| *** dimsum__ has quit IRC | 00:37 | |
| *** dims has joined #kolla | 00:44 | |
| *** sdake_ has quit IRC | 01:11 | |
| *** zhiwei has joined #kolla | 01:12 | |
| *** tobe has joined #kolla | 01:30 | |
| *** dylan has joined #kolla | 01:34 | |
| *** dims has quit IRC | 01:37 | |
| *** dims has joined #kolla | 01:38 | |
| *** erkules_ has joined #kolla | 01:42 | |
| *** dims has quit IRC | 01:43 | |
| *** erkules has quit IRC | 01:44 | |
| SamYaple | barely | 02:55 |
|---|---|---|
| *** sdake has joined #kolla | 03:01 | |
| *** sdake_ has joined #kolla | 03:03 | |
| *** sdake has quit IRC | 03:07 | |
| *** jpeeler has joined #kolla | 03:13 | |
| *** jpeeler has quit IRC | 03:14 | |
| *** jpeeler has joined #kolla | 03:15 | |
| *** diga has joined #kolla | 03:42 | |
| *** sdake_ has quit IRC | 04:11 | |
| *** sdake has joined #kolla | 04:14 | |
| *** tobe has quit IRC | 05:01 | |
| *** tobe has joined #kolla | 05:07 | |
| sdake | samyaple around? | 05:08 |
| SamYaple | yep | 05:09 |
| *** sdake_ has joined #kolla | 05:30 | |
| *** sdake_ has quit IRC | 05:33 | |
| *** sdake_ has joined #kolla | 05:33 | |
| *** sdake has quit IRC | 05:34 | |
| *** sdake_ is now known as sdake | 05:43 | |
| *** tobe has quit IRC | 05:44 | |
| *** tobe has joined #kolla | 06:03 | |
| *** sdake_ has joined #kolla | 06:10 | |
| *** dasm|afk is now known as dasm | 06:12 | |
| *** sdake has quit IRC | 06:14 | |
| *** RuiChen has left #kolla | 06:15 | |
| *** sdake has joined #kolla | 06:21 | |
| *** sdake_ has quit IRC | 06:25 | |
| sdake | kolla is only 2300 lines of code | 06:40 |
| sdake | crazy what it can do with such a small code base | 06:41 |
| *** daneyon has joined #kolla | 06:50 | |
| *** inc0 has joined #kolla | 06:50 | |
| *** daneyon has quit IRC | 07:00 | |
| *** daneyon has joined #kolla | 07:01 | |
| *** daneyon has quit IRC | 07:06 | |
| *** sdake_ has joined #kolla | 07:07 | |
| *** sdake has quit IRC | 07:10 | |
| *** sdake_ is now known as sdake | 07:16 | |
| *** gfidente has joined #kolla | 07:26 | |
| *** gfidente has joined #kolla | 07:26 | |
| *** tobe has quit IRC | 07:37 | |
| *** tobe has joined #kolla | 07:39 | |
| *** dylan has quit IRC | 07:43 | |
| *** mstachow has joined #kolla | 07:51 | |
| *** zhiwei has quit IRC | 08:10 | |
| *** sdake has quit IRC | 08:16 | |
| *** sdake has joined #kolla | 08:18 | |
| *** erkules_ is now known as erkules | 08:20 | |
| *** erkules has joined #kolla | 08:20 | |
| *** tobe has quit IRC | 08:22 | |
| *** tobe has joined #kolla | 08:23 | |
| *** athomas has joined #kolla | 08:35 | |
| *** mstachow has quit IRC | 08:45 | |
| openstackgerrit | weiyu proposed stackforge/kolla: docker-compose version is 1.3.0 https://review.openstack.org/187136 | 08:45 |
| *** sdake_ has joined #kolla | 09:16 | |
| *** sdake_ has quit IRC | 09:17 | |
| *** sdake_ has joined #kolla | 09:18 | |
| *** sdake has quit IRC | 09:20 | |
| *** sdake_ is now known as sdake | 09:20 | |
| *** pdb has quit IRC | 09:41 | |
| *** pdb has joined #kolla | 09:41 | |
| *** dims has joined #kolla | 09:55 | |
| *** openstackstatus has quit IRC | 10:20 | |
| *** openstack has quit IRC | 10:20 | |
| *** openstack has joined #kolla | 10:26 | |
| harmw | sdake: I probably won't be able to make it wednesday, if I'm lucky I can make it for the last 30 minutes though | 10:39 |
| sdake | harmw ok | 10:39 |
| sdake | samyaple if you want to review harmw's great work in designate - https://review.openstack.org/#/c/182799/ | 10:51 |
| sdake | or mandre if your around :) | 10:51 |
| SamYaple | sdake: i started digging into it, but its a good amount and i cant test it at the moment | 10:52 |
| sdake | no need to test | 10:52 |
| SamYaple | i can give it a +1 now, but i havent tested it | 10:52 |
| sdake | +2 doesn't require testing | 10:52 |
| sdake | it requires reviewing | 10:52 |
| sdake | our gate is supposed to do the testing | 10:53 |
| sdake | it doesn't atm | 10:53 |
| SamYaple | test that it builds vs test that it functions? | 10:53 |
| sdake | but it will | 10:53 |
| sdake | no testing at all required for +2/+a | 10:53 |
| sdake | just reviewing | 10:53 |
| sdake | guarding the gate = 3 human beings + automated testing | 10:53 |
| sdake | our automated testing is sort of weak atm | 10:54 |
| sdake | but our gate does test that it builds | 10:54 |
| SamYaple | ok so it wont break the gate, but are we merging things without actually testing theM/ | 10:54 |
| sdake | core reviewer's job is to guard the gate | 10:54 |
| SamYaple | functionally? | 10:54 |
| sdake | it is not mandatory for a +2 review | 10:54 |
| sdake | I don't test every review | 10:54 |
| sdake | I didn't test designate, I inspected thatit looks correct | 10:55 |
| SamYaple | fair enough. il lgive it a once over again. i review it last night | 10:55 |
| sdake | we have to trust our developers that they are doing the righ tthing | 10:55 |
| SamYaple | oh i trust them | 10:55 |
| SamYaple | i just verify | 10:55 |
| sdake | that is the job of the ci system | 10:55 |
| sdake | it doesn't do a very good job of it yet | 10:55 |
| sdake | its really a personal choice | 10:56 |
| sdake | but if you test every change you will never get reviews done and our review queue will be outrageously long :) | 10:56 |
| sdake | jpeeler is going to get the gate rolling as soon as we get single node NIC working | 10:56 |
| harmw | isn't there something to do fucntional tests in openstack? | 10:56 |
| SamYaple | harmw: DESIGNATE_BIND9_RNDC_KEY="Fkt8Qbm6nDXDcODfEeMe3g==" | 10:57 |
| sdake | that is a better place to spend time then manual testing | 10:57 |
| SamYaple | whats the base64 string? | 10:57 |
| sdake | samyaple ya i didn't like that either | 10:57 |
| sdake | harmw there are functioanl tests in each project's repo | 10:57 |
| harmw | ah, true | 10:57 |
| sdake | what we shoudl be doing in the gate is pulling the repo's tree, and running the functional tests | 10:57 |
| harmw | SamYaple: ? | 10:57 |
| sdake | but we dont do that because we only have 1 nic and we require 2 for neutron | 10:58 |
| SamYaple | harmw: why are you using a base64 string there? | 10:58 |
| harmw | SamYaple: I'd have to lookup what kind of format is required, I think I currently did hmac-md5 or something | 10:58 |
| SamYaple | right but shouldnt that be generated | 10:58 |
| sdake | make suggestions in the review, +2 or -1 so other reviewers can comment as well if they care ;) | 10:59 |
| harmw | well, since it's not used in it's original form anywhere I opted to just include the final key | 10:59 |
| SamYaple | i will comment in the review with the IRC conversation | 10:59 |
| sdake | if you +2, remember your the second reviewer so +w as well | 10:59 |
