYaple | since "master" is arbitrary i just use the first host in the group as master in ansible | 09:55 |
| *** erkules_ is now known as erkules | 10:02 | |
| *** erkules has joined #kolla | 10:02 | |
| inc0 | I'll see how keepalived works without any config out there | 10:02 |
| inc0 | maybe it'll choose random, which would work | 10:03 |
| SamYaple | you need to set the id | 10:03 |
| inc0 | you mean router id? | 10:04 |
| SamYaple | yea | 10:04 |
| inc0 | it should be the same at all nodes | 10:04 |
| SamYaple | err the vid | 10:04 |
| SamYaple | hold on i cant rememebr | 10:04 |
| inc0 | what should be different is priority | 10:05 |
| inc0 | afaik | 10:05 |
| SamYaple | yea youre correct | 10:05 |
| SamYaple | virtual_router_id matches and priority is different | 10:06 |
| *** dims_ has joined #kolla | 10:06 | |
| inc0 | yup, and I'm wondering if we set same priority, what will break:) | 10:06 |
| inc0 | we'll see I guess | 10:06 |
| SamYaple | it doesnt work well, no :/ | 10:07 |
| SamYaple | been down that road | 10:07 |
| inc0 | meh | 10:07 |
| SamYaple | i cant remember if it just plain broke, or the vips ended up in both places | 10:07 |
| SamYaple | but something bad happened | 10:07 |
| SamYaple | https://github.com/SamYaple/yaodu/blob/master/ansible/roles/keepalived/templates/keepalived.conf.j2#L12 | 10:07 |
| SamYaple | i ended up doing that where {{ loop.index }} would be unique | 10:08 |
| inc0 | hmm...I have very nasty idea | 10:09 |
| inc0 | ;) | 10:09 |
| SamYaple | lalalala im not listening | 10:09 |
| inc0 | :D | 10:09 |
| inc0 | make it random with int big enough that risk of overlap will be lesser than risk of meteorite hitting datacenter;) | 10:10 |
| SamYaple | nah | 10:10 |
| SamYaple | thats not cool with me | 10:10 |
| SamYaple | plus i think yo ucan only go to 2^16 | 10:11 |
| inc0 | nah, it's as dirty as it can possibly be | 10:11 |
| inc0 | 2^16 is still pretty big;) | 10:11 |
| SamYaple | only 65000 | 10:11 |
| inc0 | another idea - env with name of master host | 10:12 |
| inc0 | and if hostname == master: priority is bigger | 10:12 |
| SamYaple | still have to have uniques per "slaves" | 10:12 |
| inc0 | true | 10:13 |
| SamYaple | whats wrong with the ansible approach? this doesnt break setting an environment variable for the non-ansible deploy | 10:13 |
| inc0 | ansible needs to be involved, that's my worry | 10:14 |
| SamYaple | triple-o should be able to easily generate that unique | 10:14 |
| inc0 | I'd rather make env variable for ansible, tripleo or whatever to fill | 10:14 |
| SamYaple | no ansible will be generating configs dynamically | 10:15 |
| inc0 | like HOSTNAME_PRIORITY=host1:100;host2:99 | 10:15 |
| inc0 | but we're not there yet right? | 10:15 |
| SamYaple | the dynamic config is all good ot go, i just havent been able to finalize the commit since my gear crashed | 10:15 |
| inc0 | on what basis it will generate configs? env vars? | 10:15 |
| SamYaple | the yaodu link i posted is a template, that template gets filled out when its transfered to a host | 10:16 |
| SamYaple | that file on the host is unique per host | 10:16 |
| SamYaple | it is the nconverted into a variable that can be pasred dynamically in the container | 10:16 |
| SamYaple | so ansible still handles the config gneeration entirely | 10:16 |
| inc0 | yeah, that's what I'm concerned about - tripleo guys won't like that | 10:17 |
| SamYaple | tripleo guys can do it however they see fit | 10:17 |
| inc0 | also, having several config generators isn't good idea as well | 10:17 |
| SamYaple | im not complaing about HOSTNAME_PRIORITY=host1:100;host2:99 | 10:17 |
| SamYaple | its just not how ansible will do it | 10:17 |
| inc0 | I get it, but I'll want to make it not-ansible-specifiv | 10:18 |
| SamYaple | what do you mean, several config generators? | 10:18 |
| inc0 | with all my love to ansible (true and deep), this might be a problem | 10:18 |
| inc0 | trilpleo uses puppet for that | 10:18 |
| inc0 | and they have pretty advanced stuff like support for multiple physical networks and so on | 10:18 |
| SamYaple | this is more a conversation about scope and our mission statement then | 10:19 |
| inc0 | I agree | 10:19 |
| SamYaple | the container build isnt in question, correct? | 10:19 |
| inc0 | yup | 10:19 |
| inc0 | but deployment orchiestration and variable generation might | 10:19 |
| inc0 | heat can do that already | 10:20 |
