bmace | there is still the port 6080 hard coded in the nova.conf.j2. are there issues with making that a parameter? it sort of stands out as the only one in there. i guess if you wanted you could override the entire novncproxy_base_url in globals.yml? | 00:02 |
---|---|---|
SamYaple | bmace: there are issues with changing that that i havent been able to solve | 00:03 |
SamYaple | mainly because 6080 is a standard port used my libvirt/qemu rather than something openstack is doing explicitly | 00:03 |
bmace | kk. not sure if we might want to put some sort of comment in there about that. otherwise those changes look good. | 00:03 |
*** diogogmt has joined #kolla | 00:09 | |
*** echoingumesh has quit IRC | 00:10 | |
*** mfalatic has quit IRC | 00:40 | |
*** dolpher has joined #kolla | 00:42 | |
sdake | wars make not one great | 00:43 |
*** sdake_ has joined #kolla | 00:47 | |
*** jasonsb has joined #kolla | 00:47 | |
*** sdake has quit IRC | 00:50 | |
*** sdake_ is now known as sdake | 01:12 | |
*** daneyon has quit IRC | 01:16 | |
*** tobe_ has joined #kolla | 01:18 | |
*** coolsvap_ has joined #kolla | 01:24 | |
sdake | jpeeler patch looks good | 01:25 |
sdake | samyaple can you review https://review.openstack.org/#/c/202216/ pzl | 01:25 |
SamYaple | if thats teh logging review its open havent gotten to it yet | 01:26 |
sdake | it is | 01:26 |
SamYaple | probably wont til i get to work | 01:26 |
sdake | what time is that? | 01:26 |
SamYaple | couple hours | 01:26 |
sdake | not rushing just curious | 01:26 |
sdake | so 9pm pst then? | 01:27 |
SamYaple | yea | 01:27 |
sdake | brutal schedule | 01:27 |
sdake | i dont know how you handle it | 01:27 |
SamYaple | i get to not see people most of the night | 01:28 |
SamYaple | its great | 01:28 |
*** weiyu_ has joined #kolla | 01:28 | |
*** jtriley has quit IRC | 01:29 | |
*** zhiwei has joined #kolla | 01:30 | |
*** erkules_ has joined #kolla | 01:34 | |
*** coolsvap_ has quit IRC | 01:36 | |
*** erkules has quit IRC | 01:36 | |
*** tobe_ has quit IRC | 02:05 | |
*** tobe_ has joined #kolla | 02:06 | |
*** SourabhP has quit IRC | 02:13 | |
*** unicell has quit IRC | 02:15 | |
*** jtriley has joined #kolla | 02:25 | |
*** jtriley has quit IRC | 02:30 | |
*** unicell has joined #kolla | 02:34 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Added initial template for base image https://review.openstack.org/210288 | 02:40 |
*** diogogmt has quit IRC | 02:44 | |
*** klint has joined #kolla | 02:46 | |
*** tobe_ has quit IRC | 02:47 | |
*** tobe_ has joined #kolla | 02:49 | |
*** dolpher has quit IRC | 03:04 | |
SamYaple | any oracle guys up yet? | 03:25 |
SamYaple | pbourke? | 03:25 |
*** jtriley has joined #kolla | 03:30 | |
openstackgerrit | Merged stackforge/kolla: Use docker-engine in setup_gate https://review.openstack.org/211123 | 03:32 |
openstackgerrit | Merged stackforge/kolla: Add centos-src-build tox gate https://review.openstack.org/211128 | 03:34 |
*** jtriley has quit IRC | 03:35 | |
*** britthouser has quit IRC | 03:38 | |
coolsvap | morning :) | 03:55 |
coolsvap | evening :) | 03:55 |
coolsvap | SamYaple, do you mind if i add the nit pics for pep8 failure in https://review.openstack.org/#/c/210288/ | 03:56 |
SamYaple | coolsvap: can you just tell me why its breacking on the import line? | 04:00 |
SamYaple | i mean that is explicitly allowed https://www.python.org/dev/peps/pep-0008/#imports | 04:00 |
bmace | pbourke and much of the other oracle crew are GMT so probably sleeping right now :) | 04:01 |
SamYaple | bmace: i thought id try | 04:02 |
coolsvap | one import per line | 04:02 |
SamYaple | look at the link though | 04:02 |
SamYaple | pep8 explicitly says to do that as "ok" | 04:02 |
coolsvap | ./tools/build.py:36:31: H301 one import per line | 04:02 |
coolsvap | from jinja2 import Environment, FileSystemLoader | 04:02 |
coolsvap | H301 | 04:02 |
SamYaple | thats dumb | 04:03 |
SamYaple | but i dont like doing from imports anyway unless i have one thing ot import | 04:03 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Added initial template for base image https://review.openstack.org/210288 | 04:04 |
*** tobe_ has quit IRC | 04:12 | |
*** tobe_ has joined #kolla | 04:17 | |
*** achanda has quit IRC | 04:22 | |
*** daneyon has joined #kolla | 04:24 | |
*** weiyu_ has quit IRC | 04:33 | |
*** britthouser has joined #kolla | 04:41 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add ansible support for Cinder https://review.openstack.org/208868 | 05:09 |
*** tfukushima has joined #kolla | 05:17 | |
*** achanda has joined #kolla | 05:17 | |
*** tfukushima has quit IRC | 05:22 | |
coolsvap | SamYaple, can you tell me what i am missing in https://review.openstack.org/#/c/208868/9/ansible/roles/cinder/tasks/start.yml ? | 05:40 |
*** weiyu_ has joined #kolla | 05:40 | |
coolsvap | wait wait i think start.yml is not the same i updated :-/ | 05:42 |
SamYaple | coolsvap: yea see https://github.com/stackforge/kolla/blob/master/ansible/roles/keystone/tasks/start.yml | 05:43 |
coolsvap | SamYaple, yes, i updated it not sure how i missed it in the changeset | 05:44 |
SamYaple | no problem | 05:44 |
coolsvap | i accidently updated the glance start.yml with cinder updates since i was referring to the glance changeset of yours :-( | 05:49 |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add ansible support for Cinder https://review.openstack.org/208868 | 05:50 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Added initial template for base image https://review.openstack.org/210288 | 05:56 |
*** weiyu_ has quit IRC | 06:02 | |
*** SourabhP_ has joined #kolla | 06:02 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add ansible support for Cinder https://review.openstack.org/208868 | 06:17 |
*** shardy has joined #kolla | 06:24 | |
*** akwasnie has joined #kolla | 06:36 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Remove Ansible abstraction for Neutron https://review.openstack.org/211442 | 06:39 |
*** weiyu_ has joined #kolla | 06:51 | |
*** Haomeng has joined #kolla | 06:54 | |
*** achanda has quit IRC | 07:00 | |
*** achanda has joined #kolla | 07:03 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Remove Ansible abstraction for Nova https://review.openstack.org/211445 | 07:04 |
*** achanda has quit IRC | 07:05 | |
*** shardy has quit IRC | 07:06 | |
*** jmccarthy has quit IRC | 07:12 | |
*** jmccarthy has joined #kolla | 07:12 | |
*** unicell has quit IRC | 07:17 | |
*** unicell has joined #kolla | 07:17 | |
sdake | time to ptfo | 07:20 |
sdake | later | 07:20 |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add ansible support for Cinder https://review.openstack.org/208868 | 07:26 |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Remove Ansible abstraction for Neutron https://review.openstack.org/211442 | 07:27 |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Remove Ansible abstraction for Nova https://review.openstack.org/211445 | 07:28 |
coolsvap | ptfo? | 07:31 |
coolsvap | sdake, are you working on ansible-heat, i can update since i have the context with cinder patch with the updates required | 07:32 |
SamYaple | coolsvap: i have that one covered | 07:32 |
SamYaple | it needs other things too | 07:33 |
coolsvap | SamYaple, cool | 07:33 |