| sdake | sounds good | 10:59 |
| sdake | as long as what happens on irc makes it into gerrit :) | 10:59 |
| sdake | keeping context with the irc logs is hard :) | 11:00 |
| harmw | yep | 11:00 |
| SamYaple | agreed | 11:00 |
| SamYaple | harmw: the POOLID/TARGETS/NSS should be in the tools/genenv script, no? | 11:01 |
| SamYaple | https://review.openstack.org/#/c/182799/8/docker/centos/binary/designate/designate-poolmanager/start.sh,unified | 11:01 |
| harmw | they are hardcoded because upstream doesn't realy know what else to do currently | 11:01 |
| harmw | and since tools/genenv generates stuff that the user can configure, I didn't find this fitted in there | 11:01 |
| SamYaple | so thats an upstream problem, those arent configurable? | 11:01 |
| harmw | you can use other UUIDS but that wouldn't make any difference | 11:02 |
| harmw | so better not trouble $operator with those vars (is what I thought) | 11:02 |
| SamYaple | are those used in anyway as a secretkey of sorts? | 11:03 |
| harmw | not configuring the Sink service falls in that category as well, since upstream doesn't know what they want in that direction either | 11:03 |
| harmw | no secrets that I know off, you just need some UUID to have poolmanager create some stuff | 11:04 |
| harmw | so they show up in designate <somecommand-related-to-pool-manager> | 11:04 |
| SamYaple | since those are configurable I would prefer to have those randomized during the environment generation | 11:06 |
| harmw | argh | 11:07 |
| SamYaple | what does that break for oyu? | 11:07 |
| harmw | nothing, just need to add another review :) | 11:07 |
| SamYaple | :/ yea sorry | 11:07 |
| harmw | (and sucks I didn't think of it myself) | 11:07 |
| harmw | please leave notes in Gerrit though, Ill address them tonight | 11:08 |
| SamYaple | my worry is there is a security/conflict implication here i dont fully understand and i want to err on the side of cuation | 11:08 |
| SamYaple | i will | 11:08 |
| sdake | samyaple I had the same comment :) | 11:13 |
| sdake | but harmw convinced me it didn't matter | 11:14 |
| sdake | and its not a mainstream container, so I don't think its totally critical :) | 11:14 |
| sdake | mainstream being nova heat keystone mariadb glance and that other required container | 11:15 |
| sdake | brain fried | 11:15 |
| sdake | 5am ftl | 11:15 |
| SamYaple | im just worried aobut security/information leakage | 11:16 |
| SamYaple | could some attack or something be performed if oy uknow the uuid | 11:16 |
| SamYaple | i dont know since i havent dug into it | 11:16 |
| harmw | SamYaple: I will fix it tonight, even if this regards a non-mainstream (or core) container it should be solid | 11:16 |
| SamYaple | if we auto generate them its not a question | 11:16 |
| harmw | (just don't -1 again b/c it'll mess up my stats!) | 11:17 |
| harmw | :p | 11:17 |
| SamYaple | yea harmw it looks really solid other than that | 11:17 |
| SamYaple | im gonna | 11:17 |
| SamYaple | yo ucant stop meeee | 11:17 |
| harmw | damn thee, holder of superpowers | 11:17 |
| SamYaple | nah ill just leave comments | 11:17 |
| sdake | -1 are good for your stats harmw | 11:18 |
| harmw | hehe, cool :) Ill get it fixed in some 6/7hrs from now | 11:18 |
| sdake | you should -1 reviews that need changes | 11:18 |
| SamYaple | is DESIGNATE_BIND9_RNDC_KEY="Fkt8Qbm6nDXDcODfEeMe3g==" also unique to the environmet? | 11:19 |
| sdake | dont worry about stats on stackalytics | 11:19 |
| sdake | we are going to be moving very fast, our combined stats will all be in top 10% | 11:19 |
| harmw | sdake: no I don't, but stats are fun in general :> and I constantly -1 and +1 stuff | 11:19 |
| harmw | SamYaple: yes, thats why it is in genenv so $operator nknows he can (should) change it according to his env | 11:19 |
| harmw | but perhaps making it easier to get something else in there would be better | 11:20 |
| SamYaple | lets geenrate that as well | 11:20 |
| harmw | eg. add in someword and have the container translate that to base64 or whatever | 11:20 |
| SamYaple | anything generatable lets generate it | 11:20 |
| SamYaple | uuidgen | base64 | 11:20 |
| harmw | true, but what in this case I want to know what it reads so I could hook up another rndc session | 11:21 |
| harmw | yes, uuidgen :) | 11:21 |
| SamYaple | you have it stored in the variable | 11:21 |
| sdake | my stats overall for openstack are top 2% in commits ;-) | 11:21 |
| sdake | for the lifetime of openstack | 11:21 |
| harmw | showoff | 11:22 |
| * sdake gloats a bit ;) | 11:22 | |
| harmw | :P | 11:22 |
| SamYaple | if oyu need the origianly value back "echo $DESIGNATE_BIND9_RNDC_KEY | base64 -d` | 11:22 |
| sdake | I think I'm #87 out of 4500 contribs | 11:22 |
| harmw | SamYaple: true, but given ease of use I should have set the var to <blabla> instead of the base64 smuck | 11:22 |
| harmw | oh, no | 11:23 |
| sdake | stackalytics is all about getting you guys to work harder :) | 11:23 |
| harmw | aha | 11:23 |
| harmw | SamYaple: got it :) | 11:23 |
| SamYaple | whatever you want to do is cool with me as long as we can geenerate random variables | 11:23 |
| sdake | there, the secret is out :) | 11:23 |
| *** tobe has quit IRC | 11:25 | |
| sdake | hey guys diga is new to the kolla community, but I think he will be working on kolla from an ansible POV | 11:26 |
| sdake | he did alot of good work on magnum | 11:26 |
| sdake | not sure if he is still around, but if he is, say hi :) | 11:26 |
| harmw | hi! | 11:26 |
| SamYaple | can use all the help we can get, he diga | 11:26 |
| SamYaple | hey* | 11:26 |
| sdake | t-30 minutes until 4 hours of pointless agile training | 11:27 |
| harmw | then why do it? | 11:28 |
| sdake | not many things in my job are mandatory | 11:28 |
| SamYaple | to flow | 11:28 |
| sdake | this is one of em | 11:28 |
| sdake | agile = more trackers to deal with | 11:28 |
| sdake | as if I don't have enough already | 11:29 |
| harmw | https://review.openstack.org/#/c/182799/8/docker/centos/binary/designate/designate-backend-bind9-data/Dockerfile | 11:29 |
| harmw | SamYaple: that -is- the datacontainer | 11:29 |
| harmw | it stores all zonefiles | 11:29 |
| sdake | hey samyaple does your ceph container work in yaodu | 11:29 |
| SamYaple | yes indeed | 11:30 |
| sdake | how is the performance of nova boot | 11:30 |
| SamYaple | as fast as can beeeeeeee | 11:30 |
| sdake | in data containers, its pokey | 11:30 |
| sdake | doesit work single node? | 11:30 |
| SamYaple | harmw: right, but does that need to stick around if yo uremove the container? | 11:30 |
| SamYaple | my concern is on a restart/recreate of the container youll lose those zonefiles and they need to be persistent | 11:31 |