| SamYaple | sure and thats certainly a conversation to have. i dont want to block triple-o, but none of that will block ansible | 10:20 |
| inc0 | I agree, I just want to prepare containers in a way they will be agnostic to deployment tool | 10:20 |
| SamYaple | there will be a kill switch to turn off all autoconfiguration i nthe container | 10:21 |
| inc0 | personally I don't think autoconfig will be bad | 10:21 |
| SamYaple | if they community wants autoconfiguration with environment variables I am ok with that | 10:21 |
| inc0 | if we consider option like we discussed before - pass a config, ansible will pack it to dict, and then fill conf inside container with exactly these values | 10:22 |
| inc0 | thing to discuss for sure. | 10:22 |
| inc0 | for now I'll do host1:101 | 10:22 |
| inc0 | and we'll change it when we decide on something | 10:23 |
| SamYaple | the problem _I_ have with autoconfiguration, is it is not flexible. all the available options must be handcoded and i dont feel that is a good use of time and it is error prone personally | 10:23 |
| SamYaple | but i wont block whatever anyone wants to do, as long as there is a kill switch | 10:23 |
| inc0 | especially that those have to be maintained | 10:23 |
| inc0 | configs in openstack change in time | 10:23 |
| SamYaple | yep | 10:23 |
| SamYaple | in yaodu the upgrade from juno to kilo required no container build changes at all | 10:24 |
| inc0 | but, being able to pass default config + changes should be ok | 10:24 |
| *** sdake has joined #kolla | 11:02 | |
| sdake | morning | 11:06 |
| * sdake growls at 5am training 3 days in a row | 11:06 | |
| SamYaple | morning sdake | 11:07 |
| SamYaple | quit being lazy. get to work | 11:07 |
| SamYaple | so i was thinking about the "all packages from source" container that I wanted to do. I think gentoo might be the answer for that. I did an LFS container but i dont see any benefit over a gentoo container | 11:08 |
| SamYaple | plus the openstack ebuild package maintainer for gentoo works on my team now | 11:08 |
| sdake | fawk wife leaves for ansible fest for 3 days leaves 2 pepsis in the fridge | 11:11 |
| sdake | wtfawk! | 11:11 |
| inc0 | ah, that's why you all want ansible so much;) | 11:16 |
| pdb | SamYaple: surely python virtualenvs are more portable | 11:17 |
| inc0 | SamYaple, you want to make container for source installations? | 11:19 |
| inc0 | and install *everything* from source? | 11:19 |
| inc0 | (aka compile all the stuff \o/) | 11:19 |
| SamYaple | inc0: :D as a side side project, but yes | 11:19 |
| inc0 | wow. | 11:19 |
| SamYaple | i have it working, but its not "accessible" | 11:20 |
| SamYaple | bits of manula stuff in it as well | 11:20 |
| SamYaple | pdb: im not opposed to virtualenvs. I dont see the benefit given the end product being a container. Can you expand on what it would give us? | 11:21 |
| pdb | SamYaple: maybe Im not up to speed on what your goal is. We want to remove the dependency on rpms/debs right? | 11:22 |
| SamYaple | pdb: correct. | 11:23 |
| pdb | so why would you tie yourself to emerge which only works on one OS? (gentoo) | 11:23 |
| pdb | whereas virtualenvs will work pretty much anywhere | 11:23 |
| SamYaple | oh sorry for the confution pdb | 11:23 |
| SamYaple | no thats a side project. i shouldnt have brought it up | 11:24 |
| SamYaple | ignore that | 11:24 |
| pdb | ah :) | 11:24 |
| SamYaple | no i do a pip install in the container build for ubutnu/centos installs | 11:24 |
| SamYaple | i only have working for ubuntu, but havent attempted centos family yet | 11:24 |
| *** weiyu-001 has quit IRC | 11:31 | |
| *** dwalsh has joined #kolla | 11:51 | |
| sdake | I find your lack of faith disturbing :) | 11:53 |
| *** tobe has quit IRC | 11:56 | |
| *** tobe has joined #kolla | 11:57 | |
| *** tobe has quit IRC | 12:02 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Keepalived container https://review.openstack.org/187981 | 12:28 |
| inc0 | and here I am, complaining about keepalived in ha spec... | 12:29 |
| *** weiyu has joined #kolla | 12:30 | |
| *** zhiwei has joined #kolla | 12:30 | |
| *** athomas has quit IRC | 12:36 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Keepalived container https://review.openstack.org/187981 | 12:45 |
| *** athomas has joined #kolla | 12:47 | |
| *** inc0 has quit IRC | 12:52 | |
| *** dwalsh has quit IRC | 12:56 | |
| *** mstachow has joined #kolla | 13:08 | |
| mstachow | afternoon :) | 13:08 |
| sdake | your on the wrong tz mstachow :) | 13:08 |
