| *** Haomeng|2 has joined #kolla | 00:01 | |
| *** Haomeng has quit IRC | 00:04 | |
| *** achanda_ has joined #kolla | 00:07 | |
| *** achanda has quit IRC | 00:11 | |
| *** achanda_ has quit IRC | 00:12 | |
| *** dolpher has joined #kolla | 00:18 | |
| SamYaple | sdake_: yea ive seen that before | 00:19 |
|---|---|---|
| SamYaple | when did you last rebuild your kolla-ansible container? | 00:20 |
| SamYaple | harmw: you need something i can help with? | 00:20 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Implement source fetching for build.py https://review.openstack.org/211954 | 00:22 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove bootstrap.yml abstraction for rabbitmq https://review.openstack.org/212457 | 00:22 |
| sdake_ | samyaple 3 days | 00:23 |
| SamYaple | ? | 00:23 |
| sdake_ | for kolla-ansible | 00:23 |
| SamYaple | oh yea the rebuild | 00:23 |
| sdake_ | image is 3 days oh | 00:23 |
| SamYaple | id suggest building it again | 00:23 |
| sdake_ | ok i'll pull the trigger on that | 00:23 |
| SamYaple | at one point we have ansible 1.9.2 i nthe container which throws that error | 00:23 |
| SamYaple | so now its ansible-dev | 00:24 |
| SamYaple | wow the repo looks different everyday | 00:24 |
| sdake_ | ya lots of change | 00:24 |
| SamYaple | I have almost twice the commits as anyone at RAX | 00:25 |
| sdake_ | code velocity is really high | 00:25 |
| SamYaple | but i trail in reviews | 00:25 |
| sdake_ | I am #1 and #1 at csco ;-) | 00:25 |
| SamYaple | need about 150 more to come out on top | 00:25 |
| SamYaple | yea but you wouldnt be at RAX sdake | 00:25 |
| sdake_ | check out this chart | 00:26 |
| sdake_ | http://stackalytics.com/?company=cisco%20systems&metric=commits | 00:26 |
| SamYaple | yup | 00:27 |
| SamYaple | rax looks the same | 00:27 |
| sdake_ | got some pain cream for my back from prescription helps a bit | 00:27 |
| SamYaple | http://stackalytics.com/?company=rackspace&metric=commits | 00:27 |
| sdake_ | cisco looks far different then rackspace | 00:27 |
| sdake_ | I have 27% of the commits ;-) | 00:27 |
| SamYaple | go to a company that is heavy openstack and you wouldnt! | 00:27 |
| *** zhiwei has joined #kolla | 00:28 | |
| sdake_ | would you rmeind reviewing my single image ansible work samyaple | 00:28 |
| sdake_ | i wasn't quite sure the best way to do it | 00:28 |
| SamYaple | where at | 00:28 |
| SamYaple | bout to leave on a date so quickly | 00:28 |
| sdake_ | top of queuehttps://review.openstack.org/#/c/212804/ | 00:29 |
| sdake_ | date - i've given up on those | 00:29 |
| sdake_ | must be nice to be young! | 00:29 |
| sdake_ | https://review.openstack.org/#/c/212804/ | 00:30 |
| SamYaple | did you test this sdake? | 00:30 |
| openstackgerrit | Merged stackforge/kolla: All the Cinder services need to reference the correct config https://review.openstack.org/212157 | 00:30 |
| sdake_ | yes i did test it | 00:30 |
| sdake_ | and got that error i linked you earlier | 00:30 |
| SamYaple | well that is before it ever gets the the datavol stuff | 00:31 |
| SamYaple | im pretty sure CMD ["/bin/sleep infinity"] wont work | 00:31 |
| sdake_ | it actually mounts one of the data volumes | 00:31 |
| SamYaple | thats a list of commands | 00:31 |
| SamYaple | CMD /bin/sleep infinity | 00:31 |
| sdake_ | sounds good - i copied it from some other container | 00:31 |
| sdake_ | i'll adjust before rebuilding | 00:31 |
| sdake_ | just curious if the approach is correct | 00:32 |
| sdake_ | the ansible part | 00:32 |
| SamYaple | I would move it to a global defaults, but we dont have that setup yet | 00:32 |
| sdake_ | i was thinking the same exact thing | 00:33 |
| sdake_ | but i wasn't sure if that would work | 00:33 |
| SamYaple | i have some work for moving that in | 00:33 |
| sdake_ | in the meantime i'd like to get the work in so we can trim the repo ;) | 00:33 |
| SamYaple | would you mind rebasing on https://review.openstack.org/#/c/212457/ | 00:33 |
| SamYaple | since that will break your work (or viceversa) | 00:34 |
| sdake_ | sure | 00:34 |
| *** zhiwei has left #kolla | 00:34 | |
| sdake_ | btw gat epatch hs 1 =2 vote thus far | 00:35 |
| sdake_ | so should go in soon | 00:35 |
| SamYaple | yea i didnt get to ubuntu patch last night | 00:35 |
| SamYaple | babies | 00:35 |
| sdake_ | which ubuntu patch | 00:36 |
| SamYaple | gate | 00:36 |
| sdake_ | oh | 00:36 |
| sdake_ | we should have source and binary gating shortly | 00:36 |
| sdake_ | jeff has a pep8 error in his patch | 00:36 |
| sdake_ | but it should be gtg once fixed | 00:36 |
| SamYaple | ok gotta go. ttyl | 00:36 |
| sdake_ | so looks good? | 00:37 |
| SamYaple | i left review | 00:37 |
| sdake_ | ok thanks | 00:37 |
| SamYaple | mostly good, small nitpick | 00:37 |
| sdake_ | thanks | 00:38 |
| *** bmace has quit IRC | 00:41 | |
| *** diogogmt has joined #kolla | 00:52 | |
| *** bmace has joined #kolla | 00:53 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Remove data containers and adjust compose files https://review.openstack.org/212830 | 00:56 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Remove bootstrap.yml abstraction for rabbitmq https://review.openstack.org/212457 | 00:56 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use one data container for all data volumes https://review.openstack.org/212804 | 00:56 |
| *** weiyu has joined #kolla | 00:56 | |
| *** juggler has joined #kolla | 00:57 | |
| *** echoingumesh has quit IRC | 00:59 | |
| *** echoingumesh has joined #kolla | 01:00 | |
| *** SourabhP has joined #kolla | 01:00 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Remove data containers and adjust compose files https://review.openstack.org/212830 | 01:00 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use one data container for all data volumes https://review.openstack.org/212804 | 01:00 |
| *** daneyon_ has quit IRC | 01:03 | |
| *** echoingumesh has quit IRC | 01:04 | |
| *** dolpher has quit IRC | 01:08 | |
| *** SourabhP has quit IRC | 01:16 | |
| *** jtriley has quit IRC | 01:22 | |
| Kennan | hi Sdake: | 01:23 |
| *** jtriley has joined #kolla | 01:23 | |
| Kennan | sake_: ping | 01:24 |
| Kennan | sdake_ typo sorry | 01:24 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Remove data containers and adjust compose files https://review.openstack.org/212830 | 01:27 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use one data container for all data volumes https://review.openstack.org/212804 | 01:27 |
| *** dwalsh has quit IRC | 01:27 | |
| *** erkules_ has joined #kolla | 01:29 | |
| *** erkules has quit IRC | 01:32 | |
| sdake_ | sup kennan how ican i Help | 01:45 |
| Kennan | hi sdake_ | 01:49 |
| Kennan | I saw https://review.openstack.org/#/c/210694/ | 01:49 |
| Kennan | still have some discussion on that | 01:49 |
| Kennan | so compose may still be used ? | 01:49 |
| *** erkules_ has quit IRC | 01:52 | |
| *** erkules_ has joined #kolla | 01:54 | |
| sdake_ | coolsvapping me when you get aorund | 01:58 |
| sdake_ | coolsvap ping me when youget around | 01:58 |
| sdake_ | kennan even when it is reworked which it is not now | 01:58 |
