| *** unicell has quit IRC | 00:02 | |
| *** unicell has joined #kolla | 00:02 | |
| *** banix has quit IRC | 00:10 | |
| *** achanda has quit IRC | 00:11 | |
| *** weiyu has joined #kolla | 00:15 | |
| *** unicell has left #kolla | 00:21 | |
| *** unicell has joined #kolla | 00:22 | |
| *** dims has joined #kolla | 00:30 | |
| *** ccrouch has joined #kolla | 00:32 | |
| *** rhallisey has joined #kolla | 00:35 | |
| *** rhallisey has quit IRC | 00:35 | |
| *** achanda has joined #kolla | 00:48 | |
| *** tfukushima has quit IRC | 01:08 | |
| *** tfukushima has joined #kolla | 01:09 | |
| *** diogogmt has joined #kolla | 01:11 | |
| *** dolpher1 has joined #kolla | 01:14 | |
| *** sdake_ has joined #kolla | 01:21 | |
| *** sdake has quit IRC | 01:25 | |
| *** dims has quit IRC | 01:25 | |
| *** erkules_ has joined #kolla | 01:38 | |
| *** Haomeng|2 has quit IRC | 01:40 | |
| *** erkules has quit IRC | 01:41 | |
| *** britthouser has joined #kolla | 01:45 | |
| *** britthou_ has joined #kolla | 01:46 | |
| *** britthouser has quit IRC | 01:50 | |
| *** britthou_ has quit IRC | 01:52 | |
| *** ccrouch has quit IRC | 01:54 | |
| *** britthouser has joined #kolla | 01:54 | |
| *** britthou_ has joined #kolla | 01:55 | |
| *** SourabhP has quit IRC | 01:59 | |
| *** britthouser has quit IRC | 01:59 | |
| *** Haomeng has joined #kolla | 01:59 | |
| *** banix has joined #kolla | 02:06 | |
| *** dims has joined #kolla | 02:21 | |
| *** unicell has quit IRC | 02:25 | |
| *** dims has quit IRC | 02:29 | |
| *** dims has joined #kolla | 02:30 | |
| *** dims has quit IRC | 02:34 | |
| *** loth has quit IRC | 02:38 | |
| *** britthou_ has quit IRC | 02:51 | |
| *** diogogmt has quit IRC | 03:01 | |
| *** Kennan2 is now known as Kennan | 03:36 | |
| *** diga has joined #kolla | 03:38 | |
| *** banix has quit IRC | 03:49 | |
| *** achanda has quit IRC | 04:05 | |
| *** diga has quit IRC | 04:25 | |
| *** weiyu has quit IRC | 04:29 | |
| *** unicell has joined #kolla | 04:30 | |
| sdake_ | wtb samyaple | 05:02 |
|---|---|---|
| *** sdake has joined #kolla | 05:07 | |
| *** achanda has joined #kolla | 05:07 | |
| *** weiyu has joined #kolla | 05:08 | |
| *** sdake_ has quit IRC | 05:11 | |
| *** weiyu has quit IRC | 05:17 | |
| *** zhiwei has joined #kolla | 05:20 | |
| *** weiyu has joined #kolla | 05:25 | |
| *** weiyu has quit IRC | 05:42 | |
| *** inc0 has joined #kolla | 05:49 | |
| inc0 | good morning | 05:50 |
| *** weiyu has joined #kolla | 06:01 | |
| *** unicell1 has joined #kolla | 06:04 | |
| *** unicell has quit IRC | 06:05 | |
| *** inc0_ has joined #kolla | 06:12 | |
| *** inc0 has quit IRC | 06:15 | |
| *** inc0 has joined #kolla | 06:40 | |
| *** inc0_ has quit IRC | 06:43 | |
| *** inc0_ has joined #kolla | 06:56 | |
| *** inc0 has quit IRC | 06:59 | |
| *** achanda has quit IRC | 07:00 | |
| inc0_ | soo I loaded ddwrt on my linksys router | 07:12 |
| inc0_ | and dhcp didnt came on | 07:12 |
| inc0_ | damn. | 07:12 |
| *** inc0__ has joined #kolla | 07:15 | |
| *** inc0_ has quit IRC | 07:18 | |
| *** inc0_ has joined #kolla | 07:18 | |
| *** inc0__ has quit IRC | 07:22 | |
| *** shardy has joined #kolla | 07:23 | |
| *** jmccarthy has joined #kolla | 07:23 | |
| *** SourabhP has joined #kolla | 07:26 | |
| *** SourabhP_ has joined #kolla | 07:29 | |
| *** SourabhP has quit IRC | 07:32 | |
| *** SourabhP_ is now known as SourabhP | 07:32 | |
| inc0_ | SamYaple, around? | 07:33 |
| *** tfukushima has quit IRC | 07:38 | |
| SamYaple | inc0_ kinda | 07:39 |
| inc0_ | up for an argument? | 07:39 |
| inc0_ | codename: logging | 07:40 |
| SamYaple | idk maybe | 07:40 |
| inc0_ | sooo | 07:40 |
| inc0_ | we have 2 options | 07:40 |
| inc0_ | 1. we do logstash and rsync forward logs to logstash by ansible-default | 07:40 |
| inc0_ | rsyslog even | 07:41 |
| inc0_ | 2. we do data container and log to files by default - but this container can grow fat easily | 07:41 |
| inc0_ | so we would do log rotation and stuff there | 07:41 |
| SamYaple | log to files wont work | 07:41 |
| SamYaple | not everything supports that | 07:41 |
| inc0_ | rsyslog does | 07:41 |
| inc0_ | and that's what I mean - we have rsyslog server which will write files | 07:42 |
| SamYaple | right but applications dont support directly logging to files is my point | 07:42 |
| inc0_ | we want apps to use rsyslog | 07:42 |
| SamYaple | yes | 07:42 |
| inc0_ | what I mean is what we'll do with logs after that | 07:42 |
| SamYaple | initially, just write tehm to a file | 07:42 |
| SamYaple | thats the first step | 07:42 |
| inc0_ | so data container with logs | 07:43 |
| SamYaple | get them all to one rsyslog process | 07:43 |
| SamYaple | yea i dont care so much about a data container, I will be bindmounting | 07:43 |
| inc0_ | well data containers are kinda pattern in kolla, so I'll stick to that | 07:43 |
| SamYaple | yea but im saying i dont care what that is since ill be bindmounting | 07:44 |
| SamYaple | so do whatever | 07:44 |
| *** SourabhP has quit IRC | 07:44 | |
| inc0_ | you should care since it's for the betterment of whole project! | 07:44 |
| inc0_ | :P | 07:44 |
| SamYaple | i dont agree with data containers period | 07:44 |
| SamYaple | i just go along with it | 07:45 |
| inc0_ | btw, why? | 07:45 |
| inc0_ | I'm kinda in favor of bindmouinting myself (easyier to process) | 07:45 |
| SamYaple | its like a horrible compromise between persistent data and not | 07:45 |
| SamYaple | if you want the data persistent, it shouldnt be in a container | 07:45 |
| inc0_ | then why we do that in other containers? dabasase and such | 07:46 |
| SamYaple | that was done before i got here | 07:46 |
| SamYaple | i never cared for it | 07:46 |
| inc0_ | well, maybe worth changing? | 07:46 |
| SamYaple | no that arguement wont be won and i dont care enough to have it | 07:47 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use user_database_create in the register operation https://review.openstack.org/215002 | 07:47 |
| SamYaple | I only want to have teh option of bindmounting (which i have) | 07:47 |
| sdake | samyaple can you have a look at that patch please ^^ | 07:48 |
| inc0_ | SamYaple, maybe then make this option in ansible as well? | 07:48 |
| SamYaple | sdake: user_database isnt in all.yml | 07:48 |
| inc0_ | we're kinda pushing data containers with our approach | 07:48 |
| sdake | database_user? | 07:49 |
| SamYaple | oh sorry yea i just looked at the commit message | 07:49 |
| SamYaple | yea moving to defaults would expose tha problem | 07:49 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use database_user_create in the register operation https://review.openstack.org/215002 | 07:49 |
| sdake | fixed ;) | 07:49 |
| SamYaple | reasonable fix | 07:49 |
| sdake | ya master is undeployable atm | 07:50 |
| sdake | spent like 6 hours figuring out why | 07:50 |
| sdake | but learned something along the wway :) | 07:50 |
| sdake | on the plus side i can now deploy a nova vm via boot | 07:51 |
| sdake | so master seems to work | 07:51 |
| sdake | 170 plays | 07:51 |
| sdake | yay :) | 07:51 |
| openstackgerrit | Merged stackforge/kolla: Fix missing line break in memcached template https://review.openstack.org/214683 | 07:56 |
| sdake | samyaple I studieid ansible_kolla contianer today looks prtty solid | 07:57 |
| openstackgerrit | Merged stackforge/kolla: Add Dockerfile template for Gnocchi https://review.openstack.org/214452 | 07:58 |
| openstackgerrit | Merged stackforge/kolla: Fix binary builds for oraclelinux templates https://review.openstack.org/214687 | 07:58 |
| *** tfukushima has joined #kolla | 08:00 | |
| openstackgerrit | Merged stackforge/kolla: Fix small typo in task name https://review.openstack.org/214559 | 08:01 |
| *** diga has joined #kolla | 08:02 | |
| openstackgerrit | Merged stackforge/kolla: Remove defaults.yml from docs https://review.openstack.org/214147 | 08:03 |
| *** akwasnie has joined #kolla | 08:09 | |
| SamYaple | sdake: yea im not super happy about the command line wrapping, but at least its wrapping ansible native | 08:12 |
| SamYaple | and parsing json | 08:12 |
| SamYaple | overall i am reasonably happy with it | 08:12 |
| *** SourabhP has joined #kolla | 08:35 | |
| openstackgerrit | Merged stackforge/kolla: Load ip_vs module for keepalived https://review.openstack.org/214811 | 08:38 |
| SamYaple | inc0_: yes option in ansible is what it will be. havent determined how best to do that. but probably a variable override is good enough | 08:48 |
| inc0_ | you'd need to put conditionals to tasks, and that means having 2 tasks inheriting from 3rd one with simple when statement | 08:50 |
| inc0_ | am I right? | 08:50 |
| SamYaple | no | 08:50 |
| SamYaple | simpler than that | 08:50 |
| SamYaple | when specifying the volume we have a variable | 08:50 |
| SamYaple | by default it will be, say, /var/lib/mysql | 08:50 |
| SamYaple | but since its a variable, it can be overrideen to /data/mysql:/var/lib/mysql | 08:50 |
| SamYaple | no new tasks | 08:51 |
| sdake | samyaple are there reviews you wanted me to tackle before heading to bed, i've been through the review queue | 09:04 |
| *** SourabhP has quit IRC | 09:05 | |
| SamYaple | sdake: | 09:06 |
| SamYaple | https://review.openstack.org/#/c/213600/ | 09:06 |
| SamYaple | i thinks thats the last one... | 09:06 |
| *** vbel has joined #kolla | 09:07 | |
| vbel | good morning/evening | 09:14 |
| sdake | so remove ansible abstraction done then samyaple? | 09:14 |
| SamYaple | sdake: it is once you approve https://review.openstack.org/#/c/213958/ | 09:15 |
| SamYaple | thats the actual final remove | 09:15 |
| SamYaple | it was being held up by the patch oyu just merged | 09:15 |
| vbel | is it correct understanding that build.py does not support BUILDFLAGS/GIT_REF yet? | 09:15 |
| SamYaple | vbel: all I built into build.py was basic tarball fetching | 09:16 |
| SamYaple | i dont think anyone has added more yet vbel | 09:16 |
| vbel | SamYaple: thanks, just trying to understand a bit more about the script | 09:16 |
| openstackgerrit | Merged stackforge/kolla: Remove start.yml abstraction for kolla-ansible https://review.openstack.org/213600 | 09:16 |
| sdake | vbel it is incomplete at this time | 09:17 |
| sdake | vbel evening btw ;) | 09:17 |
| SamYaple | o/ | 09:17 |
| vbel | ok, evening :) | 09:17 |
| *** sdake_ has joined #kolla | 09:18 | |
| *** unicell1 has quit IRC | 09:20 | |
| *** sdake has quit IRC | 09:21 | |
| *** weiyu has quit IRC | 09:22 | |
| *** unicell has joined #kolla | 09:23 | |
| *** dolpher1 has quit IRC | 09:27 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add Dockerfile template for nova https://review.openstack.org/213478 | 09:28 |
| sdake_ | ok guys | 09:28 |
| sdake_ | quick brainstorm session about removal of docker dir | 09:28 |
| SamYaple | hey sdake_ can I rebase this for you? https://review.openstack.org/#/c/213478/ | 09:28 |
| SamYaple | oh i thought you were going to bed | 09:29 |
| sdake_ | samyaple that is a nova change that you own? | 09:29 |
| inc0_ | sdake_, are we sure noone uses old build scripts? | 09:29 |
| sdake_ | if they do I dont want them to do so :) | 09:29 |
| inc0_ | noone we care about anyway | 09:29 |