| SamYaple | if they get regenerated thats fine | 11:31 |
| SamYaple | sdake: it can | 11:31 |
| SamYaple | yaodu as a whole doesnt work single node | 11:32 |
| harmw | those zonefiles need be persistent, yes, they are stored in sql as well (per reference) | 11:32 |
| SamYaple | ceph can be setup as AIO, but its a different path | 11:32 |
| SamYaple | if they need ot be persistent they would be in a data container I believe, sdake can yo uconfirm how kolla wants this done? | 11:32 |
| sdake | t-5 months from now I think we need to get rid of data containers atleast for nova instance data | 11:32 |
| sdake | if persistent, should be in data container | 11:33 |
| harmw | so data containers are even more magic then I did here? | 11:33 |
| SamYaple | just an empty container that has that volume and runs /bin/true | 11:33 |
| sdake | your data container is right | 11:33 |
| sdake | but your running /bin/bash | 11:33 |
| sdake | it should run /bin/true | 11:34 |
| harmw | aha | 11:34 |
| SamYaple | yea | 11:34 |
| harmw | roger | 11:34 |
| sdake | I am pretty sure I commented on that | 11:34 |
| sdake | but maybe not | 11:34 |
| SamYaple | sorry should have been more clear | 11:34 |
| sdake | anything with a VOLUME tag will be persistent | 11:34 |
| harmw | sdake: can't remember if you did, but Ill fix it up tonight | 11:34 |
| harmw | yea I thought so, already :) | 11:34 |
| SamYaple | i hear docker may implement volume management which owuld be sweet | 11:34 |
| SamYaple | ie volumes sperate from containers | 11:34 |
| sdake | that would be nice | 11:34 |
| sdake | maybe nova boot wouldnt' be so pokey | 11:35 |
| SamYaple | seen a few commits about it | 11:35 |
| SamYaple | nah it will be slow still | 11:35 |
| harmw | guys, I'll be back tonight for more funtimes :) | 11:35 |
| sdake | enjoy | 11:35 |
| SamYaple | ephemeral in a container will be slow | 11:35 |
| sdake | i'll probablybe PTFOED | 11:35 |
| SamYaple | sdake to fix that you need to do something a bit different | 11:36 |
| SamYaple | its a PITA | 11:36 |
| SamYaple | and oyu wont like it | 11:36 |
| sdake | i'e got a magnum demo at 11am then I'm crashing | 11:36 |
| SamYaple | you should crash | 11:36 |
| sdake | i have training | 11:36 |
| sdake | i can't crash until 12am | 11:36 |
| sdake | 12pm | 11:36 |
| sdake | t-20 minutes until "learn how to use a tracker 101" | 11:36 |
| *** rhallisey has joined #kolla | 11:58 | |
| *** mstachow has joined #kolla | 12:03 | |
| *** dwalsh has joined #kolla | 12:05 | |
| mstachow | hello (: | 12:11 |
| SamYaple | hello mstachow | 12:11 |
| SamYaple | how are you today | 12:12 |
| inc0 | he looked alive when I was there, but since he's kinda silent now...maybe I'll go check | 12:15 |
| mstachow | I'm fine :D and how are You Sam :>? | 12:22 |
| mstachow | @inc0 don'w worry :P I'm cleaning up code and in next few minutes I will upload galera | 12:24 |
| inc0 | cool! | 12:24 |
| inc0 | you were complaining about heart problems, so that got me concerned;) | 12:24 |
| mstachow | I'm happy because now is OK | 12:25 |
| SamYaple | tired mstachow tired | 12:25 |
| dasm | mstachow: heart problems? should i call emergency? | 12:25 |
| mstachow | I have no idea what was that, but I'm in hope that It won't turn back | 12:25 |
| dasm | mstachow: maybe girlfriend? | 12:26 |
| mstachow | @dasm It's not necessary but thank You ;) | 12:26 |
| mstachow | @SamYaple That's sad :< Maybe You need some vacation ? | 12:27 |
| *** ChanServ sets mode: +o rhallisey | 12:30 | |
| rhallisey | morning | 12:30 |
| mstachow | Hi ! :) | 12:31 |
| inc0 | good morning rhallisey | 12:33 |
| SamYaple | hey rhallisey | 12:35 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder container https://review.openstack.org/170965 | 12:50 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add cinder into the docker compose setup https://review.openstack.org/173507 | 12:56 |
| inc0 | rhallisey, did you make bp for pushing conf file to container? | 13:01 |
| inc0 | if not, mind if we do this? | 13:01 |
| rhallisey | SamYaple, do you have a bp for this ^ | 13:02 |
| rhallisey | that is, allowing config to be done by passing in a config file | 13:02 |
| SamYaple | rhallisey: i have a blueprint for dynamic config generation | 13:02 |
| inc0 | I think this is a bit different | 13:03 |
| SamYaple | inc0: rhallisey so the config is changed to an environment variable and parsed back out in the container | 13:03 |
| SamYaple | inc0: are yo utalking about updating the config file live? | 13:03 |
| inc0 | SamYaple, not exactly, I'm talking about deploying container with --config-file=my_own_cinder.conf | 13:03 |
| inc0 | and content of my_own_cinder.conf will land in /etc/cinder/cinder.conf in container | 13:04 |
| inc0 | so instead of generating big, fat config each time with hundreds of env vars, you just push a file | 13:05 |
| inc0 | we don't care where this file will come from | 13:06 |
| inc0 | anyway, we'll go ahead and post this bp | 13:09 |
| inc0 | this will be useful for tripleo | 13:09 |
| SamYaple | inc0: that will cver it, hold ill walk you through it | 13:10 |
| inc0 | ok, please do | 13:10 |
| SamYaple | sorry inc0 i was in the middle of a conversation | 13:14 |
| SamYaple | basically when i say it will generate an environment variable, i mean a single one | 13:15 |
| SamYaple | ill show you my WIP https://review.openstack.org/#/c/182168/ | 13:15 |
| SamYaple | https://gist.github.com/SamYaple/444e20b067695e8ac838 | 13:15 |
| inc0 | so basically you read .ini file, turn it to envs, and build new ini file based on these envs? | 13:19 |
| SamYaple | inc0: yes | 13:20 |
| inc0 | hmm...and it needs ansible to run...which doesn't help me that much:/ | 13:21 |
| SamYaple | i actually read it with configparser and generate a dict and load it back into config parser | 13:21 |
| *** jtriley has joined #kolla | 13:21 | |
| SamYaple | inc0: for the WIP yes | 13:21 |
| SamYaple | i will be builging support for genenv | 13:21 |
| inc0 | need help with that? This functionality is kinda important to us | 13:22 |
| SamYaple | it is a blueprint that should be done by liberty-1 | 13:22 |
| *** dasm is now known as dasm|afk | 13:22 | |
| inc0 | ok | 13:22 |
| *** sami_ has joined #kolla | 13:24 | |
| *** dwalsh_ has joined #kolla | 13:28 | |
| *** inc0 has quit IRC | 13:29 | |
| *** jtriley has quit IRC | 13:33 | |
| *** jtriley has joined #kolla | 13:34 | |
| diga | Hi sdake | 13:35 |
| diga | hello All | 13:35 |
| diga | sorry I was away for sometime due to some work | 13:36 |
| rhallisey | diga, hey | 13:39 |
| diga | hi rhallisey | 13:39 |
| diga | sdake pinged me sometime ago, but I missed due to somework | 13:40 |
| SamYaple | we were saying hey | 13:40 |