| mstachow | tz :> ? | 13:09 |
| sdake | say I noticed you reassigned the from source blueprint | 13:09 |
| sdake | is that one of your cohorts? :) | 13:09 |
| *** dwalsh has joined #kolla | 13:10 | |
| sdake | hmm maybe wrong person sorry I meant inc0 | 13:10 |
| sdake | I htink ;) | 13:10 |
| dasm | mstachow: tz == timezone | 13:10 |
| sdake | hey dasm | 13:10 |
| dasm | sdake: hello o/ | 13:10 |
| mstachow | thanks dasm | 13:11 |
| sdake | team meeting today - hope folks can make it :) | 13:11 |
| dasm | today is 22UTC? | 13:11 |
| sdake | yar | 13:11 |
| dasm | so, 0:00 CET :'( | 13:11 |
| mstachow | I understand now :D I and inc0 are from the same tz :) | 13:11 |
| mstachow | dasm :'( It hurts me too | 13:12 |
| dasm | mstachow: you're a student. you're working 'different' hours :) | 13:12 |
| sdake | mstachow are you the intel intern? | 13:12 |
| sdake | lots of new faces, trying to keep it all straight :) | 13:13 |
| dasm | sdake: he is | 13:13 |
| sdake | cool interning at intel was fantasitc for my career | 13:13 |
| mstachow | dasm, I know.. I'd got couple of exams today morning :< | 13:13 |
| sdake | dasm you an inteller too? | 13:13 |
| dasm | sdake: i am. the same team with inc0, mstachow and ala | 13:14 |
| sdake | start there out of school? | 13:14 |
| dasm | ala == akwasniew | 13:14 |
| sdake | If I was doing it over again today, there is 3 places I'd start my career, and intc would be one of em :) | 13:14 |
| dasm | sdake: me? no. i've been working in several other companies. mostly in web industries. | 13:15 |
| sdake | damn nice | 13:15 |
| dasm | sdake: i'm wondering if openstack is still 'web' or not :P | 13:15 |
| dasm | if yes, i'm still in web. :D | 13:15 |
| sdake | its more like infrastructure | 13:15 |
| sdake | samyaple around? | 13:16 |
| sdake | wanted to talk about orchestration of the HA components | 13:16 |
| dasm | sdake: probably not. he was awake around 5AM CET so ~3AM UTC | 13:17 |
| dasm | but if you'll look at logs, he talked a little with inc0 about it. | 13:17 |
| sdake | thanks for the log pointer | 13:23 |
| sdake | that was helpful | 13:24 |
| mstachow | sdake: Do we need rsync as replication method for galera? | 13:24 |
| sdake | you tell me, your the expert | 13:29 |
| sdake | if you want the question answered more directly consult danehans | 13:29 |
| sdake | he is traveling atm | 13:29 |
| mstachow | You see, we've got four method of replication. 3 of them are currently working well. Rsync is tricky one for me | 13:30 |
| *** dasm is now known as dasm|afk | 13:33 | |
| *** dwalsh has quit IRC | 13:33 | |
| *** rhallisey has joined #kolla | 13:45 | |
| *** inc0 has joined #kolla | 13:47 | |
| pdb | any reason the mariadb container isn't set to log to stdout? It's failing to startup and can't get info from 'docker logs' | 13:50 |
| mstachow | mariadb as far as I know is putting every single note into log file. stdout by default is nearly empty | 13:52 |
| pdb | yes that's the problem | 13:52 |
| *** dwalsh has joined #kolla | 13:53 | |
| mstachow | I am not sure if mariadb daemon allow You to sending directly messages to stdout | 13:58 |
| sdake | https://mariadb.com/kb/en/mariadb/mysqld_safe/ | 14:00 |
| sdake | try --log-error option to mysql | 14:00 |
| sdake | mysql_safe | 14:00 |
| mstachow | yes, but it still logs into file | 14:01 |
| mstachow | workaround: run mysqld_safe in background and listen output by tailf from log file | 14:02 |
| mstachow | that requires of course some changes in mariadb-app container and in my galera container too | 14:03 |
| *** apmelton_away is now known as apmelton | 14:03 | |
| inc0 | mstachow, also this gives us another set of problems - rotating logs | 14:04 |
| pdb | it also may interfere with the container lifecycle if mysql isn't pid 1 | 14:06 |
| pdb | im sure there's an option to log to stdout, will have a play around | 14:06 |
| mstachow | that was just a quick workaround - I know that It isn't perfect :P | 14:06 |
| *** inc0_ has joined #kolla | 14:13 | |
| *** inc0_ has quit IRC | 14:13 | |
| *** weiyu has quit IRC | 14:16 | |
| *** jruano has joined #kolla | 14:26 | |
| *** dasm|afk is now known as dasm | 14:32 | |
| *** barra204 has quit IRC | 14:36 | |
| *** shakamunyi has quit IRC | 14:36 | |
| *** zhiwei has quit IRC | 14:39 | |
| *** dwalsh has quit IRC | 14:43 | |
| sdake | damn my bottle of tl1 has evaporated | 14:46 |
| inc0 | tl1? some kind of liquor? | 14:53 |
