| *** Kennan2 is now known as Kennan | 00:02 | |
| *** Haomeng|2 has joined #kolla | 00:04 | |
| *** Haomeng has quit IRC | 00:07 | |
| *** weiyu has quit IRC | 00:16 | |
| *** britthouser has joined #kolla | 00:28 | |
| *** banix has quit IRC | 00:28 | |
| *** weiyu has joined #kolla | 00:43 | |
| *** jtriley has joined #kolla | 00:57 | |
| *** jtriley has quit IRC | 01:04 | |
| *** banix has joined #kolla | 01:05 | |
| *** ccrouch has joined #kolla | 01:06 | |
| *** achanda has joined #kolla | 01:26 | |
| *** erkules_ has joined #kolla | 01:33 | |
| *** ccrouch has quit IRC | 01:34 | |
| *** ccrouch has joined #kolla | 01:35 | |
| *** ccrouch has quit IRC | 01:35 | |
| *** erkules has quit IRC | 01:35 | |
| *** stvnoyes has joined #kolla | 01:38 | |
| *** stvnoyes has left #kolla | 01:39 | |
| *** Haomeng|2 is now known as Haomeng | 01:41 | |
| *** Haomeng has quit IRC | 01:48 | |
| *** Haomeng has joined #kolla | 01:57 | |
| *** dolpher has joined #kolla | 02:00 | |
| *** Kennan has quit IRC | 02:02 | |
| *** Kennan has joined #kolla | 02:02 | |
| *** dolpher1 has joined #kolla | 02:06 | |
| *** dolpher has quit IRC | 02:08 | |
| *** britthouser has quit IRC | 02:34 | |
| Kennan | hi sdake: | 02:42 |
|---|---|---|
| Kennan | I finisihed the build image steps | 02:42 |
| Kennan | and found | 02:42 |
| Kennan | INFO:__main__:Images that failed to build | 02:42 |
| Kennan | INFO:__main__:=========================== | 02:42 |
| Kennan | ERROR:__main__:gnocchi-a Failed with status: error | 02:42 |
| Kennan | ERROR:__main__:gnocchi-s Failed with status: error | 02:42 |
| *** klint has joined #kolla | 02:58 | |
| *** Haomeng|2 has joined #kolla | 02:58 | |
| *** Haomeng has quit IRC | 03:01 | |
| SamYaple | Kennan: that is normal | 03:15 |
| SamYaple | Kennan: those images currently fail to build for centos binary | 03:16 |
| sdake | hey kennan | 03:20 |
| sdake | knochi always fails to build | 03:20 |
| sdake | that is ok | 03:20 |
| SamYaple | hey sdake, looked at the review queue recently? | 03:20 |
| SamYaple | Well forget the buildup, this should make oyu happy sdake https://review.openstack.org/#/c/216036/ | 03:21 |
| Kennan | ok Thanks | 03:25 |
| Kennan | Let me try following steps | 03:25 |
| *** openstackgerrit has quit IRC | 03:31 | |
| *** openstackgerrit has joined #kolla | 03:32 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 03:37 |
| *** sdake has quit IRC | 03:38 | |
| *** diogogmt has quit IRC | 04:03 | |
| *** coolsvap|away is now known as coolsvap | 04:06 | |
| openstackgerrit | Merged stackforge/kolla: Update Getting Started link url in README.rst https://review.openstack.org/216037 | 04:14 |
| SamYaple | coolsvap: ping? | 04:25 |
| *** weiyu has quit IRC | 04:30 | |
| coolsvap | SamYaple, pong | 04:41 |
| SamYaple | hey! | 04:41 |
| coolsvap | good evening | 04:41 |
| SamYaple | so i dont see any ansible horizon code in the review queue, last i heard you were looking at it. I happen to have alot of this code, but didn't want to stomp your work | 04:41 |
| coolsvap | i dont remember i was asked to do horizon, i am working on mongodb and ceilometer | 04:42 |
| SamYaple | oh | 04:42 |
| SamYaple | i dont remember who was doing horizon then | 04:42 |
| coolsvap | i think sdake was lookin at it | 04:42 |
| coolsvap | there is a review | 04:43 |
| coolsvap | https://review.openstack.org/#/c/211286/ | 04:43 |
| coolsvap | i think harmw updated it | 04:43 |
| SamYaple | ah | 04:43 |
| SamYaple | maybe he was working on it | 04:43 |
| SamYaple | harmw: you around? | 04:43 |
| *** jtriley has joined #kolla | 04:46 | |
| *** jtriley has quit IRC | 04:50 | |
| *** ashishb has joined #kolla | 04:51 | |
| *** tummy has joined #kolla | 05:05 | |
| *** weiyu has joined #kolla | 05:19 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 05:28 |
| *** shardy_ has joined #kolla | 05:29 | |
| *** shardy has quit IRC | 05:30 | |
| *** shardy_ has quit IRC | 05:34 | |
| *** shardy has joined #kolla | 05:34 | |
| *** Haomeng|2 is now known as Haomeng | 05:43 | |
| *** unicell has quit IRC | 05:48 | |
| *** unicell has joined #kolla | 05:48 | |
| *** weiyu has quit IRC | 05:55 | |
| *** tummy has quit IRC | 05:56 | |
| *** jtriley has joined #kolla | 06:00 | |
| *** weiyu has joined #kolla | 06:01 | |
| *** weiyu has quit IRC | 06:01 | |
| harmw | SamYaple: at work, so not rly | 06:03 |
| SamYaple | harmw: real quick | 06:03 |
| SamYaple | ive got horizon patches for ansible | 06:03 |
| harmw | based on that WIP? | 06:04 |
| SamYaple | am i stepping on your toes by summting on that patchset | 06:04 |
| SamYaple | yea | 06:04 |
| harmw | no, ofcourse not :) | 06:04 |
| SamYaple | cool :) just wanted to verify | 06:04 |
| SamYaple | ive got it working btw | 06:05 |
| harmw | sdake was going to take that but he probably was a little too tired yesterday | 06:05 |
| harmw | or this morning, whatever | 06:05 |
| harmw | how close was my patchset? | 06:05 |
| SamYaple | pretty close, a few changes, and some changes in the horizon container | 06:05 |
| SamYaple | biggest thing is the horizon apache site conf | 06:06 |
| *** jtriley has quit IRC | 06:06 | |
| SamYaple | the acutal ansible code was pretty much spot on | 06:06 |
| harmw | nice | 06:06 |
| harmw | I hate submitting patches that I didn't patch | 06:06 |
| harmw | uhm | 06:06 |
| harmw | test | 06:06 |
| SamYaple | yea i know what you mean | 06:06 |
| harmw | but please do submit, no problem :) | 06:07 |
| SamYaple | awesome! | 06:07 |
| harmw | steve will love you for it, having a working horizon :> | 06:08 |
| SamYaple | yup its working, just need to test centos | 06:09 |
| SamYaple | oh yea, i put in all the ubuntu patches so that is a first class citezen | 06:09 |
| *** weiyu has joined #kolla | 06:09 | |
| SamYaple | I would even say its our best images | 06:09 |
| SamYaple | they are by far the smallest, by >500MB in some cases | 06:10 |
| *** weiyu has quit IRC | 06:10 | |
| harmw | lol yes | 06:10 |
| harmw | ubuntu is small :) | 06:10 |
| SamYaple | yea its what i built yaodu on for a reason | 06:11 |
| SamYaple | but im glad we have a choice | 06:11 |
| SamYaple | thats amazing | 06:11 |
| SamYaple | no other deployment tool really has that | 06:11 |
| harmw | :) | 06:11 |
| harmw | right, so, back to ansibalizing ELK against amazon | 06:11 |
| SamYaple | do it | 06:12 |
| *** weiyu_ has joined #kolla | 06:14 | |
| *** diga has joined #kolla | 06:22 | |
| openstackgerrit | Tim Potter proposed stackforge/kolla: Clean /var/run/http on container start https://review.openstack.org/216112 | 06:25 |
| *** achanda has quit IRC | 06:48 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Horizon https://review.openstack.org/216026 | 06:55 |
| *** inc0 has joined #kolla | 06:57 | |
| *** juggler has quit IRC | 06:57 | |
| inc0 | good morning | 06:58 |
| SamYaple | morning inc0 | 06:59 |
| inc0 | wassup Sam? | 06:59 |
| SamYaple | got ubuntu handled yesterday and today | 06:59 |
| SamYaple | mostly done with horizon for ansible now | 07:00 |
| inc0 | cool, so you can safely deploy ubuntu now? | 07:00 |
| SamYaple | yup | 07:00 |
| SamYaple | ubuntu is the only one where all the images build too ;) | 07:00 |
| inc0 | which one doesn't build on cent? | 07:00 |
| SamYaple | also the images ar about 300MB smaller | 07:00 |
| SamYaple | 600MB for galera image | 07:00 |
| SamYaple | gnocchi | 07:01 |
| SamYaple | for centos-binary | 07:01 |
| SamYaple | for centos-source quite a few | 07:01 |
| *** banix has quit IRC | 07:07 | |
| *** weiyu_ has quit IRC | 07:15 | |
| *** dolpher1 has quit IRC | 07:25 | |
| Kennan | hi SamYaple: | 07:29 |
| Kennan | I have a machine with only one NIC | 07:29 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Build.py regex filtering https://review.openstack.org/215523 | 07:29 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: build.py refactor https://review.openstack.org/215504 | 07:29 |
| Kennan | what should I config with neutron_external_interface ? | 07:29 |
| inc0 | uhh...this again | 07:29 |
| inc0 | SamYaple, should we tell him about DiD? | 07:30 |
| SamYaple | Kennan: youll need to have two interfaces | 07:30 |
| SamYaple | DinD inc0 | 07:30 |
| inc0 | DinD, my bad | 07:30 |
| inc0 | brb | 07:30 |
| SamYaple | Kennan: if you dont have two interfaces, you can set your first interface up in a bridge | 07:30 |
| Kennan | what's the requirements for setup the bridge? | 07:31 |
| SamYaple | what do you mean? | 07:31 |
| Kennan | I am not sure if I read the doc clearly, does the doc have said two NICs is needed? | 07:31 |
| *** weiyu has joined #kolla | 07:31 | |
| SamYaple | I do not know if the docs say that | 07:32 |
| Kennan | for if I have one NIC case, what's steps for setup the bridge ? just add-br ? | 07:32 |
| SamYaple | but you will need two unless you do fancy things with your single interface | 07:32 |
| Kennan | or ovs ? | 07:32 |
| Kennan | It not tell anything about it | 07:32 |
| SamYaple | youll put your single interface in a bridge | 07:32 |
| SamYaple | then youll create a veth pair, one side of the veth pair goes in the bridge | 07:32 |
| SamYaple | the other side is your 'external_interface' | 07:32 |
| *** dolpher has joined #kolla | 07:33 | |
| Kennan | SamYaple: did you have a example how to do that ? I used a VM on cloud, I did not want to lost network connectivty | 07:33 |
| SamYaple | you will most likely lose connection | 07:34 |
| Kennan | bad news | 07:34 |
| SamYaple | its tricky to do this over the same interface you are connected | 07:34 |
| SamYaple | not impossible though | 07:34 |
| *** chandankumar has joined #kolla | 07:39 | |
| harmw | SamYaple: please get horizon up for review, so ppl can ack on that :) | 07:42 |
| harmw | we need that, as every manager would say, 'ASAP!!!!!! NOW!!!!!' | 07:43 |
| SamYaple | harmw: im testing it | 07:47 |
| SamYaple | I test things :P | 07:47 |
| SamYaple | hey inc0 | 07:53 |
| SamYaple | here is a retry patchset, i welcome your comments https://review.openstack.org/#/c/216036/ | 07:53 |
| harmw | SamYaple: dude please, thats not very nice to say :> | 07:54 |
| SamYaple | :D | 07:54 |
| SamYaple | im kidding of course | 07:55 |
| harmw | hehe | 07:55 |
| harmw | np man, just get it up for review | 07:55 |
| harmw | and testing +1 | 07:55 |
| harmw | always | 07:55 |
| Kennan | hi SamYaple: | 08:01 |
| Kennan | for kolla what os does it support (for setuping environment) ? | 08:02 |
| Kennan | Centos, Fedora, ubuntu | 08:02 |
| Kennan | Right ? | 08:02 |
| SamYaple | Kennan: I believe we only have people testing on those, yes | 08:02 |
| Kennan | ok | 08:03 |
| SamYaple | but it _should_ work on any distro | 08:03 |
| SamYaple | I use ubuntu | 08:03 |
| SamYaple | most everyone else does centos/fedora | 08:03 |