| sdake_ | it will not be usable without tripleo | 01:58 |
| sdake_ | it will require tripleo to be used | 01:58 |
| sdake_ | because we don't have a way of generating the configuration files | 01:59 |
| sdake_ | here is how config external works | 01:59 |
| sdake_ | it copies a config file into the container and runs the appropriate binary | 01:59 |
| sdake_ | kennan the only wy to rally use kolla today | 01:59 |
| sdake_ | is with ansible | 01:59 |
| *** dolpher has joined #kolla | 02:10 | |
| Kennan | sdake_ so you mean ansible would be formal and compose way may be dropped ? it maybe a long dicssuion for that? | 02:17 |
| sdake_ | compose *IS* dropped | 02:17 |
| Kennan | ok sdake_ I would read more to make sure understood it. Thanks | 02:18 |
| sdake_ | compose is single node | 02:18 |
| sdake_ | that is no good for us | 02:18 |
| sdake_ | so we replaced it with ansible | 02:19 |
| sdake_ | read the ansible-multi.spec in the specs directry | 02:19 |
| Kennan | when you talked about compose, is that docker compose ? | 02:19 |
| Kennan | docker compose, single node, do you mean docker compose only support run containers in one node ? sdake_ | 02:20 |
| sdake_ | ansible has an all in one single node deploy mode | 02:21 |
| sdake_ | its right there in the docs | 02:21 |
| sdake_ | yes docker-compose | 02:21 |
| *** achanda has joined #kolla | 02:22 | |
| *** jtriley has quit IRC | 02:23 | |
| *** jasonsb has joined #kolla | 02:23 | |
| Kennan | sdake_ I know some of chef, maybe ansible has same structure as that | 02:23 |
| Kennan | so you want containers run support multinode | 02:23 |
| Kennan | so you choose ansible | 02:23 |
| Kennan | while compose only support deploy contains on one node | 02:24 |
| Kennan | right ? | 02:24 |
| sdake_ | roger that | 02:25 |
| *** dwalsh has joined #kolla | 02:26 | |
| sdake_ | samyaple when your back need some debugging on galera help | 02:31 |
| *** dwalsh has quit IRC | 03:04 | |
| *** unicell has quit IRC | 03:14 | |
| SamYaple | sdake_: whats up | 03:15 |
| sdake_ | galera fails to start | 03:15 |
| sdake_ | /var/lib/mysql which is a bind mount is howened by root | 03:15 |
| sdake_ | I think mysql is running as nonroot | 03:15 |
| sdake_ | a chown on the bindmount does ntot chown | 03:15 |
| sdake_ | https://github.com/boot2docker/boot2docker/issues/587 | 03:16 |
| SamYaple | thats not true though | 03:16 |
| SamYaple | it does | 03:16 |
| SamYaple | its not a bindmount | 03:16 |
| SamYaple | its a volume | 03:16 |
| sdake_ | can you deploy master of a compete rebuild? | 03:17 |
| SamYaple | ill try | 03:17 |
| sdake_ | i put chgrp mysql: in cnfig_external.sh and it doesn't change to mysql | 03:18 |
| sdake_ | rather chown | 03:18 |
| SamYaple | man i cant wait to not have to uild 60 containers | 03:19 |
| sdake_ | it may be my patch stream | 03:19 |
| sdake_ | it is possible that when you remount the permissions on var/lib/mysql are set properly | 03:19 |
| sdake_ | because its already in the filesystem | 03:19 |
| sdake_ | i see there is a -u flag how woud lI specify that in the docker module | 03:21 |
| SamYaple | what do you mean | 03:21 |
| *** achanda has quit IRC | 03:21 | |
| sdake_ | -u lets you specify the user to run the container as | 03:21 |
| SamYaple | no thats not what you want | 03:21 |
| sdake_ | i wanted to try running the data container as mysql | 03:21 |
| SamYaple | that wont work and even if it did it wouldnt work like youd expect | 03:22 |
| *** achanda has joined #kolla | 03:25 | |
| *** achanda has quit IRC | 03:25 | |
| sdake_ | well here is whati get with galera | 03:30 |
| sdake_ | 2015-08-14 03:29:36 7f821d91f880 InnoDB: Operating system error number 13 in a file operation. | 03:30 |
| sdake_ | InnoDB: The error means mysqld does not have the access rights to | 03:30 |
| sdake_ | InnoDB: the directory. | 03:30 |
| sdake_ | [sdake@bigiron tools]$ sudo docker exec a0 ls -ld /var/lib/mysql | 03:31 |
| sdake_ | drwxr-xr-x 2 root root 4096 Aug 14 03:29 /var/lib/mysql | 03:31 |
| *** unicell has joined #kolla | 03:33 | |
| *** unicell has quit IRC | 03:33 | |
| *** unicell has joined #kolla | 03:33 | |
| sdake_ | master works my patch set does not | 03:34 |
| sdake_ | master's /var/lib/mysql: [sdake@bigiron tools]$ sudo docker exec bb ls -ld /var/lib/mysql | 03:35 |
| sdake_ | drwxr-xr-x 5 mysql mysql 4096 Aug 14 03:34 /var/lib/mysql | 03:35 |
| SamYaple | what patchset? | 03:43 |
| *** achanda has joined #kolla | 03:44 | |
| coolsvap | sdake_, pong | 03:46 |
| *** diogogmt has quit IRC | 04:00 | |
| coolsvap | sdake_, regarding the fedora base image are we changing this with the docker templates | 04:02 |
| coolsvap | i think we were using the f21 as base image all along | 04:02 |
| SamYaple | sdake_: all good for me on master | 04:02 |
| SamYaple | ill dig into your one data volume patch in a bit | 04:02 |
| sdake_ | ya i already said master works | 04:02 |
| sdake_ | i think i have fixed it | 04:03 |
| SamYaple | good deal | 04:03 |
| SamYaple | well im going to go hang with peeps | 04:03 |
| *** banix has joined #kolla | 04:04 | |
| coolsvap | sdake_, around | 04:10 |
| sdake_ | shoot | 04:10 |
| coolsvap | sdake_, question regarding fedora base image | 04:11 |
| coolsvap | you put on https://review.openstack.org/#/c/212289/2/docker_templates/magnum/magnum-base/Dockerfile.j2 | 04:11 |
| coolsvap | or we are just using f22 repo for magnum packages? | 04:12 |
| sdake_ | fedora base image pulls from latesst i think | 04:12 |
| sdake_ | that is hwo the templates work | 04:12 |
| sdake_ | latesst is f22 | 04:12 |
| sdake_ | it is definatey f22 | 04:12 |
| sdake_ | it prints all kinds of spam about dnf | 04:13 |
| coolsvap | its not | 04:13 |
| coolsvap | wait | 04:13 |
| coolsvap | https://github.com/stackforge/kolla/blob/master/docker/fedora/binary/base/Dockerfile#L1 | 04:13 |
| coolsvap | https://github.com/stackforge/kolla/blob/master/docker/fedora/source/base/Dockerfile#L1 | 04:14 |
| coolsvap | okay | 04:14 |
| coolsvap | missed the templates part | 04:14 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add Dockerfile template for Magnum https://review.openstack.org/212289 | 04:21 |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Update Fedora magnum images to use fedora repo https://review.openstack.org/212972 | 04:32 |
| *** weiyu has quit IRC | 04:33 | |
| *** slagle has quit IRC | 04:46 | |
| *** coolsvap has quit IRC | 04:46 | |
| *** coolsvap has joined #kolla | 04:48 | |
| *** slagle has joined #kolla | 04:50 | |
| *** banix has quit IRC | 05:00 | |
| *** diga_ has joined #kolla | 05:17 | |
| *** sdake_ has quit IRC | 05:17 | |
| *** coolsvap has quit IRC | 06:00 | |
| *** coolsvap has joined #kolla | 06:03 | |
| *** SourabhP has joined #kolla | 06:10 | |
| *** SourabhP_ has joined #kolla | 06:13 | |
| *** SourabhP has quit IRC | 06:16 | |
| *** SourabhP_ is now known as SourabhP | 06:16 | |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Update Fedora magnum images to use fedora repo https://review.openstack.org/212972 | 06:26 |
| *** weiyu has joined #kolla | 06:43 | |
| *** weiyu has quit IRC | 07:06 | |
| *** weiyu has joined #kolla | 07:07 | |