| mstachow | :O?! | 14:55 |
| bmace | hey sdake, did you get those rebuilt images pushed up to dockerhub yesterday? the ones up there still look old (8 days) | 14:56 |
| sdake | i ran it last night | 14:56 |
| sdake | it should have completed | 14:56 |
| sdake | i havem eeting now, willl be avilable in 1 hour to check the results | 14:56 |
| *** vbel has quit IRC | 14:57 | |
| *** daneyon has joined #kolla | 14:57 | |
| *** vbel has joined #kolla | 14:58 | |
| *** daneyon_ has joined #kolla | 15:01 | |
| *** dwalsh has joined #kolla | 15:02 | |
| *** daneyon has quit IRC | 15:04 | |
| *** jtriley has joined #kolla | 15:08 | |
| *** dwalsh has quit IRC | 15:15 | |
| *** dasm is now known as dasm|afk | 15:16 | |
| *** dwalsh has joined #kolla | 15:25 | |
| *** vinkman has joined #kolla | 15:32 | |
| *** vinkman has quit IRC | 15:35 | |
| *** dwalsh has quit IRC | 15:37 | |
| *** rhallisey has quit IRC | 15:38 | |
| *** vinkman has joined #kolla | 15:53 | |
| *** bmace_ has joined #kolla | 15:56 | |
| *** dwalsh has joined #kolla | 15:58 | |
| *** bmace has quit IRC | 15:59 | |
| *** vinkman has quit IRC | 15:59 | |
| sdake | bmace checking results moment | 16:04 |
| sdake | bmace my push failed because of credentail issues | 16:14 |
| sdake | sorting out now | 16:14 |
| bmace_ | sdake excellent. | 16:16 |
| *** dims_ has quit IRC | 16:18 | |
| *** dims_ has joined #kolla | 16:18 | |
| *** dims__ has joined #kolla | 16:20 | |
| *** dims_ has quit IRC | 16:24 | |
| mstachow | harmw are You here ? | 16:26 |
| *** vinkman has joined #kolla | 16:27 | |
| *** vinkman has quit IRC | 16:29 | |
| *** athomas has quit IRC | 16:35 | |
| *** cbaesema has joined #kolla | 16:51 | |
| *** rhallisey has joined #kolla | 16:51 | |
| sdake | spartans, what is your profession! | 16:51 |
| rhallisey | ah wo! | 16:51 |
| rhallisey | great movie | 16:52 |
| cbaesema | https://github.com/stackforge/kolla/tree/stable/juno/docker/swift/swift-base missing config-swift.sh ? | 16:53 |
| sdake | swift is in a state of complete disarray in terms of implementation | 16:53 |
| sdake | looking for someone to take it on :) | 16:54 |
| sdake | cbaesema welcome to #kolla btw :) | 16:54 |
| cbaesema | thx just noticed while doing some builds yesterday | 16:54 |
| *** gfidente is now known as gfidente|afk | 17:01 | |
| *** bmace_ is now known as bmace | 17:02 | |
| *** gfidente|afk is now known as gfidente | 17:20 | |
| *** gfidente is now known as gfidente|afk | 17:31 | |
| *** rhallisey has quit IRC | 17:49 | |
| *** rhallisey has joined #kolla | 17:50 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add cinder into the docker compose setup https://review.openstack.org/173507 | 17:50 |
| sdake | rhallisey did you test the mount /dev/pts in the container | 17:51 |
| sdake | I'd like to see if that solution works - tha tis my prefferred solution | 17:51 |
| rhallisey | no I didn't yet | 17:51 |
| sdake | can you add it into your queue | 17:51 |
| rhallisey | yes | 17:52 |
| sdake | mount -t devfs /dev/pts should do it I htink | 17:52 |
| rhallisey | talking with dwalsh about it | 17:52 |
| sdake | wfm | 17:53 |
| sdake | get his advice | 17:53 |
| sdake | keep in mind we can't wait until docker 1.8 | 17:53 |
| sdake | we need to stop depending on new updates to docker to do the job | 17:53 |
| sdake | or we will be waiting forever to get to a working solution | 17:53 |
| sdake | tell of victory | 17:54 |
| sdake | victory? yes my lord | 17:54 |
| sdake | this movie rocks | 17:54 |
| sdake | i'm more satisified with a workaround then waiting until docker 1.8 | 17:55 |
| sdake | i'm ok with waiting for docker .17 june 15th deadline | 17:55 |
| sdake | 1.7 that is | 17:55 |
| sdake | maybe .17 is more accurate ;-) | 17:55 |
| rhallisey | ya ha | 17:56 |
| rhallisey | well this is using 1.7 | 17:56 |
| sdake | well we are using thee dev verison | 17:56 |
| sdake | maybe dan can jam in a fix into 1.7 | 17:56 |
| sdake | but i'm not sure if they are taking bug fixes at this point | 17:57 |
| rhallisey | ya I think they are at a freeze | 17:57 |
| sdake | fuzzy-freeze | 17:57 |
| sdake | they take certain thigns and not others | 17:57 |
| sdake | but libcontainer is locked in I think | 17:57 |
| sdake | and that is where this code is contained | 17:57 |
| sdake | from my understanding of the situation | 17:57 |
| *** mstachow has quit IRC | 18:00 | |
| dwalsh | sdake, What is the problem. | 18:02 |