| SamYaple | sdake_: sorry https://review.openstack.org/#/c/212830/ | 09:29 |
| SamYaple | you referenced the nova change | 09:29 |
| SamYaple | i thought you were going to bed, but that removal patch is missing things | 09:30 |
| inc0_ | we still can't build single image with new script right?;) | 09:30 |
| SamYaple | inc0_: youre welcome to add that.... | 09:30 |
| sdake_ | ya single image - ounds like blocker | 09:30 |
| inc0_ | I'll do it then | 09:30 |
| inc0_ | just confirmation | 09:30 |
| sdake_ | need done by friday pls :) | 09:30 |
| SamYaple | inc0_: nah i dont want you to be pulled away from logging | 09:30 |
| inc0_ | it'll be few minutes of work | 09:31 |
| sdake_ | ya focus on logging | 09:31 |
| inc0_ | don't worry | 09:31 |
| SamYaple | im looking for some busy work at the moment I can add it | 09:31 |
| SamYaple | inc0_: its not a few minutes | 09:31 |
| SamYaple | it has to do parent building too | 09:31 |
| inc0_ | I don't think it has thoi | 09:31 |
| sdake_ | ok so i do want to go to bed soon 2:30am here | 09:31 |
| inc0_ | single image is mostly for development | 09:31 |
| *** weiyu_ has joined #kolla | 09:31 | |
| sdake_ | there are more use cases then dev | 09:31 |
| SamYaple | inc0_: yea it has to be able to rebuild the parents | 09:32 |
| sdake_ | ci/cd using build.py needsds single image | 09:32 |
| inc0_ | but do they all need parent building? | 09:32 |
| SamYaple | its ok like i said you stick to logging we need that more | 09:32 |
| vbel | we use docker/ and all old scripts | 09:32 |
| sdake_ | vbel docker going bye bye for liberty | 09:32 |
| inc0_ | vbel, we want to get rid of old scripts | 09:32 |
| sdake_ | build.sh going byebye for liberty | 09:32 |
| *** britthouser has joined #kolla | 09:32 | |
| SamYaple | well docker_templates will be renamed to docker I would assume... | 09:33 |
| sdake_ | recommend adjusting now | 09:33 |
| sdake_ | well that woudl probabybe the last step samyaple | 09:33 |
| vbel | if someone adds .buildconf .buildinfo to new scripts .. | 09:33 |
| sdake_ | but yes i'd like to do that | 09:33 |
| SamYaple | yea ok just making sure | 09:33 |
| SamYaple | vbel: we arent using those | 09:33 |
| SamYaple | we have the build.ini now | 09:33 |
| sdake_ | that is what build.ini is for | 09:33 |
| SamYaple | the buildconf and info were stopgaps | 09:33 |
| *** britthou_ has joined #kolla | 09:33 | |
| sdake_ | easy to adjust - use crudini | 09:33 |
| SamYaple | (as was discussed when they were commited) | 09:34 |
| akwasnie | hi all | 09:34 |
| sdake_ | ya i dont want dot files if we can help it | 09:34 |
| sdake_ | and we can help it | 09:34 |
| SamYaple | hi akwasnie! | 09:34 |
| inc0_ | hello again Ala | 09:34 |
| sdake_ | first step is to make sure docker_templates images work for everyone | 09:35 |
| sdake_ | pbourke or vbel give me an ack on oraclelinux asap :) | 09:35 |
| vbel | if I want my GIT_REF and couple of BUILDFLAGS can I use build.ini for that? | 09:36 |
| SamYaple | vbel: as soon as someone writes support for that | 09:36 |
| SamYaple | im not sure what BUILDFLAGS are here though | 09:37 |
| sdake_ | what are BUILDFLAGS exactly? | 09:37 |
| vbel | I just cannot use it at work, will be glad to add oraclelinux if it possible to compile :) | 09:37 |
| vbel | BUILDFLAGS="$BUILDFLAGS --build-env PIP_TRUSTED_HOST=x.y.z.a" | 09:37 |
| *** britthouser has quit IRC | 09:37 | |
| vbel | BUILDFLAGS="$BUILDFLAGS --build-env PIP_INDEX_URL= | 09:38 |
| sdake_ | that shouold go in build.ini | 09:38 |
| vbel | BUILDFLAGS="$BUILDFLAGS --build-env https_proxy= | 09:38 |
| vbel | ok | 09:38 |
| SamYaple | sdake_: hold on now | 09:38 |
| sdake_ | vbel i'll write support in the morning | 09:38 |
| SamYaple | what are we using https_proxy for in the source stuff? | 09:39 |
| sdake_ | for curling the rdo repos for openvswitch for one ;) | 09:39 |
| SamYaple | but thats done iniside the containers | 09:39 |
| vbel | sdake: ok, great! We run docker 1.8 with build-env patch to fetch anything from outside | 09:40 |
| sdake_ | vbel what is the use case for git vs curl? | 09:40 |
| sdake_ | trying to understand priorities here | 09:41 |
| SamYaple | sdake_: i think youre asking the wrong questions here | 09:41 |
| SamYaple | the --build-env stuff, we dont do that currently do we? | 09:41 |
| SamYaple | its not in the build-docker-image script | 09:41 |
| vbel | curl is to get tarball from outside, git is used to get stuff from internal mirrors (with or without patches) | 09:41 |
| sdake_ | vbel got it | 09:41 |
| sdake_ | everyone will want git for that use case | 09:42 |
| sdake_ | samyaple I dont know if we do that currently tbh | 09:42 |
| sdake_ | first i've heard of it | 09:42 |
| SamYaple | sdake_: but you said we can add it to the build.ini | 09:42 |
| vbel | yes, for official build process we can only use our/internal things | 09:42 |
| sdake_ | the use case is for supporting proxy environments for building | 09:43 |
| SamYaple | vbel: that doesnt need to be added yo ujust need to export that in your environment prior to building | 09:43 |
| vbel | also can be custom vendor things that are not accepted in upstream for many reasons :) | 09:43 |
| sdake_ | samyaple that sounds like a good solution - got docs? | 09:43 |
| SamYaple | sdake_: docker does | 09:43 |
| SamYaple | we dont break that funcationality | 09:44 |
| sdake_ | vbel can you research just using the environment for that | 09:45 |
| sdake_ | i'd prefer not to expand options in build.py | 09:45 |
| vbel | docker containers inside don't have proxy settings while building. E.g. we have public-yum site with our repos and we need to have http proxy settings just for build process | 09:45 |
| sdake_ | or build.ini | 09:45 |
| SamYaple | yes vbel, but yours just setting that in .buildconf right? | 09:46 |
| SamYaple | in that case you should just be setting that in your ENV before running the build command, thats not going to be added | 09:46 |
| sdake_ | buildconf just gets sourced by build.sh | 09:46 |
| vbel | yes, we have long awaited patch of docker that allows build-env flags | 09:46 |
| SamYaple | using buildconf like that is abuse from the original use | 09:46 |
| SamYaple | right but my point is that isnt affected by build.py, and build.py isnt going to read .buildconf | 09:47 |
| vbel | why it did not work with ENV is that you cannot unset it | 09:47 |
| SamYaple | not ENV in the container | 09:47 |
| sdake_ | there is a blueprint on this | 09:47 |
| SamYaple | # BUILDFLAGS="--build-flags-go-here" build.py | 09:47 |
| SamYaple | like that vbel | 09:48 |
| SamYaple | putting it in .buildconf is abuse | 09:48 |
| sdake_ | https://blueprints.launchpad.net/kolla/+spec/add-proxy-to-dockerfiles | 09:48 |
| SamYaple | sdake_: yea but I think that is fixed with 1.8 and unneeded | 09:48 |
| vbel | we have to pass those to docker | 09:48 |
| SamYaple | vbel: youre out of luck there, we use the docker-py python bindings | 09:49 |
| SamYaple | not the docker cli | 09:49 |
| sdake_ | samyaple i think what vbel is sying is docker reads BUILDFLAGS, whereas build.py does not | 09:49 |
| vbel | yes, this is one more reason we still are with Dockerfiles | 09:49 |
| SamYaple | sdake_: well thats not entirely true | 09:49 |
| inc0_ | I think you'll get buildtime envs in docker 1.8 | 09:49 |
| vbel | BUILDFLAGS are passed to "docker build" in old scheme | 09:50 |
| SamYaple | vbel: yes but thats abusing it | 09:50 |
| SamYaple | it wasnt desgned to do that | 09:50 |
| vbel | I will check if 1.8 actually passes anything to commands inside build | 09:50 |
| sdake_ | i really prefer nto to make our minimum requirements docker 1.8 | 09:51 |
| pbourke | the other idea for this which sdake had was to template ENV into the Dockerfiles | 09:51 |
| SamYaple | the issue with that pbourke is they stick around forever | 09:51 |
| vbel | yes, true | 09:51 |
| pbourke | SamYaple: true, but the workaround was to unset at the top of start.sh | 09:52 |
| SamYaple | I get the concerns here, but BUILDFLAGS was being abused before, it isnt supported in docker-py yet and wont be supported for the templates for a while | 09:52 |
| pbourke | buildflags is *out* imo | 09:52 |
| pbourke | it was never supported as part of docker | 09:52 |
| vbel | My fast hack is sed /RUN/RUN https_proxy=a.b.c.d/g | 09:52 |
| sdake_ | that definately sin't going in master | 09:53 |
| SamYaple | pbourke: no BUILDFLAGS is the name of the vairable that the build.sh used to run the docker build | 09:53 |
| pbourke | has sdake's method being dismissed? | 09:53 |
| SamYaple | like BUILDFLAGS="build --tag test ." docker ${BUILDFLAGS} | 09:53 |
| pbourke | *been | 09:53 |
| SamYaple | using it for more optiosn is abuse and not how that was designed to work | 09:53 |
| SamYaple | so we dont have that support in build.py or the templates, nor can we add it | 09:54 |
| sdake_ | i dont want to permit overloading of the build flag options to docker - likely to cause confusion | 09:54 |
| SamYaple | nor could we | 09:55 |
| SamYaple | cli vs python bindings | 09:55 |
| SamYaple | big difference | 09:55 |
| pbourke | ahem | 09:55 |
| sdake_ | if buildflags is --tag latest --tag llama that makes no sense | 09:55 |
| pbourke | templating? ENV? | 09:55 |
| SamYaple | pbourke: sure if youre ok with that stick around forever | 09:55 |
| SamYaple | I am ok with that | 09:55 |
| SamYaple | but thats not what vbel is talking about | 09:55 |
| pbourke | SamYaple: its not ideal but its the best we came up with | 09:55 |
| sdake_ | read: http://eavesdrop.openstack.org/irclogs/%23kolla/%23kolla.2015-08-11.log.html#t2015-08-11T16:11:19 | 09:55 |
| SamYaple | I am not fighting that at all pbourke | 09:56 |
| pbourke | SamYaple: ok :) | 09:56 |
| SamYaple | But i want everyone on the same page | 09:56 |
| *** dims has joined #kolla | 09:56 | |
| pbourke | vbel: save me the scroll back, is there an issue with the above approach? | 09:56 |
| sdake_ | i dont mind compomises but clearly bulidflags wont work | 09:56 |
| sdake_ | so le sjust shit can that idea now | 09:56 |
| vbel | we are for any solution that works for such environments like us :) | 09:56 |
| sdake_ | read the eavesdrop | 09:56 |
| sdake_ | that is the only working solution | 09:57 |
| SamYaple | we can add support for buildenv stuff when docker-py _and_ we bump up the requirements to the latest docker-py | 09:57 |
| SamYaple | I too want that | 09:57 |
| sdake_ | i thought buildenv got killed off? | 09:57 |
| pbourke | I wanted to discuss this in the meeting but there wasn't time | 09:57 |
| pbourke | I think either vbel or myself can submit a WIP for the above approach | 09:58 |
| sdake_ | pbourke sounds like a good course of action | 09:58 |
| sdake_ | lets bikeshed it in the review ;) | 09:58 |
| sdake_ | once we see what it looks like | 09:58 |
| sdake_ | the only downside to the proposed approach is it leaks environment data | 09:59 |
| sdake_ | build systems do that all the time | 09:59 |
| sdake_ | its normal normal behavior | 09:59 |
| *** Kennan2 has joined #kolla | 10:03 | |
| SamYaple | vbel are you using a custom docker1.8? | 10:03 |
| vbel | currently - yes. Another way to avoid it would be running a VM and kind of transparent proxy | 10:04 |