coolsvap | SamYaple, I will put next patches for templates, are you currently working on anything? | 07:35 |
SamYaple | no hold onff on that. we need ot get this approved so there are no changes | 07:39 |
coolsvap | alright | 07:40 |
coolsvap | i was looking at the remaining work for remove abstraction, can i continue with that? | 07:40 |
SamYaple | yea sure! ive been too busy at my job to do kolla :( | 07:41 |
*** SourabhP_ has quit IRC | 07:44 | |
*** gfidente^2nd has joined #kolla | 07:51 | |
*** gfidente has quit IRC | 07:58 | |
*** gfidente^2nd is now known as gfidente | 07:59 | |
*** gfidente has quit IRC | 07:59 | |
*** gfidente has joined #kolla | 07:59 | |
*** tfukushima has joined #kolla | 08:01 | |
*** inc0 has joined #kolla | 08:34 | |
akwasnie | hi | 08:40 |
akwasnie | SamYaple, thanks for you help with templates | 08:41 |
SamYaple | akwasnie: youre welcome! | 08:44 |
SamYaple | thanks for handling it, i have been swamped lately | 08:44 |
*** weiyu_ has quit IRC | 08:49 | |
inc0 | it's been a week of fighting against internal networking | 08:49 |
inc0 | I hate proxy. | 08:49 |
inc0 | I really do. | 08:49 |
inc0 | SamYaple, do you know if neutron is fixed already? | 08:50 |
inc0 | (by fixed I mean if it starts at all) | 08:51 |
zhiwei | I hate proxy too. | 08:52 |
*** inc0 has quit IRC | 08:52 | |
zhiwei | In China, I can't access too many services. Even can't access review.openstack.org:29418 | 08:52 |
zhiwei | Fuck Chinese government and GFW. | 08:53 |
SamYaple | inc0 you drop off? | 08:53 |
SamYaple | zhiwei: yea i have to deal with the china stuff too | 08:53 |
SamYaple | its a pain | 08:53 |
pbourke | SamYaple: hi | 08:54 |
SamYaple | pbourke: hey | 08:54 |
SamYaple | need to talk to you | 08:54 |
pbourke | SamYaple: did you read my comment at the top of https://review.openstack.org/#/c/210288/3/docker/docker_templates/base/Dockerfile.j2 | 08:54 |
SamYaple | yea did yo uread my review comments? | 08:55 |
SamYaple | or have you looked at the latest patchset? | 08:55 |
pbourke | yes to both | 08:55 |
*** sdake has quit IRC | 08:55 | |
pbourke | I just didnt see any reply to that one in particular | 08:55 |
SamYaple | "I only ask that we table the blocks/includes discussion until after we have a few of these implemented and see if there is a real need to do that." | 08:56 |
pbourke | ah | 08:56 |
*** weiyu_ has joined #kolla | 08:56 | |
SamYaple | im not saying no, but i do think people will see its really not that much once we get past the base | 08:57 |
pbourke | I missed that part | 08:57 |
SamYaple | i just want to see a few implemented before making the decision | 08:57 |
SamYaple | but besides that | 08:57 |
SamYaple | oraclelinux | 08:57 |
pbourke | thats fine | 08:57 |
pbourke | we wanted it that originally :) | 08:58 |
pbourke | but changed because of limits in docker name length | 08:58 |
SamYaple | ok, so oraclelinux instead of 'ol' works? | 08:58 |
pbourke | I dont think length is an issue any more though | 08:58 |
SamYaple | yea those limitations dont exist | 08:58 |
pbourke | right | 08:58 |
SamYaple | they may have at one point, but i think that was only for namespaces not the image names | 08:58 |
pbourke | no problem on name change | 08:58 |
pbourke | wrt the blocks though. I don't want them specifically | 08:59 |
pbourke | what I really do want though is the ability to add custom pieces, buildconf style | 08:59 |
SamYaple | The blocks as a usage for the proxy/preamble stuff i like for that purpose | 08:59 |
SamYaple | but i would prefer to keep it out of the normal flow to keep the file readable wihtout jumping around | 09:00 |
pbourke | I'll just give an example of the kind of scenario we have | 09:01 |
pbourke | if we're distributing kolla, we need to build everything from our own sources for liscencing reasons. so things such as cloning ansible from github we're having to replace and maintain in a fork | 09:02 |
pbourke | maybe this isn't a concern to kolla, but I think the kind of templating I'm proposing can make this kind of thing disappear | 09:03 |
pbourke | agree splitting it up too much could make things hard to follow though | 09:05 |
*** vbel has joined #kolla | 09:08 | |
pbourke | If I make an alternate patch would you have a look or are you eager to push through what's there | 09:08 |
vbel | good morning/evening | 09:09 |
SamYaple | Yea i understand what youre getting at, but right now we have a _tiny_, clean, and readable repo for everything. Introducing templating to the Dockerfiles is already a bad thing in my mind, but its a reasonable thing I think | 09:09 |
SamYaple | breaking up the blocks no longer makes it truly readable and thats my worry | 09:09 |
SamYaple | pbourke: I am more than happy to look at it, but i would like to merge that in | 09:10 |
*** weiyu_ has quit IRC | 09:10 | |
SamYaple | we can always update it | 09:10 |
*** weiyu_ has joined #kolla | 09:10 | |
SamYaple | we basically rewrite kolla a few times a year | 09:10 |
SamYaple | as it should be in my opinion | 09:10 |
pbourke | vbel: can you have a look at https://review.openstack.org/#/c/210288/6 ? | 09:10 |
vbel | yes, will do | 09:10 |
pbourke | vbel: thanks | 09:11 |
pbourke | SamYaple: agree on the docker image part, I think once ansible stabilises we should keep backward compatibility in mind though | 09:12 |
pbourke | SamYaple: i.e. how will an operator upgrade from our liberty release to mitaka | 09:12 |
SamYaple | ansible should largely remain the same i think | 09:13 |
SamYaple | but as for upgrade i believe since each major release upgrade will probbly have its own uniqueness we discussed a liberty-mintaka script | 09:13 |
SamYaple | same for m-n | 09:13 |
*** gfidente has quit IRC | 09:14 | |
vbel | pbourke: I don't see namespace arg in def createDockerfiles(self): | 09:16 |
vbel | pbourke: trying to match FROM %%KOLLA_NAMESPACE%%/%%KOLLA_PREFIX%%base:%%KOLLA_TAG%% to the function | 09:17 |
pbourke | vbel: I didn't write this patch so I dont know enough to answer these. better to stick comments on the review so there's more context | 09:18 |
vbel | sure, just wanted to raise question here because I can miss detail :) | 09:19 |
*** mfalatic has joined #kolla | 09:22 | |
*** mfalatic_ has joined #kolla | 09:23 | |
SamYaple | vbel: youre right its not there but this patch doesnt have a template file that uses it | 09:23 |
SamYaple | the next one we do would have a namespace and need that | 09:24 |
vbel | I will stick comment but won't -1 then | 09:24 |
SamYaple | however, all the code should be there already (we already have teh flags and information) | 09:24 |
vbel | is there a chance to make hierarchy of templates? like one j2 include another? | 09:25 |
SamYaple | vbel: that is exactly what we are trying to avoid | 09:26 |
vbel | it is rather hard to match ifs :) | 09:26 |
*** mfalatic has quit IRC | 09:26 | |
SamYaple | only for this large base image | 09:26 |
SamYaple | the rest are much much smaller in size | 09:27 |
SamYaple | dockerfiles are meant to be read straight through | 09:27 |
vbel | if centos/... if binary then include source.j2 else .. | 09:27 |
SamYaple | startin gto do includes and having multiple files is not in the spirit of Docker at all | 09:27 |
vbel | That's ok for me as soon as it still remains maintainable | 09:28 |