| harmw | coreos worked as a base to land the containers on, but that was some months ago | 08:04 |
| harmw | for building, the regular will do | 08:04 |
| SamYaple | its only gotten better, i bet it owuld still work harmw | 08:05 |
| harmw | absolutely, essentially it's all about running containers so I bet even boot2docker would suffice for that part :) | 08:06 |
| *** jmccarthy has joined #kolla | 08:07 | |
| *** jtriley has joined #kolla | 08:07 | |
| *** jtriley has quit IRC | 08:13 | |
| inc0 | SamYaple, https://review.openstack.org/#/c/215504/7/tools/build.py about this one...raise will kill script | 08:16 |
| inc0 | unless we did something very wrong with error handling, but we didnt afair | 08:16 |
| SamYaple | Ah sorry. I did not see you had created the exception class | 08:17 |
| SamYaple | will update the comment | 08:17 |
| inc0 | SamYaple, about https://review.openstack.org/#/c/215523/6/tools/build.py and showing unprocessed images...why you would like to show 100 containers which wasn't even touched if you used regex to build 3? | 08:21 |
| inc0 | adding new section will only make things worse - even more lines which you don't care about | 08:23 |
| SamYaple | the summary to print out everything. i want to see all the images since that is a summary | 08:23 |
| SamYaple | 'here is everything' | 08:23 |
| SamYaple | and its far from 100 images | 08:23 |
| SamYaple | feel free to wrap that summary section output it in debug output if you want | 08:24 |
| inc0 | let's wait for pbourke for that as he was one to ask precisely opposite | 08:24 |
| SamYaple | well that wont change the debug request which already sounds like the compormise everyone wants | 08:25 |
| inc0 | it can be LOG.debug if that's what you're asking | 08:26 |
| SamYaple | yup | 08:26 |
| inc0 | ok | 08:26 |
| SamYaple | gate runs debug | 08:26 |
| SamYaple | so itll show up in the gate, and thats important | 08:26 |
| SamYaple | inc0: did we just agree on something? | 08:28 |
| SamYaple | should this be marked down? | 08:28 |
| inc0 | well...we did agree few times | 08:30 |
| SamYaple | :) | 08:30 |
| SamYaple | im kidding | 08:30 |
| inc0 | beer was good idea in CA | 08:30 |
| SamYaple | we should do that more | 08:30 |
| inc0 | that we agreed upon... | 08:30 |
| inc0 | and ... yeah I got nothing more. | 08:30 |
| SamYaple | ill introduce you to the guys around here | 08:30 |
| inc0 | beer drinkers you mean? | 08:33 |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - Neutron https://review.openstack.org/215925 | 08:33 |
| SamYaple | anything alcoholic really | 08:33 |
| inc0 | btw, moving ]): to line above will break pep8, so I'm not going to do that. | 08:33 |
| inc0 | sorry. | 08:33 |
| inc0 | I'll submit alternative (I think its actually worse than all()) which is pep8 compliant | 08:36 |
| SamYaple | 'and \' | 08:37 |
| SamYaple | ;) | 08:37 |
| inc0 | not this one, I don't like backslash in python. period. It breaks line in some arbitrary point and that's not clean imho | 08:39 |
| SamYaple | i won't agrue you on that point for just that reason | 08:39 |
| *** SourabhP has joined #kolla | 08:42 | |
| *** akwasnie has joined #kolla | 08:48 | |
| *** SourabhP_ has joined #kolla | 08:48 | |
| *** SourabhP has quit IRC | 08:51 | |
| *** SourabhP_ is now known as SourabhP | 08:51 | |
| openstackgerrit | Merged stackforge/kolla: Uses the appropriate port value for Glance https://review.openstack.org/215980 | 08:53 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Horizon https://review.openstack.org/216026 | 08:56 |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - OpenVSwitch https://review.openstack.org/215922 | 08:57 |
| inc0 | SamYaple, mind turning https://review.openstack.org/#/c/215504/7 this one to +2/1 if you don't have any issues with it? | 09:00 |
| SamYaple | inc0: i do have issues | 09:01 |
| SamYaple | all the unaddressed comments in patchset 6 | 09:01 |
| inc0 | you mean turning list to {} | 09:01 |
| inc0 | ? | 09:01 |
| SamYaple | yea | 09:01 |
| SamYaple | i think that was all | 09:01 |
| SamYaple | I dont see a need for this to be a dict anywhere | 09:02 |
| SamYaple | and you constently address it like a list afterward (because it should be a list) | 09:03 |
| *** dolpher has quit IRC | 09:04 | |
| inc0 | if you look at line 274 | 09:04 |
| inc0 | I need quick lookup over this datastructure there | 09:04 |
| inc0 | and I'll be refactoring using this as list as well | 09:04 |
| SamYaple | You dont need to change it for that | 09:05 |
| inc0 | unless I want to have O(n**2 ) I do | 09:05 |
| inc0 | and we dont want it to be O(n**2) | 09:05 |
| inc0 | we need to have quick reference by fullname to turn parent fullname to pointer to actual parent | 09:06 |
| SamYaple | are you kidding me? | 09:07 |
| inc0 | and quick references by name is what dicts are for | 09:07 |
| SamYaple | even with millions of images this would finish instnatly _without_ being a dict | 09:07 |
| SamYaple | ill take readability | 09:07 |
| inc0 | hey, let's not do memory crap | 09:08 |
| inc0 | please | 09:08 |
| SamYaple | yea ok, so change it back and we wont | 09:08 |
| SamYaple | youre the one bring up bigO | 09:08 |
| inc0 | hold on, we need iteration over that to filter out deps | 09:09 |
| inc0 | so we can build regex with deps | 09:09 |
| inc0 | I'm against tier building, this should be tree datastruct, but since we have it.. | 09:09 |
| SamYaple | hey im all for a refactor there man. i went with what was quickest to write to get this down | 09:10 |
| inc0 | then don't block it;) | 09:10 |
| SamYaple | i did that tier thing in bash originally nad just ported it | 09:10 |
| SamYaple | then dont fight stupid things | 09:10 |
| inc0 | trust me I'll work on it in iterations | 09:10 |
| SamYaple | and you dont need a dict for that filter | 09:10 |
| SamYaple | whats up with the fighting everything btw? This is all functionally the same, I care nothing about syntax and scrtucture... as long as it is consistent | 09:11 |
| inc0 | you'd rather have me iterate through images each time I check whether parent is present? | 09:11 |
| SamYaple | since we treat it like a list everywhere else? sure | 09:12 |
| SamYaple | remember youre going ot rewrite that tier building anyway | 09:12 |
| SamYaple | shouldnt be a problem for long | 09:12 |
| inc0 | but that's bad programming | 09:12 |
| inc0 | I have issues with tier programming, but right now I don't have time to rewrite it | 09:12 |
| inc0 | so I'll optimize little things | 09:13 |
| inc0 | and you pushback memory optimization | 09:13 |
| inc0 | I can't really discuss with that can I? | 09:13 |
| SamYaple | ill take readability over pointless optinmization | 09:13 |
| inc0 | it doesn't really hurt readibility. | 09:13 |
| SamYaple | calling values everytime you address it is worth a few cycles | 09:13 |
| inc0 | and it's not few cycles | 09:14 |
| SamYaple | microseconds. fractions of a second due | 09:14 |
| SamYaple | even if you think im being unreasonable, you are arguing over nothing | 09:15 |
| SamYaple | this saves nothing | 09:15 |
| SamYaple | it speeds nothing up | 09:15 |
| inc0 | and .values() is just few chars more | 09:16 |
| SamYaple | ....everytime you have to address it. | 09:17 |
| inc0 | so it's microseconds vs one method | 09:17 |
| SamYaple | versus something you dont like in one area | 09:17 |
| inc0 | no, every time you did adderss this | 09:17 |
| inc0 | right now you have option to use keys | 09:17 |
| inc0 | I have issues with thinking "it's just few miliseconds" | 09:18 |
| inc0 | it might become widely used later | 09:18 |
| SamYaple | its not milliseconds | 09:18 |
| SamYaple | even at millions of objects | 09:18 |
| SamYaple | go timeit | 09:18 |
| SamYaple | its fractions of nothingness | 09:19 |
| inc0 | we should make it as good as we can. and that's an improvement | 09:19 |
| inc0 | values thing is to keep current code working, not because it's how it should be done | 09:19 |
| SamYaple | we could also optimize the if threads == 1 test and save a few microseconds | 09:19 |
| SamYaple | but we dont | 09:19 |
| SamYaple | because there is no need | 09:19 |
| inc0 | but does it hurt anyone? | 09:20 |
| SamYaple | changing the current code and losing some code knowledge? having to pointless type values after every use? yes | 09:21 |
| SamYaple | i would argue that does affect things | 09:21 |
| SamYaple | no, that isnt to hard, but neither is leaving it alone | 09:21 |
| SamYaple | as a list | 09:21 |
| inc0 | whatever. we have meeting about python stuff anyway | 09:22 |
| SamYaple | really though, tell me what we would ever need to do with that other than loop through it? | 09:23 |
| inc0 | I would really ditch this logic, because there we change parent field and put reference that | 09:24 |
| inc0 | there | 09:24 |
| inc0 | this will potentially break things if people start to change this value in one place and it will change somewhere else | 09:24 |
| SamYaple | why keep the key name as a string in the value though? | 09:25 |
| inc0 | image['parent']['status'] == "error" this line uses reference, while it could use images[image['parent']]['status'] | 09:25 |
| inc0 | its safer | 09:25 |
| SamYaple | wait. | 09:25 |
| SamYaple | we are already working with that image | 09:26 |
| SamYaple | its the only thing we have | 09:26 |
| SamYaple | thats in an entirely different class | 09:26 |
| SamYaple | you dont have the full list | 09:26 |
| SamYaple | what are you talking about | 09:26 |
| SamYaple | why would oyu ever do that anyway? | 09:26 |
| SamYaple | you already have the object reference | 09:26 |
| inc0 | because object references are harder than lambda | 09:27 |
| inc0 | and much more volitale. | 09:27 |
| SamYaple | bullshit | 09:27 |
| SamYaple | now youre trying ot say dont do something because its hard. you cant have it both ways | 09:27 |
| inc0 | SamYaple, I've never seen someone had problems with one-liner lambda | 09:28 |
| inc0 | but I've seen bugs due to references in unlikely places | 09:28 |
| harmw | SamYaple: you've checked ansible 2.0 already? | 09:28 |
| inc0 | I'm saying this isn't hard becasue you need to know python, this one is hard because you need to carefully examine what's happending with datastructures throughout code | 09:28 |
| SamYaple | harmw: yea its still mostly busted | 09:29 |
| SamYaple | inc0: i dont see how a dict uncomplicates that | 09:29 |
| SamYaple | its just more coplexity | 09:29 |
| harmw | yea, thats what I see | 09:29 |
| inc0 | it's explicitly different datastructure, not calling references | 09:29 |
| inc0 | also, this is prework to turning this into actuall tree struct | 09:32 |