| SamYaple | transparent proxy + https wont work | 10:04 |
| *** Kennan has quit IRC | 10:04 | |
| SamYaple | not without having the client cert (which would be autogenerated) | 10:04 |
| SamYaple | or the destination cert if yo ucontrol that | 10:05 |
| SamYaple | that would work too | 10:05 |
| inc0_ | I've been playing with man in the middle proxy stuff | 10:07 |
| sdake_ | ok well lets sort out a wip today plz ,I want docker dir gone aap :) | 10:07 |
| inc0_ | but that requires CA cets exposed to dockers | 10:07 |
| SamYaple | inc0_: yea i do transparent proxy at my host for https | 10:08 |
| SamYaple | basically mitm | 10:08 |
| sdake_ | pbourke this needs a review plz | 10:08 |
| SamYaple | all my device certs are controlled so i can do that | 10:08 |
| pbourke | sdake_: link? | 10:08 |
| sdake_ | https://review.openstack.org/#/c/213478/4 | 10:10 |
| sdake_ | ok so with that review docker_templates is done I think | 10:14 |
| sdake_ | we need some type of environemntal overrides to deal with proxy nonsense during builds | 10:14 |
| sdake_ | we need pull from git in build.ini | 10:15 |
| *** sdake_ is now known as sdake | 10:15 | |
| sdake | what else is needed to pull the plug on the docker dir? | 10:15 |
| vbel | GIT_REF to specify tag to pull | 10:16 |
| sdake | that is git pull from build.ini | 10:16 |
| vbel | ok | 10:16 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove start.yml and bootstrap.yml https://review.openstack.org/213958 | 10:17 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove start.yml and bootstrap.yml https://review.openstack.org/213958 | 10:17 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix up base, RabbitMQ, and MariaDB for Ubuntu https://review.openstack.org/214625 | 10:18 |
| sdake | somethign is busted with gerrit | 10:18 |
| SamYaple | yea ive noticed that too | 10:18 |
| SamYaple | probably related to the crash the other week | 10:18 |
| * SamYaple is making things up | 10:18 | |
| *** britthou_ has quit IRC | 10:20 | |
| SamYaple | sdake: the ubuntu mariadb container is 584MB | 10:22 |
| SamYaple | centos is 1.17GB | 10:22 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Add build from source and templating for Zaqar https://review.openstack.org/202441 | 10:22 |
| sdake | samyaple epic difference ;) | 10:23 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Just a test patch to test building with templates https://review.openstack.org/213538 | 10:24 |
| sdake | pbourke master is busted completely without an ack on this https://review.openstack.org/#/c/215002/ | 10:25 |
| SamYaple | sdake: its not completely busted | 10:26 |
| SamYaple | I still use it ;) | 10:26 |
| sdake | if your db is already initialized it would probably work :) | 10:26 |
| SamYaple | nah its just a config option | 10:26 |
| SamYaple | if you have database_user: root in your globals.yml itll work | 10:26 |
| sdake | which config option? | 10:26 |
| sdake | oh i see | 10:27 |
| SamYaple | i mean the patch still needs to merge, but you can workaround the issue | 10:27 |
| sdake | cool if people ask before it merges point em at that workaround :) | 10:27 |
| sdake | i have hit the wall | 10:27 |
| SamYaple | bai | 10:27 |
| sdake | bedtime | 10:27 |
| sdake | samyaple before i go to bed | 10:30 |
| openstackgerrit | Merged stackforge/kolla: Use database_user_create in the register operation https://review.openstack.org/215002 | 10:30 |
| sdake | is one of the gaps in shade your sorting out related to keystone domains? | 10:31 |
| SamYaple | yes | 10:31 |
| SamYaple | namely domain_create doesnt exist ;) | 10:31 |
| sdake | any rough idea when that will land - it is blocking heat | 10:31 |
| *** weiyu_ has quit IRC | 10:31 | |
| SamYaple | nope syncing up with the shade folks has been slow | 10:31 |
| sdake | we can't create domain in the bootstrap | 10:31 |
| SamYaple | We can, if we move to v3 only keystone | 10:32 |
| SamYaple | but that busts kilo | 10:32 |
| sdake | becuse the domain id needs to be stored in the heat engine's conf file | 10:32 |
| SamYaple | which i dont test | 10:32 |
| sdake | otherwise stack delete doesnt' work | 10:32 |
| SamYaple | yea i know the issue. i can get it to work with v3 only keystone | 10:32 |
| SamYaple | no v3 endpoints | 10:32 |
| SamYaple | no v2.0 endpoints* | 10:32 |
| sdake | how do you get the domain id into heat.conf in the engine container? | 10:34 |
| SamYaple | magic! | 10:34 |
| SamYaple | I can do a quick module to query keystone to fill that gap | 10:34 |
| SamYaple | maybe | 10:34 |
| SamYaple | no i can | 10:35 |
| SamYaple | if we have v3 endpints | 10:35 |
| SamYaple | thats what i was blocked on before | 10:35 |
| openstackgerrit | Merged stackforge/kolla: Update support to control https://review.openstack.org/213480 | 10:37 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix group name in vagrant docs https://review.openstack.org/213950 | 10:38 |
| SamYaple | /win/win 35 | 10:38 |
| sdake | don't we have v3 endpoints already? | 10:39 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: rsyslog container https://review.openstack.org/214589 | 10:39 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 10:39 |
| inc0_ | test it out please | 10:40 |
| SamYaple | inc0_: that means all the services log to files | 10:40 |
| inc0_ | to files generated by rsyslog | 10:41 |
| inc0_ | for now, until we'll have logstash | 10:41 |
| SamYaple | how do they get to rsyslog? | 10:41 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: WIP: Implement Ansible role for Swift https://review.openstack.org/215065 | 10:41 |
| inc0_ | this commits create rsyslog server | 10:42 |
| inc0_ | we'll need to turn every service to log to this one now | 10:42 |
| *** weiyu_ has joined #kolla | 10:42 | |
| inc0_ | but we need server first | 10:42 |
| inc0_ | now we need to think how to teach services to use rsyslog clients | 10:42 |
| inc0_ | without running rsyslog in every containert | 10:43 |
| sdake | i dont want services blocked on logging to rsyslog | 10:44 |
| sdake | i'd prefer they log to files and logstash jam that to rsyslog | 10:44 |
| sdake | and on that note, i'm going to bed :) | 10:44 |
| SamYaple | inc0_: understandabled | 10:44 |
| SamYaple | sdake: they will be | 10:44 |
| SamYaple | thats how it goes | 10:44 |
| SamYaple | not all services can even log to files | 10:44 |
| inc0_ | service -> rsyslog -> logstash | 10:44 |
| SamYaple | yep | 10:45 |
| SamYaple | service -> rsyslog -> file|logstash | 10:45 |
| inc0_ | later on logstash -> elasticsearch -> kibana | 10:45 |
| SamYaple | both even | 10:45 |
| *** shardy_ has joined #kolla | 10:48 | |
| *** shardy has quit IRC | 10:48 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 10:52 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: rsyslog container https://review.openstack.org/214589 | 10:52 |
| *** shardy_ has quit IRC | 10:53 | |
| *** shardy has joined #kolla | 10:54 | |
| *** unicell1 has joined #kolla | 10:57 | |
| inc0_ | SamYaple, how about installing rsyslog on host and having docker log driver use it? | 10:57 |
| inc0_ | this will solve everything logging to stdin | 10:57 |
| *** unicell has quit IRC | 10:58 | |
| inc0_ | which is...well pretty much everything | 10:59 |
| SamYaple | we dont touch host | 10:59 |
| SamYaple | and we arent using the docker log driver | 10:59 |
| SamYaple | not all things print to stdout/stderr | 10:59 |
| SamYaple | which rules that out | 10:59 |
| inc0_ | what doesn't? | 11:00 |
| SamYaple | even the openstack services dont all work well with stdout | 11:00 |
| SamYaple | but keepalived haproxy only work with rsyslog | 11:00 |
| SamYaple | openvswitch wont log to stdout | 11:00 |
| SamYaple | libvirt wont do debug to stdout | 11:01 |
| SamYaple | and again, openstack is buggy with that too | 11:01 |
| inc0_ | then how would you give them access to rsyslog withot having rsyslog in container? | 11:01 |
| SamYaple | what do you mean? | 11:01 |
| SamYaple | we want rsyslog in a container | 11:01 |
| inc0_ | we don't want rsyslog process in every container right? | 11:01 |
| SamYaple | no | 11:01 |
| SamYaple | one rsyslog container | 11:01 |
| SamYaple | all services logging to it | 11:01 |
| inc0_ | my question - how? | 11:02 |
| SamYaple | /dev/log | 11:02 |
| inc0_ | so by exposing /dev/log it works? | 11:02 |
| SamYaple | you have to have /dev/log from the rsyslog container be /dev/log in the other containers | 11:02 |
| inc0_ | I'll test it | 11:02 |
| SamYaple | not /dev/log on the host | 11:02 |
| inc0_ | well, pids will get messy anyway | 11:03 |
| inc0_ | since several pid 1 services will write to same /dev/log | 11:03 |
| inc0_ | I'm not sure rsyslog can handle that well | 11:03 |
| SamYaple | we can provide a /dev/log for each service | 11:03 |
| SamYaple | /dev/log-nova /dev/log-neutron | 11:04 |
| *** unicell1 has quit IRC | 11:04 | |
| inc0_ | uhh...it will be /dev/log-nova-scheduler and /dev/log-nova-api and so on | 11:04 |
| inc0_ | which means we'll need to create all that | 11:04 |
| *** unicell has joined #kolla | 11:04 | |
| SamYaple | yup | 11:04 |
| inc0_ | prior to deploying rest of stack | 11:04 |
| inc0_ | so ansible has to know what will be deployed on given host | 11:05 |
| SamYaple | there will be an rasyslog server on each host dude | 11:05 |
| inc0_ | at begginging of play | 11:05 |
| inc0_ | server? why? | 11:05 |
| inc0_ | don't we want few servers and clients on each host? | 11:05 |
| *** weiyu_ has quit IRC | 11:06 | |
| SamYaple | client cant do what we need | 11:06 |
| SamYaple | everything logs to the server | 11:06 |
| SamYaple | hence each server has a rsyslogserver | 11:06 |
| *** shardy_ has joined #kolla | 11:06 | |
| inc0_ | wait, so we don't do remote rsyslog? | 11:07 |
| inc0_ | we wanted centralized logging | 11:07 |
| SamYaple | what do you think is providing /dev/log | 11:07 |
| SamYaple | am i mistaken here in my understanding | 11:08 |
| *** shardy has quit IRC | 11:08 | |
| SamYaple | client wont provide /dev/log, yes? | 11:08 |
| inc0_ | I'm not sure, it might | 11:08 |
| SamYaple | it doesnt | 11:08 |
| SamYaple | its a client | 11:08 |
| inc0_ | yeah, but server listens on port | 11:08 |
| SamYaple | server listens | 11:08 |
| inc0_ | so client has to get these logs somehow right? | 11:09 |
| SamYaple | port and device | 11:09 |
| inc0_ | question is, how | 11:09 |
| SamYaple | server creates /dev/log | 11:09 |
| SamYaple | each node writes to /dev/log | 11:09 |
| SamYaple | we arent doing central logging yet | 11:09 |
| SamYaple | thats what ELKstack is for | 11:09 |
| inc0_ | we can do this with rsyslog as well | 11:09 |
| *** zhiwei has quit IRC | 11:10 | |
| SamYaple | yea but we arent... | 11:10 |
| inc0_ | why | 11:10 |
| inc0_ | ? | 11:10 |
| SamYaple | thats what elkstack was for | 11:10 |
| SamYaple | we arent reinventing that stuff | 11:10 |
| inc0_ | elkstack was to be pluggable from the begginging | 11:10 |
| SamYaple | we want all logging to be accissible in one place... per node | 11:10 |
| inc0_ | and we're not, rsyslog is there for some time | 11:10 |
| SamYaple | initially we are not doing anythign with central logging | 11:11 |