| SamYaple | hey! | 13:40 |
| diga | Hi SamYaple | 13:41 |
| diga | SamYaple: I am reading ansible for now | 13:45 |
| diga | will be in good shape to contribute before our wednesday's meeting | 13:45 |
| SamYaple | diga: there isnt much atm | 13:45 |
| SamYaple | trying to finalize the way we build it | 13:46 |
| diga | SamYaple: do you mean we dont have enough left on the ansible side ?? | 13:47 |
| *** blahRus has joined #kolla | 13:47 | |
| SamYaple | diga: no i mean not alot of code :) | 13:48 |
| SamYaple | lots of work dont worry :D | 13:48 |
| diga | :) | 13:48 |
| diga | Good | 13:48 |
| sdake | hey diga, I was just introducing you to the team :) | 13:51 |
| diga | I am setting up kolla locally on my vm | 13:51 |
| sdake | or part of the team | 13:51 |
| diga | sdake: Okk :) | 13:52 |
| SamYaple | debian 8 defaulted to systemd | 13:52 |
| SamYaple | i like systemd | 13:52 |
| diga | I also like systemd | 13:54 |
| SamYaple | i like networking in systemd | 13:55 |
| diga | :) | 13:55 |
| SamYaple | also journal logging, though that took a while to get used to | 13:55 |
| *** dwalsh_ has quit IRC | 13:55 | |
| diga | hmm | 13:55 |
| diga | going to visit hospital, not feeling well today | 13:56 |
| diga | will be back in an hour | 13:57 |
| sdake | diga rhaillisey is part of the core team too, say hi when you get back :) | 13:57 |
| SamYaple | 13:39 <@rhallisey> diga, hey | 13:58 |
| SamYaple | 13:39 < diga> hi rhallisey | 13:58 |
| SamYaple | sdake ^^ | 13:58 |
| sdake | nice :) | 13:58 |
| *** jruano has joined #kolla | 14:01 | |
| *** jtriley has quit IRC | 14:11 | |
| *** jtriley has joined #kolla | 14:13 | |
| sdake | diga the code base for ansible is about 1kloc - but its alot to digest | 14:19 |
| *** jtriley has quit IRC | 14:27 | |
| *** jtriley has joined #kolla | 14:27 | |
| *** v_b has joined #kolla | 14:42 | |
| *** sami_ has quit IRC | 14:42 | |
| *** sami_ has joined #kolla | 14:45 | |
| *** sami_ has quit IRC | 14:50 | |
| *** absubram has quit IRC | 14:50 | |
| openstackgerrit | Michal Stachowski proposed stackforge/kolla: Galera container https://review.openstack.org/187225 | 14:59 |
| sdake | nice work michal! | 14:59 |
| *** dwalsh has quit IRC | 15:00 | |
| mstachow | @sdake thanks :D | 15:00 |
| mstachow | please review it ! :) It's time for feedback | 15:00 |
| mstachow | ;) | 15:00 |
| sdake | in training atm | 15:02 |
| SamYaple | mandre: whats up with the command for galeradbdata? | 15:02 |
| sdake | when i wrap up i'll get to it | 15:02 |
| SamYaple | mandre: ignore | 15:02 |
| SamYaple | mstachow: ^^ | 15:02 |
| SamYaple | command: cat /dev/random > /dev/null' | 15:02 |
| mstachow | yeah, my fault. On my workspace galeradb-data shuts down few seconds after spawning | 15:02 |
| *** sami_ has joined #kolla | 15:03 | |
| SamYaple | its cool, leaving notes | 15:03 |
| mstachow | this command prevents this | 15:03 |
| SamYaple | mstachow: its supposed to | 15:04 |
| SamYaple | thats the design | 15:04 |
| SamYaple | its just a placeholder | 15:04 |
| *** jtriley has quit IRC | 15:05 | |
| SamYaple | if you have other things to add and it not read for merge, mark it WIP mstachow | 15:05 |
| SamYaple | looking at your git message is see TODO | 15:05 |
| mstachow | sure, it's my first commit into community | 15:05 |
| mstachow | It's more stressful than I thought :S | 15:06 |
| SamYaple | mstachow: its a good one :D | 15:06 |
| SamYaple | thank you | 15:06 |
| mstachow | heh, it's nice to read this :) | 15:07 |
| SamYaple | mstachow: you know how to change and resubmit a patch? | 15:10 |
| *** jtriley has joined #kolla | 15:11 | |
| mstachow | I have no idea ... yet | 15:12 |
| SamYaple | just make your changes, then `git commmit -a --amend` | 15:12 |
| SamYaple | afterward `git review` | 15:12 |
| SamYaple | thats it | 15:12 |
| mstachow | thanks ! next commit within few minutes :) | 15:14 |
| diga | Hi rhallisey | 15:16 |
| rhallisey | diga, hey whats up? | 15:16 |
| diga | sdake told to introduce my self to you | 15:18 |
| rhallisey | nice to meet you! | 15:18 |
| diga | :) | 15:19 |
| openstackgerrit | Michal Stachowski proposed stackforge/kolla: Galera container https://review.openstack.org/187225 | 15:21 |
| *** jtriley has quit IRC | 15:29 | |
| SamYaple | hello mstachow, you around? | 15:30 |
| mstachow | yes, I am here | 15:30 |
| SamYaple | are you planning on extending this patch, or trying to get it merged as-is? | 15:30 |
| mstachow | really - I'm not sure. If we need rsync as a sst method, probably I will change it. In my opinion we can do this without rsync. We've got working mysqldump, xtremebackup and new xtremebackup(v2) | 15:33 |
| mstachow | but this is only my opinion | 15:33 |
| SamYaple | well as-is it is missing changes to genenv | 15:33 |
| SamYaple | and you have TODO in your commit message | 15:33 |
| *** jtriley has joined #kolla | 15:33 | |
| mstachow | I will add this envs to genenv - that's no problem for me. I am going to do tests on fedora and this bullet will also could be removed from TODO list | 15:36 |
| SamYaple | ok | 15:36 |
| *** daneyon has joined #kolla | 15:36 | |
| *** dims has quit IRC | 15:41 | |
| mstachow | @Sam Do I must add extra genenv properties for non-required envs? | 15:42 |
| *** dims_ has joined #kolla | 15:42 | |
| *** bmace_ has joined #kolla | 15:42 | |
| SamYaple | thats a good question mstachow | 15:43 |
| SamYaple | i say yes since the options wont be documented anywhere else | 15:43 |
| SamYaple | i am open to suggestions though | 15:44 |
| SamYaple | sdake: opinion? ^^ | 15:44 |
| *** bmace has quit IRC | 15:45 | |
| *** absubram has joined #kolla | 15:54 | |
| mstachow | maybe it's good to select most important envs and expose them by genenv | 15:56 |
| *** dwalsh has joined #kolla | 15:56 | |
| *** jtriley has quit IRC | 15:56 | |
| mstachow | I won't expose every single option, but also every envs should be documented | 15:58 |
| sdake | jesus that was painful - not as bad as the dentist | 16:01 |
| mstachow | oh sorry, maybe That was too hard(I'm not English native user and my English skills aren't great :<) | 16:04 |
| mstachow | but | 16:04 |
| sdake | I was talking about training session I just suffered through | 16:04 |
| mstachow | oh, ok I see :D I don't like them too :) | 16:05 |
| *** mstachow has quit IRC | 16:05 | |
| *** mstachow has joined #kolla | 16:05 | |
| *** jtriley has joined #kolla | 16:05 | |
| *** dasm|afk is now known as dasm | 16:12 | |
| sdake | mr samyaple still awake? | 16:21 |
| harmw | sdake: good to see you survived :p | 16:22 |
| sdake | ya now I need some serious r&r after that | 16:23 |
| *** dwalsh has quit IRC | 16:32 | |