| inc0 | because I can implement tree on it, I can't do it in any good way without being able to find parent in single call | 09:33 |
| SamYaple | you can do that already... | 09:33 |
| *** SourabhP has quit IRC | 09:33 | |
| SamYaple | image['parent'] | 09:33 |
| SamYaple | done | 09:33 |
| SamYaple | youre making things up now | 09:33 |
| inc0 | find me all child images of given parent | 09:33 |
| SamYaple | now thats different. but a dict wont help you solve that any better | 09:34 |
| inc0 | [img for img in images if img['parent'] is self] - that's wrong way to do it | 09:34 |
| inc0 | anyway, whatever | 09:34 |
| inc0 | I'll just do it your way | 09:35 |
| *** shardy_ has joined #kolla | 09:53 | |
| *** shardy has quit IRC | 09:54 | |
| *** shardy_ has quit IRC | 09:58 | |
| *** shardy has joined #kolla | 09:59 | |
| openstackgerrit | Kai Qiang Wu(Kennan) proposed stackforge/kolla: Make doc more readable and easy to follow https://review.openstack.org/216175 | 10:06 |
| *** Kennan has quit IRC | 10:09 | |
| *** Kennan has joined #kolla | 10:10 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Ansible waits nova bootstrap container on all hosts https://review.openstack.org/216176 | 10:12 |
| *** diga has quit IRC | 10:23 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Ansible shouldn't wait for neutron bootstrap on all hosts https://review.openstack.org/216179 | 10:24 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Ansible shouldn't wait for neutron bootstrap on all hosts https://review.openstack.org/216179 | 10:25 |
| *** sdake has joined #kolla | 10:26 | |
| sdake | morning | 10:27 |
| SamYaple | morning sdake | 10:28 |
| sdake | hey SamYaple | 10:28 |
| SamYaple | sdake: i have ansible-horizon almost done | 10:30 |
| SamYaple | some css issues that im resolving | 10:30 |
| sdake | nice ! | 10:30 |
| SamYaple | aslo, you may be interested in https://review.openstack.org/#/c/216036/ | 10:30 |
| sdake | sweet retries | 10:30 |
| *** jtriley has joined #kolla | 10:37 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Wait for bootstrap containers on the same hosts they run https://review.openstack.org/216182 | 10:39 |
| *** jtriley has quit IRC | 10:44 | |
| *** vbel has joined #kolla | 10:44 | |
| harmw | SamYaple: there is a fix for css quirks in one of the (perhaps old) Dockerfiles | 10:44 |
| harmw | running some cmd | 10:44 |
| SamYaple | harmw: yea ive got it fixed. bad symlink | 10:45 |
| sdake | harmw tht patch was merged | 10:46 |
| vbel | good morning/evening | 10:48 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add ansible support for Horizon https://review.openstack.org/211286 | 10:48 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Horizon https://review.openstack.org/216026 | 10:48 |
| SamYaple | working for ubuntu ^^ | 10:48 |
| SamYaple | ill check centos now | 10:48 |
| harmw | great | 10:49 |
| SamYaple | likely there will be some path differcnes, but ill check it out | 10:49 |
| SamYaple | ah right centos is /usr/lib | 10:50 |
| SamYaple | duh | 10:50 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Horizon https://review.openstack.org/216026 | 10:51 |
| *** akscram has quit IRC | 10:56 | |
| SamYaple | hey sdake | 10:56 |
| SamYaple | that gating patch | 10:56 |
| SamYaple | got issues | 10:56 |
| SamYaple | docker 1.8.1 seems to make the gate take _forever_ | 10:57 |
| SamYaple | the template scripts time out | 10:57 |
| SamYaple | and the functional-f21 test takes 80 minutes | 10:57 |
| sdake | functionl-f21 is running build.sh | 10:57 |
| SamYaple | i know | 10:57 |
| sdake | sometimes th gate is slower when running eithe on ra or hp | 10:57 |
| sdake | rax or hp | 10:57 |
| SamYaple | but its consistently longer | 10:57 |
| sdake | i'm not sure which one is slower | 10:57 |
| SamYaple | ive rechecked and had 2 different patchsets | 10:58 |
| SamYaple | its _always_ consistently slower | 10:58 |
| SamYaple | and like i said, the build.py stuff takes over an hour, it times out | 10:58 |
| SamYaple | no errors, just slowness it seems | 10:58 |
| sdake | may not be docker 1.8.1 | 10:59 |
| SamYaple | i even put in the format of the ephemeral disk for speedand it only helped a little | 10:59 |
| sdake | but something else with th egate | 10:59 |
| SamYaple | maybe, but i dont know what | 10:59 |
| SamYaple | docker 1.8.1 is the only change that makes a difference | 10:59 |
| sdake | oh you tried using 1.7.1? | 10:59 |
| SamYaple | yea you can see in the history | 11:00 |
| SamYaple | for the record i run 1.8.1 and its zippy | 11:00 |
| SamYaple | its taken 2 minutes off my builds | 11:00 |
| sdake | so the gate prints whther its running on hp cloud or rax cloud duringstartup | 11:00 |
| SamYaple | ill check | 11:00 |
| sdake | one is definatley slower | 11:00 |
| SamYaple | rax is both the fast and slow one | 11:01 |
| SamYaple | (difference is 1.8.1) | 11:01 |
| sdake | doesn't the current gat use 1.8.1? | 11:02 |
| SamYaple | http://logs.openstack.org/36/216036/2/check/gate-kolla-functional-f21/2570deb/console.html | 11:02 |
| SamYaple | http://logs.openstack.org/97/213397/13/check/gate-kolla-functional-f21/4a331ea/ | 11:02 |
| SamYaple | both rax | 11:02 |
| SamYaple | ~30 min difference | 11:02 |
| SamYaple | sdake: i would be shocked, it is like a week old | 11:02 |
| SamYaple | but maybe | 11:02 |
| sdake | th current gate uses curl iirc | 11:02 |
| sdake | that uses docker 1.8.1 | 11:02 |
| SamYaple | oh current gate is 17.1 | 11:03 |
| *** akscram has joined #kolla | 11:03 | |
| SamYaple | current gate doesnt use curl | 11:03 |
| SamYaple | does it? | 11:03 |
| sdake | i dont know | 11:03 |
| sdake | sec | 11:03 |
| SamYaple | idk anyway the changes i said i would make are there but its slow for seeminlgy unrelated reasons | 11:03 |
| sdake | gate uses docker-engine 1.7.1 | 11:04 |
| SamYaple | yup | 11:04 |
| sdake | presently | 11:04 |
| sdake | try running a tet patch with only docker engine 1.8.1 aas the only chnge | 11:06 |
| sdake | that will tell you if its docker 1.8.1 orsomthing else with your change | 11:06 |
| SamYaple | yea i did that in a seperate gate hacking patch | 11:07 |
| SamYaple | also slow | 11:07 |
| SamYaple | but why is the question | 11:07 |
| SamYaple | its got to be using like the loopback-lvm driver | 11:07 |
| SamYaple | i bet thats what it is | 11:07 |
| SamYaple | let me check that | 11:07 |
| *** Haomeng|2 has joined #kolla | 11:12 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 11:12 |
| *** Haomeng has quit IRC | 11:15 | |
| *** weiyu has quit IRC | 11:17 | |
| *** rhallisey has joined #kolla | 11:19 | |
| sdake | samyaple does that horizon patch work on centos? | 11:29 |
| SamYaple | im testing these things now | 11:29 |
| sdake | ok | 11:30 |
| SamYaple | my setup.py install from source things aren't going as planned | 11:30 |
| SamYaple | but i have a fix | 11:30 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Build.py regex filtering https://review.openstack.org/215523 | 11:32 |
| openstackgerrit | Merged stackforge/kolla: Wait for bootstrap containers on the same hosts they run https://review.openstack.org/216182 | 11:35 |
| sdake | samyaple re docker dir and horizon, if you have to make any changes to docker_files, can yu change the docker dir files for hirozn binary as well? I'd like to keep them in sync until we remove the dockerdir entirely | 11:41 |
| *** diogogmt has joined #kolla | 11:41 | |
| SamYaple | they are way out of sync right now | 11:43 |
| SamYaple | there is alot of code that only exists in templates currently sdake | 11:44 |
| SamYaple | from several different (merged) patches | 11:44 |
| sdake | not asking for all the new stuf to be merged | 11:44 |
| sdake | only horizon plz | 11:44 |
| SamYaple | im not making any binary centos chnages fyi | 11:44 |
| SamYaple | thatll probably still be broken | 11:44 |
| sdake | https://review.openstack.org/#/c/211286/ | 11:52 |
| sdake | it says outdated dependency | 11:52 |
| sdake | i think gerrit wont merge it once approved | 11:53 |
| sdake | SamYaple ^^ | 11:53 |
| SamYaple | http://logs.openstack.org/97/213397/14/check/gate-kolla-build-images-centos-binary/e6c21ac/console.html | 11:53 |
| SamYaple | it _is_ using the /dev/loop0 device | 11:53 |
| SamYaple | sdake: im actively working on these | 11:54 |
| sdake | ok let me know when you want a review | 11:54 |
| SamYaple | i will, i just put them back in WIP | 11:54 |
| sdake | i usually just put wip in th commit header so ppl know | 11:54 |
| sdake | better then the wip flag because new git reviews rset it :( | 11:54 |
| SamYaple | well i thought they were good before | 11:54 |
| SamYaple | but then centos | 11:55 |
| SamYaple | i mean it works for ubuntu | 11:55 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Ceph container https://review.openstack.org/214706 | 11:56 |
| SamYaple | rhallisey: are you making the ansible patch a seperate patch? | 11:58 |
| rhallisey | ya | 11:58 |
| SamYaple | awesome | 11:58 |
| SamYaple | is that cpeh one ready for review then? | 11:58 |
| rhallisey | it's likely going to be reworked as I add things to the ansible side | 11:59 |
| rhallisey | but you can look at it if you want | 11:59 |
| SamYaple | oh | 11:59 |
| SamYaple | ok in that case | 11:59 |
| SamYaple | can i repatch that with ubuntu support? | 11:59 |
| SamYaple | i was going to work on that after that metrged | 11:59 |
| rhallisey | ya that's fine | 12:00 |
| SamYaple | cool. ill probably do that tomorrow them. shouldn't be more than one patch from me then the rest should sync up | 12:01 |
| rhallisey | kk | 12:01 |
| *** ccrouch has joined #kolla | 12:02 | |
| sdake | rhallisey please hit the patch queue re cinder | 12:03 |
| sdake | cinder is busted atm | 12:03 |
| rhallisey | what happened? | 12:03 |
| sdake | aftr my patch set lvm works properly withouot binding /dev for example | 12:03 |
| sdake | it wasn't tested would be my guess ;-) | 12:03 |
| sdake | iscsi is still busted though | 12:03 |
| SamYaple | we may just need to abandon iscsi in containers | 12:04 |
| SamYaple | its not very container like anyway | 12:04 |
| sdake | what other way would we do cinder? | 12:04 |
| rhallisey | ya we will. Weird it worked for me | 12:04 |
| SamYaple | sdake: ceph | 12:04 |
| sdake | some operators may not be willing to commit to ceph | 12:05 |
| *** diogogmt has quit IRC | 12:05 | |
| rhallisey | we will have 2 storage options | 12:05 |
| *** shardy_ has joined #kolla | 12:05 | |
| SamYaple | fair, but iscsi isnt exactly stable | 12:05 |
| rhallisey | we need a local storage and ceph | 12:05 |
| sdake | agree iscsi is a pos | 12:05 |
| SamYaple | if they are using kolla they should be using ceph | 12:05 |
| SamYaple | iscsi gonna suck for them | 12:05 |
| sdake | i'm not confident ceph will be ready by wednesday ;) | 12:05 |