| SamYaple | that was already agree | 11:11 |
| SamYaple | we can have that converastion after | 11:11 |
| SamYaple | we just need each server to have a central point for forwarding | 11:11 |
| SamYaple | rsyslog on each server that collects and writes all logs to file | 11:11 |
| inc0_ | also, that doesn't change anything | 11:11 |
| inc0_ | we still need to get logs to rsyslog | 11:11 |
| *** shardy_ has quit IRC | 11:11 | |
| inc0_ | remote or not | 11:11 |
| SamYaple | /dev/log | 11:11 |
| inc0_ | having different /dev/log per every container seems ugly to me. | 11:12 |
| *** britthouser has joined #kolla | 11:12 | |
| *** shardy has joined #kolla | 11:12 | |
| SamYaple | well test a single /dev/log | 11:12 |
| SamYaple | you said it wouldnt work, i wasnt going to qeustion you | 11:12 |
| inc0_ | I will, I expect it to fail | 11:13 |
| inc0_ | but I'll explore other ideas as well | 11:13 |
| *** unicell has quit IRC | 11:13 | |
| SamYaple | ok, but it does need to be a /dev/ since that is the only thing all these servcies have in common to log to | 11:13 |
| *** unicell has joined #kolla | 11:13 | |
| inc0_ | because I don't want to redeploy/reconfigure rsyslog container each time we start/stop new container | 11:13 |
| SamYaple | well you wouldnt | 11:14 |
| inc0_ | how then new /dev/log-whatever would be created? | 11:14 |
| SamYaple | clearly we would (if we had to go down that road) create all the possible devices first | 11:14 |
| inc0_ | that's ugly and will cause tons of problems | 11:14 |
| SamYaple | it wont cause any problems | 11:15 |
| SamYaple | it is ugly | 11:15 |
| inc0_ | remembering that we need to add a device each time we add a container to kolla will break | 11:15 |
| inc0_ | it will cause problems on development cycle | 11:15 |
| *** dims has quit IRC | 11:15 | |
| SamYaple | .... for that container... one time | 11:15 |
| inc0_ | still, trust me, people will forget about that | 11:15 |
| inc0_ | we want this to be pluggable | 11:15 |
| SamYaple | sure... until they look for the logs | 11:15 |
| *** dims has joined #kolla | 11:16 | |
| SamYaple | which is like the first thing anyone does adding a new service | 11:16 |
| SamYaple | to see if it works | 11:16 |
| inc0_ | in which case we'll get tons of bugs "omg omg my logs are gone, world is ending" | 11:16 |
| SamYaple | no it wont ever merge see above | 11:16 |
| SamYaple | but this isnt a conversation i need to have | 11:16 |
| SamYaple | try testing it first | 11:16 |
| SamYaple | i think youll be fine with a single /dev/log | 11:16 |
| inc0_ | yeah, I'll try to get cleaner solution | 11:16 |
| inc0_ | hopefully | 11:17 |
| *** dims has quit IRC | 11:20 | |
| *** rhallisey has joined #kolla | 11:25 | |
| *** dims has joined #kolla | 11:33 | |
| *** weiyu_ has joined #kolla | 11:44 | |
| *** rhallisey has quit IRC | 11:49 | |
| *** rhallisey has joined #kolla | 11:49 | |
| *** britthouser has quit IRC | 11:55 | |
| *** weiyu_ has quit IRC | 12:00 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add kolla-ansible for ubuntu https://review.openstack.org/215112 | 12:13 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add kolla-ansible for ubuntu https://review.openstack.org/215112 | 12:14 |
| SamYaple | pbourke: hey | 12:16 |
| SamYaple | i figured something out | 12:16 |
| SamYaple | i figured out why the docker cache is always busted | 12:16 |
| SamYaple | it uses more than just a file checksum | 12:16 |
| SamYaple | it uses mtime and even inode in the case of ADD | 12:17 |
| SamYaple | how wierd is that | 12:17 |
| *** sdake has quit IRC | 12:17 | |
| *** shardy_ has joined #kolla | 12:19 | |
| *** shardy has quit IRC | 12:20 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add kolla-ansible for ubuntu https://review.openstack.org/215112 | 12:20 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix up base, RabbitMQ, and MariaDB for Ubuntu https://review.openstack.org/214625 | 12:20 |
| *** shardy_ has quit IRC | 12:24 | |
| *** shardy has joined #kolla | 12:25 | |
| SamYaple | im writing a python patch to set the atime and mtime to 0 epoch so the cache wont break when building | 12:29 |
| *** banix has joined #kolla | 12:32 | |
| *** unicell1 has joined #kolla | 12:34 | |
| *** unicell has quit IRC | 12:34 | |
| *** britthouser has joined #kolla | 12:36 | |
| *** banix has quit IRC | 12:40 | |
| *** diga has quit IRC | 12:44 | |
| pbourke | SamYaple: you mean for the source tars? | 12:49 |
| SamYaple | I mean for everything, but yes that too | 12:50 |
| SamYaple | this will fix the cache for everything | 12:50 |
| pbourke | sounds great! | 12:50 |
| SamYaple | so far its working fantastic! | 12:50 |
| pbourke | what else was the cache not working for though out of interest? | 12:50 |
| SamYaple | most everything, esecially after a git pull | 12:50 |
| SamYaple | basically if you used `vi` to open a file, you break the cache | 12:51 |
| pbourke | I wonder is there a reason for that though | 12:51 |
| pbourke | or just an oversight | 12:51 |
| SamYaple | no there are some bugs for it | 12:51 |
| SamYaple | its 'working as designed' | 12:51 |
| SamYaple | some people might care about atime and mtime | 12:51 |
| SamYaple | we do not | 12:51 |
| pbourke | fair enough | 12:51 |
| pbourke | btw | 12:52 |
| pbourke | have you have problems with the config merging in ansible breaking when the template isn't an ini file? | 12:52 |
| SamYaple | yea, it doesnt work | 12:52 |
| SamYaple | specifically that | 12:52 |
| pbourke | so just no augment files in that case? | 12:52 |
| SamYaple | the config merging is a module i wrote to merge ini files | 12:52 |
| SamYaple | yea | 12:53 |
| pbourke | kk | 12:53 |
| SamYaple | the plan is to have a full override file | 12:53 |
| SamYaple | like, 'this stomps the ansible template' overrie | 12:53 |
| SamYaple | rabbitmq, keepalived, haproxy, dnsmasq, are the uses i see | 12:53 |
| SamYaple | all the others can merge | 12:53 |
| pbourke | I was wondering why haproxy had a different style config.yml actually | 12:54 |
| pbourke | now i know | 12:54 |
| *** rhallisey_ has joined #kolla | 13:00 | |
| *** rhallisey has quit IRC | 13:02 | |
| inc0_ | soo...logging to simple /dev/log works as I thought...from rsyslog perspective it's all same pid = same service | 13:04 |
| SamYaple | uhh inc0_ thats what the rules are for though | 13:08 |
| SamYaple | you have to config that stuff correctly | 13:08 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add ansible role for memcached https://review.openstack.org/214593 | 13:08 |
| *** stvnoyes has joined #kolla | 13:10 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 13:14 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 13:14 |
| *** rhallisey_ is now known as rhallisey | 13:17 | |
| *** apuimedo has joined #kolla | 13:29 | |
| apuimedo | SamYaple: ping | 13:29 |
| SamYaple | apuimedo: pong | 13:29 |
| apuimedo | that was fast! | 13:30 |
| apuimedo | thanks SamYaple | 13:30 |
| SamYaple | what did i do? | 13:30 |
| SamYaple | did I win a prize? | 13:30 |
| apuimedo | you were picked from the mirrorlist | 13:30 |
| apuimedo | as the fastest server | 13:30 |
| apuimedo | of course | 13:30 |
| apuimedo | :-) | 13:30 |
| SamYaple | thats what my wife tells me | 13:30 |
| apuimedo | SamYaple: I have never tried kolla in multinode, is it being tested like that? | 13:30 |
| SamYaple | apuimedo: thats almost all i do | 13:31 |
| apuimedo | SamYaple: speed and efficiency, what more can she ask for? | 13:31 |
| SamYaple | i use AIO everynow and then just to maek sure it works | 13:31 |
| SamYaple | :) | 13:31 |
| apuimedo | SamYaple: do we have some link for a guide to the multinode deployment? | 13:33 |
| SamYaple | AIO and multinode is exactly the same deploymentwise | 13:33 |
| SamYaple | the only difference is the inventory file for ansible | 13:34 |
| apuimedo | SamYaple: just using swarm with compose, then? | 13:34 |
| SamYaple | no we dont use compose | 13:34 |
| SamYaple | not with ansible | 13:34 |
| SamYaple | ansible starts the appropriate containers | 13:34 |
| SamYaple | on the appropriate hosts | 13:34 |
| apuimedo | mmm, I thought I had seen compose the last time with the all in one | 13:34 |
| SamYaple | yea that was the old way, but never with ansible | 13:35 |
| apuimedo | SamYaple: cool | 13:35 |
| SamYaple | the docs here work apuimedo https://github.com/stackforge/kolla/blob/master/docs/ansible-deployment.md | 13:35 |
| apuimedo | thanks | 13:35 |
| *** dims_ has joined #kolla | 13:35 | |
| SamYaple | for multinode look at https://github.com/stackforge/kolla/blob/master/ansible/inventory/multinode | 13:35 |
| SamYaple | basically you just setup those first 4 groups with the servers you want (the same server can be in multiple groups) | 13:36 |
| apuimedo | I hope with this I'll be able to convince my team to use kolla for CI (which in turn will add midonet to kolla) | 13:36 |
| SamYaple | Here is what I will say, Kolla is in heavy development. We will have a stable working branch for Liberty, but its bumpy right now | 13:37 |
| SamYaple | you have to build your images, host them somewhere (local docker registry), etc | 13:37 |
| SamYaple | so we wont have a 'it just works' deploy until LIberty branch | 13:37 |
| SamYaple | (2 months) | 13:37 |
| *** diogogmt has joined #kolla | 13:39 | |
| *** dims has quit IRC | 13:39 | |
| *** dims_ has quit IRC | 13:44 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 13:45 |
| *** dims has joined #kolla | 13:46 | |
| pbourke | SamYaple: thanks for reviews | 13:48 |
| SamYaple | pbourke: no problem | 13:49 |
| SamYaple | i think we zeroing in on the memcache one | 13:49 |
| SamYaple | ill help you tweak the swift one to consume memcache properly | 13:50 |
| pbourke | great | 13:51 |
| pbourke | on your latest comment | 13:51 |
| pbourke | do we need a group memcache? I removed it from the inventories as it just gets picked up as swift dependency | 13:51 |
| pbourke | or maybe we dont want memcache on every swift node | 13:52 |
| *** jtriley has joined #kolla | 13:53 | |
| apuimedo | SamYaple: thanks for that info! | 13:55 |
| *** tfukushima has quit IRC | 13:57 | |
| *** dwalsh has joined #kolla | 14:02 | |
| *** diogogmt has quit IRC | 14:04 | |
| jpeeler | SamYaple: did you get around to making build.py build single image? i was actually working on it too | 14:04 |
| *** rhallisey has quit IRC | 14:04 | |
| *** inc0_ has quit IRC | 14:04 | |
| *** rhallisey has joined #kolla | 14:05 | |
| jpeeler | i have a doctor apt me in a few, so it won't get done immediately. but i'll finish it up if i hear you don't have a patch started. | 14:07 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 14:08 |
| *** jtriley_ has joined #kolla | 14:09 | |
| *** jtriley has quit IRC | 14:11 | |
| pbourke | SamYaple: for comment here https://review.openstack.org/#/c/215065/1/ansible/roles/swift/tasks/config.yml | 14:13 |
| pbourke | which part is not consistent? | 14:13 |
| *** diogogmt has joined #kolla | 14:17 | |
| SamYaple | jpeeler: just tonight i was working on it, i moved onto other things though not alot of progress there | 14:22 |
| SamYaple | jpeeler: you do it! | 14:22 |