vbel | it is nicely commented so I like it | 09:28 |
SamYaple | yea maintainability and readability are my biggest concerns | 09:34 |
SamYaple | I feel like includes are not good for readability and would like to get by wihtout them | 09:34 |
vbel | I try to understand how do I stick per distro build env into it | 09:35 |
SamYaple | mainly i just want to get the code in and see how it goes | 09:35 |
vbel | I thought those build env things would be includes | 09:35 |
vbel | if it can be done later I am ok with that | 09:35 |
vbel | otherwise we patch docker :) | 09:36 |
pbourke | patching docker is the biggest pita | 09:44 |
pbourke | its not even on the table for non dev users | 09:45 |
vbel | hire a go programmer? ;) | 09:46 |
*** inc0 has joined #kolla | 09:54 | |
*** zhiwei has quit IRC | 10:02 | |
SamYaple | pbourke: for the record here, this blueprint was never meant to add customization... | 10:13 |
SamYaple | we have a blueprint that depnds on us templating al lthe files | 10:14 |
SamYaple | TEHN we add customization | 10:14 |
SamYaple | you are holding that up by not voting fyi | 10:14 |
pbourke | no im specifically not holding it up | 10:14 |
pbourke | I didnt vote -1 | 10:14 |
*** weiyu_ has quit IRC | 10:15 | |
*** gfidente has joined #kolla | 10:15 | |
SamYaple | -1 wouldn't hold it up, not voting does delay it | 10:16 |
pbourke | umm | 10:16 |
pbourke | ok | 10:16 |
SamYaple | I mean you want the proxy stuff, you are delaying us getting that if that is the reaso nyou are not voting | 10:18 |
SamYaple | I dont understand the reasons you have t obe neutral here is my point | 10:18 |
SamYaple | pbourke: i wasn't saying vote negative one. im having several conversations at the moment | 10:21 |
SamYaple | im trying to understand why you are trying to be neutron on the matter | 10:22 |
SamYaple | this has to be done first to implement the other blueprints | 10:22 |
*** tfukushima has quit IRC | 10:23 | |
SamYaple | oh jeez | 10:23 |
SamYaple | s/neutron/neutral/ | 10:23 |
pbourke | dont worry I got that :) | 10:23 |
*** tfukushima has joined #kolla | 10:24 | |
inc0 | SamYaple, remember? I moved the very same topic;) | 10:24 |
inc0 | adding proxy stuff to dockerfiles | 10:24 |
SamYaple | i am perfectly ok with that | 10:24 |
SamYaple | i like that | 10:24 |
SamYaple | but... its a dependant blueprint | 10:24 |
pbourke | SamYaple: its simple, just like I explained. I didn't want to hold up the change, just to comment on it. My worry was if I voted -1, as a core developer, it would hold things up | 10:24 |
SamYaple | we have to do this first | 10:24 |
inc0 | really (and I mean FUCKNIG REALLY) that would help | 10:24 |
pbourke | *core reviewer | 10:24 |
pbourke | I understand this has to be done first, but I would do it differently. Majority rules, and that's fine with me | 10:25 |
inc0 | btw added --reuse flag to build script | 10:25 |
inc0 | in case you don't want to rebuild everything | 10:25 |
SamYaple | what would reuse do inc0? | 10:25 |
inc0 | again, for example if you tinkered your base image to use proxy | 10:25 |
inc0 | which I ded | 10:25 |
*** tfukushima has quit IRC | 10:26 | |
inc0 | also it adds retry mechanism by default | 10:26 |
*** tfukushima has joined #kolla | 10:26 | |
inc0 | if n-th image raises, you just run script again | 10:26 |
SamYaple | inc0: to your question earlier, no neutron-server still not working for me | 10:29 |
SamYaple | dunno why | 10:29 |
SamYaple | well i think i know why im just not positive | 10:30 |
pbourke | SamYaple: when doing ansible templates, do you have any method for arriving at the minimal required options | 10:41 |
SamYaple | yea | 10:43 |
SamYaple | digging really deep into the code | 10:43 |
SamYaple | the deafults are easy to find | 10:44 |
openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Reuse option for build.py https://review.openstack.org/211495 | 10:45 |
inc0 | btw while testing this thing use case for it pop out itself | 10:47 |
inc0 | I ran out of disk on vm while building and after cleaning up files I could just carry on;) | 10:47 |
pbourke | inc0: how does this differ from docker caching | 10:48 |
SamYaple | thats really my question | 10:49 |
inc0 | well, docker cache didn't seem to work for me, but that might just be my mistake | 10:50 |
inc0 | I'll check it out once more | 10:50 |
inc0 | well it failed on parent image builds for me | 10:51 |
inc0 | unless I used this one | 10:51 |
pbourke | inc0: source or binary? | 10:52 |
inc0 | binary | 10:52 |
SamYaple | inc0: that packaged oslo.messaging is breaking neutron server | 10:53 |
SamYaple | 2.2.0 works file | 10:53 |
inc0 | meh | 10:53 |
inc0 | why do we always need to find packaging fuckups | 10:53 |
SamYaple | ima submit a patch | 10:53 |
SamYaple | i want source | 10:53 |
SamYaple | pbourke: do you test on that? | 10:54 |
SamYaple | i know you lead that i havent kept up | 10:54 |
pbourke | SamYaple: I use it and binary about 50/50 | 10:54 |
inc0 | I think we have most of source stuff done? | 10:54 |
pbourke | we do, I just find binary faster as it caches better | 10:55 |
pbourke | that's the main drawback to source imo | 10:55 |
*** Haomeng|2 has joined #kolla | 10:58 | |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Glance bootstrap container does not create tables and runs forever https://review.openstack.org/211500 | 10:58 |
vbel | Can we get this ugly bug fixed soon please? ^^^ | 10:59 |
SamYaple | pbourke: can you ack that as well? | 11:01 |
*** Haomeng has quit IRC | 11:01 | |
vbel | SamYaple, pbourke: thanks! | 11:03 |
SamYaple | strangely enough vbel it doesnt hang forever for me | 11:03 |
SamYaple | it didn't bootstrap the database, but it existed :/ | 11:04 |
SamYaple | exited* | 11:04 |
pbourke | I have a question around that var actually | 11:04 |
SamYaple | shoot | 11:04 |
pbourke | is "KOLLA_BOOTSTRAP:" | 11:04 |
pbourke | enough to set it? | 11:04 |
vbel | yes, the var was ok - it was None inside | 11:04 |
pbourke | if so, why not make it most explicit, "KOLLA_BOOTSTRAP: true" | 11:04 |
pbourke | *more | 11:04 |
SamYaple | that give th false impression it can be set to false | 11:04 |
SamYaple | which is not true | 11:05 |
pbourke | ah | 11:05 |
vbel | yes, it is a bit tricky :) | 11:05 |
SamYaple | there is a comment block in all the bootstrap start.sh scripts explaining this | 11:05 |
pbourke | something something including unset :) | 11:05 |
SamYaple | really all it comes down to is having KOLLA_BOOTSTRAP will trigger it | 11:06 |
SamYaple | no matter the content | 11:06 |
vbel | another bug I am going to open - glance base container misses 'glance' user :) | 11:07 |
vbel | I mean centos-source | 11:07 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Update neutron confs and ports https://review.openstack.org/210962 | 11:08 |
vbel | su -s /bin/sh -c "glance-manage db_sync" glance | 11:08 |
vbel | will fail | 11:09 |
SamYaple | yea for all source installs we should be creating the user | 11:09 |
SamYaple | pbourke: have you not run into that? | 11:09 |
SamYaple | you said yo uused source alot | 11:09 |
pbourke | its possible this was done in config-internal | 11:10 |
pbourke | and I've only used glance-binary with ansible | 11:10 |
SamYaple | i see | 11:10 |
vbel | I will check/fix it | 11:12 |