| *** dasm is now known as dasm|afk | 16:37 | |
| sdake | mstachow which non-required envs? | 16:37 |
| sdake | an environment variable is either required or not required | 16:37 |
| sdake | either it must be defined or it does not exist | 16:37 |
| sdake | mstachow we are taking optional enviornment variables out of our containers globally | 16:38 |
| *** mstachow has quit IRC | 16:38 | |
| harmw | hm, taking a deep-dive into Bind9 sources just to find out what makes up that rndc key bit :| | 16:40 |
| harmw | entropy, duh | 16:43 |
| sdake | nice first contribution from mstachow ! :) | 16:47 |
| harmw | which is? | 16:47 |
| sdake | the mysql ha container | 16:47 |
| harmw | oh cool! | 16:47 |
| harmw | I noticed another cinder review can be done, ill have a look at that galera thingy after that | 16:48 |
| harmw | all after designate is done, that is :p | 16:48 |
| *** gfidente has quit IRC | 16:48 | |
| *** gfidente has joined #kolla | 16:48 | |
| *** gfidente has joined #kolla | 16:48 | |
| SamYaple | sdake: im still at work even! | 16:54 |
| sdake | wow | 16:54 |
| rhallisey | harmw, I'm going to change it a little it's not complete | 16:57 |
| harmw | so I'd rather wait with reviewing? | 16:58 |
| rhallisey | ya I'll update it again in a minute | 16:59 |
| rhallisey | need to test removing tgt stuff | 16:59 |
| jpeeler | sdake: do you know how far out sorting the networking with one interface is? | 16:59 |
| sdake | samyaple is on that | 17:00 |
| sdake | I don't think he has started | 17:00 |
| jpeeler | SamYaple: status update? | 17:00 |
| SamYaple | jpeeler: it works for me when using subinterfaces like vlan (even if the traffic never leaves the node) or vxlan ones | 17:01 |
| jpeeler | is it too early to put up a review? | 17:03 |
| SamYaple | it is in manual form at the moment. not code worthy | 17:04 |
| *** dwalsh has joined #kolla | 17:05 | |
| * jpeeler would love to get to a point of not having to install openstack to work on kolla | 17:06 | |
| SamYaple | jpeeler: :) | 17:06 |
| sdake | well samyaple can always stay up another 24 hours ;) | 17:06 |
| sdake | lol | 17:06 |
| SamYaple | sorry a bit slow at the moment for lots of reasons | 17:07 |
| sdake | man I need to hit the rack | 17:07 |
| sdake | jpeeler why do you ahve o install openstack? | 17:08 |
| harmw | I was going to ask the same thing :) | 17:08 |
| jpeeler | well, since the networking work is not done (and i do NOT understand openstack networking), i thought getting everything up and going required the heat generated devenv | 17:09 |
| * harmw just uses a coreos vm and communicates with that using DOCKER_HOST= | 17:09 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Add the Designate DNSaaS services. https://review.openstack.org/182799 | 17:11 |
| jpeeler | hrm, maybe i'll give it a go then on baremetal. i just didn't think it would work quite right yet | 17:12 |
| sdake | it does work if you have two networking interfaces and both have external routing ;-) | 17:12 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add cinder into the docker compose setup https://review.openstack.org/173507 | 17:12 |
| jpeeler | sdake: i do on this machine apparently, but normally i wouldn't. you just mean external routing on the first nic right? | 17:13 |
| sdake | both nics | 17:14 |
| sdake | its hard to get setup | 17:14 |
| sdake | the heat template does the job | 17:14 |
| jpeeler | yeah i was just going to ask if there are any docs. probably best to just do the openstack thing then until the network changes are completed | 17:14 |
| sdake | i am hopeful the networking changes will jsut be some kind of shell script people can run | 17:15 |
| sdake | rather then changes to the container | 17:15 |
| jpeeler | right | 17:15 |
| sdake | samyaple thoughts? | 17:15 |
| SamYaple | i cannot think at the moment | 17:16 |
| SamYaple | trying to finish up task before crash | 17:16 |
| sdake | changing containers? | 17:16 |
| sdake | for single nic? | 17:16 |
| sdake | if you have hi the wall go to bed :) | 17:16 |
| SamYaple | should be able to do it witout container change | 17:16 |
| sdake | nice | 17:16 |
| sdake | please do :) | 17:16 |
| *** mstachow has joined #kolla | 17:33 | |
| *** sdake_ has joined #kolla | 17:34 | |
| *** dwalsh has quit IRC | 17:35 | |
| *** sdake has quit IRC | 17:38 | |
| *** rhallisey has quit IRC | 17:50 | |
| *** bmace_ is now known as bmace | 17:54 | |
| *** rhallisey has joined #kolla | 17:55 | |
| harmw | SamYaple: sdake_, could you guys checkout designate a final time? | 17:57 |
| harmw | rhallisey: https://review.openstack.org/#/c/173507/ | 17:58 |
| harmw | what happens should that get approved before your massive cinder changeset? | 17:58 |
| rhallisey | I'll -1 it | 17:58 |
| harmw | (eg. I think I like those 2 changesets in 1) | 17:58 |
| rhallisey | it won't really make a big difference though | 17:59 |
| harmw | it would put some (at that point) useless stuff in the integration guide and tools/genenv, right? | 18:00 |
| rhallisey | ya that patch just updates the tools | 18:02 |
| rhallisey | it won't break anything | 18:02 |
| harmw | not break, no, but functionally it belongs with the other cinder work (can't have one without the other, since that would look odd) | 18:03 |
| rhallisey | ya I know, it should depend on it | 18:04 |
| harmw | (or be included :p) | 18:04 |
| harmw | can we expect another big one from cinder btw? | 18:04 |
| rhallisey | I'm about to update the bigger cinder patch | 18:04 |
| rhallisey | in a moment | 18:04 |
| harmw | wicked | 18:04 |
| rhallisey | trying to remove tgt | 18:04 |
| rhallisey | since it's unused | 18:04 |
| rhallisey | and anything else | 18:05 |
| rhallisey | harmw, thanks for the review it pointed out a lot of old stuff I had lying around | 18:06 |
| harmw | I noticed LVM57, is that number intentionaly? | 18:06 |
| harmw | np :) | 18:06 |
| *** sami_ has quit IRC | 18:08 | |
| rhallisey | ya thats the default | 18:11 |
| rhallisey | I think | 18:11 |
| rhallisey | from what I've seen | 18:11 |
| harmw | any idea whre the 57 comes from? | 18:11 |
| rhallisey | no clue | 18:11 |
| *** dwalsh has joined #kolla | 18:26 | |
| *** athomas_ has quit IRC | 18:26 | |
| *** jpeeler has quit IRC | 18:39 | |
| *** sami_ has joined #kolla | 18:42 | |
| *** jpeeler has joined #kolla | 18:48 | |
| *** jpeeler has quit IRC | 18:48 | |
| *** jpeeler has joined #kolla | 18:48 | |
| openstackgerrit | Michal Stachowski proposed stackforge/kolla: WIP: Galera container https://review.openstack.org/187225 | 18:54 |
| *** mstachow has quit IRC | 18:58 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder container https://review.openstack.org/170965 | 19:00 |