| SamYaple | pbourke: using variable name in the group | 14:23 |
| pbourke | SamYaple: i thought that, is that all though? | 14:23 |
| SamYaple | pbourke: memcache is a tricky thing, you dont want it running on your swift nodes, no | 14:23 |
| SamYaple | pbourke: yea thats all | 14:23 |
| pbourke | SamYaple: it seems like there might be two complaints in that comment | 14:24 |
| pbourke | ok cool | 14:24 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Allow Docker to make better use of cache https://review.openstack.org/215149 | 14:27 |
| SamYaple | pbourke: ^^ it works! | 14:27 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Allow Docker to make better use of cache https://review.openstack.org/215149 | 14:27 |
| *** banix has joined #kolla | 14:31 | |
| *** diogogmt has quit IRC | 14:34 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix up base for ubuntu https://review.openstack.org/215157 | 14:40 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix RabbitMQ for Ubuntu https://review.openstack.org/215158 | 14:40 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix up base for ubuntu https://review.openstack.org/215157 | 14:40 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix MariaDB for ubuntu https://review.openstack.org/215159 | 14:40 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix up base for ubuntu https://review.openstack.org/215157 | 14:43 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix MariaDB for ubuntu https://review.openstack.org/215159 | 14:43 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix RabbitMQ for Ubuntu https://review.openstack.org/215158 | 14:43 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add kolla-ansible for ubuntu https://review.openstack.org/215112 | 14:44 |
| *** ccrouch has joined #kolla | 14:49 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix keystone for Ubuntu https://review.openstack.org/215163 | 14:50 |
| *** ccrouch has quit IRC | 14:50 | |
| *** ccrouch has joined #kolla | 14:50 | |
| *** inc0 has joined #kolla | 14:52 | |
| pbourke | SamYaple: maybe related to discussion earlier, are you against build options going into build.ini | 14:52 |
| *** jasonamyers has joined #kolla | 14:52 | |
| SamYaple | pbourke: what do you mean? | 14:53 |
| *** shardy_ has joined #kolla | 14:53 | |
| pbourke | SamYaple: im looking again at the neutron third party plugins (https://review.openstack.org/#/c/207545/) | 14:53 |
| pbourke | im thinking an alternative would be to have something like {% if enable_plugins %} in the template | 14:53 |
| SamYaple | like global build options? rather than having to specify flags? | 14:53 |
| pbourke | as far as I can see this code has to be part of the neutron-server | 14:54 |
| pbourke | I may be wrong | 14:54 |
| *** shardy has quit IRC | 14:54 | |
| SamYaple | im ok with extra stuff being in the agents and server container to behonest | 14:55 |
| SamYaple | i dont actually like this data vol plugin thing | 14:55 |
| SamYaple | i would prefer it just be available | 14:55 |
| pbourke | it just seems like a lot of bloat | 14:55 |
| pbourke | no? | 14:55 |
| SamYaple | define alot? | 14:55 |
| SamYaple | i dont know how big is it? | 14:55 |
| pbourke | like 10-15 repos | 14:55 |
| pbourke | each are reasonably small but I reckon 3/4 of people wont use them | 14:55 |
| SamYaple | I mean if we are talking 5-10mb i say go for it | 14:56 |
| SamYaple | 100-500mb thats a different story | 14:56 |
| SamYaple | whats the size? | 14:56 |
| pbourke | i'll check | 14:56 |
| pbourke | on the same topic does that mean you'd be ok with adding the murano plugin directly into the horizon dockerfile? | 14:57 |
| pbourke | no conditional | 14:57 |
| SamYaple | the issue there is it will then show up in the dashboard, correct? | 14:57 |
| pbourke | yes | 14:57 |
| *** pbourke has quit IRC | 14:57 | |
| SamYaple | yes so i wouldnt be ok there | 14:57 |
| SamYaple | ther neutron stuff is just dormat files | 14:57 |
| SamYaple | correct me if im wrong there | 14:58 |
| *** pbourke has joined #kolla | 14:58 | |
| *** shardy_ has quit IRC | 14:58 | |
| pbourke | not sure if I missed a line there | 14:58 |
| SamYaple | the neutron stuff is just dormat files, right? not active unless configured? | 14:59 |
| *** shardy has joined #kolla | 14:59 | |
| pbourke | correct | 14:59 |
| SamYaple | yea thats the difference for me on the neutron vs horizon stuff | 14:59 |
| pbourke | makes sense, I'll have another think wrt the murano add on | 15:00 |
| *** jruano has joined #kolla | 15:01 | |
| SamYaple | hot damn this works across git checkouts very well too | 15:01 |
| SamYaple | https://review.openstack.org/#/c/215149/ | 15:02 |
| SamYaple | lets get that merged in! | 15:02 |
| pbourke | how do you use build.py day to day without single image building? | 15:04 |
| pbourke | I've not been using it for that reason | 15:05 |
| inc0 | let me get this commit up and I'm going to do this single img build | 15:06 |
| inc0 | this is pain. | 15:06 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 15:06 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Templates for rsyslog https://review.openstack.org/215181 | 15:06 |
| pbourke | inc0: I'll owe you a drink :) | 15:06 |
| *** blahRus has joined #kolla | 15:08 | |
| SamYaple | inc0: jpeeler just poped on and said he had code already for that | 15:08 |
| pbourke | SamYaple: Im going to need git for binary neutron now also | 15:08 |
| SamYaple | inc0: he said hell be pushing it later today | 15:08 |
| pbourke | SamYaple: would you move it to the base? | 15:08 |
| pbourke | in your patch | 15:09 |
| SamYaple | pbourke: git is only installed for ubuntu | 15:09 |
| pbourke | inc0: looks like jpeeler gets the drink instead | 15:09 |
| inc0 | waiting on jpeeler then | 15:09 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 15:10 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Templates for rsyslog https://review.openstack.org/215181 | 15:10 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: rsyslog container https://review.openstack.org/214589 | 15:10 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Ceph container https://review.openstack.org/214706 | 15:10 |
| inc0 | SamYaple, mind removing -1 from first patch since templates are done in different one? | 15:11 |
| pbourke | SamYaple: will need it for all though | 15:11 |
| SamYaple | no put the templates in the first one | 15:11 |
| SamYaple | its already causing review problems | 15:11 |
| SamYaple | you are duping start.sh | 15:11 |
| pbourke | SamYaple: to clone the plugin repos | 15:11 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 15:12 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: rsyslog container https://review.openstack.org/214589 | 15:12 |
| SamYaple | pbourke: hmm that needs to be discussed then | 15:12 |
| SamYaple | pbourke: git clone is less than ideal | 15:13 |
| pbourke | SamYaple: tarballs.openstack.org? | 15:13 |
| SamYaple | do they exist there? | 15:13 |
| pbourke | yeah | 15:13 |
| pbourke | is size the worry? | 15:13 |
| SamYaple | oh yea yea definetely. But the size should be figured out first | 15:13 |
| SamYaple | and we need to talk about this more boradly, this seems like a good mailing list conversation | 15:14 |
| *** dwalsh has quit IRC | 15:14 | |
| pbourke | I'll send one. Please respond with your concerns as although usually valid they're not always clear to me :) | 15:15 |
| SamYaple | :) | 15:15 |
| *** diogogmt has joined #kolla | 15:16 | |
| *** kfox1111 has quit IRC | 15:19 | |
| *** diogogmt has quit IRC | 15:19 | |
| *** diogogmt has joined #kolla | 15:22 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Allow Docker to make better use of cache https://review.openstack.org/215149 | 15:35 |
| *** dwalsh has joined #kolla | 15:35 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 15:45 |
| *** ccrouch has quit IRC | 15:45 | |
| *** ccrouch has joined #kolla | 15:45 | |
| *** ccrouch has quit IRC | 15:57 | |
| *** ccrouch has joined #kolla | 15:57 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix keystone for Ubuntu https://review.openstack.org/215163 | 15:59 |
| *** jasonamyers has quit IRC | 16:00 | |
| *** SourabhP has joined #kolla | 16:02 | |
| *** unicell has joined #kolla | 16:03 | |
| *** unicell1 has quit IRC | 16:05 | |
| *** dwalsh has quit IRC | 16:05 | |
| *** SourabhP_ has joined #kolla | 16:05 | |
| *** sdake has joined #kolla | 16:06 | |
| *** SourabhP has quit IRC | 16:08 | |
| *** SourabhP_ is now known as SourabhP | 16:08 | |
| sdake | morning | 16:10 |
| SamYaple | morning sdake | 16:10 |
| SamYaple | i have good news | 16:10 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix keystone for Ubuntu https://review.openstack.org/215163 | 16:13 |
| sdake | shoot | 16:14 |
| sdake | morning btw | 16:14 |
| sdake | still waking up so don't give me any brain teasers pl zz :) | 16:14 |
| *** SourabhP has quit IRC | 16:15 | |
| SamYaple | I have fixed the docker cache! It works in a bunch of cases where it didnt before! | 16:15 |
| SamYaple | https://review.openstack.org/#/c/215149/ | 16:15 |
| SamYaple | from source uses the cache properly now | 16:15 |
| SamYaple | too | 16:16 |
| sdake | surprised that works | 16:18 |
| SamYaple | there are some complaints around it | 16:19 |
| SamYaple | but for the docker folks they say if a timestamp changes, the file change and cache is bad | 16:19 |
| SamYaple | we dont care about timestamps and want the cache more than the timestamp | 16:19 |
| sdake | docker cache still seems broke from binary builds | 16:21 |
| sdake | if i build and rebuild it takes same time and does same thing | 16:21 |
| SamYaple | it doesnt take the same time as a full rebuild | 16:21 |
| SamYaple | it may take the same time because it is failing to build an image | 16:21 |
| SamYaple | there are also still some thing we have in there that break the cache i believe (like VOLUME that breaks it i think) | 16:22 |
| sdake | those need to go anyway | 16:22 |
| SamYaple | I asked to rebase your patch to remove that stuff last night | 16:22 |
| sdake | this is not the volume you were looking for | 16:22 |
| sdake | rembase which one | 16:23 |
| sdake | i dont recall editing anything with volumes | 16:23 |
| SamYaple | https://review.openstack.org/#/c/212830/ | 16:23 |
| sdake | vote -1 on it so i know you want hat | 16:24 |
| sdake | i must have missed that commentary | 16:24 |
| SamYaple | hey pbourke or rhallisey would you mind merging the last of the docker_templates for the main services? https://review.openstack.org/#/c/213478/ | 16:24 |
| sdake | that patc has been through the rebase ringer | 16:24 |
| SamYaple | yea i know :/ | 16:25 |
| SamYaple | also, guys I know no one is using ubuntu, but I need eyes on the ubuntu reviews | 16:25 |
| SamYaple | ill be using that as my primary once its ready | 16:26 |
| SamYaple | i should have all the patches in within the week | 16:26 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 16:28 |
| openstackgerrit | Merged stackforge/kolla: Add Dockerfile template for nova https://review.openstack.org/213478 | 16:28 |
| sdake | i'll probably end up trying it out once its done | 16:29 |