pbourke | when I said binary:source was 50:50 its probably more like 60:40 | 11:12 |
pbourke | I just can't stand waiting for them to build | 11:12 |
pbourke | there's also the fact that binary seems to take precendence in general | 11:13 |
pbourke | i.e. if implementing a new service binary is done first, source second | 11:13 |
vbel | $ grep -r useradd . | 11:13 |
vbel | ./barbican/Dockerfile: && useradd --user-group barbican \ | 11:13 |
vbel | ./ceilometer/ceilometer-base/Dockerfile: && useradd --user-group ceilometer \ | 11:13 |
vbel | ./cinder-app/cinder-base/Dockerfile: && useradd --user-group cinder \ | 11:13 |
vbel | ./designate/designate-base/Dockerfile: && useradd --user-group designate \ | 11:13 |
vbel | ./gnocchi/gnocchi-base/Dockerfile: && useradd --user-group gnocchi \ | 11:13 |
vbel | ./heat/heat-base/Dockerfile: && useradd --user-group heat \ | 11:13 |
vbel | ./horizon/Dockerfile: && useradd --user-group horizon \ | 11:13 |
vbel | ./keystone/Dockerfile: && useradd --user-group keystone \ | 11:13 |
vbel | ./magnum/magnum-base/Dockerfile: && useradd --user-group magnum \ | 11:13 |
SamYaple | vbel: pastebin! | 11:13 |
vbel | ./neutron/neutron-base/Dockerfile: && useradd --user-group neutron \ | 11:13 |
vbel | ./nova/nova-base/Dockerfile: && useradd --user-group nova \ | 11:13 |
pbourke | :D | 11:13 |
vbel | ./swift/swift-base/Dockerfile: && useradd --user-group swift \ | 11:13 |
vbel | ok, :) | 11:13 |
SamYaple | pbourke: id only use source if it was working | 11:13 |
vbel | http://paste.openstack.org/show/412530/ | 11:14 |
SamYaple | its not binary comes first, its its the only one working | 11:14 |
pbourke | SamYaple: it does work? | 11:14 |
SamYaple | well clearly not | 11:14 |
pbourke | there's some bugs there like the ones vlad is finding but that's just because people aren't using it enough | 11:14 |
pbourke | there's plenty of bugs in binary also | 11:14 |
SamYaple | i get it but that means its never worked for config-external | 11:14 |
SamYaple | no need to get defensive. I think the source work is outstanding, i just dont have the cycles to switch and fix the bugs to get it functional | 11:15 |
pbourke | no its fine | 11:16 |
pbourke | why are data containers created bootstrap rather than start ? | 11:16 |
pbourke | *created in | 11:16 |
SamYaple | in case bootstrap touches files in those areas (which it shouldnt) | 11:17 |
SamYaple | but in the case of rabbitmq and mariadb its required | 11:17 |
SamYaple | so consistency | 11:17 |
SamYaple | i have no strong feelings on it | 11:18 |
*** tfukushima has quit IRC | 11:18 | |
pbourke | grand | 11:18 |
SamYaple | ? | 11:18 |
pbourke | i mean, makes sense | 11:19 |
SamYaple | ah | 11:19 |
pbourke | http://www.urbandictionary.com/define.php?term=grand&defid=333480 | 11:20 |
pbourke | ;) | 11:20 |
SamYaple | thanks | 11:20 |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Glance source based container misses 'glance' user https://review.openstack.org/211512 | 11:21 |
vbel | That's not urban, that's Irish :) | 11:22 |
vbel | I am checking containers one by one locally so can meet unnoticed things | 11:25 |
*** rhallisey has joined #kolla | 11:25 | |
vbel | would be nice to get it fixed too ( https://review.openstack.org/211512 ) . Other containers have users. If I miss someone I will add new review | 11:39 |
*** diga has quit IRC | 11:45 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix oslo.messaging in neutron-server https://review.openstack.org/211518 | 11:48 |
SamYaple | fyi to all cores ^^ without that patch neutron-server will not start. we should probably merge it sooner rather than later | 11:49 |
*** coolsvap has quit IRC | 11:50 | |
*** tobe_ has quit IRC | 11:50 | |
*** coolsvap has joined #kolla | 11:51 | |
openstackgerrit | Merged stackforge/kolla: Glance bootstrap container does not create tables and runs forever https://review.openstack.org/211500 | 11:51 |
*** tobe_ has joined #kolla | 11:52 | |
*** sdake has joined #kolla | 11:54 | |
sdake | morning | 11:54 |
*** diga has joined #kolla | 11:54 | |
vbel | sdake: morning | 11:56 |
*** sdake_ has joined #kolla | 11:56 | |
*** tobe_ has quit IRC | 11:57 | |
sdake_ | pbourke good news about ansible swift! | 11:58 |
*** sdake has quit IRC | 12:00 | |
*** openstackgerrit_ has joined #kolla | 12:02 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix oslo.messaging in neutron-server https://review.openstack.org/211518 | 12:10 |
SamYaple | sdake_: ^ that fixes neutron and gets us a working deploy again | 12:10 |
SamYaple | its only been a week | 12:10 |
SamYaple | :/ | 12:10 |
sdake_ | just curious which version is installed from delorean | 12:11 |
sdake_ | i would like to report it upstream | 12:12 |
sdake_ | is that version in the patch the master of oslo.messaging or latest release? | 12:12 |
SamYaple | 2.2.1-dev5 | 12:15 |
SamYaple | 2.2.0 is the stable | 12:15 |
SamYaple | 2.2.1-dev5 is what is installed | 12:15 |
SamYaple | _and_ it only breaks neutron-server | 12:15 |
SamYaple | so i dont blame oslo | 12:15 |
sdake_ | that must be master | 12:15 |
SamYaple | it probably is | 12:16 |
*** gfidente has quit IRC | 12:16 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Remove Ansible abstraction for Nova https://review.openstack.org/211445 | 12:18 |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed stackforge/kolla: Remove Ansible abstraction for Neutron https://review.openstack.org/211442 | 12:18 |
*** tfukushima has joined #kolla | 12:18 | |
*** tfukushima has quit IRC | 12:23 | |
*** thumpba has joined #kolla | 12:25 | |
sdake_ | sweet my 845 tubes worked out of the box this time | 12:31 |
sdake_ | wow sound fantastic too | 12:32 |
*** thumpba has quit IRC | 12:45 | |
coolsvap | sdake_, i didnt get :-/ | 12:46 |
sdake_ | 845 tubes are for my tube amp | 12:47 |
sdake_ | last time I go tthtem one was doa | 12:47 |
sdake_ | they cost 160$ to ship back to the vendor... | 12:47 |
* sdake_ groans | 12:47 | |
pbourke | sdake_: whats that, the priority change? | 12:47 |
*** tfukushima has joined #kolla | 12:48 | |
coolsvap | oh | 12:48 |
sdake_ | huh pbourke? | 12:48 |
pbourke | the good news | 12:48 |
sdake_ | no that you are doing it | 12:48 |
pbourke | ah | 12:48 |
pbourke | was very confused there for a moment! | 12:49 |
pbourke | was like, I know | 12:49 |
SamYaple | good news about your tubes sdake! | 12:51 |
*** openstackgerrit_ has quit IRC | 12:51 | |
*** thumpba has joined #kolla | 12:52 | |
sdake_ | ya i was going to be sadpanda if theywere busted again | 12:52 |
pbourke | are they for guitar? | 12:52 |
sdake_ | what sucks is they cost $700 so I pretty much had to ship them back | 12:52 |
sdake_ | nah for a tube amp http://www.stevehuffphoto.com/line-magnetic-219ia-integrated-tube-amp-review-300b-and-845-tube-magic/ | 12:53 |
sdake_ | i bought some o ther tubes for 120 bucks (4 pack) and 2 were doa | 12:53 |
sdake_ | but the shipping cost is too high to replace them | 12:53 |
*** sdake_ is now known as sdake | 12:54 | |
*** dwalsh has joined #kolla | 12:54 | |
openstackgerrit | Merged stackforge/kolla: Added initial template for base image https://review.openstack.org/210288 | 12:55 |
sdake | dual mono western electric SET design | 12:56 |
sdake | design been around since the 60s probably | 12:56 |