| SamYaple | but im fine with keeping it | 12:05 |
| SamYaple | is that a challenge i hear... | 12:06 |
| SamYaple | remember i have all the code already in yaodu | 12:06 |
| sdake | rather have heat working ;) | 12:06 |
| SamYaple | say the word and find two reviews and its merged | 12:06 |
| *** shardy has quit IRC | 12:06 | |
| sdake | i am confident we could merge ceph from yaodu by wednesday in working order - not sure i coul get everyhing setup with it by then tho | 12:07 |
| sdake | rhallisey this changeset and dependent patches: https://review.openstack.org/#/c/215975/2 | 12:08 |
| sdake | note this bookmark is what I use to look at kolla reviews: https://review.openstack.org/#/q/is:watched+status:open+project:stackforge/kolla,n,z | 12:10 |
| *** shardy_ has quit IRC | 12:10 | |
| *** weiyu has joined #kolla | 12:11 | |
| *** shardy has joined #kolla | 12:11 | |
| SamYaple | oh goody | 12:17 |
| SamYaple | all my patches require rebases | 12:17 |
| SamYaple | :/ | 12:17 |
| sdake | SamYaple I don't know if you caught this has a -1 on it https://review.openstack.org/#/c/215924/ | 12:18 |
| SamYaple | the one i have responded too sdake ? | 12:18 |
| sdake | oh didn't see a response | 12:20 |
| sdake | (didn't check forone) | 12:20 |
| SamYaple | it was about a minute before you posted that :) | 12:21 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Build.py regex filtering https://review.openstack.org/215523 | 12:23 |
| sdake | rhallisey this needs review https://review.openstack.org/#/c/215965/1 | 12:25 |
| sdake | and this https://review.openstack.org/#/c/215969/ | 12:25 |
| sdake | and this https://review.openstack.org/#/c/215970/ | 12:26 |
| sdake | ;-) | 12:26 |
| sdake | all 1 liners iirc | 12:26 |
| *** weiyu has quit IRC | 12:27 | |
| SamYaple | sdake: all my patches need rebases | 12:29 |
| SamYaple | i do not comprehend | 12:29 |
| sdake | all your bases are belong to us! | 12:30 |
| sdake | ok got guidance from ttx about version #'s | 12:31 |
| SamYaple | 1.0.0 | 12:31 |
| sdake | he said we should start with 0.1.0 so z can be bumped | 12:31 |
| sdake | he said if we are ready to commit to abi stability in the containers, we shoud lbump to 1.0.0 | 12:31 |
| SamYaple | liberty becomes 1.0.0? | 12:31 |
| openstackgerrit | Merged stackforge/kolla: automaton renamed to python-automaton https://review.openstack.org/215965 | 12:31 |
| *** weiyu has joined #kolla | 12:32 | |
| sdake | so if https://blueprints.launchpad.net/kolla/+spec/replace-config-external is finished we go to 1.0.0 | 12:32 |
| openstackgerrit | Merged stackforge/kolla: Add retries to build.py https://review.openstack.org/216036 | 12:32 |
| openstackgerrit | Merged stackforge/kolla: Cinder reads non-Kolla generated config files https://review.openstack.org/215969 | 12:32 |
| openstackgerrit | Merged stackforge/kolla: Use cinder-volumes as name of LVM VG https://review.openstack.org/215970 | 12:32 |
| sdake | note samyaple on a rebase if there are no conflicts and already two +2's after gerrit rebases, your free to apply a +a if you like | 12:33 |
| sdake | if the +2's are lost, we have to go throug hthe review process again - groan | 12:33 |
| SamYaple | yea the rebase is going to kill them | 12:33 |
| SamYaple | or it always has for me | 12:33 |
| sdake | if you use gerrit to rebase they are retained iirc | 12:33 |
| SamYaple | it wont gerrit rebase | 12:34 |
| sdake | bummer | 12:34 |
| SamYaple | its ok. i can use this to clean up the source installs anyway | 12:34 |
| SamYaple | unify them too | 12:34 |
| * SamYaple wants bright side | 12:34 | |
| sdake | this is why I always sdo my shit as one big straem ;) | 12:34 |
| sdake | easier to do one big rebase op | 12:35 |
| SamYaple | you still end up rebasing alot | 12:35 |
| sdake | then 20 different ones | 12:35 |
| SamYaple | but that wouldnt help here | 12:35 |
| SamYaple | they are not dependant patches | 12:35 |
| sdake | ya | 12:35 |
| sdake | btw nice work on retries | 12:35 |
| sdake | the gate actually worked on your patch ;) | 12:35 |
| harmw | sdake: sattisfied about horizon? | 12:36 |
| SamYaple | it was super simple when i realized where we could inject it | 12:36 |
| sdake | harmw samyaple has been working on it - i haven' treid it | 12:36 |
| harmw | i know :) | 12:36 |
| harmw | ok | 12:36 |
| SamYaple | harmw im having some install trouble with centos | 12:36 |
| harmw | why? | 12:36 |
| SamYaple | pip installing is different for the two distros | 12:36 |
| harmw | interesting | 12:36 |
| *** jtriley has joined #kolla | 12:44 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: WIP: Remove data containers and adjust compose files https://review.openstack.org/212830 | 12:46 |
| *** sdake_ has joined #kolla | 12:47 | |
| *** shardy_ has joined #kolla | 12:48 | |
| *** shardy has quit IRC | 12:49 | |
| *** sdake__ has joined #kolla | 12:49 | |
| *** sdake has quit IRC | 12:50 | |
| sdake__ | samyaple re that json work | 12:50 |
| sdake__ | given all the other stuff on your plate, will that finish before the 25th? | 12:50 |
| SamYaple | tomorrow? | 12:52 |
| SamYaple | or sept? | 12:52 |
| sdake__ | sorry sept | 12:53 |
| SamYaple | probably | 12:53 |
| *** shardy_ has quit IRC | 12:53 | |
| SamYaple | but ill let others handle if they want | 12:53 |
| sdake__ | it would be nice to lock in 1.0.0 | 12:53 |
| *** sdake_ has quit IRC | 12:53 | |
| SamYaple | i need to catch up on reviews sdake. youre pulling away! | 12:53 |
| sdake__ | openstack policy in x (the 1.0.0 x.y.z) is that api changes, x changes | 12:54 |
| *** jtriley has quit IRC | 12:54 | |
| *** shardy has joined #kolla | 12:54 | |
| SamYaple | thats hardly and api | 12:54 |
| SamYaple | an* | 12:54 |
| SamYaple | well i guess comparitevely it might be | 12:54 |
| *** sdake__ is now known as sdake | 12:59 | |
| sdake | got mor eguidance from ttx | 12:59 |
| sdake | 1.0.0 to 1.1.0 should not break apis | 12:59 |
| sdake | if you break apis you ahve to bump x to 2.0.0 | 12:59 |
| sdake | which looks a bit odd for a young project | 12:59 |
| sdake | the only api in our system imo is that container interface | 13:00 |
| sdake | possibly build.py is an api | 13:00 |
| SamYaple | i dont know | 13:00 |
| *** banix has joined #kolla | 13:01 | |
| SamYaple | biuld.py wont affect deploy | 13:01 |
| sdake | right | 13:01 |
| SamYaple | if a container _requires_ new ansible code, that would be api to me | 13:01 |
| sdake | could you expand with an example | 13:02 |
| sdake | 6am - brain not warme up yet ;) | 13:02 |
| SamYaple | container changes and requires a new mount/new bootstrap code/new permissions to work when it was working previously that would be 'api' change to me | 13:03 |
| SamYaple | if its a bug then its a bug | 13:03 |
| sdake | ya upgrade will be a new api for example | 13:03 |
| sdake | I think we need a new environment variable for that like BOOTSTRAP but UPGRADE instead | 13:03 |
| sdake | or in addition | 13:03 |
| SamYaple | yea i cant think about htis at the moment | 13:04 |
| SamYaple | goots thoughts though i like that | 13:04 |
| sdake | ok thanks | 13:04 |
| sdake | goots? | 13:04 |
| SamYaple | good | 13:04 |
| SamYaple | idk | 13:04 |
| SamYaple | just trying to get horizon out | 13:05 |
| sdake | thanks stay focused on that :) | 13:05 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Horizon https://review.openstack.org/216026 | 13:06 |
| SamYaple | that code _should_ be centos and ubuntu compatible | 13:06 |
| SamYaple | and the template that im using for the rest of the source installs | 13:06 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add ansible support for Horizon https://review.openstack.org/211286 | 13:07 |
| inc0 | https://github.com/stackforge/kolla/blob/master/docker/centos/binary/neutron/neutron-agents/Dockerfile#L5 guys is there any reason why we don't use pip here | 13:10 |
| inc0 | ? | 13:10 |
| sdake | SamYaple or pbourke can you guys review this bad boy and all his children ;) | 13:11 |
| SamYaple | inc0: thats fixed in templates | 13:11 |
| *** britthouser has joined #kolla | 13:11 | |
| SamYaple | inc0: https://github.com/stackforge/kolla/blob/master/docker_templates/neutron/neutron-agents/Dockerfile.j2 | 13:11 |
| sdake | docker directory going away asap | 13:12 |
| sdake | we just need to sort out what gaps remain to make that happen | 13:12 |
| SamYaple | i think git install... | 13:12 |
| *** dims has joined #kolla | 13:12 | |
| sdake | we sorted out the environment thing with the preeamble section right? | 13:13 |
| sdake | vbel ^^ | 13:13 |
| vbel | sdake: I will check how it builds now | 13:14 |
| sdake | the single image build is ready to go once it passes teh gate | 13:15 |
| sdake | that was the other gap we had | 13:15 |
| sdake | hey look ma it passed the gate ;-) | 13:15 |
| *** shardy_ has joined #kolla | 13:15 | |
| *** britthou_ has joined #kolla | 13:16 | |
| sdake | nice job inc0 | 13:16 |
| *** shardy has quit IRC | 13:17 | |
| inc0 | thanks | 13:17 |
| inc0 | it was costly to my mental health. I think I'll drink alkohol tonight | 13:18 |
| sdake | nic e;) | 13:18 |
| sdake | thats sure to improve it ;) | 13:18 |
| inc0 | alcohol even | 13:18 |
| *** britthouser has quit IRC | 13:18 | |
| *** banix has quit IRC | 13:19 | |
| sdake | SamYaple for multinode deploy do I need to run kolla-deploy as sudo or does it run as my uuid? | 13:19 |
| sdake | uid thatis | 13:20 |
| SamYaple | ansible can run as non-root, it must ssh to root | 13:21 |
| *** shardy_ has quit IRC | 13:21 | |
| *** apmelton has quit IRC | 13:21 | |
| sdake | in the muitnode deploy inventory file you have somehting ansible_ssh_use = sam iirc | 13:21 |
| sdake | whatis that about? | 13:21 |
| SamYaple | that is about something we didnt implement yet | 13:21 |
| *** shardy has joined #kolla | 13:21 | |
| harmw | sudo! | 13:21 |
| harmw | :> | 13:21 |
| sdake | should I change it to sdake? | 13:22 |
| harmw | either that, or | 13:22 |
| harmw | use -e ansible_ssh_user=root | 13:22 |
| SamYaple | sdake: yea just remove it | 13:22 |
| harmw | it's about the user to connect with | 13:22 |
| SamYaple | or change it root | 13:22 |
| SamYaple | either one | 13:22 |
| harmw | yep | 13:22 |
| harmw | he SamYaple ansible sucks balls | 13:22 |
| harmw | why do I keep copy/pasting blocks of ansible code | 13:23 |
| harmw | throughout all my playbooks and tasks | 13:23 |
| *** weiyu has quit IRC | 13:23 | |
| SamYaple | case you dont know how to abstract! | 13:23 |
| *** dims has quit IRC | 13:23 | |
| harmw | explain me how I should go about abstracting some ec2 module that registers some variable | 13:23 |
| inc0 | SamYaple, I wouldn't say that if we wouldn't have big sets of patches to remove start.yml inheritance | 13:23 |