| sdake | and of course i redview several times a day | 16:29 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 16:29 |
| *** akwasnie has quit IRC | 16:29 | |
| *** sdake_ has joined #kolla | 16:30 | |
| *** ccrouch has quit IRC | 16:30 | |
| *** ccrouch has joined #kolla | 16:30 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add Ansible role for memcached https://review.openstack.org/214593 | 16:31 |
| *** sdake has quit IRC | 16:33 | |
| SamYaple | looks good pbourke :) | 16:34 |
| pbourke | SamYaple: thanks! I noticed the port thing before I saw your comment | 16:34 |
| *** sdake_ is now known as sdake | 16:34 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: WIP: Implement Ansible role for Swift https://review.openstack.org/215065 | 16:37 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: WIP: Implement Ansible role for Swift https://review.openstack.org/215065 | 16:38 |
| *** diga has joined #kolla | 16:39 | |
| sdake | i had a dream my porsche was being reposessed | 16:43 |
| sdake | which is odd because my bill paying on it is automatic | 16:44 |
| sdake | bad way to start out the day | 16:44 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: WIP: Implement Ansible role for Swift https://review.openstack.org/215065 | 16:44 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: WIP: Implement Ansible role for Swift https://review.openstack.org/215065 | 16:45 |
| *** shardy has quit IRC | 16:45 | |
| sdake | nice swift1 | 16:45 |
| sdake | will that make it by wednesday paul? | 16:45 |
| sdake | pbourke will that be in running form by wednesday? | 16:45 |
| pbourke | sdake: its running now, but the rings are issue | 16:46 |
| pbourke | *an issue | 16:46 |
| sdake | we decided to handle the rings in rc1 right? | 16:46 |
| pbourke | hmm you did say that yeah | 16:46 |
| pbourke | that works for me then | 16:46 |
| sdake | how are rings loaded now? | 16:46 |
| pbourke | you have to supply them, if they're there, config-external loads them in | 16:47 |
| sdake | if th eanswer is ther eis no way to get them into the system, that needs tobe fixed :) | 16:47 |
| sdake | where do you supply them? | 16:47 |
| sdake | and what do they look like | 16:47 |
| pbourke | the usual /opt/kolla/swift | 16:47 |
| pbourke | they're gzipped files | 16:48 |
| sdake | groan that is super hacky | 16:48 |
| pbourke | I wouldnt say super | 16:48 |
| *** unicell has quit IRC | 16:48 | |
| sdake | /opt/kolla/configs/swift is super hacky | 16:48 |
| sdake | that is an internal directory | 16:48 |
| pbourke | ops can supply any config this way | 16:48 |
| sdake | can you atlaest copy them from /etc in some way? | 16:48 |
| pbourke | https://review.openstack.org/#/c/215065/5/docker/common/swift/swift-object-server/config-external.sh | 16:49 |
| SamYaple | sdake: we need to change everything over to /etc | 16:49 |
| SamYaple | so we can ignore that for now | 16:49 |
| SamYaple | itll get switched over with everything else | 16:49 |
| pbourke | SamYaple: missed your comments somehow, looking now | 16:49 |
| sdake | etc shouldn't be writetn by programs in geeneral I think | 16:49 |
| sdake | that is what /var/lib is for | 16:50 |
| *** Haomeng|2 has joined #kolla | 16:50 | |
| SamYaple | pbourke: you didnt miss them, you upped two patchsets while i was writing them | 16:50 |
| pbourke | kk | 16:50 |
| SamYaple | sdake: /etc/ is where the ansible configs will go | 16:50 |
| SamYaple | since they would be modifyable from there | 16:50 |
| SamYaple | COPY_ALWAYS | 16:50 |
| *** Kennan2 has quit IRC | 16:50 | |
| sdake | i'm good with that after the container is started | 16:50 |
| SamYaple | sdake: no we are having different conversations again | 16:51 |
| sdake | copy always still uses cp to /etc | 16:51 |
| SamYaple | outside the container vs inside | 16:51 |
| SamYaple | outside the container we save to /opt/kolla/contig | 16:51 |
| SamYaple | inside the container we BINDMOUNT to /opt/kolla/{{ service }} | 16:51 |
| SamYaple | we need ot change both of these | 16:52 |
| sdake | yes that is what pbourke is talking about putting the ring files there | 16:52 |
| SamYaple | but they would be cp over to etc like veverything else | 16:52 |
| sdake | change outside the contianer to what? | 16:52 |
| SamYaple | /etc | 16:52 |
| *** Haomeng has quit IRC | 16:52 | |
| sdake | and just not copy? | 16:53 |
| sdake | how does that preserve immutability | 16:53 |
| SamYaple | no no no | 16:53 |
| SamYaple | ugh | 16:53 |
| SamYaple | ill just do the patches | 16:53 |
| SamYaple | well talk about it there so you see whats going on | 16:53 |
| sdake | ok well back in my world i don't want to put a ring file in /etc/opt/kolla/config to launch swift ;) | 16:54 |
| SamYaple | thats not what happens | 16:54 |
| sdake | pbourke could you verify | 16:54 |
| SamYaple | its like every other file in /opt/kolla/config, it gets copied to /etc | 16:54 |
| SamYaple | sdake: https://review.openstack.org/#/c/215065/5/docker/common/swift/swift-object-server/config-external.sh | 16:54 |
| SamYaple | its right there | 16:55 |
| pbourke | sdake: its like every other config we do currently, the only thing missing is we dont generate them automatically | 16:55 |
| sdake | yes what i'm saying is the swift ansible code shoudl copy from /etc to /opt/kolla/config | 16:55 |
| sdake | pbourke its different in that its not an ini file | 16:55 |
| SamYaple | no it shoudlnt | 16:55 |
| *** SourabhP has joined #kolla | 16:55 | |
| SamYaple | ansible is distributing that file | 16:55 |
| SamYaple | its not unique per host | 16:55 |
| pbourke | sdake: that doesn't matter :) lots of other services don't have ini files | 16:55 |
| sdake | pbourke so lets say i have a 3 ring swift node, do i login to all 3 nodes, createa a tar gz of the swift ring on each node, and then launch swift? | 16:56 |
| SamYaple | no | 16:56 |
| SamYaple | tahts not how swift works | 16:56 |
| SamYaple | you generate teh ring in one location and distribute | 16:57 |
| pbourke | SamYaple is correct. The only thing is we currently aren't doing the distrubution | 16:57 |
| pbourke | *distribution | 16:57 |
| SamYaple | pbourke: we can with ansible | 16:57 |
| sdake | that is the part i just asked for! | 16:57 |
| SamYaple | pbourke: you already have the config-external.sh code | 16:58 |
| pbourke | SamYaple: yes but I think what sdake is asking is how do they get onto the nodes | 16:58 |
| SamYaple | just have ansible fetch the file from /etc on the deploy host | 16:58 |
| pbourke | ansible can do this, but the current patchset does not | 16:58 |
| sdake | samyaple yes i'm pretty sure that is what I was saying :) | 16:58 |
| SamYaple | you never said deploy host sdake | 16:58 |
| SamYaple | you said container and thats not deploy host | 16:58 |
| * sdake remebers to use real world examples next time to cut tot he chase :) | 16:59 | |
| sdake | roger its hard to communicate this stuff | 16:59 |
| SamYaple | /etc from the deploy host ->> /opt/kolla/config on the destination node | 16:59 |
| sdake | pbourke can you fix that up then with a file in /etc | 16:59 |
| SamYaple | then config-external takes over no problem | 16:59 |
| sdake | yes i am good with that model samyaple | 16:59 |
| SamYaple | cool | 16:59 |
| SamYaple | crisis averted | 16:59 |
| sdake | there has got to be a simple copy operation in asnible | 16:59 |
| SamYaple | its called copy | 16:59 |
| pbourke | there is but im not sure we want to add that | 17:00 |
| pbourke | its would be non standard | 17:00 |
| SamYaple | how so? | 17:00 |
| SamYaple | copy and template are interchangable to me | 17:00 |
| pbourke | yes but we're not supplying them | 17:00 |
| SamYaple | template is only used when we replace variables | 17:00 |
| sdake | pbourke we are adding more copys for other types of overrides I believe | 17:00 |
| pbourke | we supply templates | 17:00 |
| SamYaple | pbourke: we still do the distrbition | 17:01 |
| sdake | pbourke although samyaple may implement that a different way not certain | 17:01 |
| SamYaple | thats perfectly fine | 17:01 |
| SamYaple | I am more than happy ammend the patchset with this model pbourke | 17:01 |
| pbourke | no i can do it | 17:01 |
| sdake | pbourke thanks you rock :) | 17:01 |
| SamYaple | ok ill be here to review :) | 17:01 |
| pbourke | i dont mind if you do but im sure you have better things to do | 17:01 |
| sdake | ya sam busy busy | 17:02 |
| sdake | best to let him focus :) | 17:02 |
| SamYaple | ok ill be here to review :)yea but thats like 5 minutes at most and then he can take back over | 17:02 |
| sdake | jumping around ruins focus | 17:02 |
| pbourke | ok so my plan is to get some of the supporting services in before moving this patch from wip | 17:02 |
| SamYaple | im not volunteerign to finish it | 17:02 |
| pbourke | the ones there will make it work but you dont get replication and all the other consistency stuff | 17:02 |
| SamYaple | sdake: where do you want templates (non-modifyable files that have to persisnt on the destination host)? currently they are in "/usr/share/kolla/templates" | 17:02 |
| sdake | pbourke I am after a dead chicken implementation by wednesday in gerrit | 17:03 |
| sdake | samyaple let me think a bit just woke up | 17:03 |
| SamYaple | youre arguing pretty hard for not thinking | 17:03 |
| pbourke | googled dead chicken implementation | 17:03 |
| sdake | samyaple we have two options /var/lib/kolla and /usr/share/kolla | 17:03 |
| pbourke | came up empty | 17:03 |
| SamYaple | well they are already in /usr/shade | 17:03 |
| SamYaple | /usr/share | 17:03 |
| sdake | dead chicken testing is when you wave a dead chicken over the box, and if the chicken doesn't come alive, it passed the test | 17:04 |
| sdake | they are not user modifiable? | 17:04 |
| sdake | only part of the deploy system | 17:04 |
| sdake | I think /usr/share is right | 17:04 |
| *** dwalsh has joined #kolla | 17:04 | |
| SamYaple | the templates are not, they simply exist to keep idepontency | 17:05 |
| sdake | kolla scripts need to go in there too | 17:05 |
| SamYaple | no no no | 17:05 |
| SamYaple | on the physical host | 17:05 |
| SamYaple | outside the container | 17:05 |
| *** Haomeng has joined #kolla | 17:05 | |
| *** achanda has joined #kolla | 17:05 | |
| sdake | right i was talking about the kolla scripts outside the container too that are copied to /opt currently | 17:05 |
| sdake | btw, there is abug for this akwasnie has assigned herself to | 17:05 |
| SamYaple | there are no scripts outside the container | 17:05 |
| *** Kennan has joined #kolla | 17:05 | |
| sdake | your right | 17:06 |
| *** diga has quit IRC | 17:06 | |
| sdake | fuck brain not operational let me wake up | 17:06 |
| SamYaple | outside the container, we write configs to /opt/kolla/config, I suggest /etc/kolla/ | 17:07 |
| sdake | ok templates on deployed host should go in /usr/share/kolla/templates or something similar | 17:07 |
| SamYaple | done | 17:07 |
| *** SourabhP has quit IRC | 17:08 | |
| sdake | the problem with writing straight to /etc/kolla is AIO | 17:08 |
| sdake | i want aio to not implode :) | 17:08 |
| SamYaple | no it woulnt | 17:08 |
| sdake | /etc/kolla/configs might be better | 17:08 |