| harmw | ah time to get to work again :> | 19:06 |
| rhallisey | harmw, ya I pulled out supervisord | 19:07 |
| rhallisey | I'm going to next try and extract iscsi | 19:07 |
| harmw | ah yes, I didn't even comment on that :) | 19:07 |
| harmw | *being there | 19:07 |
| *** dwalsh has quit IRC | 19:07 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder container https://review.openstack.org/170965 | 19:16 |
| rhallisey | ignore that commit, just changed the commit msg | 19:16 |
| *** mstachow has joined #kolla | 19:27 | |
| *** dwalsh has joined #kolla | 19:32 | |
| harmw | sdake_: I have a comment in https://review.openstack.org/#/c/187225/ asking for other ppls opinions :) (sux I can't link to it) | 19:38 |
| *** daneyon_ has joined #kolla | 19:39 | |
| harmw | rhallisey: that Cinder patch... Its scaring me | 19:39 |
| *** daneyon has quit IRC | 19:42 | |
| rhallisey | harmw, what about it is scaring you? | 19:43 |
| *** dperaza has joined #kolla | 19:46 | |
| Slower | harmw: it is scary | 19:48 |
| Slower | harmw: I still wonder what ceph would look like and if it would be less scary | 19:48 |
| Slower | we may need to actually have cinder-volumes-lvm cinder-volumes-ceph etc | 19:48 |
| Slower | just because of the different requirements for mounts etc | 19:48 |
| harmw | rhallisey: its size :> | 19:50 |
| harmw | Slower: yup, something like that seems appropriate | 19:51 |
| harmw | having different containers per backend (that what designate does as well) | 19:51 |
| *** sdake_ has quit IRC | 19:55 | |
| *** dperaza has quit IRC | 19:57 | |
| *** dwalsh has quit IRC | 19:57 | |
| harmw | rhallisey: poke me tomorrow around 3 hours before now and Ill review it :) | 19:57 |
| rhallisey | harmw, sure! | 19:57 |
| rhallisey | thanks | 19:57 |
| harmw | np, I just need to powerup before going over it again :> (plus, that nice Galera stuff was tiring as well) | 19:59 |
| *** dperaza has joined #kolla | 19:59 | |
| *** mstachow has quit IRC | 20:00 | |
| *** dwalsh has joined #kolla | 20:09 | |
| *** dperaza has left #kolla | 20:11 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Access volume groups created by Cinder from the host https://review.openstack.org/187313 | 20:18 |
| *** jruano has quit IRC | 20:21 | |
| *** jpeeler has quit IRC | 20:42 | |
| *** jruano has joined #kolla | 20:42 | |
| *** jpeeler has joined #kolla | 20:47 | |
| *** rhallisey has quit IRC | 20:51 | |
| *** jruano has quit IRC | 21:04 | |
| *** sdake has joined #kolla | 21:15 | |
| sdake | hey guys | 21:30 |
| sdake | bck from a power nap | 21:30 |
| sdake | harmw whats up bro | 21:30 |
| *** jtriley has quit IRC | 21:37 | |
| *** jruano has joined #kolla | 21:40 | |
| bmace | hey sdake, messing around with kolla start based on the dev environment setup, on latest. seems like there are some magnum dependencies that are perhaps not tagged right in dockerhub? was looking for latest, it doesn't have that tag, only kilo tagged image available, so i changed the compose file for the magnum-api to get past it | 21:43 |
| sdake | bmace so docker pull on latest doesnt' work for you? | 21:44 |
| sdake | rather dkolla pull? | 21:44 |
| sdake | (or start) | 21:44 |
| bmace | let me make sure i was 100 up to date, but on the pull i did last wasn't, which was at most a couple of days old, no | 21:44 |
| bmace | on a new pull i don't see any changes that would have fixed anything | 21:45 |
| sdake | well keep in mind we don't actually cotninually push latest | 21:45 |
| sdake | so its best to build | 21:46 |
| *** daneyon_ has quit IRC | 21:47 | |
| bmace | if you mean the container, there just is no latest, but the compose references a latest.. | 21:47 |
| sdake | ya, you said the kolla magnum containers are not tagged correctly in dockerhub, i think htat is possible | 21:47 |
| sdake | we haven't yet solved how to tag altest on every commit tot the repo | 21:48 |
| sdake | this atm is a manual step | 21:48 |
| sdake | and was probably not done right | 21:48 |
| bmace | at a minimum the default magnum-api container should probably just reference kilo then, or at least something that won't choke when it tries to pull down form dockerhub? | 21:48 |
| *** absubram has quit IRC | 21:51 | |
| sdake | bmace what should happen is the magnum-api should be built for latest | 21:55 |
| sdake | I don't understand why that didn't happen | 21:55 |
| sdake | (and pushed) | 21:55 |
| sdake | so a workaround is to use build-all-docker-iamges | 21:55 |
| sdake | note someone (rhallisey perhaps?) reports galnce on master is broken | 21:55 |
| sdake | I would highly recommend if you are giving kolla a spin for the first tim to use stable/kilo branch (git checkout stable/kilo) to minimize problems with master churn. | 21:56 |
| sdake | Our CI system is imperfect, but jpeeler is doing a rockin job of making it perfect so we will have a high degree of confidence in our changes in the future | 21:57 |
| bmace | well, i'm trying to work towards containers working on ol7, so not sure work against an older set of code would be optimal, since a lot of what is happening isn't planned for backport, and reasonably enough, backports are a pain :) | 21:57 |
| sdake | oh roger that | 21:57 |
| sdake | brb need to switch networks | 21:58 |
| *** sdake_ has joined #kolla | 21:58 | |
| *** sdake has quit IRC | 22:02 | |
| *** openstackgerrit has quit IRC | 22:07 | |
| *** openstackgerrit has joined #kolla | 22:07 | |
| sdake_ | bmace if thats the case, yes you definately want to use master | 22:11 |
| sdake_ | if you want master to run on kilo, you would probably be better off at this point taking master and making the few changes necessary to get it running kilo rpms ;) | 22:12 |
| sdake_ | rather then backporting | 22:12 |
| bmace | sdake_ yup. just bringing things up while traveling down this path to make sure people are away of what the general experience is for people trying to get running on master at this point. | 22:12 |
| sdake_ | (just to address your comments about backporting isn't fun:) | 22:12 |
| sdake_ | cool | 22:12 |
| bmace | away = aware | 22:12 |
| sdake_ | what can I do to help? | 22:12 |
| sdake_ | my first recommendation is build-all-docker-images is the proper way to proceed | 22:13 |
| sdake_ | we need to debug the problem with glance | 22:13 |
| sdake_ | rhallisey reported it, is it still a problem rhallisey? | 22:13 |
| bmace | oh, well, on the magnum side i just tweaked the compose file to use the kilo version and i'm not seeing any bad behavior yet. the status is showing most services up, minum nova scheduler for some reason. | 22:13 |
| sdake_ | sudo docker ps in your favorite paste service | 22:14 |