| harmw | used in a playbook I call multiple times | 13:23 |
| inc0 | but I will...we do copy a lot of code | 13:24 |
| sdake | i think the remove start.yml inherentance was a good change personally ;) | 13:24 |
| harmw | I can't register with variable names | 13:24 |
| inc0 | sdake, it was, but still not very DRY;) | 13:24 |
| *** apmelton has joined #kolla | 13:24 | |
| sdake | dry is engineering 100 | 13:25 |
| sdake | magic pushbutton | 13:26 |
| sdake | now there is something to avoid ;-) | 13:26 |
| sdake | magic pushbutton is the reason heat exists | 13:26 |
| SamYaple | inc0: that was different, that was an ansible but with run_once | 13:26 |
| harmw | SamYaple: or how would I go registering something in a hashmap | 13:26 |
| harmw | you've got some good abstraction examples? | 13:27 |
| inc0 | sdake, if I wouldn't know better, I'd say you don't like heat:P | 13:27 |
| SamYaple | config.yml harmw | 13:27 |
| sdake | i like heat | 13:27 |
| SamYaple | thats decent abstraction | 13:27 |
| sdake | rht had a product called "Aeolus" | 13:27 |
| sdake | it was one big magic pushbutton | 13:28 |
| harmw | lets see, if that is usefull for what I'm after | 13:28 |
| sdake | heat was written to take the bottom half of the magic pushbutton and put it behind an api | 13:28 |
| inc0 | let me guess....Eoghan created this name? | 13:28 |
| sdake | no i dont think he was with rht when it started | 13:28 |
| sdake | this was prior to the creation of oepnstack | 13:28 |
| sdake | asalkeld came up with the name heat actually :) | 13:28 |
| sdake | aelous no idea where that came from | 13:29 |
| sdake | terribad name | 13:29 |
| inc0 | heat is pretty cool wordplay;) | 13:29 |
| sdake | heat is supposed to make the "clouds rise" | 13:29 |
| inc0 | I'm heavily into meteorology, I know how cumulus are created | 13:30 |
| inc0 | and I also know what happends where there is too much heat around;) | 13:30 |
| sdake | i later found out heat makes clouds lower ;) | 13:30 |
| inc0 | sdake, actually heated vapour goes up to so called dew point, and that's where clouds appear | 13:31 |
| inc0 | lower clouds anyway | 13:31 |
| inc0 | upper, like cirrus are bit different | 13:31 |
| sdake | my wife came up with the name kolla | 13:31 |
| inc0 | yeah, glue in greek, we spoke about that;) | 13:32 |
| sdake | originally itwas a project to glue all theshit together | 13:32 |
| sdake | so it was apprapos | 13:32 |
| sdake | now it is not so much ;) | 13:32 |
| SamYaple | inc0: you gotta rebase | 13:32 |
| SamYaple | https://review.openstack.org/#/c/215523/ | 13:32 |
| SamYaple | gerrit being dumbs | 13:32 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Build.py regex filtering https://review.openstack.org/215523 | 13:33 |
| inc0 | anyway, where is too much heat, we get cumulunimbus and storms;) thing to consider;) | 13:34 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ironic From Source https://review.openstack.org/216027 | 13:34 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Mongodb https://review.openstack.org/216025 | 13:34 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - memcached https://review.openstack.org/216024 | 13:35 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Gnocchi https://review.openstack.org/216017 | 13:35 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Murano https://review.openstack.org/216016 | 13:35 |
| SamYaple | sdake: rebase isnt giving me a 400 error anymore | 13:35 |
| SamYaple | wooo | 13:35 |
| harmw | perhaps we should rename the project to Crane, since all we do is generally just moving around containers from ship to shore :> | 13:35 |
| sdake | sometimes it does sometimes it doesn't | 13:36 |
| sdake | not renaming anything - too painful | 13:36 |
| SamYaple | sdake: it was earlier | 13:36 |
| harmw | :P | 13:36 |
| SamYaple | oh im -20000 on a name change | 13:36 |
| SamYaple | i _just_ got people to know what kolla is | 13:36 |
| sdake | SamYaple that is over 9000! | 13:36 |
| SamYaple | indeed | 13:37 |
| SamYaple | i mean it |absolutely| is | 13:37 |
| SamYaple | if you catch my drift | 13:37 |
| harmw | renames suck :) | 13:37 |
| sdake | the only time to rename a project is whn you want to bury the dead | 13:37 |
| harmw | Ill just save Crane for some other time | 13:37 |
| sdake | we have no dead to bury :) | 13:37 |
| harmw | so quantum had dead bodies to bury? | 13:38 |
| sdake | ya trademark bodies :) | 13:38 |
| harmw | ya ok | 13:39 |
| SamYaple | sdake how to set dokcer storage driver with centos? | 13:39 |
| SamYaple | /etc/default/docker ? | 13:39 |
| harmw | yep | 13:39 |
| SamYaple | cool | 13:39 |
| SamYaple | thats not sdake | 13:39 |
| sdake | so centos is busted wrt reading /etc/sysconfig/docker | 13:39 |
| harmw | docker-engine however isn't reading that file | 13:39 |
| harmw | (check the systemd file please) | 13:40 |
| sdake | you have to modify the systemd file manually | 13:40 |
| SamYaple | damn | 13:40 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add an environment variable to allow args to be set on startup. https://review.openstack.org/214799 | 13:40 |
| *** klint has quit IRC | 13:40 | |
| SamYaple | ill just start it by hand | 13:40 |
| SamYaple | screw that | 13:40 |
| harmw | SamYaple: the vagrant file modifies it iirc | 13:40 |
| sdake | you could probably just sed the systemd file ;) | 13:40 |
| harmw | there is a sed line in bootstrap.sh | 13:40 |
| SamYaple | i know i could but its the principal of the matter | 13:40 |
| SamYaple | thanks harmw ill steal that | 13:40 |
| sdake | someone should file a bugo n that | 13:41 |
| SamYaple | i dont think systemd typically reads default files though | 13:41 |
| SamYaple | i dont know that it is a 'bug' | 13:41 |
| sdake | systmd does read files - look at how sshd works | 13:41 |
| harmw | the file is shipped in the rpm | 13:41 |
| SamYaple | sdake: i dont know man | 13:41 |
| SamYaple | i used arch | 13:41 |
| harmw | but not set as environmentsomething | 13:41 |
| harmw | in systemd | 13:42 |
| harmw | the .service | 13:42 |
| SamYaple | sed -i -r 's,(ExecStart)=(.+),\1=/usr/bin/docker -d --insecure-registry operator.local:5000 --registry-mirror=http://operator.local:5000,' /usr/lib/systemd/system/docker.service | 13:42 |
| sdake | the service file doesn't read /etc/sysconfig/docker | 13:42 |
| SamYaple | that one right harmw | 13:42 |
| *** jtriley has joined #kolla | 13:42 | |
| harmw | yep | 13:42 |
| SamYaple | hot diggity | 13:42 |
| SamYaple | thanks man | 13:42 |
| sdake | sshd does read /etc/sysconfig/sshd | 13:42 |
| harmw | but don't paste the registry crap ;) | 13:42 |
| SamYaple | ill do what i want | 13:43 |
| harmw | whahevah! | 13:43 |
| SamYaple | :) | 13:43 |
| harmw | screw you guys, I'm going home | 13:43 |
| harmw | (for real) | 13:43 |
| sdake | harmw jpeeler is working on libvirt support for vagrant | 13:43 |
| harmw | :> | 13:43 |
| sdake | later harmw ;) | 13:43 |
| harmw | ah cool | 13:43 |
| harmw | yow | 13:43 |
| sdake | then i'll actually be able to use it ;-) | 13:43 |
| sdake | our heat template needs love | 13:44 |
| sdake | right now its all bulit around compose | 13:44 |
| sdake | harmw dude | 13:45 |
| sdake | review my cinder patches since everyone else is slacking! | 13:45 |
| sdake | they are 1 liners | 13:45 |
| sdake | harmw https://review.openstack.org/#/c/215975/ and all its merry children | 13:47 |
| vbel | I would like to know more kilo vs liberty information - what works well in L but fails in K. E.g. cinder, neutron ... | 13:48 |
| vbel | plugins? | 13:48 |
| vbel | I know myself that cinder 2015.1.1 does not work at all :) https://bugs.launchpad.net/cinder/+bug/1487542 | 13:49 |
| openstack | Launchpad bug 1487542 in Cinder "cinder-api 2015.1.1 fails to respond to client" [Undecided,New] | 13:49 |
| sdake | vbel can you fix up https://review.openstack.org/#/c/207925/ | 13:51 |
| *** dims_ has joined #kolla | 13:52 | |
| vbel | sdake: now I can test it :) | 13:52 |
| sdake | vbel nobody i know has tried kolla with kilo | 13:52 |
| sdake | vbel if you find out i would like to know as well | 13:52 |
| sdake | vbel please let me know re build.sh removal | 13:54 |
| sdake | what gaps you have besides git pull installation method | 13:54 |
| vbel | we test with kilo :) So far I can say keystone and glance do work, cinder does not, nova parts running but need to check with VMs | 13:54 |
| sdake | now that we have the preeamble sectino | 13:54 |
| sdake | vbel no changes needed from msater for all that? | 13:54 |
| vbel | yes, since you use kilo configs, right? | 13:55 |
| sdake | we are using master configs i think ;-) | 13:55 |
| vbel | or is there master docs? OK | 13:55 |
| sdake | there is no master docs | 13:55 |
| sdake | but if something is deprecated we mvoe to the new config option | 13:55 |
| sdake | but have no record of what those things are | 13:55 |
| vbel | so far we did not change anything but one simple option - it was for glance store | 13:56 |
| sdake | our users are going to want to run with kilo | 13:57 |
| sdake | i dont think we want to implement kilo support in our tree | 13:57 |
| sdake | but documenting how to get kilo running might be helpful | 13:57 |
| sdake | although completely unsupported in any way as just declared with the implemented part ;) | 13:57 |
| vbel | I hope it is not needed to run kilo (besides checking out release by tag) | 13:58 |
| vbel | :) | 13:58 |
| sdake | vbel can you throw your testing into an etherpad | 13:59 |
| sdake | of what you needed to do to get kilo rolling? | 13:59 |
| sdake | call it "kolla-kilo-running" | 13:59 |
| vbel | ok | 13:59 |
| sdake | I would support putting docs in the repo for it with a big warning :) | 13:59 |
| sdake | vbelnote cinder is broken in general in master | 14:00 |
| rhallisey | sdake, I don't think cinder was fully functionally in compose until liberty-1 | 14:00 |
| rhallisey | iirc | 14:00 |
| sdake | ya | 14:00 |
| sdake | well its busted in master as well as soon as somoen approves my patches ill sort out the rest of the problems ;) | 14:01 |
| rhallisey | either way that error is new | 14:01 |
| sdake | nova attach isnt working | 14:01 |
| rhallisey | I already did need another core | 14:01 |
| sdake | harmw pbourke SamYaple wtb 1 liner reviews | 14:01 |
| sdake | mandre around? | 14:02 |
| sdake | where is jpeeler vacation? | 14:02 |
| sdake | coolsvap around? | 14:02 |
| sdake | mountain of core reviewers around at this hour typically ;) | 14:02 |
| SamYaple | hey guys | 14:03 |
| SamYaple | what storage backend should the gate use | 14:03 |
| SamYaple | for docker | 14:03 |
| SamYaple | ubuntu pretty much has to use btrfs since we cant change kernels | 14:04 |