*** openstackgerrit_ has joined #kolla | 12:56 | |
*** klint has quit IRC | 12:57 | |
sdake | first set amp i've sen implemented in dual mono with 300bs in set with 845s tho | 12:57 |
sdake | also hand wired, all the junk made today is circuit boards | 13:01 |
sdake | which imo are noisy | 13:01 |
sdake | i cheap stereo from the 80s is better then most stuff made today | 13:01 |
sdake | anyone have a systems camera they could recommend | 13:07 |
*** jmccarthy has quit IRC | 13:07 | |
vbel | sdake: do you have spedial starter for your amp? To avoid electricity outage in neighborhood :)) | 13:08 |
sdake | its only 24 watts | 13:09 |
vbel | ..special... ok | 13:09 |
sdake | i doubt it draws mor then 100 watts | 13:09 |
sdake | i can keep it on with the microwave | 13:09 |
sdake | my printer however i hae to shut off or the circuit blows ;( | 13:10 |
vbel | my friend had to buy starter for his lamp stereo | 13:10 |
sdake | you mean soft on ? | 13:10 |
sdake | thats built in to this amp | 13:10 |
vbel | yes | 13:10 |
*** jmccarthy has joined #kolla | 13:10 | |
*** gfidente has joined #kolla | 13:11 | |
sdake | i think if it allowed the tubes to play before warming up the tubes could pop | 13:11 |
sdake | so they have a 30 second soft on circuit | 13:11 |
sdake | but i am not sure why there is a soft on, that is justspeculation | 13:12 |
sdake | quick google search shows tube life is the reason for soft on | 13:17 |
sdake | the only problem with my tube amp is the tubes are off the hook expensive | 13:18 |
sdake | i've spent about 1800 on tubes or so | 13:18 |
sdake | just to retube the amp from stock | 13:18 |
sdake | the stock tubes kind of blow :( | 13:18 |
sdake | but atleast I have full backups in case of failure ;-) | 13:18 |
openstackgerrit | Merged stackforge/kolla: Fix oslo.messaging in neutron-server https://review.openstack.org/211518 | 13:19 |
openstackgerrit | Merged stackforge/kolla: Glance source based container misses 'glance' user https://review.openstack.org/211512 | 13:20 |
sdake | fun filled meeting day today | 13:21 |
* sdake groans | 13:21 | |
sdake | 4 hours of negotiation on the container track - can't wait for that | 13:21 |
* sdake ughs | 13:21 | |
*** cfarquhar has joined #kolla | 13:22 | |
sdake | i dont get why neo hits ctrl-x | 13:23 |
sdake | why not ctrl-c? | 13:23 |
rhallisey | sdake, maybe he uses emacs | 13:28 |
sdake | is ctrl-x an emacs thing? | 13:29 |
* sdake loses all respect for neo | 13:30 | |
sdake | real programmers use vi ;-) | 13:30 |
rhallisey | ctrl-x is usually part of an emacs command | 13:30 |
sdake | i'm glad they didnt teach emacs at my university | 13:30 |
sdake | it seems overly complicated | 13:30 |
sdake | vi is wierd and totally not intuitive but if you have been using it for 20 years, its almost magical what canbe done without even thinkingabout it ;-) | 13:31 |
sdake | I have to type in the commands i want to tell othe people how to use vi | 13:31 |
sdake | because I dont even know what to type, just how to type it for what I want ;-) | 13:31 |
*** inc0 has quit IRC | 13:32 | |
sdake | i've spend several years using microsoft's c compiler toolchain as well, their editor is pretty solid | 13:32 |
sdake | i think alled visual studio | 13:32 |
sdake | that was 20 year ago so who knwos what has changed | 13:33 |
*** jtriley has joined #kolla | 13:34 | |
*** diogogmt has joined #kolla | 13:40 | |
pbourke | SamYaple: wrt ansible swift, do you think https://github.com/stackforge/kolla/blob/master/docker/common/swift/swift-base/build-swift-ring.py should become an ansible module? | 13:41 |
*** gfidente has quit IRC | 13:42 | |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Glance source based containers cannot write to /var/log/glance https://review.openstack.org/211580 | 13:47 |
sdake | vlad are you interested in joining the kolla-drivers team - these folks manage the launchpad tracker | 13:49 |
sdake | rather vbel | 13:49 |
vbel | sdake: yes (vlad is short for vladislav :) | 13:49 |
pbourke | SamYaple: ah wait, scratch that, it has to run within the swift container | 13:50 |
sdake | vbel what is your login | 13:50 |
sdake | on launchpad that is | 13:50 |
sdake | oh i see vlad-belogrudov | 13:50 |
vbel | yes | 13:50 |
vbel | I have 2 more things to fix to get glance going - https://review.openstack.org/211580 plus one more bug against ansible code (bootstrap container mounts wrong config directory) | 13:51 |
sdake | vbel where are you located timezone wise | 13:51 |
vbel | moscow - UTC+4 (always summer time since recently) | 13:52 |
*** ccrouch has quit IRC | 13:52 | |
vbel | any special duties for kolla-drivers? | 13:53 |
sdake | bug triage when you feel up to it | 13:54 |
sdake | which means if a bug is in new state, set to triaged, set priority, target to liberty-3 | 13:54 |
sdake | brb | 13:54 |
*** sdake_ has joined #kolla | 13:55 | |
*** openstackgerrit_ has quit IRC | 13:55 | |
*** openstackgerrit_ has joined #kolla | 13:56 | |
vbel | sdake: ok, + learning on the job :) | 13:57 |
*** sdake__ has joined #kolla | 13:58 | |
*** sdake has quit IRC | 13:58 | |
*** sdake_ has quit IRC | 14:02 | |
*** slagle has quit IRC | 14:02 | |
*** diogogmt has quit IRC | 14:02 | |
*** slagle has joined #kolla | 14:04 | |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Glance bootstrap container mounts wrong config folder https://review.openstack.org/211587 | 14:10 |
vbel | sdake: moscow is my timezone, not location :) | 14:12 |
*** jmccarthy has quit IRC | 14:12 | |
*** jmccarthy has joined #kolla | 14:13 | |
*** openstackgerrit_ has quit IRC | 14:17 | |
*** openstackgerrit_ has joined #kolla | 14:18 | |
*** ccrouch has joined #kolla | 14:19 | |
pbourke | in ansible, I'm thinking we should allow two conditions when Swift rings are rebalanced. 1) is when we detect the config data has changed, 2) is when a force flag is set to true | 14:19 |
pbourke | is the only way to do to write out the config to a marker file in the container and then check this on subsequent runs? | 14:20 |
pbourke | another option is to just generate them the first time and have the operator take care of it after that, as it can be an expensive operation afaik | 14:23 |
*** banix has joined #kolla | 14:24 | |
vbel | I believe we shoud leave re-balancing swift to operators | 14:24 |
pbourke | just wondering how this fits in with the whole docker best practice of not modifying running containers | 14:25 |
*** openstackgerrit_ has quit IRC | 14:31 | |
pbourke | hmm I got an ATC code for tokyo | 14:36 |
pbourke | is that for kolla because we're in the big tent now? | 14:36 |
sdake__ | roger | 14:37 |
*** sdake__ is now known as sdake | 14:37 | |
pbourke | :) | 14:37 |
rhallisey | I still don't have ATC | 14:37 |
rhallisey | wtf | 14:37 |
sdake | rhallisey you should get a code | 14:37 |
rhallisey | :( | 14:37 |
sdake | rhallisey if you didn't get a code I can make contact with fungi on your behalf or you can harass him | 14:37 |
sdake | it sounds like the codes are coming out right now | 14:38 |
rhallisey | right I will hold for now and see | 14:38 |
sdake | i did make sure kolla cats will get atc codes | 14:38 |
sdake | anyone that made commits that is | 14:38 |
vbel | i got ATC earlier for some neutron bug fixes | 14:40 |