| sdake | dwalsh /dev/pts/ptmx ends up with 00 permissiosn | 18:02 |
| dwalsh | <dwalsh> <dwalsh> RIght but libvirt in the container is running as root. with privs turned off. | 18:02 |
| dwalsh | <dwalsh> <dwalsh> ls -lZ /dev/pts/ptmx | 18:02 |
| dwalsh | <dwalsh> <dwalsh> crw-rw-rw-. root root system_u:object_r:svirt_sandbox_file_t:s0:c642,c891 /dev/pts/ptmx | 18:02 |
| dwalsh | <dwalsh> <dwalsh> Permissions in a normal container. | 18:02 |
| dwalsh | <dwalsh> <dwalsh> ls -lZ /dev/pts/ptmx | 18:02 |
| dwalsh | <dwalsh> <dwalsh> c---------. 1 root root system_u:object_r:devpts_t:s0 5, 2 Jun 2 15:14 /dev/pts/ptmx | 18:02 |
| dwalsh | <dwalsh> <dwalsh> Weird. But the only reason a root process would not be allowed to ouch /dev/pts/ptmx is if it did not have DAC_OVERRIDE. | 18:02 |
| dwalsh | <dwalsh> <dwalsh> I can change the permisions from within the container. I guess I need to see the failure. | 18:02 |
| dwalsh | sdake, Right but that is what it is on my host system, even with libvirtd installed and it handles it. | 18:03 |
| dwalsh | Not sure why it is 666 inside the container. | 18:03 |
| sdake | it is 666 inside the container because rhallisey chmoded it in start.sh | 18:03 |
| sdake | I am not sure what it should be on a fresh install | 18:03 |
| dwalsh | But why is it needed are you running libvirtd as a non privileged user? | 18:04 |
| sdake | i agree selinux handles it, we only see a problem when we bindmount /dev | 18:04 |
| sdake | rather let me restart | 18:04 |
| dwalsh | If a process is running as root and has DAC_OVERRIDE it should be able to handle it. | 18:04 |
| sdake | I am not running with selinux | 18:04 |
| dwalsh | RIght | 18:04 |
| sdake | the process is running as root | 18:04 |
| dwalsh | Then the process should have access. | 18:04 |
| *** mstachow has joined #kolla | 18:04 | |
| sdake | is DAC_OVERRIDE a default permission? | 18:04 |
| sdake | in the capabilties for root | 18:05 |
| dwalsh | in a --privileged container yes | 18:05 |
| sdake | this is a privileged container | 18:05 |
| dwalsh | Then it should have access, unless something else is breaking. | 18:05 |
| sdake | only way to know for sur eis to run libvirt in strace I suspect | 18:05 |
| dwalsh | I just rand a --privileged -v /dev:/dev and I can modify the permissions on it. | 18:05 |
| sdake | the permissions canbe modified | 18:06 |
| sdake | but the default permissions dont' allow libvirt to create its monitor device | 18:06 |
| dwalsh | Is libvirtd putting something in its log files? | 18:06 |
| sdake | I haven't pulled down rhallisey's change and debugged personally | 18:06 |
| sdake | so I'm the wrong guy to ask :) | 18:06 |
| rhallisey | dwalsh, ya http://pastebin.test.redhat.com/287739 | 18:06 |
| sdake | rhallisey change the libvirt container to run in strace | 18:07 |
| sdake | and get that output of that | 18:07 |
| sdake | that will tell us precisely what is going on | 18:07 |
| sdake | one with bindmounting one without | 18:07 |
| dwalsh | If libvirtd can handle 000 on ptmx on the host, it should be able to handle it in the container. | 18:07 |
| sdake | i hear what your saying and agree | 18:07 |
| rhallisey | dwalsh, agreed that's what confuses me too | 18:08 |
| sdake | but it is not | 18:08 |
| rhallisey | it get's that above error | 18:08 |
| rhallisey | libvirtError: internal error: process exited while connecting to monitor: 2015-06-03T17:58:30.071758Z qemu-kvm: -chardev pty,id=charserial1: Failed to create chardev | 18:08 |
| sdake | rhallisey how about putting that on a public pastebin | 18:08 |
| sdake | so I can access it :) | 18:08 |
| rhallisey | ya hold on ha :D | 18:08 |
| sdake | I've dealt with /dev/pts/ptmx since the beginning of this damn project :( | 18:08 |
| rhallisey | http://paste.fedoraproject.org/228551/14333549/ | 18:09 |
| rhallisey | that's comes out of nova-compute | 18:09 |
| sdake | we need the libvirt logs | 18:09 |
| rhallisey | that* | 18:09 |
| sdake | docker exec livirt-container cat /var/log/libvirt/instances/000x.log | 18:09 |
| sdake | whatever 000x is | 18:09 |
| sdake | you can determine that by using the ls command | 18:10 |
| sdake | sicne docker doesn't do shell expansion on exec inside the container | 18:10 |
| rhallisey | http://paste.fedoraproject.org/228552/55075143/ | 18:11 |
| sdake | ya i've seen that before | 18:11 |
| sdake | I was never able to root cause it | 18:11 |
| sdake | i've seen itt before *ALOT* | 18:11 |
| sdake | but it seems to never happen now | 18:11 |