| SamYaple | and devicemapper has a bug | 14:04 |
| SamYaple | centos im good with whatever i guess | 14:04 |
| sdake | whatever is fastest | 14:05 |
| sdake | and works | 14:05 |
| sdake | lvm-devicemapper is probably the best bet | 14:05 |
| sdake | the non loopback model | 14:05 |
| SamYaple | centos has an anchient kernel, so btrfs is out there. i dont trust it | 14:05 |
| SamYaple | i can try and setup the devicemapper direct | 14:05 |
| SamYaple | i guess it might work for centos | 14:05 |
| sdake | ya devicemapper directis pretty fast | 14:05 |
| SamYaple | ubuntu it doesnt work because ubuntu has old packages | 14:05 |
| SamYaple | i cant see centos being better | 14:06 |
| sdake | right | 14:06 |
| sdake | centos is fedora 21 | 14:06 |
| sdake | with older kernel | 14:06 |
| *** mandre_ has joined #kolla | 14:06 | |
| SamYaple | so new packages? | 14:06 |
| sdake | the libraries are about the same | 14:06 |
| sdake | ya new packages | 14:06 |
| sdake | old kernel | 14:06 |
| SamYaple | its the dynamically linked libraries that trips up ubuntu | 14:06 |
| SamYaple | so centos should be fine with that then | 14:06 |
| sdake | i will let you know what storage backend works for me ass soon as I get my other two centos boxes deployed | 14:07 |
| sdake | overlayfs doesn't work on rht anyway | 14:07 |
| sdake | so thats out even if the kernel was 3.19 | 14:07 |
| SamYaple | wht do you mean? | 14:07 |
| sdake | rpm + overlayfs are incompatible in a contianer | 14:07 |
| *** mandre has quit IRC | 14:07 | |
| *** mandre_ is now known as mandre | 14:07 | |
| SamYaple | overlayfs is in the kernel as of 3.18 | 14:07 |
| SamYaple | no thatsw only for building | 14:08 |
| sdake | ya 3.18 3.19 whatevere ;) | 14:08 |
| SamYaple | deploying it works great | 14:08 |
| sdake | right only for building | 14:08 |
| SamYaple | same with ubuntu | 14:08 |
| sdake | but we build kernels | 14:08 |
| sdake | rather build containers | 14:08 |
| SamYaple | im thinking we recommend overlayfs because its just SOOOO fast for deployying | 14:08 |
| SamYaple | _and_ it works with ext4 backend | 14:08 |
| sdake | i'd be good with that but for current docs, i think before we geet to optimization recommendations we get the basics down | 14:09 |
| sdake | like getting the heat template dev environment working for example | 14:09 |
| SamYaple | agreed for sure | 14:09 |
| *** jasonamyers has joined #kolla | 14:09 | |
| sdake | 7 minutes for single node deploy from registry | 14:11 |
| sdake | wonder what multinode will be | 14:11 |
| SamYaple | if that was with fresh pulled images, about the same | 14:13 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 14:14 |
| jmccarthy | Anyone else getting 'mkdir: cannot create directory '/var/lib/rabbitmq/mnesia': Permission denied' in docker logs for rabbitmq container, with vagrant/ansible all-in-one ? | 14:14 |
| *** sdake_ has joined #kolla | 14:15 | |
| vbel | jmccarthy: are you up to date? | 14:17 |
| vbel | jmccarthy: it has been fixed in ansible bootstrap.yml for rabbitmq | 14:18 |
| jmccarthy | Erm since before lunch ? .. Ok will double check | 14:18 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 14:19 |
| *** sdake has quit IRC | 14:19 | |
| *** inc0 has quit IRC | 14:19 | |
| *** ashishb has quit IRC | 14:19 | |
| vbel | jmccarthy: kolla]$ grep "Waiting" ansible/roles/rabbitmq/tasks/bootstrap.yml ? | 14:20 |
| jmccarthy | [vagrant@operator kolla]$ grep "Waiting" ansible/roles/rabbitmq/tasks/bootstrap.yml | 14:22 |
| jmccarthy | - name: Waiting for bootstrap container | 14:22 |
| jmccarthy | [vagrant@operator kolla]$ | 14:22 |
| SamYaple | Who here is using centos source? | 14:24 |
| vbel | jmccarthy: can you remove rabbitmq_data and rabbitmq containers before running it again? | 14:24 |
| SamYaple | I need to rewrite all of the source stuff to make it compatible between both ubuntu and centos | 14:24 |
| SamYaple | i need to know if it currently works | 14:24 |
| *** sdake has joined #kolla | 14:25 | |
| *** jasonsb has quit IRC | 14:27 | |
| jmccarthy | vbel: I'm recreating at the minute with the latest to be sure | 14:27 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Prevents keystone from listening on port 80 https://review.openstack.org/216278 | 14:28 |
| *** sdake_ has quit IRC | 14:28 | |
| SamYaple | hey everyone that bug above is interfering with my horizon work | 14:29 |
| sdake | does keystone listne on port 80 on centos binary? | 14:29 |
| SamYaple | https://bugs.launchpad.net/kolla/+bug/1488121 | 14:30 |
| openstack | Launchpad bug 1488121 in kolla "apache2 is binding to port 80 for keystone" [High,In progress] - Assigned to Sam Yaple (s8m) | 14:30 |
| SamYaple | ^^ it doesnt appear so | 14:30 |
| sdake | roger | 14:30 |
| SamYaple | would be nice to have confirmation | 14:30 |
| SamYaple | i only see 80 bound with ubuntu | 14:30 |
| sdake | rhallisey ^^ please ack that patch | 14:30 |
| sdake | samyaple if it does i'll submit a patch to fix it | 14:31 |
| SamYaple | thanks | 14:31 |
| sdake | samyaple i'll find out next time I do a deploy | 14:31 |
| SamYaple | you will cause horizon will crash ;) | 14:31 |
| sdake | workingon changing the meeting times atm | 14:31 |
| vbel | we don't have port 80 for any containers | 14:31 |
| vbel | centos/ol source | 14:31 |
| SamYaple | vbel: it was a default configuration in apache2 | 14:32 |
| vbel | I mean no one listens it (just checked netstat). SamYaple: ok | 14:33 |
| vbel | I have httpd on 5000/35357 | 14:33 |
| SamYaple | awesome! thanks for confirming vbel | 14:34 |
| vbel | you are welcome :) | 14:34 |
| *** jpeeler has joined #kolla | 14:34 | |
| sdake | jpeeler can you ack https://review.openstack.org/#/c/215975/ and its children plz | 14:36 |
| *** jpeeler has quit IRC | 14:36 | |
| sdake | https://review.openstack.org/216283 | 14:39 |
| sdake | jpeeler just went back to bed ;( | 14:39 |
| sdake | yo samyaple can you PLZ approv those patches | 14:39 |
| *** zhiwei has joined #kolla | 14:41 | |
| SamYaple | WHY DOES CENTOS INSTALL IN /usr/bin/python2.7 instead of /usr/local/bin/python2.7 | 14:41 |
| *** jpeeler has joined #kolla | 14:41 | |
| *** jpeeler has joined #kolla | 14:41 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Run cinder-volume with privileged access https://review.openstack.org/215971 | 14:42 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Run cinder-volume with privileged access https://review.openstack.org/215971 | 14:42 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Run cinder-volume with privileged access https://review.openstack.org/215971 | 14:42 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Make cinder work by removing udev+lvm integration https://review.openstack.org/215975 | 14:42 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add a Cinder guide https://review.openstack.org/215976 | 14:42 |
| sdake | did those all need rebases? | 14:42 |
| * sdake beats gerrit with a stick | 14:43 | |
| SamYaple | I updated the commit i didnt realize it was a new merge | 14:43 |
| SamYaple | i should have guessed :/ | 14:43 |
| SamYaple | you had semi-colon causing the bug tracking stuff to break | 14:43 |
| *** zhiwei has left #kolla | 14:46 | |
| sdake | so our gating is running much more reliably with the retries patch | 14:46 |
| sdake | yay :) | 14:46 |
| SamYaple | cool! | 14:46 |
| SamYaple | im glad | 14:46 |
| SamYaple | im glad it was that clean too | 14:46 |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - Mongodb https://review.openstack.org/216025 | 14:47 |
| sdake | i want to see it over the next couple of days | 14:47 |
| SamYaple | i had nasty code envisioned | 14:47 |
| openstackgerrit | Merged stackforge/kolla: Build.py regex filtering https://review.openstack.org/215523 | 14:47 |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - memcached https://review.openstack.org/216024 | 14:47 |
| openstackgerrit | Merged stackforge/kolla: Ironic From Source https://review.openstack.org/216027 | 14:47 |
| SamYaple | w00t | 14:47 |
| SamYaple | regex | 14:47 |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - Gnocchi https://review.openstack.org/216017 | 14:47 |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - Gnocchi https://review.openstack.org/216017 | 14:47 |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - Murano https://review.openstack.org/216016 | 14:47 |
| SamYaple | you see that? two Ubuntu - Gnocchi | 14:47 |
| sdake | ugh | 14:47 |
| sdake | better check the repo | 14:47 |
| SamYaple | its cool looks fine | 14:48 |
| sdake | ok | 14:48 |
| sdake | vbel | 14:48 |
| SamYaple | its just a message | 14:48 |
| sdake | build gaps, besides git what else is needed now | 14:48 |
| *** diogogmt has joined #kolla | 14:49 | |
| SamYaple | sdake: would oyu mind reviewing https://review.openstack.org/#/c/215910/ | 14:51 |
| SamYaple | before it needs a rebase ;) | 14:51 |
| *** achanda has joined #kolla | 14:52 | |
| sdake | oh thought i had | 14:54 |
| sdake | will do | 14:54 |
| vbel | ? | 14:54 |
| vbel | I can add git support to build.py | 14:55 |
| vbel | will add optional git_ref to ini | 14:55 |
| vbel | and git.Git().clone .. | 14:56 |
| *** pbourke has quit IRC | 14:59 | |
| *** pbourke has joined #kolla | 14:59 | |
| *** achanda has quit IRC | 15:02 | |
| openstackgerrit | Merged stackforge/kolla: Adjust pip install for kolla-ansible https://review.openstack.org/215910 | 15:04 |
| sdake | vbel is that the only gap? | 15:05 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 15:05 |
| sdake | vbel you may have told me but i dc alot to check email and may ahve missed yur comments sorry | 15:05 |
| rhallisey | ^not sure how to completely kill all the threads, but it would be a nice addition to that patch | 15:06 |
| rhallisey | threading module says threads can't be killed :/ | 15:06 |
| vbel | sdake: no problem, we have covered pypi and proxy, may be include_footer? | 15:06 |
| sdake | vbel pypi and proxy is solved by the include header right? | 15:06 |
| vbel | sdake: and clone / checkout by tag - this I can add | 15:07 |
| vbel | right | 15:07 |
| sdake | what is requirement for include_footer? | 15:07 |
| sdake | use case? | 15:07 |
| vbel | nothing special, just clean waste :) | 15:07 |
| SamYaple | rhallisey: cant you just sys.exit()? I mean it wil ldo a full cleanup | 15:07 |
| rhallisey | SamYaple, didn't work, unless I didn't wait long enough | 15:07 |
| vbel | like unsetting those private addresses / proxies | 15:08 |
| *** diga has joined #kolla | 15:08 | |
| vbel | do we have _exit() like in C? :) | 15:08 |
| sdake | vbel so just confirming the only gap atm is include_footer and git clone/checkout | 15:09 |
| sdake | and we never had an include_footer before, so thats not really a gap | 15:09 |