| *** fangfenghua has joined #kolla | 07:19 | |
| *** achanda has quit IRC | 07:19 | |
| *** bmace has quit IRC | 07:39 | |
| *** SourabhP has quit IRC | 07:47 | |
| *** fangfeng_ has joined #kolla | 07:48 | |
| *** fangfenghua has quit IRC | 07:52 | |
| *** bmace has joined #kolla | 07:53 | |
| *** erkules_ is now known as erkules | 07:54 | |
| *** erkules has joined #kolla | 07:54 | |
| *** hoonetorg_ has quit IRC | 08:04 | |
| *** jmccarthy has quit IRC | 08:15 | |
| *** jmccarthy has joined #kolla | 08:16 | |
| *** dolpher has quit IRC | 08:18 | |
| *** fangfeng_ has quit IRC | 08:21 | |
| *** fangfenghua has joined #kolla | 08:26 | |
| *** gfidente has joined #kolla | 08:27 | |
| *** fangfeng_ has joined #kolla | 08:33 | |
| *** fangfenghua has quit IRC | 08:37 | |
| pbourke | can anyone give me background on these 'source incorrectly places config' reviews? | 08:41 |
| pbourke | they all seem wrong as far as I can see | 08:41 |
| *** tfukushima has joined #kolla | 08:42 | |
| pbourke | as in the configs are in the right place | 08:42 |
| *** athomas has joined #kolla | 08:47 | |
| vbel | which component was it? | 08:49 |
| pbourke | vbel: there's reviews for almost all | 08:51 |
| pbourke | I just looked at the heat one and it does actually need to be changed | 08:51 |
| pbourke | so maybe it applies to some, will double check | 08:51 |
| vbel | :) ever-changing animal :) | 08:51 |
| pbourke | vbel: much luck with the ol templates? | 08:54 |
| vbel | not, yet, have some urgent stuff to do internally :) | 08:55 |
| openstackgerrit | Merged stackforge/kolla: Fix heat configuration. https://review.openstack.org/212706 | 08:55 |
| openstackgerrit | Merged stackforge/kolla: Designate install from source incorrectly places config https://review.openstack.org/212729 | 08:57 |
| *** athomas has quit IRC | 08:57 | |
| openstackgerrit | Merged stackforge/kolla: Cinder install from source incorrectly places config https://review.openstack.org/212718 | 08:58 |
| openstackgerrit | Merged stackforge/kolla: Ceilometer install from source incorrectly places config https://review.openstack.org/212727 | 08:59 |
| openstackgerrit | Merged stackforge/kolla: Barbican install from source incorrectly places config https://review.openstack.org/212724 | 09:00 |
| openstackgerrit | Merged stackforge/kolla: Magnum install from source incorrectly places config https://review.openstack.org/212740 | 09:03 |
| openstackgerrit | Merged stackforge/kolla: Gnocchi install from source incorrectly places config https://review.openstack.org/212730 | 09:03 |
| openstackgerrit | Merged stackforge/kolla: Murano install from source incorrectly places config https://review.openstack.org/212771 | 09:03 |
| *** dwalsh has joined #kolla | 09:04 | |
| openstackgerrit | Merged stackforge/kolla: Remove bootstrap.yml abstraction for rabbitmq https://review.openstack.org/212457 | 09:05 |
| *** vbel has quit IRC | 09:09 | |
| *** vbel has joined #kolla | 09:09 | |
| *** fangfeng_ has quit IRC | 09:18 | |
| *** fangfenghua has joined #kolla | 09:18 | |
| *** diga has quit IRC | 09:26 | |
| *** fangfeng_ has joined #kolla | 09:30 | |
| *** tfukushima has quit IRC | 09:31 | |
| *** fangfenghua has quit IRC | 09:31 | |
| *** tfukushima has joined #kolla | 09:33 | |
| openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Update Fedora magnum images to use fedora repo https://review.openstack.org/212972 | 09:43 |
| *** tfukushima has quit IRC | 09:53 | |
| *** tfukushima has joined #kolla | 09:55 | |
| *** fangfeng_ has quit IRC | 10:01 | |
| *** fangfenghua has joined #kolla | 10:02 | |
| *** Haomeng has joined #kolla | 10:07 | |
| *** Haomeng|2 has quit IRC | 10:09 | |
| *** fangfeng_ has joined #kolla | 10:10 | |
| *** fangfenghua has quit IRC | 10:13 | |
| *** tfukushima has quit IRC | 10:15 | |
| *** tfukushima has joined #kolla | 10:17 | |
| *** fangfeng_ has quit IRC | 10:20 | |
| *** weiyu has quit IRC | 10:22 | |
| *** diga_ has quit IRC | 10:35 | |
| jmccarthy | Is this due to the data container discussion that's taking place at the moment ? | 10:38 |
| jmccarthy | Pulling mariadbdata (kollaglue/centos-rdo-mariadb-data:latest)... | 10:38 |
| jmccarthy | Pulling repository kollaglue/centos-rdo-mariadb-data | 10:38 |
| jmccarthy | ... | 10:38 |
| jmccarthy | compose.progress_stream.StreamOutputError: Tag latest not found in repository kollaglue/centos-rdo-mariadb-data | 10:38 |
| pbourke | jmccarthy: no there's no latest tags on dockerhub | 10:39 |
| pbourke | jasonsb: you have to build yourself | 10:39 |
| pbourke | jasonsb: meant the above for jmccarthy | 10:40 |
| jmccarthy | Hrm ok 'docker-compose -f compose/mariadb.yml up -d' used to work for me ? (stuff built with --release) | 10:41 |
| jmccarthy | holidays = everything forgotten :D Ok I'll try that | 10:43 |
| *** diga has joined #kolla | 10:51 | |
| *** weiyu has joined #kolla | 10:58 | |
| *** diga has quit IRC | 11:00 | |
| *** tfukushima has quit IRC | 11:10 | |
| *** tfukushima has joined #kolla | 11:11 | |
| *** vbel has quit IRC | 11:11 | |
| *** achanda has joined #kolla | 11:11 | |
| *** achanda has quit IRC | 11:15 | |
| *** weiyu has quit IRC | 11:26 | |
| *** tfukushima has quit IRC | 11:32 | |
| *** tfukushima has joined #kolla | 11:32 | |
| *** echoingumesh has joined #kolla | 11:33 | |
| *** tfukushima has quit IRC | 11:34 | |
| *** Kennan2 has joined #kolla | 11:36 | |
| *** echoingumesh has quit IRC | 11:36 | |
| *** echoingumesh has joined #kolla | 11:37 | |
| *** Kennan has quit IRC | 11:37 | |
| pbourke | SamYaple: I'm of the opinion exceptions should be used for exceptional circumstances | 11:40 |
| pbourke | SamYaple: I think its probably one of those debatable things that differs between people | 11:41 |
| SamYaple | i agree with that | 11:41 |
| SamYaple | the debateable part | 11:41 |
| *** echoingumesh has quit IRC | 11:41 | |
| SamYaple | the thing is, I have two options here, loop throughall the config options each time OR catch the exception | 11:41 |
| SamYaple | catchign the exception is cleaner to me | 11:41 |
| SamYaple | but i understand your viewpoint | 11:41 |
| SamYaple | to me it feels like duplicating work to loop through the options | 11:42 |
| SamYaple | I mean, this is kinda what the try/except block is for. Its really the only purpose for it | 11:43 |
| pbourke | could we just check if the section exists? | 11:46 |
| SamYaple | This is how you check... | 11:46 |
| pbourke | right, but its more explicit, imo, to say 'if self.config.hassection(...):' (or whatever the method is called) | 11:47 |
| SamYaple | thier is no hasection function is my point | 11:47 |
| pbourke | oh! | 11:47 |
| SamYaple | https://docs.python.org/2/library/configparser.html | 11:48 |
| SamYaple | this is literally docs code | 11:48 |
| SamYaple | try: except | 11:48 |
| *** dwalsh has quit IRC | 11:55 | |
| *** sdake_ has joined #kolla | 12:11 | |
| sdake_ | morning | 12:16 |
| jmccarthy | o/ | 12:18 |
| *** tfukushima has joined #kolla | 12:23 | |
| *** tfukushima has quit IRC | 12:29 | |
| *** tfukushima has joined #kolla | 12:31 | |
| *** britthouser has joined #kolla | 12:33 | |