sdake | the policy atm (which is subject to change) is anyone who made a commit to a project in the governance projects.yaml file with correctly listed repos will receive an atc pass if they made a commit past april 30th | 14:40 |
sdake | 31 of this month I think is the cutoff from what fungi said | 14:40 |
sdake | but its really up to when the foundation does the cut off ;) | 14:40 |
sdake | rhallisey did you get a code for atc for VC ODS? | 14:42 |
sdake | coolsvap pingola | 14:50 |
coolsvap | sdake, pongila | 14:50 |
sdake | mind I ask what your working on and are you LFW | 14:50 |
pbourke | vbel: do you think we should have operators generate the rings locally somehow, then copy them across in the same way as we do the configs? | 14:50 |
pbourke | vbel: the only problem is it's not so user friendly | 14:51 |
coolsvap | sdake, i was going to work on the docker template stuff and some remaining ansible stuff | 14:51 |
coolsvap | trove i have started but slow progress | 14:51 |
sdake | trove was pushed to ibrty-3 | 14:52 |
sdake | rather to mitaka | 14:52 |
coolsvap | i started working on trove from l2, but not enough time | 14:52 |
sdake | main reason is to focus on the basic or "core" services as defcore calls them | 14:52 |
coolsvap | given to it | 14:52 |
vbel | pbourke: it's a bit of math and guess, not sure what we do | 14:53 |
sdake | ansible is th eonly one that knows the machine ids right? | 14:53 |
coolsvap | i was waiting on the base template merging, i will push next change sets | 14:53 |
sdake | there isn't a self discovery mechanism? | 14:53 |
coolsvap | at least for core services | 14:53 |
coolsvap | tonight and tomorrow mostly | 14:54 |
sdake | coolsvap I was looking for someone to take https://blueprints.launchpad.net/kolla/+spec/ansible-ceilometer | 14:55 |
sdake | that is 1 of our 2 unassigned blueprints | 14:55 |
coolsvap | sdake, i will do | 14:55 |
sdake | now the containers are done and working | 14:55 |
sdake | and IIRC you did them, so you have relevant experience with the config options | 14:55 |
*** achanda has joined #kolla | 14:56 | |
sdake | coolsvap we can probably spread the templating out among the community so everyone can do 1-2 patches each | 14:56 |
coolsvap | sure np | 14:57 |
sdake | as soon as samyaple does keystone or akwasnie does keystone we can break it up | 14:57 |
sdake | i'd like everyone to have 1 experience making a new template | 14:57 |
*** gfidente has joined #kolla | 14:58 | |
*** gfidente has joined #kolla | 14:58 | |
*** diogogmt has joined #kolla | 15:00 | |
*** tfukushima has quit IRC | 15:01 | |
*** tfukushima has joined #kolla | 15:01 | |
sdake | maybe i'll do keystone today if I can find a break in my meetings | 15:03 |
*** tfukushima has quit IRC | 15:03 | |
*** tfukushima has joined #kolla | 15:04 | |
*** achanda has quit IRC | 15:05 | |
coolsvap | sure | 15:05 |
coolsvap | please also populate the https://etherpad.openstack.org/p/kolla-dockerfile-template | 15:06 |
*** sdake_ has joined #kolla | 15:09 | |
*** sdake has quit IRC | 15:13 | |
*** jmccarthy has quit IRC | 15:16 | |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Glance-registry is missing a messaging config https://review.openstack.org/211616 | 15:17 |
*** blahRus has joined #kolla | 15:20 | |
sdake_ | anyone know off hand how many containers we have? | 15:20 |
openstackgerrit | Jeff Peeler proposed stackforge/kolla: Improve logging and prepare for gate https://review.openstack.org/202216 | 15:25 |
bmace | was there ever an l2 tag done on a set of images on dockerhub? | 15:26 |
bmace | so that people had a reliable set of images / deployment to mess with? | 15:26 |
pbourke | bmace: yes | 15:27 |
pbourke | should be there? | 15:27 |
sdake_ | bmace roger | 15:27 |
*** jruano has joined #kolla | 15:28 | |
bmace | i'm looking at the tags for images that should have it, and i don't see l2 tags | 15:28 |
bmace | only centos tagged? | 15:29 |
pbourke | oh, it seems so | 15:34 |
pbourke | I didnt actually know other OSes were being pushed at all | 15:34 |
bmace | might just be fedora? i don't think i saw any ol up there | 15:36 |
bmace | just depends on who does the pushes :) | 15:36 |
pbourke | in fairness there are only about 2 ol images upstream right now ;( | 15:48 |
pbourke | but that's on us | 15:50 |
pbourke | on that kind of topic, I was just discussing with sdake_ about how concerns of kolla in a corp environment | 15:51 |
sdake_ | liberty2 centos only | 15:51 |
sdake_ | did you want me to build other distros as well? | 15:51 |
* sdake_ not mindreader ;) | 15:51 | |
*** sdake_ is now known as sdake | 15:51 | |
sdake | bmace those fedora images are super old, I never pushed fedora either | 15:52 |
sdake | just centos | 15:52 |
sdake | if someone makes a super-push script and merges in repo, I'll build at each major milestone | 15:52 |
sdake | the super-push script should build eash source/binary distro combo | 15:52 |
pbourke | sdake: I dont think we need to say too much more about the proxy/customisation stuff though, as SamYaple indicated he had a plan and I trust that he does :) | 15:53 |
pbourke | im just more kind of putting it on the radar | 15:54 |
*** sdake has quit IRC | 15:57 | |
*** sdake has joined #kolla | 15:57 | |
*** juggler__ has joined #kolla | 15:58 | |
sdake | pbourke reading scrollback, what is busted about corporate environments, the docker-proxy thing? | 15:58 |
sdake | people complain about that alot with magnum ;) | 15:58 |
*** daneyon has quit IRC | 15:58 | |
*** daneyon has joined #kolla | 15:59 | |
bmace | ah, yup, and needing to run stuff as root right now is tough in corp environments. i know that we have the power to make sudo work, and that will help a bit. since this is a trusted / remote .ssh key enabled user the less permissions we can give them the better. | 15:59 |
pbourke | sdake: any line in a Dockerfile that needs to access the internet ignores the proxy | 16:00 |
sdake | you mean during build? | 16:00 |
pbourke | yes | 16:00 |
sdake | so the problem we need to solve is only build behind a corporate firwall? | 16:00 |
sdake | or is there more to it | 16:00 |
pbourke | that's one of them | 16:00 |
pbourke | but probably the main one atm | 16:01 |
sdake | magnum people complain about pull not build | 16:01 |
*** juggler_ has quit IRC | 16:01 | |
pbourke | pull is ok as you can set proxy args to the docker daemon | 16:01 |
pbourke | but it wont use those inside the image when doing RUN | 16:01 |
sdake | I'll be 1million% honest, I have no idea how docker proxying works :) | 16:01 |
pbourke | I'll tell you how it works, it doesn't :p | 16:01 |
pbourke | harmw: ping | 16:01 |
sdake | how on earth do you setup a proxy for runs then? | 16:02 |
sdake | like we have pip and curl and other things | 16:02 |
pbourke | so in pre kolla days we were doing something like this | 16:02 |
sdake | must proxying be set for each service individually? | 16:02 |
pbourke | echo http_proxy >> /proxy | 16:02 |
pbourke | then for every command it was RUN source /proxy && yum ... | 16:02 |
pbourke | finally rm /proxy | 16:02 |
pbourke | ugly but it worked | 16:02 |
sdake | groan | 16:02 |
sdake | super-groan | 16:03 |
pbourke | yup | 16:03 |
sdake | something must be done for curl and pip as well | 16:03 |
pbourke | yeah same thing | 16:03 |