| rhallisey | tried os._exit exit sys,exit | 15:09 |
| rhallisey | tried to prevent more threading from starting | 15:09 |
| SamYaple | rhallisey: its blocking on pool.join() im sure | 15:09 |
| SamYaple | rhallisey: i added a comment, with those two things it should close quickly | 15:09 |
| rhallisey | kk | 15:09 |
| SamYaple | well one thing | 15:09 |
| SamYaple | the other is just a new status | 15:09 |
| vbel | sdake: yes, that's all | 15:09 |
| bmace | is it still necessary to run as root on all the nodes in multi-node or does sudo work ok now? | 15:10 |
| SamYaple | youll need to update teh bad statuses throughout i guess | 15:10 |
| SamYaple | maybe just keep it as error | 15:10 |
| sdake | vbel are you goin to implement the git reference thing? | 15:10 |
| vbel | sdake: yes | 15:11 |
| *** jruano has joined #kolla | 15:12 | |
| sdake | vbel when is that planned for completion? | 15:15 |
| sdake | vbel will that depend on a python library of some sort? | 15:15 |
| SamYaple | there are not very many good git python libraries that ive seen | 15:15 |
| vbel | today/tomorrow. GitPython - import git | 15:15 |
| sdake | vbel license on GitPython is what? | 15:16 |
| SamYaple | sdake: https://github.com/gitpython-developers/GitPython/blob/master/LICENSE | 15:16 |
| *** chandankumar has quit IRC | 15:16 | |
| SamYaple | MIT it looks like | 15:16 |
| sdake | bsd 3 clause actually | 15:16 |
| SamYaple | well whatevs | 15:16 |
| sdake | that is a usable license for a dependency | 15:16 |
| vbel | https://github.com/gitpython-developers/GitPython/blob/master/LICENSE, yes, looks like bsd | 15:17 |
| vbel | otherwise I would go to system() :) | 15:17 |
| sdake | jinja2 is 3 clause bsd licnese an in global requirements.txt | 15:18 |
| rhallisey | sdake, re the ovs config change. I pretty sure that only that container will require this change | 15:18 |
| rhallisey | sdake, we are working around it now by building it locally | 15:19 |
| sdake | vbel what are licenses on these: https://github.com/gitpython-developers/GitPython/blob/master/requirements.txt | 15:19 |
| sdake | rhallisey leavea commentin the review pleae | 15:19 |
| rhallisey | kk | 15:19 |
| vbel | https://github.com/gitpython-developers/gitdb/blob/master/LICENSE But they have | 15:21 |
| vbel | The files at | 15:21 |
| vbel | gitdb/test/fixtures/packs/pack-11fdfa9e156ab73caae3b6da867192221f2089c2.idx | 15:21 |
| vbel | and | 15:21 |
| vbel | gitdb/test/fixtures/packs/pack-11fdfa9e156ab73caae3b6da867192221f2089c2.pack | 15:21 |
| vbel | are licensed under GNU GPL as part of the git source repository, | 15:21 |
| *** blahRus has joined #kolla | 15:21 | |
| vbel | see http://en.wikipedia.org/wiki/Git_%28software%29 for more information. | 15:21 |
| vbel | They are not required for the actual operation, which is why they are not found | 15:21 |
| vbel | in the distribution package | 15:21 |
| sdake | groan | 15:21 |
| vbel | the same bsd license | 15:22 |
| vbel | I can use simple git invocation - the code will be even smaller :) | 15:22 |
| sdake | vbel ok go ahead and proceed with the implementaiton using GitPython, if we can't get it in global requirements we may have to change to a system() based implementation | 15:22 |
| vbel | sdake: ok | 15:23 |
| sdake | I have never submitted a change for global requirements, so I don't really know what the rquirements are so to speak ;) | 15:23 |
| *** mfalatic has joined #kolla | 15:24 | |
| sdake | may mfalatic | 15:24 |
| mfalatic | hey. | 15:24 |
| rhallisey | SamYaple, how should I unblock on pool.join. Empty the queue? | 15:34 |
| *** jasonsb has joined #kolla | 15:35 | |
| *** dims__ has joined #kolla | 15:36 | |
| *** britthou_ has quit IRC | 15:36 | |
| SamYaple | rhallisey: no the command i gave you will do it | 15:37 |
| SamYaple | well, it will finish the task | 15:37 |
| SamYaple | it will still try to loop through all images | 15:37 |
| SamYaple | how long does the timeout take? | 15:37 |
| rhallisey | I'll let you know when it's done | 15:39 |
| *** dims_ has quit IRC | 15:39 | |
| *** SourabhP has joined #kolla | 15:42 | |
| *** SourabhP_ has joined #kolla | 15:44 | |
| sdake | two nodes have centos installed yay | 15:44 |
| sdake | 1 to go ;) | 15:44 |
| openstackgerrit | Merged stackforge/kolla: Prevents keystone from listening on port 80 https://review.openstack.org/216278 | 15:45 |
| sdake | samyaple how much longer ar eyou going to be around today? | 15:45 |
| SamYaple | dude i dont know | 15:45 |
| SamYaple | im still at work | 15:45 |
| sdake | more then 2 hours? | 15:45 |
| SamYaple | maybe | 15:45 |
| SamYaple | probably not | 15:46 |
| *** diga has quit IRC | 15:46 | |
| *** unicell1 has joined #kolla | 15:46 | |
| *** SourabhP has quit IRC | 15:47 | |
| *** SourabhP_ is now known as SourabhP | 15:47 | |
| SamYaple | rhallisey: you know you can hard kill that script right? | 15:47 |
| *** unicell has quit IRC | 15:47 | |
| SamYaple | if ^C doesnt work ^\ will for sure | 15:47 |
| rhallisey | SamYaple, I know | 15:47 |
| SamYaple | ok just making sure | 15:47 |
| rhallisey | :) I figured it would be nicer if it failed on it's owne | 15:48 |
| rhallisey | but it doesn't want to | 15:48 |
| rhallisey | I'm fine with just a better error message | 15:48 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 15:49 |
| SamYaple | with your patch rhallisey it will _never_ fail | 15:49 |
| SamYaple | well the first one | 15:49 |
| rhallisey | ya lol | 15:49 |
| SamYaple | it will block forever | 15:49 |
| SamYaple | thats why you have to do the task_done thing | 15:49 |
| rhallisey | it still never fails, but atleast the message gets out | 15:50 |
| rhallisey | I'm cool with that for now | 15:50 |
| rhallisey | better that than have to wait 15min | 15:50 |
| rhallisey | btw it's not me that came up with add this. I saw someone wanted to add a note in the doc to check this | 15:51 |
| SamYaple | that regex patch is working great | 15:51 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Run cinder-volume with privileged access https://review.openstack.org/215971 | 16:00 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Make cinder work by removing udev+lvm integration https://review.openstack.org/215975 | 16:00 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add a Cinder guide https://review.openstack.org/215976 | 16:00 |
| * sdake hates computers | 16:00 | |
| *** chandankumar has joined #kolla | 16:01 | |
| SamYaple | indeed | 16:01 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 16:03 |
| *** thumpba has joined #kolla | 16:12 | |
| *** dims__ has quit IRC | 16:12 | |
| *** dims has joined #kolla | 16:13 | |
| *** dims is now known as dims__ | 16:13 | |
| *** SourabhP has quit IRC | 16:15 | |
| *** unicell1 has quit IRC | 16:18 | |
| *** sdake_ has joined #kolla | 16:33 | |
| *** sdake has quit IRC | 16:37 | |
| *** absubram has joined #kolla | 16:42 | |
| *** dims__ has quit IRC | 16:43 | |
| *** dims__ has joined #kolla | 16:45 | |
| openstackgerrit | venkatamahesh proposed stackforge/kolla: Extension .md is changed to .rst https://review.openstack.org/216333 | 16:47 |
| *** jtriley has quit IRC | 16:50 | |
| openstackgerrit | Suhail Syed proposed stackforge/kolla: Make Kolla installable via PIP https://review.openstack.org/216341 | 16:57 |
| *** jtriley has joined #kolla | 16:57 | |
| *** unicell has joined #kolla | 17:02 | |
| *** sdake has joined #kolla | 17:02 | |
| *** sdake_ has quit IRC | 17:06 | |
| openstackgerrit | Merged stackforge/kolla: Run cinder-volume with privileged access https://review.openstack.org/215971 | 17:06 |
| openstackgerrit | Merged stackforge/kolla: Make cinder work by removing udev+lvm integration https://review.openstack.org/215975 | 17:06 |
| openstackgerrit | Merged stackforge/kolla: Add a Cinder guide https://review.openstack.org/215976 | 17:06 |
| *** britthouser has joined #kolla | 17:07 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 17:14 |
| sdake | SamYaple that install patch looks pretty swet | 17:15 |
| sdake | missing build.py install tho | 17:15 |
| SamYaple | sdake: yea | 17:16 |
| sdake | think simply copyign build.y insto /usr/bin not ideal | 17:16 |
| sdake | i think when i did this work on conductor i used a console script thingy | 17:17 |
| SamYaple | well it would be /usr/bin/kolla-build or something like that | 17:17 |
| SamYaple | it wouldnt be build.py | 17:17 |
| sdake | could you add a comment in the review to let suhail know what it shoud be named | 17:18 |
| SamYaple | im fried atm | 17:18 |
| SamYaple | im around for like 10 min | 17:18 |
| sdake | ok nm then :) | 17:18 |
| SamYaple | i can do so tomorrow if neede | 17:19 |
| *** akwasnie has quit IRC | 17:19 | |
| sdake | samyaple cna you gvie me a quick update on the statue of horizon | 17:19 |
| SamYaple | ubuntu and centos with the exact same pip install is dumb :( | 17:19 |
| SamYaple | thats your update | 17:19 |
| SamYaple | and centos horizon source doesnt seem to build | 17:19 |
| sdake | huh | 17:19 |
| SamYaple | i havent checked master | 17:19 |
| sdake | so the ansible code is workignbut not sure on the containers? | 17:20 |
| SamYaple | the ansible _and_ the container worked for ubuntu at one point, but that wouldnt work for centos due to path issues | 17:20 |
| SamYaple | some im back to the base trying to get the pip to install the same on ubuntu and centos | 17:20 |
| sdake | can you push your latest work in cae i get a chance to look at it today | 17:22 |
| sdake | i've got like 12 hours ahead of me of work day | 17:22 |
| SamYaple | yea but the issue is going to be handled in the base since it affects all source installs | 17:23 |
| SamYaple | so thats not going to help centos much | 17:23 |
| sdake | yes but if you push your work i can possibly get centos binary installs working | 17:26 |
| sdake | which is really what i'm after | 17:26 |
| SamYaple | the horizon work is up | 17:26 |
| sdake | cool | 17:26 |
| SamYaple | its specific to ubuntu though | 17:26 |
| sdake | the ansible parts are specific to ubuntu? | 17:27 |
| SamYaple | one of the file copies over has paths in it | 17:27 |
| sdake | that could be changed with some kind of conditional right? | 17:29 |
| sdake | rather enhanced | 17:29 |
| SamYaple | yea go for it | 17:29 |
| SamYaple | i was just going to make it all match | 17:29 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 17:30 |
| sdake | and symlink it or something? | 17:30 |
| SamYaple | sdake: no python seems to be stomping my symlinks | 17:31 |
| *** achanda has joined #kolla | 17:33 | |
| bmace | hey SamYaple, looks like the rabbitmq_user is defined in both rabbitmq/defaults/main.yml and group_vars/all.yml. i believe the intent is to have it defined in only one place, and overridden in globals.yml when needed. | 17:43 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 17:46 |