| bmace | glance is showing as Up but that doesn't mean it is working right :) | 22:14 |
| sdake_ | I can help you debug | 22:14 |
| sdake_ | it wasn't staying operational previously | 22:14 |
| bmace | ah, well, all the glance components are showing as Up for me | 22:16 |
| sdake_ | so the docker ps in the paste service ;) | 22:17 |
| bmace | yeah.. that is tricky in my setup. i am doing all this work on a separate system than i am chatting with, to make things as complicated as possible. :) | 22:18 |
| sdake_ | cut and paste should do it :) | 22:18 |
| sdake_ | if I can't see the states of yoru system it will be hard for me to debug ;) | 22:18 |
| bmace | hrm, if i cut out a piece from my monitor and try to paste it on top of my laptop screen i bet you still couldn't be able to see it ;) | 22:19 |
| sdake_ | i see | 22:19 |
| sdake_ | ok, so you said nova conductor is theo nly service not starting? | 22:19 |
| bmace | would a problem with novascheduler cause keystone to not start up correctly? | 22:20 |
| sdake_ | so keystone restarts in a loop? | 22:20 |
| sdake_ | is it your novascheduler service that isnt' starting up? | 22:20 |
| bmace | yeah, novascheduler.. and the name of it is odd, starts with a uuid from the looks of it, not like the rest of the container naming | 22:21 |
| bmace | on my last kolla start it said there was a naming conflict when doing the recreate step... | 22:22 |
| sdake_ | ok sudo cleanup-contaienrs -f | 22:23 |
| sdake_ | then sudo kolla start | 22:23 |
| sdake_ | try a fresh start | 22:23 |
| bmace | kk | 22:24 |
| sdake_ | as to the container recreate step failling, no advice to offer there other then capture logs next time and file in the docker-compose community | 22:25 |
| sdake_ | which version of docker compose are you using? | 22:25 |
| sdake_ | master? | 22:25 |
| bmace | yeah, pretty much straight up following the dev environment directions | 22:27 |
| sdake_ | nice | 22:28 |
| sdake_ | well I haven't tried dorker-compoe master lately | 22:28 |
| sdake_ | possible they have had some kind of regression | 22:28 |
| bmace | that is a great name for it :) | 22:32 |
| bmace | much better. Up across the board.. just had to give it a whack, fonzi style. | 22:35 |
| sdake_ | ya docker compose sometimes fails recreates | 22:38 |
| sdake_ | notsure why | 22:38 |
| sdake_ | never seen coductor fail, usually its libvirt | 22:38 |
| sdake_ | arey ou usingdocker .1.7.0-dev? | 22:38 |
| bmace | no, i'm on 1.6. that is still what that page suggests. | 22:39 |
| sdake_ | or docker 1.6.0? | 22:39 |
| sdake_ | right | 22:39 |
| sdake_ | I have a review otu to roll to 1.7.0 | 22:39 |
| sdake_ | but it hasn't been reiewed | 22:39 |
| sdake_ | the reason I ask is docker 1.7.0 may be broken in some subtle way | 22:40 |
| sdake_ | but docker 1.6.0 seemed very solid during the months of testing we went through? | 22:40 |
| sdake_ | after they fixed the bugs we reported upstream | 22:40 |
| sdake_ | recreate failures usually happened because of docker 1.6.0 upstream bugs | 22:40 |
| sdake_ | usually involving volumes_from | 22:43 |
| sdake_ | this results in a container with a wierd name running like /bin/bash or something | 22:43 |
| bmace | ports mapped through to the host via docker i would expect to still show up appropriately in netstat? the keystone client is looping on connecting to port 5k and i can't hit horizon on https.. but i thought i had gone through the same steps on ubuntu and saw it work. | 22:43 |
| sdake_ | you mean kolla start is looking waiting for keystone endpoint-list ti finish? | 22:44 |
| bmace | right | 22:44 |
| sdake_ | looping | 22:44 |
| sdake_ | ya netstat will show the ports | 22:44 |
| bmace | and netstat -a doesn't show anything listening on 5k | 22:44 |
| bmace | but the kolla status shows everything as looking good. | 22:45 |
| sdake_ | sudo docker ps | grep keystone | 22:47 |
| sdake_ | sudo docker logs first 3 #'s for keystone contianer | 22:47 |
| bmace | well, i just do logs compose_keystone_1 no need to use the #s if you have a good tag :) | 22:47 |
| sdake_ | whatever works | 22:48 |
| sdake_ | did it finish the job on starting the keystone container? | 22:48 |
| sdake_ | or is it still doing db dsetup? | 22:48 |
| bmace | interesting, looks like db issues, i'll dig into the db container | 22:48 |
| sdake_ | or is it in a fail_unless_db_avaialble loop | 22:49 |
| sdake_ | before doing that | 22:50 |
| sdake_ | why do you think it looks like a db issue | 22:50 |
| bmace | well, keystone is looping checking connectivity to the db and failing.. database mysql is not available @ 10.0.2.15 | 22:51 |
| sdake_ | telnet 10.0.2.15 3306 | 22:51 |
| *** openstackgerrit has quit IRC | 22:51 | |
| sdake_ | from the host | 22:51 |
| *** openstackgerrit has joined #kolla | 22:51 | |
| sdake_ | 10.0.2.15 is your host ip address? | 22:51 |
| bmace | hrm, yup, and telnet worked | 22:52 |
| sdake_ | is 3306 the mysql port | 22:52 |
| sdake_ | i think it is.. but just woke u pfrom powernap so brain not quite humming yet | 22:52 |
| bmace | yup, yeah, i think i know :) | 22:53 |
| bmace | local mysql fighting w/ mariadb container | 22:53 |
| *** dims__ has joined #kolla | 22:53 | |
| sdake_ | oh | 22:53 |
| sdake_ | ya that would be it | 22:53 |
| sdake_ | make sure to turn off libvirt too | 22:53 |
| bmace | yeah, i did that since the directions told me to do so.. usually i don't install the db stuff by default, shouldn't have on this vm setup but i did :/ | 22:54 |
| *** dims___ has joined #kolla | 22:56 | |
| bmace | going to try starting from scratch again on the containers, for good measure. got the local libvirt / mysql disabled via systemctl | 22:56 |
| *** dims_ has quit IRC | 22:56 | |
| *** dims__ has quit IRC | 22:58 | |
| *** dwalsh has quit IRC | 22:58 | |
| *** jruano has quit IRC | 22:59 | |
| *** jtriley has joined #kolla | 22:59 | |
| sdake_ | bmace I may be repeating myself, but i can't keep straight who I have told what bout our design choices | 23:01 |
| sdake_ | we use --net=pid, which means no docker proxy | 23:01 |
| sdake_ | the ports should always show up | 23:02 |
| bmace | seems to be looking sort of better, not complaining about not being able to connect to the host / port | 23:05 |
| sdake_ | if your running with neutron | 23:08 |
| sdake_ | neutron agent-show | 23:08 |
| sdake_ | should show 4 agents | 23:08 |
| sdake_ | that is a good sign things are operating correctly | 23:08 |
| sdake_ | or agent-list | 23:08 |
| sdake_ | rather | 23:08 |
| bmace | with the admin / steakfordinner user? | 23:11 |
| sdake_ | source openrc | 23:12 |
| sdake_ | via cmd line run neutron agent-list | 23:12 |
| sdake_ | not sureh ow to do via horizon | 23:12 |