sdake | is http_proxy some kind of globally defined environment variable? | 16:04 |
pbourke | its kind of bizarre the upstream docker team haven't done something here | 16:04 |
pbourke | sdake: you can set it globally with ENV | 16:04 |
pbourke | sdake: but then you can't unset it | 16:04 |
pbourke | so its stuck in the final image | 16:04 |
sdake | and what is wrong with that precisely | 16:04 |
pbourke | customers/users download images from dockerhub with our corp proxy set inside it | 16:05 |
bmace | yup. build proxies != run proxies for a container | 16:05 |
*** tfukushima has quit IRC | 16:05 | |
sdake | and that has what impact besides leaking data? | 16:05 |
sdake | does that override the run proxy? | 16:05 |
pbourke | well, I guess it screws internal access for containers if you're outside the proxy | 16:05 |
bmace | wouldn't it cause stuff just to plain not work? i know when proxies are set on my laptop and i'm off the vpn.. things don't work well :) | 16:06 |
pbourke | s/internal/external | 16:06 |
pbourke | yes | 16:06 |
pbourke | :) | 16:06 |
*** tfukushima has joined #kolla | 16:06 | |
sdake | your certain the runtime proxy is overridden if ENV proxy_http is set? | 16:06 |
*** tfukushima has quit IRC | 16:06 | |
pbourke | let me check... | 16:06 |
pbourke | to be sure | 16:06 |
*** tfukushima has joined #kolla | 16:06 | |
sdake | I totally wouldn't expect that to be the cae thinking of how they would implement build vs pull | 16:06 |
*** openstackgerrit_ has joined #kolla | 16:08 | |
pbourke | sdake: http://paste.openstack.org/show/412582/ | 16:08 |
*** pbourke has quit IRC | 16:08 | |
*** pbourke has joined #kolla | 16:09 | |
sdake | how about running with -e http_proxy=specialsauce | 16:09 |
sdake | this line : bourke@paul-laptop:foo$ docker run -it 180832a994d3 bash | 16:09 |
pbourke | -e overrides it yes | 16:10 |
*** wshao has joined #kolla | 16:10 | |
sdake | ok if its overriden to "" does that rewsult in a non-proxy curl operation? | 16:10 |
pbourke | so that's something | 16:10 |
pbourke | it does | 16:11 |
sdake | sounds like a complete solution to the problem to me | 16:11 |
pbourke | hmm ok, so we could maybe template in ENV, and add -e params to ansible | 16:11 |
pbourke | it doesn't solve the data leak but its better than what we have now | 16:11 |
pbourke | sdake: https://blueprints.launchpad.net/kolla/+spec/add-proxy-to-dockerfiles | 16:13 |
pbourke | I'll add this info there | 16:13 |
sdake | that is the best I can come up with ;) | 16:14 |
pbourke | thanks for the help | 16:14 |
sdake | you leak data with your model as well I think | 16:14 |
sdake | (the one you mentioned earlier) | 16:14 |
sdake | its just harder to access | 16:14 |
sdake | but I bet you could get at it with a docker save operation | 16:14 |
sdake | I am not super keen on passing an environment variable to the built images to override to null proxy but its better then not working at all | 16:17 |
pbourke | yeah I think some folk would have a thing or two to say about the leak too | 16:17 |
pbourke | sdake: do you mind if I bring this into l3 | 16:18 |
sdake | there is already a blueprint iirc | 16:18 |
sdake | but no I dont mind | 16:18 |
sdake | sec let me look at the tracker | 16:18 |
pbourke | yeah but its filed under M I think | 16:18 |
sdake | ya lets jam it in | 16:18 |
sdake | will be annoying to implement | 16:18 |
sdake | but easy | 16:18 |
sdake | as far as the data leak goes, all kinds of stuff gets leaked during build process of rpms for example | 16:19 |
sdake | on purpose | 16:19 |
sdake | so I dont think we are totally out of our minds here - following what others have done | 16:20 |
sdake | granted this is not the build server but the proxy machine for a corporate environment | 16:20 |
pbourke | whitebaord/topic | 16:21 |
sdake | got a link? | 16:21 |
pbourke | ignore ^ | 16:21 |
pbourke | :) | 16:21 |
sdake | say does the ENV override only need to go in the base image? | 16:22 |
pbourke | someday my clipboard is going to paste something I really regret | 16:22 |
vbel | for me keystone container does not work if i leave ENV with proxy inside | 16:22 |
pbourke | vbel: even if you unset it via -e | 16:22 |
pbourke | ? | 16:22 |
sdake | -e docker_proxy="" | 16:22 |
sdake | rather http_proxy | 16:22 |
vbel | it goes to that non-existing proxy | 16:23 |
vbel | I tried similar with ENV http_proxy="" | 16:23 |
pbourke | :( | 16:23 |
vbel | it should be unset | 16:23 |
vbel | people did not find workaround yet therefore many want that --build-env patch | 16:24 |
sdake | i have solution | 16:24 |
sdake | if [ $http_proxy="" ]; then | 16:24 |
sdake | unset the shit | 16:25 |
sdake | fi | 16:25 |
sdake | in every start file | 16:25 |
sdake | vbel can you try that | 16:25 |
*** wshao has quit IRC | 16:25 | |
*** daneyon has quit IRC | 16:26 | |
vbel | yes, per start.sh it would work, may be stick to kolla-common.sh | 16:26 |
sdake | ya kolla-common.sh wfm | 16:26 |
vbel | to hide corporate proxy we would have to define the very end / child container and ENV http_proxy="nowhere" there | 16:27 |
*** thumpba has quit IRC | 16:27 | |
vbel | this small hack solves http_proxy... we still have pypi mirrors :) | 16:27 |
sdake | i think there would possiblybe a leak there | 16:27 |
vbel | yes, with layers probably | 16:28 |
sdake | ya docker save | 16:28 |
sdake | there is another command, can't remember which which shows the layer info | 16:28 |
*** thumpba has joined #kolla | 16:28 | |
sdake | not sure if ENV is stored in the docker file format sounds like it is since it gets set | 16:29 |
sdake | pypi isn't http? | 16:29 |
pbourke | sdake: no he means package mirrors in general | 16:29 |
pbourke | as in the ability to use them | 16:30 |
sdake | you mean the problem that we get data from like 5 different sources | 16:30 |
vbel | yes, we can say via env about which pypi servers to use | 16:30 |
vbel | like external or internal ones | 16:30 |
sdake | oh i see mirror server overrides | 16:30 |
pbourke | yeah | 16:30 |
* sdake groans | 16:30 | |
pbourke | :D | 16:30 |
pbourke | kolla also needs a suit and tie | 16:31 |
vbel | :)))) | 16:31 |
sdake | well that pypi mirror problem only is needed during build | 16:31 |
sdake | not runtime | 16:31 |
sdake | its runtime where I don't want a bunch of -e overrides | 16:31 |
*** unicell has quit IRC | 16:31 | |
pbourke | yeah I think we can just template that in | 16:31 |
sdake | it makes the ansible implementation untidy | 16:31 |
vbel | it would be build time environment | 16:32 |
sdake | rhallisey ping | 16:33 |
vbel | proxies, repos, registries, special compiler flags .... | 16:33 |
sdake | we just morphed into gentoo ;) | 16:33 |
vbel | debug info, instrumentation :) | 16:33 |
sdake | oh speaking of debug info | 16:33 |
sdake | i found a wacky problem | 16:33 |
sdake | if a process segfaults, it saves the segfault IN THE HOST OS | 16:33 |
sdake | rather the core file | 16:34 |
*** gfidente has quit IRC | 16:38 | |
sdake | pbourke heads up when setting bluerptins need to set the milestone target too so it shows up in teh tracker | 16:43 |
pbourke | sdake: gotcha | 16:43 |