| rhallisey | this is the first i've seen it | 18:12 |
| sdake | I'm not sure what fixed it | 18:12 |
| rhallisey | docker probably | 18:12 |
| rhallisey | :) | 18:12 |
| sdake | oh I remember what fixed it | 18:12 |
| sdake | brain just cooperated :) | 18:12 |
| sdake | I stopped bindmounting /dev ;-) | 18:12 |
| rhallisey | will a volume mount work in this case? | 18:13 |
| dwalsh | So this works with 666 correct? Does not work with 600? | 18:13 |
| sdake | not sure but I'd definately prefer 600 | 18:13 |
| * rhallisey checking | 18:13 | |
| sdake | and over that I'd prefer mount -t devfs /dev/pts | 18:13 |
| sdake | maybe our default libvirt config is running with a specific uid/gid | 18:14 |
| dwalsh | Won't that cause a problem with libvirt exiting and restarting, where it could not see other pts devices it created. | 18:14 |
| sdake | dwalsh you may be correct | 18:15 |
| sdake | lets go with 600 if that works | 18:15 |
| sdake | i'd be satisified with that solution | 18:16 |
| sdake | although something is still busted somewhere | 18:16 |
| sdake | samyaple's thinking got me thinking it may have something to do with the new mount propogation feature | 18:16 |
| sdake | but I recall removing /dev from the bindmounting in the libvirt container fixed this d/ev/pts/ptmx problem | 18:16 |
| sdake | so this was prior to the rwork to add /dev bindmounting | 18:16 |
| sdake | (in docker 1.5) | 18:16 |
| rhallisey | gotta run | 18:19 |
| rhallisey | I'll beck back later | 18:19 |
| *** rhallisey has quit IRC | 18:19 | |
| sdake | well dwalsh since rhallisey just disappared | 18:21 |
| sdake | and I have no idea how to use cinder | 18:21 |
| sdake | I guess we will have to revisit this permission issue | 18:21 |
| loth | SamYaple: you around? | 18:24 |
| *** mstachow has quit IRC | 18:39 | |
| *** jtriley has quit IRC | 18:43 | |
| *** jtriley has joined #kolla | 18:54 | |
| *** dwalsh has quit IRC | 19:03 | |
| *** shardy_ has joined #kolla | 19:06 | |
| *** shardy has quit IRC | 19:08 | |
| harmw | wasn't thre a meeting tonight? | 19:09 |
| harmw | like, an hour ago? | 19:09 |
| *** cbaesema has quit IRC | 19:11 | |
| *** cbaesema has joined #kolla | 19:11 | |
| sdake | 3 hours harmw | 19:12 |
| *** shardy_ has quit IRC | 19:12 | |
| *** shardy has joined #kolla | 19:12 | |
| harmw | oh, it's an even week | 19:13 |
| sdake | ya confusign huh | 19:16 |
| sdake | if only the earth didn't rotate ;) | 19:16 |
| harmw | yea, fixing that is still on my TODO | 19:16 |
| sdake | blow up the moon | 19:16 |
| sdake | problem solved ;-) | 19:16 |
| *** dwalsh has joined #kolla | 19:19 | |
| *** rhallisey has joined #kolla | 19:24 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add the Designate DNSaaS services. https://review.openstack.org/182799 | 19:26 |
| harmw | fingers crossed I got *all* crap this time :P | 19:27 |
| harmw | SamYaple: sdake mandre , check it out :) | 19:27 |
| sdake | harmw otp with wife | 19:28 |
| harmw | so? hangup | 19:28 |
| sdake | she already wants to fire me | 19:28 |
| harmw | hehe | 19:28 |
| *** jtriley has quit IRC | 19:29 | |
| *** rhallisey has quit IRC | 19:37 | |
| *** rhallisey has joined #kolla | 19:37 | |
| *** jtriley has joined #kolla | 19:38 | |
| *** ryansb has left #kolla | 19:53 | |
| *** inc0 has quit IRC | 19:53 | |
| *** daneyon_ has quit IRC | 20:08 | |
| *** daneyon has joined #kolla | 20:08 | |
| *** daneyon_ has joined #kolla | 20:11 | |
| *** daneyon has quit IRC | 20:14 | |
| *** loth has left #kolla | 20:16 | |
| sdake | yo danyeon glad to see you back from off the road ;-) | 20:17 |
| sdake | BBL dentist torture session | 20:17 |
| *** dwalsh has quit IRC | 20:19 | |
| *** pdb-mobile has joined #kolla | 20:53 | |
| *** pdb-mobile has quit IRC | 20:59 | |
| *** jtriley has quit IRC | 21:06 | |
| *** dasm has joined #kolla | 21:12 | |
| *** jtriley has joined #kolla | 21:14 | |
| SamYaple | hey guys | 21:29 |
| dasm | welcome back SamYaple. o/ | 21:30 |
| SamYaple | yea dasm | 21:30 |
| SamYaple | loth you here? | 21:31 |
| *** jruano has quit IRC | 21:34 | |
| SamYaple | most of my replacement gear just arrived | 21:36 |
| *** jtriley has quit IRC | 21:38 | |
| *** loth has joined #kolla | 21:47 | |
| *** mstachow has joined #kolla | 21:49 | |
| * mstachow feel drowsy | 21:50 | |
| sdake | boy its hot in arizona | 21:50 |
| dasm | mstachow: be a man. | 21:51 |
| sdake | irc meeting in t-15 minutes | 21:51 |