| SamYaple | ive thought of everything | 17:08 |
| *** Haomeng|2 has quit IRC | 17:08 | |
| SamYaple | on AIO, it wouldnt affect it at all | 17:08 |
| SamYaple | you would have /etc/kolla/nova-api /etc/kolla/neutron-agents | 17:08 |
| SamYaple | etc | 17:08 |
| SamYaple | no overlap | 17:08 |
| sdake | i guess it woudl be a little confusing for aio but we could document it | 17:09 |
| sdake | and its more tidy on multinode deploys | 17:09 |
| sdake | which is what we are really after | 17:09 |
| sdake | so yup go with that | 17:09 |
| SamYaple | agreed, but we want that | 17:09 |
| SamYaple | yea | 17:09 |
| sdake | its a config management tool so writing to /etc is acceptable | 17:10 |
| sdake | i.e. meets lsb | 17:10 |
| *** Haomeng|2 has joined #kolla | 17:11 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Change default config location https://review.openstack.org/215250 | 17:11 |
| SamYaple | inside the container is a whole different thing | 17:12 |
| SamYaple | but i suggest /usr/share inside the container since its mounted read-only and they never change | 17:12 |
| SamYaple | (from the containers perspective) | 17:13 |
| sdake | /usr/share for which? | 17:13 |
| sdake | scripts? | 17:13 |
| SamYaple | for the bindmount | 17:13 |
| SamYaple | and scripts i suppose | 17:13 |
| sdake | cool make it so #1 :) | 17:13 |
| SamYaple | alright | 17:14 |
| *** Haomeng has quit IRC | 17:14 | |
| SamYaple | this update-configs blueprint is really getting a workout | 17:14 |
| sdake | /usr/share/kolla tho pls :) | 17:14 |
| SamYaple | yea | 17:14 |
| *** Kennan has quit IRC | 17:14 | |
| sdake | i filed a bug on this | 17:14 |
| sdake | make sure to mark it invalid or something | 17:14 |
| sdake | if your going to do all this work | 17:14 |
| *** absubram has joined #kolla | 17:14 | |
| sdake | it has "sed" in the titel | 17:14 |
| SamYaple | ill closes the bug too | 17:15 |
| sdake | do before akawasnie spends mucht ime on it | 17:15 |
| sdake | i hate for her to waste time on something that is wrong | 17:15 |
| SamYaple | I dont think she has done anything on it, inc0 is she still around? | 17:15 |
| SamYaple | mmmmmm nice | 17:16 |
| SamYaple | an empty /opt | 17:16 |
| *** unicell has joined #kolla | 17:16 | |
| sdake | inc0 usually not here that late | 17:16 |
| sdake | this late | 17:16 |
| SamYaple | yea | 17:16 |
| sdake | so I've been thinking, we need a design.md | 17:16 |
| sdake | becuase the core reviewer team understands the design of the system | 17:17 |
| sdake | but other people will want to know as well | 17:17 |
| SamYaple | make it so sdake | 17:17 |
| sdake | yes i will | 17:17 |
| sdake | probably later next week | 17:17 |
| sdake | i'm going to beat heat into gerrit-submission | 17:18 |
| SamYaple | HOW https://review.openstack.org/#/c/213958/ | 17:18 |
| SamYaple | HOW IS THAT IN MERGE CONFLICT | 17:18 |
| sdake | samyaple dude i dont get it | 17:18 |
| sdake | gerrit has been acting up badly | 17:18 |
| sdake | please hold, I will ask on openstack-infra | 17:19 |
| *** Haomeng has joined #kolla | 17:19 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove start.yml and bootstrap.yml https://review.openstack.org/213958 | 17:20 |
| *** Kennan has joined #kolla | 17:20 | |
| *** Haomeng|2 has quit IRC | 17:21 | |
| *** Haomeng|2 has joined #kolla | 17:22 | |
| pbourke | is there usually a delay on sending mails to the ML? | 17:22 |
| pbourke | send one wrt neutron plugins about 2 hours ago | 17:22 |
| pbourke | but not seeing it | 17:22 |
| sdake | pbourke jsut saw on #openstac-infnra mailman is lagged up | 17:22 |
| pbourke | sdake: ah ok, thanks! | 17:23 |
| pbourke | keep an eye out for it :) would like to get that bp worked out | 17:23 |
| sdake | ok so re the rebases complaining | 17:23 |
| sdake | the gate and gerrit use "jgit" whatever that is | 17:24 |
| sdake | apparently its not as pro as our c git client at sorting out merge | 17:24 |
| sdake | s | 17:24 |
| *** Kennan has quit IRC | 17:24 | |
| *** Haomeng has quit IRC | 17:24 | |
| sdake | SamYaple the reason that file was in merge conflict was because those files were edited (by my patch last night) | 17:24 |
| sdake | one of em atleast | 17:24 |
| sdake | bootstrap.yml iirc | 17:24 |
| SamYaple | ah ok | 17:24 |
| SamYaple | at least that makes sense | 17:25 |
| sdake | and a delete after an edit requires a rebase with jgit | 17:25 |
| sdake | you wouldn't think a delete after an edit would require a rebase - git sorts it out for us | 17:25 |
| sdake | but jgit does not :) | 17:25 |
| *** achanda has quit IRC | 17:27 | |
| SamYaple | mmmmm that feel when you have the review queue up-to-date and have a dozen patches waiting for reviews | 17:29 |
| SamYaple | its great | 17:29 |
| SamYaple | im going to go eat and sleep | 17:29 |
| *** ccrouch has quit IRC | 17:30 | |
| *** ccrouch1 has joined #kolla | 17:30 | |
| * sdake gets a bucket of -1's | 17:30 | |
| SamYaple | :) | 17:30 |
| SamYaple | I expect some -1 on my ubuntu stuff | 17:31 |
| *** tfukushima has joined #kolla | 17:31 | |
| SamYaple | if people really want to get on my good side, get this reviewed ;) https://review.openstack.org/#/c/215149/ | 17:32 |
| *** jtriley_ has quit IRC | 17:32 | |
| *** jtriley has joined #kolla | 17:35 | |
| *** loth has joined #kolla | 17:36 | |
| openstackgerrit | Merged stackforge/kolla: The requirements.txt file isn't correct https://review.openstack.org/213398 | 17:39 |
| inc0 | back | 17:39 |
| inc0 | I'll tell Ala tomorrow that bug is invalid, afaik she didn't spent too much time on it | 17:40 |
| *** Kennan has joined #kolla | 17:40 | |
| *** dwalsh has quit IRC | 17:43 | |
| sdake | rhallisey the ceph container has many comments in it ;) | 17:54 |
| rhallisey | I'll make my way back to it shortly | 17:54 |
| *** achanda has joined #kolla | 17:55 | |
| *** jasonsb has quit IRC | 17:57 | |
| *** bmace has quit IRC | 18:05 | |
| *** SourabhP has joined #kolla | 18:10 | |
| sdake | pbourke nice job on swift - bunch of comments in the review - needs a bit more work before mergable | 18:17 |
| sdake | but looking fantastic | 18:17 |
| sdake | if you could tell me how to get it started, i'd like to try it out | 18:18 |
| sdake | specificlaly how do i generte a ring file and what is /srv/node (a filesystem?) | 18:18 |
| *** achanda has quit IRC | 18:24 | |
| openstackgerrit | Merged stackforge/kolla: Allow Docker to make better use of cache https://review.openstack.org/215149 | 18:26 |
| openstackgerrit | Merged stackforge/kolla: Fix group name in vagrant docs https://review.openstack.org/213950 | 18:28 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix MariaDB for ubuntu https://review.openstack.org/215159 | 18:30 |
| sdake | off to pick up mail bbl | 18:31 |
| *** dwalsh has joined #kolla | 18:31 | |
| SamYaple | sdake, please change your vote https://review.openstack.org/#/c/215112 | 18:32 |
| openstackgerrit | Merged stackforge/kolla: Change default config location https://review.openstack.org/215250 | 18:34 |
| openstackgerrit | Merged stackforge/kolla: Remove start.yml and bootstrap.yml https://review.openstack.org/213958 | 18:42 |
| *** inc0 has quit IRC | 18:44 | |
| *** jruano has quit IRC | 19:08 | |
| *** achanda has joined #kolla | 19:08 | |
| *** tfukushima has quit IRC | 19:12 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Ceph container https://review.openstack.org/214706 | 19:13 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Ceph container https://review.openstack.org/214706 | 19:14 |
| sdake | samyaple apologies for that | 19:17 |
| *** jasonsb has joined #kolla | 19:18 | |
| *** tfukushima has joined #kolla | 19:19 | |
| *** dims_ has joined #kolla | 19:26 | |
| *** tfukushima has quit IRC | 19:27 | |
| *** dims has quit IRC | 19:28 | |
| *** achanda has quit IRC | 19:34 | |
| *** dwalsh has quit IRC | 19:34 | |
| *** jruano has joined #kolla | 19:42 | |
| *** bmace has joined #kolla | 19:58 | |
| *** bmace has quit IRC | 20:09 | |
| *** achanda has joined #kolla | 20:13 | |
| *** achanda has quit IRC | 20:19 | |
| *** jruano has quit IRC | 20:24 | |
| *** jruano has joined #kolla | 20:24 | |
| *** vinkman has joined #kolla | 20:26 | |
| *** achanda has joined #kolla | 20:32 | |
| *** vinkman has left #kolla | 20:33 | |
| sdake | pbourke you happen to still be around | 20:36 |
| sdake | pbourke if you see this later, please submit some kind of docs on how to setup and use swift for noobs like me - it doens't have to be perfect - just put in gerrit queue with code review -2 - doesn't even have to be formatted, just a quick get started guide targeted at someone who understands the architecture in detail (just for me:) | 20:38 |
| sdake | pbourke like what needs to go in /srv/node (just an empty directory) and how to generate the ring file and some basic swift client commands | 20:39 |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Add options for selective build and listing images https://review.openstack.org/215323 | 20:42 |
| sdake | holy crap someone finally wrote that | 20:43 |
| * sdake cheers | 20:43 | |
| sdake | jpeeler could you go into more detail what -l does? | 20:43 |
| jpeeler | in the commit or here | 20:43 |
| sdake | irc is good | 20:44 |
| jpeeler | i just traverses the directories and shows the image names for the available Dockerfiles | 20:44 |
| sdake | what is -d? | 20:44 |
| jpeeler | debug, not necessary but i use it sometimes | 20:45 |
| sdake | so sam had suggested not having an option like -i | 20:47 |
| sdake | but instead anything at the end would be built | 20:47 |
| sdake | if nothing was specified everything is built | 20:47 |
| sdake | eg build.py -T 1000 heat | 20:47 |
| sdake | would build anything that matches *heat* | 20:48 |
| sdake | or build.py -T 1000 heat_base | 20:48 |
| sdake | or build.oy -T 1000 heat swift ceilometer_base | 20:48 |
| sdake | is that easy or hard to change around? | 20:48 |
| sdake | I also prefer that model | 20:48 |
| sdake | jpeeler re gating, i spoke with the infra team about mutiple vms | 20:49 |
| sdake | here is the deal | 20:49 |
| sdake | they have 2 vm nodepool jobs | 20:49 |
| sdake | 2 vms atm is hte limit | 20:49 |
| jpeeler | i'm not sure how hard it'll be, probably not hard, but i don't know how to parse it off the top of my head | 20:49 |
| sdake | they said if we can get 1 node AIO stable, to go to 2 node AIO | 20:49 |
| sdake | if we get 2 node AIO (1 controller 1 compute) stable they will give us more vms | 20:50 |
| sdake | i think the blocker for aio deploy atm is the lack of a --retry flag in build.py | 20:50 |
| sdake | i have a meeting in 10 minutes but i'll reviwe yur patch after | 20:51 |
| sdake | if you can see how to get the non option arguments out of the argparser that would probably do the job | 20:51 |
| jpeeler | i really need to get going on ironic, but haven't forgotten about the retry flag | 20:51 |
| *** jruano has quit IRC | 20:51 | |
| sdake | I dont know off the top of my head | 20:51 |
| jpeeler | yeah i'll look | 20:51 |
| sdake | we cn finish ironic in rc1 if need be | 20:51 |
| sdake | anything that misses liberty3 will be essential or high priority for rc1 | 20:52 |