| *** jpeeler has quit IRC | 12:33 | |
| *** britthou_ has joined #kolla | 12:34 | |
| *** sdake_ has quit IRC | 12:37 | |
| *** britthouser has quit IRC | 12:38 | |
| *** sdake has joined #kolla | 12:55 | |
| *** britthou_ has quit IRC | 12:57 | |
| *** dwalsh has joined #kolla | 12:57 | |
| *** britthouser has joined #kolla | 12:58 | |
| *** rhallisey has joined #kolla | 12:58 | |
| *** sdake_ has joined #kolla | 12:59 | |
| *** sdake has quit IRC | 13:02 | |
| *** diogogmt has joined #kolla | 13:05 | |
| *** sdake has joined #kolla | 13:17 | |
| sdake | morning | 13:17 |
| *** tfukushima has quit IRC | 13:18 | |
| *** diogogmt has quit IRC | 13:19 | |
| *** sdake_ has quit IRC | 13:20 | |
| sdake | inc0 around | 13:20 |
| sdake | samyaple around | 13:29 |
| *** sdake_ has joined #kolla | 13:34 | |
| *** sdake has quit IRC | 13:38 | |
| *** jtriley has joined #kolla | 13:40 | |
| *** jmccarthy has quit IRC | 13:44 | |
| *** jpeeler has joined #kolla | 13:45 | |
| *** jmccarthy has joined #kolla | 13:48 | |
| sdake_ | anyone experience a problem with keepalive being busted and restarting | 14:01 |
| sdake_ | we have 24 containers in an ansible deploy | 14:02 |
| sdake_ | quite a bit | 14:02 |
| *** dwalsh has quit IRC | 14:02 | |
| rhallisey | sdake_, which do you prefer | 14:03 |
| rhallisey | http://fpaste.org/255159/60937143/ | 14:03 |
| rhallisey | or | 14:03 |
| rhallisey | http://fpaste.org/255160/60972143/ | 14:03 |
| rhallisey | the nested /etc directory is what was throwing off source builds | 14:03 |
| sdake_ | the first one will error out if heat/etc/heat isn't present | 14:04 |
| sdake_ | that is preferrable | 14:04 |
| sdake_ | (erroring vs not copying unexpectedly) | 14:05 |
| sdake_ | because debugging a wrong copy will be extremely hard | 14:05 |
| rhallisey | ya I agree | 14:05 |
| rhallisey | some of the services don't have nested /ect (swift keystone) | 14:05 |
| sdake_ | right | 14:05 |
| rhallisey | so I'l abandon those | 14:05 |
| sdake_ | lets be consistent | 14:05 |
| sdake_ | keepalive, why not working!! | 14:06 |
| rhallisey | k I'll throw up a new set of fixes | 14:06 |
| rhallisey | sdake_, should I make separate bugs for these again? | 14:09 |
| sdake_ | not new bugs | 14:09 |
| sdake_ | just make one bug | 14:09 |
| sdake_ | and one patch | 14:09 |
| rhallisey | k | 14:09 |
| sdake_ | since they are all one liners | 14:09 |
| rhallisey | easier | 14:09 |
| sdake_ | less work for you | 14:09 |
| sdake_ | less work for reviewers | 14:09 |
| sdake_ | abandon all your changes plz | 14:10 |
| sdake_ | all the ones reelevant to this tho | 14:10 |
| sdake_ | btw folks propoed coolsvap for core reviewer on ml | 14:14 |
| sdake_ | please vote, or dont vote, but read the email | 14:15 |
| sdake_ | rhallisey does keepalived start up for you with ansible? | 14:26 |
| *** absubram has joined #kolla | 14:30 | |
| *** unicell1 has joined #kolla | 14:36 | |
| *** unicell has quit IRC | 14:37 | |
| *** absubram has quit IRC | 14:38 | |
| rhallisey | I'll try it.. | 14:40 |
| *** absubram has joined #kolla | 14:41 | |
| *** echoingumesh has joined #kolla | 14:57 | |
| *** absubram has quit IRC | 14:57 | |
| *** echoingumesh has quit IRC | 14:58 | |
| *** echoingumesh has joined #kolla | 14:58 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Refactor where source builds place config files https://review.openstack.org/213184 | 14:59 |
| rhallisey | sdake_, that will fix up source | 14:59 |
| *** diogogmt has joined #kolla | 15:01 | |
| *** echoingumesh has quit IRC | 15:02 | |
| *** absubram has joined #kolla | 15:06 | |
| *** banix has joined #kolla | 15:14 | |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Use inheritance for overriding of builder args https://review.openstack.org/212814 | 15:14 |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Use inheritance for overriding of builder args https://review.openstack.org/212814 | 15:16 |
| sdake_ | jpeeler | 15:16 |
| sdake_ | could you add a couple more classes to that patch | 15:17 |
| sdake_ | for --template | 15:17 |
| sdake_ | BuildTestTemplateCentosBinary | 15:17 |
| sdake_ | BuildTestTemplateCentosSource | 15:17 |
| jpeeler | those are there? | 15:17 |
| sdake_ | and change the current tox source building to build with templates (but keep binary as is) | 15:17 |
| sdake_ | which are where | 15:17 |
| jpeeler | i'm confused what you want added | 15:18 |
| sdake_ | there is a build --template option | 15:18 |
| sdake_ | it buidsl from docker_templates dir instead | 15:18 |
| sdake_ | I want to gate the template building | 15:18 |
| sdake_ | atleasst nonvoting | 15:18 |
| sdake_ | i also want to run the regular build gate | 15:19 |
| sdake_ | rather build.py but for docker directory | 15:19 |
| jpeeler | ok i see, so it's just --template? | 15:19 |
| sdake_ | thats it | 15:19 |
| sdake_ | simple huh :) | 15:19 |
| sdake_ | so keep waht you have | 15:20 |
| sdake_ | and add BuildTestTemplateCentosBinary and Source | 15:20 |
| sdake_ | an just add more args | 15:20 |
| sdake_ | its a shame we have to do this in code | 15:20 |
| sdake_ | rather then as options to the testr | 15:20 |
| sdake_ | but we went down that road and it didn't work | 15:20 |
| jpeeler | well... if we used a different test runner maybe we could | 15:21 |
| sdake_ | openstack uses testr overall right? | 15:21 |
| sdake_ | I'd like to stick with that | 15:21 |
| *** dwalsh has joined #kolla | 15:21 | |
| jpeeler | believe so | 15:21 |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Use inheritance for overriding of builder args https://review.openstack.org/212814 | 15:23 |
| jpeeler | i really want to finish gate stuff soon | 15:24 |
| sdake_ | boy our gate is coming along | 15:25 |
| sdake_ | I dont know what is taking inra cats so long | 15:25 |
| sdake_ | i'll go harass them a gain right now brb :) | 15:25 |
| *** echoingumesh has joined #kolla | 15:26 | |
| jpeeler | pbourke: if you could look one last time :) need it to merge ideally before the gate tries to start using it | 15:28 |
| *** openstackgerrit has quit IRC | 15:31 | |
| *** openstackgerrit has joined #kolla | 15:31 | |
| *** absubram has quit IRC | 15:32 | |
| *** blahRus has joined #kolla | 15:34 | |
| *** SourabhP has joined #kolla | 15:38 | |
| sdake_ | jpeeler | 15:38 |
| sdake_ | our patches are in the +a state | 15:38 |
| *** absubram has joined #kolla | 15:38 | |
| *** SourabhP_ has joined #kolla | 15:39 | |
| sdake_ | for the gate | 15:41 |
| *** SourabhP has quit IRC | 15:43 | |
| *** SourabhP_ is now known as SourabhP | 15:43 | |
| jpeeler | right, i guess the worst the happens if the above patch and depependcy doesn't merge then the gate will run all combinations twice | 15:43 |
| sdake_ | jpeeler its merged | 15:45 |
| jpeeler | the gate change? | 15:45 |
| sdake_ | yup | 15:45 |
| sdake_ | lets get a review up | 15:45 |
| sdake_ | to test it out | 15:45 |
| jpeeler | dang, better fix this up fast | 15:45 |
| sdake_ | is there a problem? | 15:46 |
| jpeeler | the other changes to use the proper tox targets and such have not merged | 15:46 |
| sdake_ | link | 15:46 |
| sdake_ | lets get it merged | 15:46 |