| loth | sdake: native? :) | 21:51 |
| sdake | yar | 21:51 |
| sdake | born and raised | 21:51 |
| sdake | lived in flagstaff during university | 21:51 |
| sdake | and for about 6 years in my middle 30s | 21:51 |
| loth | pfft should be immune :) | 21:51 |
| sdake | but needed to live near a proper airport | 21:52 |
| sdake | its not like I can put on an air conditioned jacket ;-) | 21:52 |
| mstachow | I woke up ~15 hours ago after 4 hours of sleeping. I'm looking forward on my bed ^^ | 21:52 |
| dasm | sdake: what does it mean 'hot'? | 21:52 |
| sdake | 105f? | 21:52 |
| sdake | nto sure | 21:52 |
| dasm | mstachow: you can sit in bed, while chatting on irc :) | 21:52 |
| sdake | its hot | 21:52 |
| sdake | dasm thats how I roll ;-) | 21:52 |
| loth | 97 now | 21:53 |
| dasm | sdake: let me check how much is it. | 21:53 |
| mstachow | dasm in this case I will fall asleep soon ;) | 21:53 |
| dasm | 40.5 *C | 21:53 |
| dasm | pretty hot :D | 21:53 |
| sdake | it gets up to118f | 21:53 |
| SamYaple | hows the humidity | 21:53 |
| sdake | in august | 21:53 |
| sdake | 0% typical | 21:53 |
| sdake | 10-20 sometimes | 21:54 |
| SamYaple | psshhh | 21:54 |
| SamYaple | try 100% humidity in 100F weather | 21:54 |
| sdake | been there | 21:54 |
| SamYaple | its awful | 21:54 |
| loth | people here just hop from ac sources | 21:54 |
| sdake | agree | 21:54 |
| SamYaple | you feel like youre drowning | 21:54 |
| sdake | my panamera needs more ac power | 21:54 |
| sdake | damn germans and their thinking people don't need ac | 21:55 |
| dasm | SamYaple: where are described conditions? 100% + 100F? | 21:55 |
| sdake | texas and luiousana | 21:55 |
| sdake | and carolinas | 21:55 |
| sdake | maybe elsewhere | 21:55 |
| dasm | sdake: ok, noted. this is place where i don't want to leave. when i'll move to US, it won't be there :P | 21:56 |
| SamYaple | i live in TX atm, but im from SC dasm | 21:56 |
| mandre | hey guys, I'm going to make it this time for the meeting | 21:56 |
| mandre | good morning :) | 21:56 |
| sdake | mandre morning ;) | 21:56 |
| dasm | mandre: good evening | 21:57 |
| sdake | what time is it in jp mandre? | 21:57 |
| mandre | 7AM | 21:57 |
| mstachow | hi ! mandre | 21:57 |
| dasm | or even good midnight, mstachow | 21:57 |
| SamYaple | hey mandre | 21:57 |
| SamYaple | i too will maek this one | 21:57 |
| dasm | s/mstachow/mandre :P | 21:57 |
| mandre | it will be war zone soon when the kids get ready for school | 21:57 |
| sdake | ya thats how my household rolls in the morning during the school year | 21:57 |
| sdake | chaotic | 21:57 |
| dasm | so, we have jp with 7AM, pl with 12AM and us around 4PM? | 21:59 |
| sdake | 3pm here | 21:59 |
| sdake | 7pm on east coast | 21:59 |
| dasm | it would be much simpler to stick just to utc :P | 21:59 |
| SamYaple | 6pm east coast | 21:59 |
| *** pdb-mobile has joined #kolla | 21:59 | |
| *** akwasnie has joined #kolla | 22:02 | |
| *** vinkman has joined #kolla | 22:07 | |
| *** pdb-mobile has quit IRC | 22:39 | |
| *** jruano has joined #kolla | 22:46 | |
| *** jtriley has joined #kolla | 22:52 | |
| *** smeisner has joined #kolla | 22:56 | |
| *** pdb-home has joined #kolla | 22:56 | |
| dasm | so, do we know already, what's the mission of Kolla? ;) | 22:57 |
| SamYaple | and in the docker bind them | 22:57 |
| sdake | pretty sure we don't know :) | 22:57 |
| sdake | that is what this exercise was all about | 22:57 |
| sdake | finish up voting and we will select 3 top votes to merge | 22:58 |
| sdake | top 3 voted missions | 22:58 |
| bmace | to see the competition driven before us, and hear the lamentation their developers? | 22:58 |
| sdake | bmace huh? | 22:59 |
| bmace | conan reference? | 22:59 |
| bmace | :/ | 22:59 |
| sdake | folks if people could vote on SamYaplels proposal again, he fixed the wordness | 22:59 |
| *** jruano has quit IRC | 22:59 | |
| SamYaple | no need to revote i dont think, im no wordsmith. i like others better | 23:00 |
| SamYaple | they better capture what i mean | 23:00 |
| sdake | when you have finished voting can you pin in the channel | 23:00 |
| sdake | ping voting done | 23:00 |
| bmace | voting done | 23:01 |
| dasm | voting done | 23:02 |
| sdake | folks remember one of the rule we set for voting, if you dont' like the exact wording, but agee with the mission please vote +1 and we can edit as necessary | 23:02 |
| rhallisey | sdake, done | 23:03 |
| rhallisey | sdake, I have to go though | 23:03 |