*** daneyon has joined #kolla | 16:44 | |
*** jasonsb has quit IRC | 16:49 | |
*** jasonsb has joined #kolla | 16:50 | |
*** daneyon_ has joined #kolla | 16:51 | |
*** daneyon has quit IRC | 16:53 | |
*** jasonsb has quit IRC | 16:54 | |
*** dwalsh has quit IRC | 16:57 | |
rhallisey | sdake, hey | 17:01 |
sdake | hey dude | 17:01 |
sdake | we had a conversatino about proxy for http | 17:01 |
sdake | solution we invented was to pass -e http_proxy="" to the container if there is no proxy | 17:01 |
sdake | or http_proxy="address" if there is one | 17:02 |
sdake | but it will always have to be set | 17:02 |
sdake | you good with that from tripleo's perspective? | 17:02 |
rhallisey | ya we can use -e also | 17:03 |
*** unicell has joined #kolla | 17:04 | |
*** dwalsh has joined #kolla | 17:12 | |
*** jruano has quit IRC | 17:16 | |
*** blahRus has quit IRC | 17:19 | |
*** blahRus has joined #kolla | 17:19 | |
*** thumpba has quit IRC | 17:22 | |
rhallisey | sdake, k got ATC | 17:29 |
rhallisey | it was in my junk :P | 17:29 |
*** jasonsb has joined #kolla | 17:33 | |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add ansible support for Cinder https://review.openstack.org/208868 | 17:53 |
sdake | rhallisey arey ou syncing up with coolsvap on those cinder changes | 17:53 |
sdake | I think he is pushing reviews too | 17:53 |
rhallisey | ^ that works except getting database error | 17:53 |
rhallisey | sdake, ya | 17:53 |
sdake | cool | 17:53 |
rhallisey | that's good to merge | 17:54 |
*** achanda has joined #kolla | 17:54 | |
rhallisey | I'll fix the database issue in a following patch since there is a bug file for it | 17:54 |
sdake | anyone know a way to display a git diff from a patch in a patch stream without running git format-patch? | 17:55 |
sdake | I typically do git format-patch -3 | 17:55 |
sdake | or whatever how many patches I have in my stream | 17:55 |
sdake | but I suspect there is a better way | 17:55 |
jpeeler | you just want to see the diff? | 17:59 |
sdake | right | 17:59 |
jpeeler | git show <sha> | 17:59 |
sdake | cool | 17:59 |
sdake | thanks ;) | 17:59 |
jpeeler | or if you just want to see everything | 17:59 |
sdake | always wondered how to do that | 17:59 |
jpeeler | git log -p | 17:59 |
sdake | git show is what i was after | 17:59 |
*** tfukushima has quit IRC | 18:06 | |
*** openstackgerrit_ has quit IRC | 18:06 | |
*** tfukushima has joined #kolla | 18:07 | |
*** rhallisey has quit IRC | 18:14 | |
*** mfalatic has joined #kolla | 18:16 | |
*** dwalsh has quit IRC | 18:18 | |
*** mfalatic_ has quit IRC | 18:20 | |
*** jruano has joined #kolla | 18:22 | |
*** mfalatic has quit IRC | 18:23 | |
*** mfalatic has joined #kolla | 18:24 | |
*** tfukushima has quit IRC | 18:25 | |
*** mfalatic_ has joined #kolla | 18:34 | |
*** mfalatic has quit IRC | 18:36 | |
*** mfalatic has joined #kolla | 18:39 | |
*** mfalatic_ has quit IRC | 18:42 | |
harmw | pbourke: sup | 18:46 |
sdake | fawk 91$ bill form rax | 18:55 |
* sdake groan | 18:55 | |
sdake | s | 18:55 |
*** tfukushima has joined #kolla | 18:59 | |
*** mfalatic has quit IRC | 19:00 | |
*** mfalatic has joined #kolla | 19:02 | |
*** jruano has quit IRC | 19:06 | |
*** mfalatic has quit IRC | 19:14 | |
*** mfalatic has joined #kolla | 19:17 | |
*** akwasnie has quit IRC | 19:43 | |
*** mfalatic has quit IRC | 20:15 | |
*** echoingumesh has joined #kolla | 20:46 | |
*** jruano has joined #kolla | 20:48 | |
*** tfukushima has quit IRC | 20:51 | |
*** jruano has quit IRC | 20:54 | |
*** tfukushima has joined #kolla | 21:07 | |
*** tfukushima has quit IRC | 21:09 | |
*** tfukushi_ has joined #kolla | 21:09 | |
*** tfukushi_ has quit IRC | 21:21 | |
*** tfukushima has joined #kolla | 21:22 | |
*** tfukushima has quit IRC | 21:26 | |
*** tfukushima has joined #kolla | 21:26 | |
*** tfukushima has quit IRC | 21:27 | |
*** tfukushima has joined #kolla | 21:28 | |
*** jtriley has quit IRC | 21:37 | |
*** juggler__ is now known as juggler | 21:38 | |
*** tfukushima has quit IRC | 21:48 | |
*** tfukushima has joined #kolla | 21:48 | |
*** mfalatic has joined #kolla | 21:59 | |
*** Haomeng has joined #kolla | 21:59 | |
*** Haomeng|2 has quit IRC | 21:59 | |
sdake | here is mostly our list of projects to integrate ;) | 22:02 |
sdake | http://governance.openstack.org/reference/projects/index.html | 22:02 |
*** jruano has joined #kolla | 22:09 | |
*** diogogmt has quit IRC | 22:12 | |
*** diogogmt has joined #kolla | 22:14 | |
*** Haomeng has quit IRC | 22:21 | |
*** Haomeng has joined #kolla | 22:22 | |
jasonsb | sdake: you forgot kolla | 22:40 |
bmace | integrate kolla into kolla? very inception like | 22:40 |
jasonsb | :) | 22:41 |
sdake | jasonsb it was actually a problem with the gating software which pushes the documentation | 22:42 |
sdake | eventually consistent documentation is what fungi said ;) | 22:42 |
jasonsb | i was figgering that | 22:42 |
sdake | jasonsb two jobs pushed at about the same time, index.html came from the pervious job | 22:42 |
*** achanda has quit IRC | 22:50 | |
*** jruano has quit IRC | 22:54 | |
*** achanda has joined #kolla | 23:02 | |
sdake | <Team> (top commit % | top review % | top core review % | top core reviewer %) | 23:27 |
sdake | (top 2 commit % | top 2 review % | top 2 core review % | top 2 core reviewer %) | 23:27 |
sdake | I18n Unable to determine diversity because | 23:27 |
sdake | this project lacks a deliverable | 23:27 |
sdake | Kolla (36.25% | 24.73% | 29.75% | 25.00%) | 23:27 |
sdake | (55.00% | 45.08% | 50.47% | 50.00%) | 23:27 |
sdake | yay ;) | 23:27 |
bmace | when does the official move happen? | 23:29 |
sdake | bmace its not scheduled yet | 23:32 |
sdake | however, the stackforge namespace has been deprecated, I'd expect the infra team will probably just manually move every project over | 23:32 |
sdake | there is a mail from james blair on the mailing list about it | 23:32 |
sdake | but he doesn't say what will happen with existing projects | 23:33 |
*** ccrouch has quit IRC | 23:36 | |
*** ccrouch has joined #kolla | 23:40 | |
*** jruano has joined #kolla | 23:40 | |
*** ccrouch has quit IRC | 23:40 | |
*** ccrouch1 has joined #kolla | 23:40 | |
SamYaple | hola | 23:43 |
*** echoingumesh has quit IRC | 23:45 | |
sdake | hey samyaple | 23:45 |
sdake | morning | 23:45 |
SamYaple | howdy | 23:45 |
*** echoingumesh has joined #kolla | 23:45 | |
sdake | pretty sweet the stackforge namespace is being deprecated | 23:46 |
SamYaple | idk i always liked it | 23:46 |
SamYaple | it seemed like good seperation | 23:46 |
*** tfukushima has quit IRC | 23:46 | |
sdake | be a little confusing to newcomers about what is an actual big tent project | 23:46 |
SamYaple | not if moves happened quickly ;) | 23:46 |
sdake | it was good for a long time but it doesn't scale imo | 23:46 |
sdake | there arelike 800 repos in openstack iirc | 23:47 |
sdake | ya agree if moves were a single operation | 23:48 |
*** ccrouch1 has quit IRC | 23:48 | |
*** ccrouch has joined #kolla | 23:48 | |
sdake | but its like a 20 step process ;) | 23:48 |
*** ccrouch has quit IRC | 23:48 | |
*** ccrouch has joined #kolla | 23:48 | |
sdake | all the while activity is going on in various places | 23:48 |
sdake | that is why they shut down the infra during moves | 23:48 |
*** echoingumesh has quit IRC | 23:51 | |
*** blahRus has quit IRC | 23:54 | |
*** thumpba has joined #kolla | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!