| sdake_ | note the gate doesn't have "Template" yet | 15:46 |
| sdake_ | i'll submit those today | 15:46 |
| sdake_ | so dont worry about that right now | 15:46 |
| jpeeler | i missed a blank line | 15:46 |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Use inheritance for overriding of builder args https://review.openstack.org/212814 | 15:46 |
| sdake_ | rhallisey please approve that patch | 15:47 |
| openstackgerrit | Merged stackforge/kolla: Implement source fetching for build.py https://review.openstack.org/211954 | 15:48 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Remove data containers and adjust compose files https://review.openstack.org/212830 | 15:57 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use one data container for all data volumes https://review.openstack.org/212804 | 15:57 |
| openstackgerrit | Merged stackforge/kolla: Refactor where source builds place config files https://review.openstack.org/213184 | 15:58 |
| *** kevsi has quit IRC | 16:02 | |
| *** SourabhP has quit IRC | 16:08 | |
| *** banix has quit IRC | 16:09 | |
| *** pbourke has quit IRC | 16:09 | |
| *** pbourke has joined #kolla | 16:10 | |
| *** dwalsh has quit IRC | 16:11 | |
| *** unicell1 has quit IRC | 16:11 | |
| *** banix has joined #kolla | 16:13 | |
| *** jasonsb has quit IRC | 16:15 | |
| *** bmace has quit IRC | 16:17 | |
| *** daneyon has joined #kolla | 16:23 | |
| *** unicell has joined #kolla | 16:39 | |
| openstackgerrit | Merged stackforge/kolla: Use inheritance for overriding of builder args https://review.openstack.org/212814 | 16:48 |
| sdake_ | the fetch from source patch breaks build.py | 16:52 |
| sdake_ | samyaple ^^ | 16:52 |
| *** Haomeng has quit IRC | 16:59 | |
| *** jtriley has quit IRC | 17:04 | |
| *** Kennan2 has quit IRC | 17:05 | |
| *** jasonsb has joined #kolla | 17:11 | |
| *** dwalsh has joined #kolla | 17:12 | |
| *** jasonsb has quit IRC | 17:12 | |
| *** jasonsb has joined #kolla | 17:13 | |
| *** kfox1111 has joined #kolla | 17:15 | |
| kfox1111 | is kolla planning on supporting whats in the docker hub or are those just for testing? | 17:16 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Remove data containers and adjust compose files https://review.openstack.org/212830 | 17:16 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use one data container for all data volumes https://review.openstack.org/212804 | 17:16 |
| *** achanda has joined #kolla | 17:16 | |
| sdake_ | kfox1111 our project does not really provide support | 17:17 |
| sdake_ | that is a downstream's job | 17:17 |
| sdake_ | its not part of our mission to support openstack | 17:17 |
| sdake_ | kfox1111 that said, we want the images to work for the purposes of people being able to evaluate the software without having to build their own images | 17:17 |
| *** rhallisey has quit IRC | 17:18 | |
| kfox1111 | understood, but I meant more like, if rdo releases an update, does the registry get it quickly, | 17:18 |
| kfox1111 | or do us ops resign ourselves to always building the containers from source ourselves? | 17:18 |
| kfox1111 | rdo's the source of support, | 17:19 |
| kfox1111 | but kolla being the distribution channel. | 17:19 |
| sdake_ | it does not get it quickly | 17:19 |
| sdake_ | we release on milestones only | 17:20 |
| kfox1111 | bummer. so we have to deal with the extra layer of complexity ourselves. :/ | 17:20 |
| kfox1111 | ok. | 17:20 |
| sdake_ | if i could get em into the ci system of openstack i would | 17:20 |
| sdake_ | but pushes takes like 5 hours per distro+src or binary | 17:20 |
| kfox1111 | ah. that would be awesome. :) | 17:20 |
| *** Kennan has joined #kolla | 17:21 | |
| sdake_ | infra has a 90 minute limit | 17:21 |
| *** pmartini_ has joined #kolla | 17:21 | |
| sdake_ | kfox1111 have you tried out the ansible deploy bits? | 17:21 |
| kfox1111 | no. all the clouds I have right now are production, so I'm just starting to kick the tires now on the redone external config stuff. I've got one cloud that doens't have trove on it yet, and am considering pulling just that in. | 17:22 |
| kfox1111 | It is looking very useful though. | 17:22 |
| sdake_ | you could try just all in one | 17:23 |
| sdake_ | looking for critical feedback - anything wrong or needs attention | 17:23 |
| kfox1111 | yeah. unfortuantly I have very little time. :( | 17:23 |
| sdake_ | i get the implementaiton ins't complete | 17:23 |
| *** achanda has quit IRC | 17:23 | |
| sdake_ | but beyond that looking for ways to improve | 17:23 |
| kfox1111 | playing with kolla containers to add trove's even questionable right now. :/ | 17:23 |
| kfox1111 | but I figured I could justify it by saying we'd get percona mysql out of it. | 17:23 |
| sdake_ | trove is descoped for lilberty | 17:23 |
| kfox1111 | ugg. | 17:24 |
| kfox1111 | k. | 17:24 |
| sdake_ | look at the tracker and you will see why :) | 17:24 |
| sdake_ | we have a mountain o fwork left | 17:24 |
| kfox1111 | yeah. I understand. | 17:24 |
| sdake_ | we may add services asyncrounsly - our governance documentaiton allows for it | 17:24 |
| kfox1111 | I may just use the rabbit container then. | 17:24 |
| kfox1111 | since my current rabbit cant be exposed for trove's consumption. | 17:24 |
| *** banix has quit IRC | 17:25 | |
| kfox1111 | ah. cool. | 17:25 |
| *** gfidente has quit IRC | 17:25 | |
| sdake_ | the only integration with config-external at this point is ansible | 17:25 |
| kfox1111 | I can do it manually with compose files. | 17:25 |
| sdake_ | if your planning on rolling your own deploy tool take care to look at the ansible model - the containers are fairly tiglty integrated | 17:25 |
| sdake_ | bad news there | 17:26 |
| sdake_ | compose hsa not been ported to config-external at this time | 17:26 |
| kfox1111 | no. I'd write my own. | 17:26 |
| sdake_ | that would work | 17:26 |
| sdake_ | but i tss a two phase bringup - bootstrap followed by container deploy | 17:26 |
| sdake_ | in the old containers bootstrap and bringup happened every time | 17:26 |
| kfox1111 | yeah. for this, I'd need a single rabbit that doesn't conflict with the system rabbit, so a container would work there. and I can install trove just on the local system then, so I think that would work. | 17:26 |
| sdake_ | config-internal i scompletely gone | 17:27 |
| kfox1111 | yeah. definately like the new split. the old combined thing made me really nervious. | 17:27 |
| kfox1111 | awesome. :) | 17:27 |
| sdake_ | ws just waiting for tripleo cats to see the light | 17:27 |
| sdake_ | now they have ;) | 17:27 |
| kfox1111 | I was hoping tripleO really didn't need it after all. | 17:27 |
| kfox1111 | yeah. :) | 17:27 |
| pmartini_ | Good morning all. I'm trying to test kolla out according to dev-quickstart, and ./tools/kolla-compose start fails because nothing in kollaglue has a latest tag. Any suggestions? | 17:28 |
| sdake_ | compose is gone | 17:28 |
| * sdake_ needs to update the docs | 17:28 | |
| sdake_ | use kolla-ansible | 17:28 |
| *** dwalsh has quit IRC | 17:28 | |
| sdake_ | there are 4 required config variables | 17:29 |