| sdake | ack | 23:03 |
| rhallisey | I'll check back later | 23:03 |
| *** rhallisey has quit IRC | 23:03 | |
| jpeeler | same | 23:03 |
| sdake | we are done after voting :) | 23:03 |
| sdake | its my birday and after that I have to drink heaviy | 23:03 |
| pdb-home | done | 23:03 |
| daneyon_ | done | 23:03 |
| pdb-home | happy birthday | 23:03 |
| SamYaple | one foot in the grave now | 23:04 |
| daneyon_ | sdake happy birthday... finally you are the legal age to drink!!! | 23:04 |
| mstachow | voting done | 23:04 |
| bmace | happy bday | 23:04 |
| sdake | which cat is blahrus | 23:04 |
| pdb-home | adding a few +1s based on above rule | 23:04 |
| mstachow | sdake happy birthday ! | 23:04 |
| sdake | thanks ;-) | 23:04 |
| sdake | blahrus if your gonna vote once, you gotta vote on all the proposals :) | 23:05 |
| mstachow | Ok, meeting is over and I'm almost sleeping on my keyboard. Goodnight everyone ! (: | 23:05 |
| blahRus | sdake: I'm dog ;) | 23:05 |
| blahRus | sdake: I'll go through the rest now | 23:05 |
| sdake | blahrus thanks | 23:05 |
| loth | i think im done | 23:05 |
| sdake | thanks guys | 23:05 |
| sdake | I knwo that was painful | 23:06 |
| sdake | please don't edit from here | 23:06 |
| sdake | after you finish voting | 23:06 |
| sdake | I'd like folks to have a look over peoples comments | 23:06 |
| sdake | and votes | 23:06 |
| sdake | and understand what people think we should be doing | 23:06 |
| pdb-home | scanning the statements/comments, it seems people are a little unclear on the scope of the deployment tooling | 23:08 |
| bmace | the most disagreement seems to be pretty clearly around the role of deployment of containers in kolla. as far as i can tell most are for some sort of deployment / ha ability. | 23:08 |
| bmace | stop reading my mind pdb! ;) | 23:08 |
| dasm | so, are we there yet? which one has the most +1? | 23:08 |
| pdb-home | i.e. are we just a reference/demo/poc type thing with the intent to offload to third party, or will the kolla bundled tool be the recommended method | 23:08 |
| sdake | these are questions for the next meeting | 23:09 |
| bmace | well, i don't even know if we need to say "recommended" method, versus actually having "a" method, period. | 23:09 |
| sdake | just read and absorb | 23:09 |
| pdb-home | kk | 23:09 |
| blahRus | sdake: done | 23:10 |
| SamYaple | i like the idea of a working supported deployment too that can be consumed as a template for downstream projects | 23:11 |
| *** mstachow has quit IRC | 23:12 | |
| sdake | pdb why do you think i wanted to mention Ansible? | 23:12 |
| loth | I don't agree with mentioning specifics in a major mission statement, 6-12 months from now we may have chef/puppet/ansible methods | 23:13 |
| bmace | ansible is good, but to make something that actually works for deployment downstream projects would all need to do the same sort of custom work to seed the playbook data. | 23:13 |
| SamYaple | loth: im eyeing saltstack | 23:14 |
| loth | watch your sodium levels :) | 23:14 |
| SamYaple | thats saltlick, also delicious | 23:14 |
| pdb-home | im glad the mission statement is being looked at as it provides at least some level of stability regardless of underlying tech used | 23:15 |
| *** jtriley has quit IRC | 23:15 | |
| bmace | puppet is too heavy imho.. or at least heavier than is needed for the container approach. | 23:15 |
| SamYaple | pdb-home: thats a must for me "production" usage | 23:15 |
| dasm | guys, got to go. it's 1AM here. next meeting at wednesday? | 23:16 |
| *** dasm is now known as dasm|afk_ | 23:16 | |
| loth | re: "danehans -1 they can be used for dev, CI, etc.. too" IMO production usability implies dev usability etc | 23:16 |
| pdb-home | same, bed time. apologies if I was unfair to any mission statements, all were good and it was pretty hectic :) | 23:16 |
| *** pdb-home has quit IRC | 23:17 | |
| *** dasm|afk_ has quit IRC | 23:18 | |
| loth | missing some votes | 23:20 |
| sdake | goodnight folks that stayed up late - appreciate it :) | 23:28 |
| sdake | hope you enjoyed that - I thought it was fun :) | 23:29 |
| *** jtriley has joined #kolla | 23:29 | |
| SamYaple | see ya sdake | 23:29 |
| *** sdake has quit IRC | 23:35 | |
| *** vinkman has quit IRC | 23:44 | |
| *** dims_ has joined #kolla | 23:51 | |
| *** dims__ has quit IRC | 23:55 | |
| *** diga_ has joined #kolla | 23:56 | |
| *** blahRus has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!