| sdake_ | but admin/steakfordinner ist he default | 23:12 |
| bmace | publicURL endpoint for network service not found :/ | 23:13 |
| sdake_ | run keystone endpoint-list | 23:15 |
| bmace | looks like in the neutron server logs there is a failure after "database is active @ 10.0.2.15" about the "pkg_resources.DistributionNotFound: argparse" | 23:16 |
| *** blahRus has quit IRC | 23:16 | |
| *** jtriley has quit IRC | 23:16 | |
| bmace | got a result from endpoint-list. RegionOne | 23:16 |
| sdake_ | that is all you get from endpoint-list? | 23:17 |
| bmace | yeah, just one entry | 23:17 |
| sdake_ | there should be a table of etnries | 23:17 |
| sdake_ | look at the glance log plz | 23:17 |
| sdake_ | glance-api | 23:17 |
| sdake_ | lets start there | 23:17 |
| sdake_ | what service is registered with keystone ? | 23:17 |
| bmace | same argparse error | 23:18 |
| bmace | right after keystone is active @ http://10.0.2.15:5000/v2.0 | 23:18 |
| sdake_ | docker exec container-id rpm -qa | grep argparse | 23:18 |
| bmace | from file /usr/bin/crux | 23:18 |
| bmace | a bunch of the containers including glance are now marked as "Restarting" but kolla status still shows them up | 23:22 |
| sdake_ | can you run the docker exec command pls | 23:23 |
| bmace | but the exec fails saying no container with the id exists.. and i used both the tag name and the actual container id | 23:23 |
| sdake_ | oh right | 23:23 |
| sdake_ | run it on the keystone container | 23:23 |
| sdake_ | brb need to switch to my lab network | 23:23 |
| *** sdake has joined #kolla | 23:24 | |
| sdake | bmace your running inside a eol container correct? | 23:25 |
| sdake | oel i mean | 23:26 |
| bmace | i have a vm running ol latest, yeah.. but the containers are still just regular centos | 23:27 |
| sdake | ok so this is not a porting issue then | 23:28 |
| *** sdake_ has quit IRC | 23:28 | |
| bmace | nope | 23:28 |
| sdake | was argparse in the keystone container? | 23:28 |
| bmace | i don't see anything odd in the keystone logs other than suggestions to use other query strategies that are more performant :) | 23:28 |
| sdake | docker exec container-id rpm -ql | grep argparse | 23:29 |
| sdake | actually | 23:29 |
| sdake | docker exec container-id yum list | grep argparse | 23:29 |
| bmace | no results | 23:29 |
| bmace | nodejs-argparse.norach is the only result from the yum list | 23:30 |
| bmace | noarch | 23:31 |
| sdake | docker exec ls -l /usr/lib64/python*/argparse* | 23:31 |
| sdake | actually shell expansion doesn't work with exec | 23:32 |
| sdake | try docker exec ls -l /usr/lib64/python2.7/argparse.py | 23:32 |
| sdake | building master atm | 23:32 |
| bmace | its there | 23:32 |
| sdake | are all of the containers barfing on crux with dist not found argparse minus the ones tht are up | 23:33 |
| sdake | (look at logs) | 23:33 |
| bmace | every one i have looked at so far, yes, glance-api, nova-api, neutron-server | 23:35 |
| sdake | docker images | grep centos | 23:36 |
| sdake | tell me age of image | 23:36 |
| sdake | is it crux that is throwing the error? | 23:36 |
| bmace | yeah, i posted that a ways up.. crux seems to be the crux :) | 23:37 |
| bmace | most are 6 days, keystone 6 weeks, glance-registry 8 weeks. | 23:38 |
| sdake | i mean the centos base image | 23:38 |
| sdake | argparse is a builtin | 23:38 |
| sdake | distribution not found is a python error | 23:38 |
| sdake | this leads me to believe the centos container image is damged in some way | 23:39 |
| sdake | hence the desire to know the age | 23:39 |
| bmace | sure, well, these images are all pulled from hub, so isn't the base info sort of baked in there? i don't even have a local centos base image. | 23:40 |
| sdake | ack give me a moment | 23:40 |
| sdake | what about magnum:kilo | 23:41 |
| sdake | is that barfing argparse? | 23:41 |
| bmace | one sec | 23:41 |
| bmace | completely different error. ERROR: missing MAGNUM_KEYSTONE_USER for both | 23:42 |
| sdake | can you change the glance compose file from latest to kilo | 23:43 |
| sdake | sudo dockercompose -f glancefile.yml up -d | 23:44 |
| bmace | sure thing, then i assume clear the image and fresh start | 23:44 |
| sdake | and see if it barfs out a crux errror | 23:44 |
| bmace | ah, ok, that is faster :) | 23:44 |
| sdake | no | 23:44 |
| sdake | keep system in place | 23:44 |
| bmace | looks happier. got another keystone endpoint now | 23:47 |
| sdake | ok so that tells us that kilo works and latest doesn't | 23:48 |
| sdake | it doesn't tell us if its the centos base image or the packaging on top | 23:48 |
| sdake | to determine that do the following | 23:48 |
| sdake | revert the change to the compose files for glance | 23:48 |
| sdake | actually | 23:49 |
| sdake | let me think how to root cause | 23:49 |
| bmace | so, you guys all build from scratch though right, and things seem to be working ok? would that infer there is some difference between the process you use and whatever process is used in the images on dockerhub? | 23:49 |
| sdake | same process | 23:50 |
| sdake | kilo works, latest doesn't | 23:50 |
| sdake | as i'm pretty sure I said before re glance ;) | 23:50 |
| sdake | looks like the problem has somethign to do with argparse | 23:50 |
| bmace | sure, but it seems like more than glance :) | 23:50 |
| sdake | agree | 23:50 |
| sdake | everything is broken | 23:51 |
| sdake | that depends on argparse | 23:51 |
| sdake | still building lateast | 23:51 |
| bmace | right. could come down to a tweak in the dockerfile to make sure that ends up in there.. | 23:51 |
| sdake | its part of python-libs | 23:52 |
| bmace | since it is broken this same way for everyone, i can just do the full image builds locally and see if i can help work out the problem. | 23:52 |
| sdake | that would be helpful, thats what i'm doing atm, but full build takes 45 minutes | 23:53 |
| bmace | since it seems to be the same breakage in many containers, that should also mean one fix will make everything a happy wonderland again :) | 23:53 |
| sdake | I think we just elimited oel kernel as the culprit | 23:55 |
| sdake | via that kilo vs lateast test | 23:56 |
| sdake | bit of a stretch there | 23:56 |
| bmace | sure, plus, that is just the host os... if that had a big effect then containerization is terrible ;) | 23:56 |
| bmace | i haven't even gotten to the ol to replace centos part yet | 23:56 |
| sdake | kernel does matter | 23:56 |
| bmace | i was just running kolla on top of ol7, which, exactly, shouldn't matter | 23:57 |
| sdake | you will see pam+noaudit hack in the centos container | 23:57 |
| sdake | that was to support running on debuntian | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!