| sdake_ | kolla-ansible deploy will default to all in one deployment | 17:29 |
| sdake_ | tools/build.py -T 1000 to build images | 17:29 |
| sdake_ | we only ahe liberty-2 images in the repo | 17:29 |
| sdake_ | in the dockerhub that is | 17:29 |
| sdake_ | I know everyone wants latest images in the dockerhub but the ci guys wont let me suckup a build machine for 6 hours for a push operation | 17:30 |
| kfox1111 | where does centos-rdo-rabbit for liberty2 come from? | 17:30 |
| kfox1111 | kilo's repo? | 17:30 |
| sdake_ | i built it and pushed it on my machine | 17:30 |
| sdake_ | its based upon delorean | 17:30 |
| kfox1111 | ah. ok. | 17:30 |
| sdake_ | delorean is redhat's "master" | 17:30 |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Change tox targets to use new build.py https://review.openstack.org/212821 | 17:32 |
| *** jtriley has joined #kolla | 17:32 | |
| pmartini_ | sdake_ - when you said use kolla-ansible, do you mean just run './tools/kolla-ansible deploy' ? That also fails because of latest. Trying to do the tools/build now | 17:32 |
| sdake_ | jpeeler build.py is busted | 17:32 |
| sdake_ | would you mind fixing it | 17:33 |
| sdake_ | i verified its sam's latest fetch source patch | 17:33 |
| sdake_ | tools/build.py -T 100 | 17:33 |
| sdake_ | should take about 20 minuts | 17:33 |
| sdake_ | you need to copy kolla/etc/kolla to /etc/kolla | 17:33 |
| sdake_ | and configure 4 variables | 17:34 |
| sdake_ | it is documented in the ansible doc in the doc repo | 17:34 |
| sdake_ | dev-qucikstart and ansible doc need to be mered/reworked for new model | 17:34 |
| pmartini_ | k, rechecking | 17:34 |
| sdake_ | i'll have new docs today | 17:34 |
| sdake_ | atm working on gating | 17:35 |
| sdake_ | pmartini_ I'd highly recommend reading the ansible docs in the docs directory - you need to install three deps (docker-py, docker, and ansible) | 17:36 |
| *** jtriley has quit IRC | 17:37 | |
| pmartini_ | sdake_ - I have docker and ansible, not sure about docker-py; if I didn't have those it wouldn't have gotten to the point of complaining that it couldn't fetch latest :-) | 17:37 |
| pmartini_ | I sett up /etc/kolla now, and changed globals to have "openstack_release: "liberty2" and now it's moving along. Reading through the rest of the docs to see what else I've broken | 17:39 |
| sdake_ | pmartini_ run pip show docker-py and paste | 17:39 |
| sdake_ | i doubt liberty2 tag will work with mster of ansible | 17:39 |
| sdake_ | bt it may | 17:39 |
| sdake_ | the problem is master of liberty has added services, and we haven't pushed those yet | 17:39 |
| sdake_ | rahter mastr of kolla | 17:40 |
| sdake_ | you need to configure 4 other variables as well pmartini_ | 17:40 |
| pmartini_ | hah, nevermind, it just failed differently | 17:40 |
| sdake_ | or your system will implode | 17:40 |
| sdake_ | run pip show docker-py - you may not have the correct version | 17:40 |
| pmartini_ | I have 1.2.3; what's the correct version? | 17:40 |
| sdake_ | that will work | 17:40 |
| sdake_ | if you have a paste of the failure i can help you debug | 17:41 |
| sdake_ | master deploys for me atm, but keepalived is busted | 17:41 |
| sdake_ | debugging now | 17:41 |
| sdake_ | master of build.py is also busted | 17:41 |
| * sdake_ WTB MOAR GATING | 17:41 | |
| pmartini_ | I'm reading through the docs now before I ask any more questions | 17:41 |
| pmartini_ | Particularly to see what those 4 variables are | 17:42 |
| openstackgerrit | Robyn Bergeron proposed stackforge/kolla: Port openvswitch to docker_templates https://review.openstack.org/213267 | 17:43 |
| openstackgerrit | Robyn Bergeron proposed stackforge/kolla: Port openvswitch to docker_templates https://review.openstack.org/213267 | 17:45 |
| sdake_ | brb need to check email | 17:47 |
| *** sdake has joined #kolla | 17:48 | |
| *** sdake_ has quit IRC | 17:49 | |
| *** sdake_ has joined #kolla | 17:50 | |
| sdake_ | jpeeler able to look at that build.py problem or want me to look | 17:50 |
| jpeeler | i'm just now starting to look | 17:51 |
| *** rhallisey has joined #kolla | 17:53 | |
| *** sdake has quit IRC | 17:53 | |
| jpeeler | if just accessing the key is all that's required, then i'll have a patch up in a minute | 17:53 |
| jpeeler | seriously need to stop changing the build script before it gets in the gate | 17:55 |
| jpeeler | since we're so close | 17:55 |
| *** dwalsh has joined #kolla | 17:55 | |
| *** achanda has joined #kolla | 17:55 | |
| kfox1111 | ah.. yeah. this exteral config stuff's already sooo much nicer. :) | 17:55 |
| kfox1111 | I think the container will work for the trove use case. I can move both rabbit default port and erlang port as part of the external config. | 17:56 |
| kfox1111 | navive + containered rabbit on the same host shoudl work. :) | 17:56 |
| sdake_ | pmartini_ if you build your images, take care to change the openstack_release field in the globals.yml to latest | 17:56 |
| pmartini_ | Ah, thanks! | 17:57 |
| sdake_ | once i fix keepalived, I'll sort out better docs | 17:57 |
| sdake_ | step by step instructions | 17:57 |
| sdake_ | as dev-quickstart is | 17:58 |
| sdake_ | the ansible directions are not really step by step | 17:58 |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Safely check for dictionary key presence https://review.openstack.org/213277 | 17:58 |
| sdake_ | rhallisey pbourke need an ack on https://review.openstack.org/#/c/213277/ | 18:03 |
| sdake_ | good fix jpeeler I wouldn't have thought of that | 18:04 |
| jpeeler | now if everything will just merge | 18:08 |
| sdake_ | atm wea re waiting on gearman | 18:17 |
| sdake_ | we need to get rid of he keepalived pid file | 18:19 |
| kfox1111 | nice. kolla's rabbit's working. :) | 18:21 |
| kfox1111 | thanks. Liberty's kolla is orders of magnagude better. :) | 18:21 |
| openstackgerrit | Merged stackforge/kolla: Change tox targets to use new build.py https://review.openstack.org/212821 | 18:23 |
| sdake_ | the container partis improving | 18:23 |
| sdake_ | we are templating them atm | 18:24 |
| jpeeler | ok sdake_ as soon as that last patch gets merged and the gearman (what's that anyway) gets resolved, gate should be in a good state | 18:25 |
| sdake_ | gearman attaches the build machines to the jobs | 18:25 |
| sdake_ | i think | 18:26 |
| *** jtriley has joined #kolla | 18:26 | |
| *** achanda has quit IRC | 18:33 | |
| sdake_ | this is a problem | 18:36 |
| sdake_ | http://ur1.ca/nfqtl -> http://paste.fedoraproject.org/255277/39577386 | 18:36 |
| sdake_ | jpeeler can you try an aio install and see if it works for you | 18:41 |
| sdake_ | i want to see if master is busted on my machine or all machines | 18:41 |
| jpeeler | it'll take me a while since i either need to reprovision my dev machine or get vagrant working | 18:43 |
| *** dwalsh has quit IRC | 18:44 | |
| sdake_ | just use kolla-ansible | 18:46 |
| sdake_ | and copy etc/kolla to /etc/kolla | 18:46 |
| sdake_ | and configure the 4 keys | 18:46 |
| sdake_ | throw out vagrant fo rthe moment | 18:46 |
| pmartini_ | Ugh, now build isn't working, looks like due to this: https://bugs.launchpad.net/mos/+bug/1484485 | 18:47 |