| jpeeler | looks like arg_parse supports the model you are wanting | 20:54 |
| sdake | also not sure we need a -l flag | 20:55 |
| sdake | what is the use case for that? | 20:55 |
| sdake | to list out what is available to build? | 20:55 |
| jpeeler | yes | 20:55 |
| sdake | cool make a separate patch for that plz | 20:55 |
| sdake | jpeeler i am not keen on adding stuff to the cli - because clis are like APIs :) | 20:56 |
| sdake | but i can see the value in that type of operation | 20:56 |
| *** achanda has quit IRC | 20:57 | |
| *** achanda_ has joined #kolla | 20:57 | |
| *** ccrouch1 has quit IRC | 21:17 | |
| *** ccrouch has joined #kolla | 21:20 | |
| *** ccrouch has quit IRC | 21:20 | |
| *** ccrouch1 has joined #kolla | 21:20 | |
| *** jtriley has quit IRC | 21:23 | |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Add options for selective build and listing images https://review.openstack.org/215323 | 21:24 |
| *** dwalsh has joined #kolla | 21:25 | |
| openstackgerrit | Jeff Peeler proposed stackforge/kolla: Add options for selective build and listing images https://review.openstack.org/215323 | 21:28 |
| *** bmace has joined #kolla | 21:34 | |
| bmace | SamYaple around? | 21:36 |
| *** absubram has quit IRC | 21:42 | |
| *** ccrouch1 has quit IRC | 21:46 | |
| *** absubram has joined #kolla | 21:47 | |
| *** ccrouch has joined #kolla | 21:50 | |
| bmace | sdake around? | 21:54 |
| *** britthouser has quit IRC | 21:54 | |
| *** ccrouch has quit IRC | 21:56 | |
| *** ccrouch1 has joined #kolla | 21:56 | |
| *** ccrouch1 has quit IRC | 21:56 | |
| *** ccrouch has joined #kolla | 21:57 | |
| *** diogogmt_ has joined #kolla | 22:00 | |
| *** diogogmt has quit IRC | 22:00 | |
| *** diogogmt_ is now known as diogogmt | 22:00 | |
| *** ccrouch has quit IRC | 22:01 | |
| sdake | bmance wound me | 22:04 |
| sdake | bmace wound me | 22:04 |
| sdake | bmace whats up bro | 22:08 |
| sdake | sorry was OTP | 22:08 |
| sdake | my life is so much improved since i put my main phone number on permanent voice mail - no ringing phone all day long | 22:09 |
| bmace | np, had a question about the change that went in around where the config files end up being put. i see they are going into /etc/kolla/ previously that was going to be the location of the "original" files on the node you push from | 22:09 |
| sdake | right | 22:09 |
| bmace | so is there some sort of separation of the /etc/kolla space now between the ansible node and the control / compute / etc nodes? | 22:10 |
| sdake | and on the deplooyed nodes as well | 22:10 |
| bmace | since you might deploy to the same node you are on.. which i think was the reason for the separation? | 22:10 |
| sdake | the separation is that each /etc/kolla directory on deployed node swil end up with a nova-api directory nova-compute directory etc | 22:10 |
| sdake | bmace we thought through all in one - I had that same concern | 22:11 |
| bmace | directly under /etc/kolla? so /etc/kolla/config still has the stuff that gets pushed out? | 22:11 |
| sdake | /etc/kolla/config is gone | 22:11 |
| sdake | there was never an /etc/kolla/config afaik | 22:11 |
| sdake | there was an /opt/kolla/config | 22:12 |
| bmace | ah, i think you are right.. and that is still going to be there? | 22:12 |
| sdake | no /opt/kolla/config got moved to /etc/kolla | 22:12 |
| bmace | so /opt/kolla is for all the stuff on the deployment node? | 22:12 |
| sdake | there will be no /opt/kolla going forward | 22:12 |
| sdake | there will be /usr/share/kolla | 22:13 |
| *** rhallisey has quit IRC | 22:13 | |
| sdake | and /etc/kolla | 22:13 |
| sdake | on both deploy and deployed nodes | 22:13 |
| bmace | hrm.. so /usr/share/kolla will have all the 'usually empty' config files? something like /usr/share/kolla/configs or whatever? | 22:14 |
| sdake | the empty config files wil lgo in /etc/kolla | 22:14 |
| sdake | /usr/share/kolla/tempaltes | 22:14 |
| sdake | will contain the template files for use with ansible | 22:14 |
| sdake | only on deployed nodes | 22:15 |
| bmace | how about all the role stuff? all the bits under git:kolla/ansible need to be somewhere. | 22:15 |
| sdake | the structure of the /etc directory on the deployment node is not changing at all | 22:15 |
| sdake | ansible should go in /usr/share/kolla/ansible imo | 22:15 |
| sdake | but we haven't worked out packaging for kolla yet | 22:15 |
| bmace | kk.. /usr/share it is | 22:16 |
| sdake | strangely i think the inventory file should go in /etc/kolla | 22:17 |
| bmace | so the empty configs go in /etc/kolla/configs, from git:kolla/etc/kolla/configs | 22:17 |
| sdake | but its in the ansible directory at the moment for unknown to me reasons | 22:17 |
| sdake | let me look at the repo | 22:17 |
| bmace | well, we are having the inventory file dynamically generated right now, so it doesn't need to live anywhere :) | 22:18 |
| sdake | right /etc/kolla/config is where the configurable deployed-from node stuff goes | 22:18 |
| bmace | kk | 22:18 |
| sdake | bmace there is some concern from pbourke and vbel on the removal of docker dir | 22:19 |
| bmace | thanks sdake :) | 22:19 |
| bmace | yeah, i think that is around the build process. | 22:19 |
| sdake | ltes see if we can get that sorted out sooner rather then later | 22:19 |
| sdake | the directory needs to go | 22:19 |
| sdake | how do you dynamicallygenerate the inventory file | 22:19 |
| bmace | something about that getting nuked making it harder to use some custom bits or sources during the build than the new build process. | 22:19 |
| sdake | bmace ya I dont want to maintain two build systems or two sets of docker files | 22:20 |
| sdake | need both out of the repo as soon as feasible | 22:20 |
| sdake | defintely before liberty | 22:20 |
| *** jruano has joined #kolla | 22:20 | |
| bmace | we have our own file that is a superset of the information in the inventory file, and we just generate json that ansible-playbook can use. if you pass an executable as the param to playbook -i it will run it and take the output as its inventory file. | 22:20 |
| sdake | pbourke and vbel said they would sort out a patch to bikeshed on | 22:21 |
| sdake | interesting | 22:21 |
| sdake | so this is the dynamic inventory thing people talk about - running a binary via -i? | 22:23 |
| bmace | yeah, working pretty well so far. all the basic inventory / property management / deploy stuff is in. | 22:23 |
| bmace | yeah, and i don't know if it can take a normal inventory format at that point.. i know we pass in a json structured file. | 22:23 |
| bmace | was able to deploy with custom properties and inventory without ever running vi ;) | 22:24 |
| sdake | where do you store hte inventory data? | 22:25 |
| sdake | you said in a superset file? | 22:25 |
| bmace | yeah, .json formatted file. now that the /opt/kolla stuff is moving i'll move it, but something like /opt/kolla/kollacli/ansible/inventory.json.. or some such | 22:26 |
| sdake | why not use yaml? | 22:26 |
| sdake | what manages the json file, cli tools? | 22:26 |
| bmace | just some python code. and the ansible-playbook looks like it wants the dynamic inventory in json anyway, so closer to what we are passing to ansible to consume | 22:27 |
| bmace | ansible is really ADD about what technologies it wants to use. yaml for some stuff, their inventory file, which isn't valid yaml and they use a bunch of proprietary parsing code, and json for the dynamic inventory consumption.. | 22:28 |
| sdake | that json file should definatley go in /var/lib/kolla ;) | 22:28 |
| bmace | hrm.. what else is in /var/lib/kolla? | 22:28 |
| sdake | i see - wasn't aware | 22:28 |
| sdake | nothing but the json file is basically a database | 22:28 |
| sdake | I mean its up to you but thats where i'd put it :) | 22:28 |
| sdake | does it contain stuff for gobals.yml? | 22:29 |
| bmace | definitely no, all the property stuff is done via directly touching the ansible property files, and the only one every directly changed is that globals file | 22:29 |
| sdake | cool would like to see a video demo if you ever come up for air :) | 22:30 |
| bmace | i am getting short on air :) | 22:30 |
| sdake | would be nice to get some of those ideas upstream but understand the turnaround time on reviews + overload of work atm in upstream queue | 22:30 |
| bmace | i'm hoping it can just be super awesome for M.. the full monty :) | 22:31 |
| sdake | ;-) | 22:31 |
| sdake | well we dont have alot of work planned for m yet | 22:31 |
| sdake | most of liberty I'd almost call "complete" | 22:32 |
| *** dwalsh has quit IRC | 22:32 | |
| sdake | and not in need of alot of refactoring | 22:32 |
| sdake | just service expansion | 22:32 |
| bmace | yeah, for M maybe just more polish / images.. improved testing / stability / etc.. | 22:32 |
| sdake | as in adding more services, beyond that i dont have alot of ideas for new dev work | 22:32 |
| sdake | clearly we need to go full hog on making the gate working well in m | 22:33 |
| bmace | not super exciting but very important for people who want to use it :) | 22:33 |
| sdake | i'm trying to get some basic gating in place by conclusion of liberty | 22:33 |
| sdake | so i know i'm asking twice but just for my satisfication | 22:35 |
| sdake | bmace | 22:35 |
| bmace | sdake | 22:35 |
| sdake | if --inventory is specified, and hte file has +x permissions, it gets executed and the results are fed in as the invetory file? | 22:35 |
| bmace | yup | 22:35 |
| sdake | does it have arguments? | 22:36 |
| bmace | hrm.. not sure if it can or not, ours doesn't. it knows where the db is, consumes it, and dumps the json inventory data to stdout | 22:36 |
| *** tfukushima has joined #kolla | 22:37 | |
| *** jruano has quit IRC | 22:38 | |
| *** jruano has joined #kolla | 22:38 | |
| *** jruano has quit IRC | 22:38 | |
| *** ccrouch has joined #kolla | 22:50 | |
| *** ccrouch has quit IRC | 22:51 | |
| *** ccrouch1 has joined #kolla | 22:51 | |
| *** tfukushima has quit IRC | 22:55 | |
| *** jasonsb has quit IRC | 23:02 | |
| *** ccrouch1 has quit IRC | 23:03 | |
| *** jasonsb has joined #kolla | 23:03 | |
| *** achanda_ has quit IRC | 23:05 | |
| *** jasonsb has quit IRC | 23:07 | |
| *** diogogmt has quit IRC | 23:12 | |
| *** diogogmt has joined #kolla | 23:14 | |
| *** ccrouch has joined #kolla | 23:15 | |
| *** ccrouch has quit IRC | 23:15 | |
| *** ccrouch has joined #kolla | 23:15 | |
| *** stvnoyes has quit IRC | 23:16 | |
| *** achanda has joined #kolla | 23:16 | |
| *** blahRus has quit IRC | 23:18 | |
| *** dwalsh has joined #kolla | 23:21 | |
| *** dwalsh has quit IRC | 23:29 | |
| *** banix has quit IRC | 23:33 | |
| *** SourabhP has quit IRC | 23:43 | |
| openstackgerrit | Duane DeCapite proposed stackforge/kolla: Workaround for existing keepalived.pid file https://review.openstack.org/215380 | 23:45 |
| openstackgerrit | Duane DeCapite proposed stackforge/kolla: Workaround for existing keepalived.pid file https://review.openstack.org/215380 | 23:53 |
| openstackgerrit | Duane DeCapite proposed stackforge/kolla: Workaround for existing keepalived.pid file https://review.openstack.org/215380 | 23:56 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!