| *** loth has joined #kolla | 17:49 | |
| SamYaple | bmace: that sounds right. since it is referenced in many locations it should be removed from rabbitmq/defaults/main.yml | 17:53 |
| bmace | yup, that is what i thought. | 17:53 |
| *** jasonsb_ has joined #kolla | 17:54 | |
| * sdake predicts a patch from bmace coming ;) | 17:54 | |
| *** jasonsb has quit IRC | 17:54 | |
| *** vinkman has joined #kolla | 17:54 | |
| SamYaple | sdake i have a horrible patch coming in you need to approve :) | 17:57 |
| sdake | depends what it is :) | 17:58 |
| SamYaple | a long ass sed command | 17:58 |
| openstackgerrit | Merged stackforge/kolla: Extension .md is changed to .rst https://review.openstack.org/216333 | 18:05 |
| SamYaple | ahhhh | 18:08 |
| SamYaple | for all you image builders out there.... | 18:08 |
| SamYaple | so it turns out you cant retag an image that you are using | 18:09 |
| SamYaple | so building a new image wont... actually... set.. the. tag. | 18:09 |
| SamYaple | assuming you also have your test cluster up and running | 18:09 |
| *** ashishb has joined #kolla | 18:09 | |
| *** sdake_ has joined #kolla | 18:14 | |
| *** sdake has quit IRC | 18:17 | |
| SamYaple | YEA | 18:18 |
| SamYaple | SUCCESS | 18:18 |
| *** sdake has joined #kolla | 18:18 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add ansible support for Horizon https://review.openstack.org/211286 | 18:20 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Horizon https://review.openstack.org/216026 | 18:20 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Sync directories https://review.openstack.org/216369 | 18:20 |
| SamYaple | sdake: ^^^ working | 18:20 |
| *** sdake_ has quit IRC | 18:21 | |
| *** jtriley has quit IRC | 18:22 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Ubuntu - Haproxy https://review.openstack.org/215924 | 18:24 |
| *** jtriley has joined #kolla | 18:28 | |
| *** jasonsb_ has quit IRC | 18:31 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 18:33 |
| SamYaple | hey is there a way to change teh patchset owner so it shows up in my queue? | 18:34 |
| SamYaple | im loosing patches that i am suppose to be working on | 18:34 |
| *** inc0 has joined #kolla | 18:35 | |
| inc0 | good evening | 18:35 |
| SamYaple | hey inc0 | 18:35 |
| SamYaple | regex patch is working awesome! | 18:35 |
| SamYaple | question though | 18:36 |
| SamYaple | i cant seem to use regex with it like ^horizon-base$ | 18:36 |
| *** absubram has quit IRC | 18:36 | |
| openstackgerrit | Merged stackforge/kolla: Ubuntu - Nova https://review.openstack.org/215912 | 18:42 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 18:44 |
| rhallisey | see if you agree with that one SamYaple | 18:44 |
| SamYaple | rhallisey: no because that is checking the parent | 18:46 |
| SamYaple | no need to log 'is docker running' there | 18:46 |
| rhallisey | the problem is logging that it will fail on the first run | 18:46 |
| rhallisey | and probably never been seen | 18:47 |
| SamYaple | what do you mean? | 18:47 |
| SamYaple | oh i follow | 18:47 |
| SamYaple | no wait i dont | 18:47 |
| rhallisey | because it will always fail on the first attempt, it will fail on all 73 images after | 18:47 |
| SamYaple | this breaks the retry loop | 18:47 |
| SamYaple | yea but fail with parent_error | 18:47 |
| rhallisey | and get lost in a load of log info | 18:48 |
| SamYaple | which is correct | 18:48 |
| SamYaple | thats why the connection_error status | 18:48 |
| SamYaple | logging that message there is not correct at all | 18:48 |
| SamYaple | ever | 18:48 |
| SamYaple | that would even print on the image that _did_ have teh connection error | 18:48 |
| SamYaple | wont* | 18:49 |
| rhallisey | well technically they all would have it | 18:49 |
| rhallisey | I can move it back up | 18:49 |
| SamYaple | not the one that had the error | 18:49 |
| SamYaple | it wont have it | 18:49 |
| SamYaple | it would have moved past that section | 18:49 |
| SamYaple | that section is only for catching that the parent had an error | 18:50 |
| SamYaple | ad breaking early | 18:50 |
| rhallisey | well I'm fine with catching the parent error as long as it's obvious that something broke | 18:50 |
| rhallisey | that was the point | 18:50 |
| rhallisey | it just won't be obvious until you scroll up a mile | 18:51 |
| SamYaple | can you print out the parent error code then? | 18:51 |
| SamYaple | parent erred with status 'connection_error' or 'error' | 18:51 |
| rhallisey | ok | 18:51 |
| SamYaple | that would print out a bunch of connection_error in the log if you do that | 18:51 |
| rhallisey | wfm | 18:52 |
| SamYaple | so you can use the iamge['parent']['status'] as a variable there | 18:52 |
| SamYaple | hey we cant be turning of udev_sync like that in cinder | 18:55 |
| SamYaple | thats really really really bad | 18:55 |
| SamYaple | the major and minor numbers of the device nodes can get out of sync | 18:55 |
| SamYaple | you end up reading and writing to the wrong block devices | 18:55 |
| SamYaple | thats a really big bug and security whole right now | 18:56 |
| *** jasonsb has joined #kolla | 18:59 | |
| *** tummy has joined #kolla | 18:59 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 19:07 |
| rhallisey | SamYaple, I'm happy with that. That's a little more friendly | 19:08 |
| SamYaple | rhallisey: yea but you still have the duplicate elif block ;) | 19:09 |
| rhallisey | :/ I'm all tangled up | 19:10 |
| rhallisey | SamYaple, I split it out to 3 to catch each error | 19:10 |
| rhallisey | need to recompress | 19:10 |
| SamYaple | ok. i only care about the if block, but past that im good :) | 19:11 |
| SamYaple | will be really truthfully sleeping now so ill ack when i get to work tomorrow | 19:11 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 19:12 |
| *** ashishb has quit IRC | 19:12 | |
| inc0 | rhallisey, shame on you | 19:17 |
| rhallisey | eh? | 19:17 |
| inc0 | not running flake8 over your code | 19:17 |
| rhallisey | what failed? | 19:19 |
| inc0 | I've seen indents not being mutliple of 4 | 19:20 |
| inc0 | it's still here, but after \ | 19:21 |
| inc0 | also, but that's problem of current file as well, imports are in incorrect order | 19:21 |
| rhallisey | I thought I addressed that | 19:21 |
| inc0 | I;ve been looking at older patchset now, but I'll nitpick another one:P | 19:22 |
| rhallisey | oh oops | 19:22 |
| rhallisey | no I didn't | 19:22 |
| inc0 | SamYaple, https://review.openstack.org/#/c/216026/10/docker/common/horizon/config-external.sh does this even work on centos? | 19:27 |
| inc0 | ah it might, carry on | 19:27 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 19:28 |
| rhallisey | inc0, thanks there we go | 19:28 |
| inc0 | you've commited pycs now;) | 19:29 |
| inc0 | line 96 is still off-by-one from multiple of 4 | 19:30 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix regression in build-script https://review.openstack.org/216395 | 19:32 |
| rhallisey | inc0, are you sure I have it at 12 | 19:33 |
| rhallisey | I guess so | 19:34 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 19:36 |
| inc0 | well image is under image so it's not that bad | 19:37 |
| inc0 | I just have different style of indenting this stuff | 19:38 |
| inc0 | still, we need to fix imports. they're just wrong | 19:38 |
| rhallisey | just tell me where to put it because I don't see where it goes | 19:40 |
| inc0 | well, requests is 3rd party library | 19:40 |
| inc0 | so order should be - import stdlib, from stdilib import ... , import 3rd party, from 3rd import ... | 19:40 |
| inc0 | just put it there and I'll fix it when it merges | 19:41 |
| *** sdake_ has joined #kolla | 19:58 | |
| *** jasonamyers has quit IRC | 19:59 | |
| *** sdake has quit IRC | 20:01 | |
| *** achanda has quit IRC | 20:07 | |
| *** inc0 has quit IRC | 20:07 | |
| *** chandankumar has quit IRC | 20:08 | |
| *** absubram has joined #kolla | 20:21 | |
| sdake_ | rhallisey what precisely is the problem with turning udev off? | 20:27 |
| rhallisey | sdake_, I don't know of an issue other what sam mentioned | 20:31 |
| *** sdake_ is now known as sdake | 20:34 | |
| sdake | i sort of missed it | 20:34 |
| sdake | he was tlaking about what precisely? | 20:34 |
| rhallisey | http://fpaste.org/258688/48487144/ | 20:34 |
| *** jasonsb_ has joined #kolla | 20:36 | |
| *** jasonsb has quit IRC | 20:36 | |
| *** tummy has quit IRC | 20:38 | |
| *** achanda has joined #kolla | 20:40 | |
| *** achanda has quit IRC | 20:41 | |
| sdake | the kernel manages major/minors not udev | 20:41 |
| *** achanda has joined #kolla | 20:41 | |
| *** daneyon has joined #kolla | 20:42 | |
| *** sdake has quit IRC | 20:44 | |
| *** daneyon has quit IRC | 20:47 | |
| *** daneyon has joined #kolla | 20:47 | |
| *** shardy has quit IRC | 20:50 | |
| *** daneyon_ has joined #kolla | 20:55 | |
| *** daneyon has quit IRC | 20:58 | |
| *** unicell has quit IRC | 21:04 | |
| *** unicell1 has joined #kolla | 21:04 | |
| *** absubram has quit IRC | 21:23 | |
| *** rhallisey has quit IRC | 21:28 | |
| *** jtriley has quit IRC | 21:31 | |
| *** jruano has quit IRC | 21:45 | |
| *** Slower has quit IRC | 22:00 | |
| *** jruano has joined #kolla | 22:24 | |
| *** sdake has joined #kolla | 22:29 | |
| *** jasonsb_ has quit IRC | 22:40 | |
| *** thumpba has quit IRC | 22:42 | |
| *** diogogmt has quit IRC | 23:03 | |
| *** unicell1 has quit IRC | 23:15 | |
| *** dims_ has joined #kolla | 23:16 | |
| *** dims__ has quit IRC | 23:18 | |
| *** blahRus has quit IRC | 23:19 | |
| *** dims__ has joined #kolla | 23:19 | |
| *** unicell has joined #kolla | 23:21 | |
| *** dims_ has quit IRC | 23:21 | |
| *** achanda_ has joined #kolla | 23:31 | |
| *** achanda has quit IRC | 23:34 | |
| *** achanda_ has quit IRC | 23:35 | |
| openstackgerrit | Suhail Syed proposed stackforge/kolla: Make Kolla installable via PIP https://review.openstack.org/216341 | 23:39 |
| openstackgerrit | Suhail Syed proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 23:39 |
| *** diogogmt has joined #kolla | 23:41 | |
| openstackgerrit | Suhail Syed proposed stackforge/kolla: Make Kolla installable via PIP https://review.openstack.org/216341 | 23:41 |
| openstackgerrit | Suhail Syed proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 23:41 |
| *** jtriley has joined #kolla | 23:46 | |
| *** alisonh has quit IRC | 23:46 | |
| sdake | harmw ping | 23:47 |
| *** sdake_ has joined #kolla | 23:51 | |
| *** alisonh has joined #kolla | 23:52 | |
| *** sdake has quit IRC | 23:55 | |
| *** jasonsb has joined #kolla | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!