| openstack | Launchpad bug 1484485 in Mirantis OpenStack "RPM-build fails during installing systemd" [High,Confirmed] - Assigned to Fuel DevOps (fuel-devops) | 18:47 |
| pmartini_ | (I'd manually done the same thing as the patch above - 213277 - but now the base image won't build because it can't unpack systemd) | 18:48 |
| sdake_ | thats a kernel bug | 18:49 |
| sdake_ | are you using ubuntu? | 18:49 |
| pmartini_ | Yes | 18:49 |
| sdake_ | need to install the hwe 3.19 kernel | 18:50 |
| sdake_ | another option is to use device mapper | 18:50 |
| sdake_ | i'veheard dev mapper doesn't work on ubuntu becasue of library dependency problems however | 18:51 |
| pmartini_ | You wouldn't happen to have the package name I'd need to install (on Trusty) to get that, would you? The likely candidates I see are 'linux-hwe-generic-trusty' and 'linux-hwe-virtual-trusty', and I don't know which (if either) would be right | 18:53 |
| sdake_ | i haven't personally tested these instructions | 18:53 |
| sdake_ | http://askubuntu.com/questions/618575/how-do-i-find-the-current-stable-kernel-hwe-for-14-04-2 | 18:53 |
| sdake_ | tldr command is sudo apt-get install --install-recommends linux-generic-lts-utopic | 18:53 |
| pmartini_ | Thanks, will see if that blows me up <grin> | 18:54 |
| sdake_ | i run fedora so i don't run into these ubuntu kernel bugs | 18:54 |
| sdake_ | fedora always packages latest kernel | 18:54 |
| sdake_ | 3.19 or later is what you need | 18:54 |
| sdake_ | samyaple I am pretty sure dind wont work because of /run bindmounting | 18:55 |
| pmartini_ | Ah, alright then. Just upgrading kernel is a sight more comfortable | 18:55 |
| sdake_ | i have no idea what the --install-recommends task does | 18:55 |
| sdake_ | when you run kolla-ansible it needs to be under sudo as well | 18:56 |
| *** pmartini_ has quit IRC | 19:11 | |
| *** pmartini has joined #kolla | 19:16 | |
| pmartini | I actually jumped straight to vivid instead of utopic, and now it's building! | 19:33 |
| *** achanda has joined #kolla | 19:46 | |
| ccrouch | sdake_: reading scrollback, I was going to do an AIO install with ansible, on a VM | 19:50 |
| ccrouch | what VM host opsys should i used? | 19:50 |
| ccrouch | is F20 good enough? | 19:50 |
| *** achanda has quit IRC | 19:59 | |
| harmw | ccrouch: why not 21? | 20:00 |
| harmw | plus, you could go multihost using the Vagrantfile in our repo :) | 20:01 |
| *** dwalsh has joined #kolla | 20:01 | |
| *** daneyon_ has joined #kolla | 20:02 | |
| *** daneyon has quit IRC | 20:05 | |
| ccrouch | harmw: just because i've got a F20 instance already available | 20:06 |
| ccrouch | no downloading required | 20:06 |
| harmw | makes sense | 20:06 |
| ccrouch | should F20 work? | 20:06 |
| harmw | from what I've seen most people use 21, but the only major things are kernel+docker versions | 20:07 |
| harmw | if you can run ansible on it aswell, it'll work | 20:07 |
| harmw | your kernel is probably fine | 20:08 |
| harmw | since 21 is at 4, 20 would probably be at something 3.whateverwasverynice | 20:08 |
| harmw | 3.19 perhaps | 20:08 |
| jpeeler | my aio attempt is complaining about DEFAULT_DOCKER_API_VERSION not being defined | 20:13 |
| jpeeler | with docker-py 1.2.3 | 20:14 |
| harmw | hm, I've seen that before | 20:15 |
| harmw | due to some other (something with an x) package being outdated | 20:15 |
| jpeeler | i thought that was why it was defined in defaults.yml | 20:15 |
| jpeeler | six? | 20:16 |
| harmw | nah, this is different | 20:16 |
| harmw | yes! | 20:16 |
| jpeeler | ok tell me what version i need | 20:16 |
| harmw | pip --upgrade six | 20:16 |
| jpeeler | have 1.9,0 | 20:16 |
| harmw | oh no, let me check | 20:16 |
| harmw | got my operator node still running I hope | 20:16 |
| harmw | got that to | 20:17 |
| harmw | both versions match my env | 20:17 |
| harmw | websocket-client? | 20:18 |
| jpeeler | 0.32.0 | 20:18 |
| harmw | hm, thats fine | 20:18 |
| jpeeler | after applying this fix(?) https://github.com/ansible/ansible-modules-core/commit/dbdbcb33ccdbc14a6eb805181e7a1403399badc0 it reports docker-py isn't installed | 20:22 |
| jpeeler | but it is. | 20:22 |
| jpeeler | upgrade to 1.3.1, no difference | 20:23 |
| openstackgerrit | Merged stackforge/kolla: Safely check for dictionary key presence https://review.openstack.org/213277 | 20:24 |
| *** sdake has joined #kolla | 20:27 | |
| jpeeler | harmw: did you ever figure out the ansible_eth0 thing with vagrant? | 20:30 |
| *** sdake_ has quit IRC | 20:31 | |
| harmw | hm no | 20:31 |
| harmw | I too had that, right? | 20:31 |
| jpeeler | i thought you said you were seeing it | 20:31 |
| harmw | yea, you filed a bug | 20:32 |
| harmw | and Sam said something, about fixing, rather trivial | 20:32 |
| harmw | so first thing to do: git pull :) | 20:32 |
| harmw | lets spin up the good ol' Vagrant stuff | 20:34 |
| jpeeler | after a git pull, i see the same error. but it could be network related? but since you said you saw it then i thought maybe it wasn't the network | 20:35 |
| jpeeler | so ansible just doesn't work for me, aio or multinode | 20:36 |
| harmw | could you paste the buglink again? so I can read through what the actual errors were.. | 20:36 |
| jpeeler | https://bugs.launchpad.net/kolla/+bug/1482406 | 20:36 |
| openstack | Launchpad bug 1482406 in kolla "ansible deploy yielding undefined variable " [Undecided,New] | 20:36 |
| harmw | I'm launching Vagrant, so going to see what happens in a while :) | 20:37 |
| jpeeler | i have some silly questions | 20:37 |
| harmw | silly questions don't exists | 20:37 |
| harmw | only silly answers :) | 20:37 |
| jpeeler | the local registry that's deployed. the images must be pulled to each node right? | 20:37 |
| *** bmace has joined #kolla | 20:37 | |
| harmw | this regards Vagrant? | 20:38 |
| jpeeler | well i assume every multinode setup, but i'm using vagrant | 20:38 |
| harmw | but yes, each node pulls images to localhost using the registry | 20:38 |
| jpeeler | oh you mean you already set that up? it didn't happen for me | 20:39 |
| harmw | it's by no means specific to Vagrant btw :) (but you got that) | 20:39 |
| harmw | vagrant is launching the operator node which starts a docker-registry | 20:39 |
| harmw | you only need to configure the ansible config to actually use it | 20:39 |
| harmw | but then again, you also need to configure some other tiny bits inside the kolla tree | 20:39 |
| jpeeler | i think i missed this tag-and-push script | 20:40 |
| jpeeler | but seeing it now it seems that it answered my other question | 20:40 |
| harmw | excellent :) | 20:40 |
| harmw | there are docs btw, docs/vagrant.md iirc | 20:41 |
| jpeeler | yeah i saw them, i might add more if i could ever get my set up working. | 20:42 |
| * jpeeler wonders if VirtualBox and libvirt would play nicely together | 20:42 | |
| harmw | just disable libvirt, not needed when using VBox :) | 20:43 |
| *** achanda has joined #kolla | 20:44 | |
| harmw | hm, the vagrant nodes should probably be changed to not download all those pip dependencies | 20:44 |
| harmw | since we now have a ansible container to do configs | 20:45 |
| *** achanda has quit IRC | 20:51 | |
| jpeeler | pip install -U websocket fixed the docker version issue | 20:52 |
| harmw | yes, I knew it was websocket :p | 20:52 |
| harmw | but thats not what pip freeze told me (without -client, that is) | 20:53 |
| harmw | glad you resolved it though :) | 20:53 |
| *** echoingumesh has quit IRC | 20:59 | |
| *** echoingumesh has joined #kolla | 21:00 | |
| harmw | hm, its probably a good idea to rebuild images... | 21:00 |
| *** echoingu_ has joined #kolla | 21:02 | |
| *** echoingumesh has quit IRC | 21:04 | |
| *** achanda has joined #kolla | 21:04 | |
| *** jasonsb has quit IRC | 21:05 | |
| *** jasonsb has joined #kolla | 21:06 | |
| *** SourabhP has joined #kolla | 21:22 | |
| *** jtriley has quit IRC | 21:22 | |
| jpeeler | sdake: something has happened with the pam+noaudit package preventing the base image to be rebuilt | 21:25 |
| jpeeler | Rpmdb checksum is invalid: dCDPT(pkg checksums): pam+noaudit.x86_64 0:1.1.8-16.1.el7.centos - u | 21:25 |
| jpeeler | at first i thought it was due to build.py being more strict, but the old way won't continue either | 21:25 |
| jpeeler | might be a local docker issue and not the package | 21:30 |
| sdake | jpeeler let me build base | 21:34 |
| *** sdake_ has joined #kolla | 21:35 | |
| *** pmartini has left #kolla | 21:36 | |
| *** sdake has quit IRC | 21:39 | |
| sdake_ | jpeeler try --no-cache build option | 21:42 |
| sdake_ | why does mac have a spinning wheel | 21:42 |
| * sdake_ groans | 21:42 | |
| sdake_ | its unix ffs | 21:42 |
| jpeeler | this is on a new host, so i would think any caching options wouldn't make a difference | 21:42 |
| *** bmace has quit IRC | 21:45 | |
| sdake_ | 2 minutes 18 seconds to deploy current ansible implementation | 21:47 |
| sdake_ | on AIO | 21:48 |
| sdake_ | glance and nova work, but neutron does not register its endpoint | 21:49 |
| jpeeler | i assume base built fine for you then? | 21:53 |
| *** echoingu_ has quit IRC | 21:55 | |
| sdake_ | getting there | 21:58 |
| sdake_ | moment | 21:58 |
| sdake_ | i was trying to fix my deployment | 21:58 |
| sdake_ | forgot to configure a config option | 21:58 |
| sdake_ | still neutron no worky | 21:58 |
| sdake_ | cinder no worky | 21:59 |
| sdake_ | jpeeler your new box | 22:00 |
| sdake_ | is it fedora? | 22:00 |
| jpeeler | yeah f21 | 22:01 |
| sdake_ | df | fpaste | 22:01 |
| jpeeler | http://paste.fedoraproject.org/255346/58971014 | 22:01 |
| sdake_ | well your partitioining is correct | 22:02 |
| sdake_ | uusaully fedora borks /home at 90% of the disk | 22:02 |
| sdake_ | are you using docker-engine? | 22:02 |
| jpeeler | yep | 22:02 |
| sdake_ | i am buildign now | 22:03 |
| jpeeler | docker-engine-1.8.1-1.fc21.x86_64 | 22:03 |
| sdake_ | I am using 1.7.1 | 22:03 |
| sdake_ | base builds for me | 22:05 |
| sdake_ | with --no-cache | 22:05 |
| jpeeler | i'm going to switch to 1.7.1 | 22:05 |
| sdake_ | make sure to delete /var/lib/docker | 22:05 |
| sdake_ | is there any disk error or segfault in dmesg? | 22:06 |
| jpeeler | no | 22:06 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: test patch - do not merge https://review.openstack.org/213355 | 22:07 |
| sdake_ | lets see if the gate is workign yet | 22:07 |
| jpeeler | still says not_registered | 22:10 |
| jpeeler | something must be wrong, shouldn't take this long | 22:10 |
| jpeeler | ok either i did something wrong or 1.8.1 is screwy | 22:12 |
| *** rhallisey has quit IRC | 22:12 | |
| *** dwalsh has quit IRC | 22:17 | |
| sdake_ | jpeeler i got with clarkb on the not registered thing | 22:34 |
| sdake_ | he says its because of the gate using distro names in the title to try to determine what node to run on | 22:35 |
| sdake_ | jpeeler https://review.openstack.org/#/c/213363/1/jenkins/jobs/projects.yaml | 22:40 |
| sdake_ | we are going to have to use the curl install method with setup_gate.sh unless we want further delay and multiple gate scripts | 22:41 |
| *** jasonsb has quit IRC | 22:56 | |
| *** Haomeng has joined #kolla | 22:58 | |
| *** bmace has joined #kolla | 23:00 | |
| *** blahRus has quit IRC | 23:11 | |
| sdake_ | samyaple when you get done with your date i want to undersstand the tehcnial difference between internal address and external address as it relates to load balancing and other things | 23:14 |
| *** jasonsb has joined #kolla | 23:22 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Register neutron as the network service https://review.openstack.org/213374 | 23:27 |
| *** jasonsb has quit IRC | 23:28 | |
| jpeeler | are there more centos7 hosts than f21 or something? | 23:28 |
| * jpeeler disappears | 23:28 | |
| jpeeler | it's fine to do the curl thing i guess... | 23:29 |
| *** Haomeng|2 has joined #kolla | 23:29 | |
| sdake_ | no but we will running the functional tests on centos | 23:29 |
| sdake_ | might as well get em on the same host | 23:29 |
| sdake_ | harmw when you getaround i want help getting vagrant rolling | 23:30 |
| *** Haomeng has quit IRC | 23:32 | |
| sdake_ | rhallisey cinder is not showing up as an endpoint | 23:34 |
| sdake_ | well master working again with that last patch | 23:43 |
| sdake_ | not sure how that got past the reviewers | 23:43 |
| sdake_ | or got put in a patch for that matter | 23:43 |
| ccrouch | sdake_: so my attempt to run an AIO from current master is doomed to failure? | 23:46 |
| sdake_ | you can jut cherrypick the last patch on top of your repo | 23:47 |
| sdake_ | and it works like a champ | 23:47 |
| sdake_ | cinder appaers not to work | 23:47 |
| sdake_ | rest works | 23:47 |
| sdake_ | click the cherrypick tab on this patch https://review.openstack.org/213374 | 23:47 |
| ccrouch | last patch = https://review.openstack.org/213374 ? | 23:47 |
| sdake_ | cut and paste into your kolla directory | 23:47 |
| sdake_ | note I Just noticed we leak /run/haproxy.pid and /run/keepalived.pid so those need to be deletd if they are on the system | 23:48 |
| sdake_ | jpeeler what vagrant command do Itype to kick this thing of | 23:49 |
| * sdake_ WtB DOCUMENTATION | 23:49 | |
| ccrouch | sdake_: but isnt that patch dependent on https://review.openstack.org/#/c/212830/6 ? | 23:50 |
| sdake_ | no | 23:50 |
| sdake_ | i was just lazy | 23:50 |
| ccrouch | :-) | 23:50 |
| ccrouch | ok great | 23:50 |
| sdake_ | its a one liner | 23:50 |
| ccrouch | BTW i'm doing an AIO in a VM, on my mac | 23:51 |
| ccrouch | what extra network adaptor do i need to add, which i can then reference from neutron_external_interface | 23:51 |
| sdake_ | [sdake@bigiron vagrant]$ sudo vagrant up | 23:52 |
| sdake_ | The provider 'libvirt' could not be found, but was requested to | 23:52 |
| sdake_ | back the machine 'operator'. Please use a provider that exists. | 23:52 |
| sdake_ | just add any type | 23:52 |
| sdake_ | are you using virtualbo? | 23:52 |
| ccrouch | vmware fusion | 23:52 |
| ccrouch | dont scream :-) | 23:53 |
| sdake_ | ruby looks interesting | 23:56 |
| sdake_ | i dont understand why its written in ruby | 23:56 |
| sdake_ | god ruby is a disaster for dependencies | 23:56 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!