| srwilkers | alright. im going to take it easy the remainder of the evening. ill chat with everyone tomorrow \o | 00:01 |
|---|---|---|
| *** srwilkers is now known as srwilkers_zzzz | 00:01 | |
| SamYaple | mattmceuen: what task is erroring? | 00:04 |
| mattmceuen | SamYaple: TASK [nova : Copying over nova.conf] | 00:05 |
| mattmceuen | ******************************************* | 00:05 |
| mattmceuen | fatal: [192.168.1.3]: FAILED! => {"failed": true, "msg": "{{ hostvars[inventory_hostname]['ansible_' + api_interface]['ipv4']['address'] if orchestration_engine == 'ANSIBLE' else '0.0.0.0' }}: 'dict object' has no attribute u'ansible_eno1'"} | 00:05 |
| mattmceuen | SamYaple would I need to /remove/ the default from globals.yml, and specify consistently per-host in the inventory? | 00:06 |
| SamYaple | mattmceuen: its been a while since ive done what you are doing, but im pretty sure you cant mix and match | 00:06 |
| SamYaple | you must remove from globals.yml and specify the options for _every_ host in inventory | 00:06 |
| mattmceuen | Thanks - lemme try w/ consistency. | 00:07 |
| SamYaple | mattmceuen: if I recall, i think the way we load that globals.yml thing gives it absolute athority, even overriding inventory variables | 00:08 |
| SamYaple | thats the reason | 00:08 |
| sbezverk | srwilkers: have a good one! | 00:08 |
| *** yuanying has quit IRC | 00:08 | |
| *** Pavo has quit IRC | 00:14 | |
| *** Pavo has joined #openstack-kolla | 00:18 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helm-entrypoint git hooks https://review.openstack.org/412238 | 00:20 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 00:20 |
| mattmceuen | You rock SamYaple, that did the trick. My heterogenous node got added into my cluster like a champ :) thanks! | 00:23 |
| SamYaple | mattmceuen: awesome. nice to see all the work on the project come together like that for you | 00:33 |
| *** fragatina has joined #openstack-kolla | 00:40 | |
| kfox1111 | starting to see a trend in the loopback gate failures. it so far seem to be always rax. | 00:40 |
| *** yingjun has joined #openstack-kolla | 00:41 | |
| SamYaple | dont look at me. i only work there | 00:42 |
| kfox1111 | SamYaple: not blaming anyone. just trying to figure out why it fails sometimes. something's different in the rax vm's. | 00:45 |
| kfox1111 | maybe an ephemeral partition or something.... | 00:45 |
| SamYaple | kfox1111: rax uses xen | 00:45 |
| kfox1111 | ah. | 00:45 |
| kfox1111 | I'm seeing this: | 00:46 |
| *** yingjun has quit IRC | 00:46 | |
| kfox1111 | http://logs.openstack.org/38/412238/3/check/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/3bb65f0/console.html#_2016-12-19_00_27_30_310407 | 00:46 |
| SamYaple | looking | 00:46 |
| SamYaple | ah. we probably arent swapoffing or something | 00:46 |
| kfox1111 | gota rerun partprobe to pickup the loopback partitions for ceph. | 00:46 |
| SamYaple | there is a bunch of debug in there from last time i had to fix this. ill check it | 00:46 |
| kfox1111 | and its failing. | 00:46 |
| *** sayantan_ has joined #openstack-kolla | 00:47 | |
| kfox1111 | this is the kolla-kubernetes gate so different set of logs probably. | 00:47 |
| SamYaple | oh well thats your own fault | 00:47 |
| kfox1111 | ;) | 00:48 |
| SamYaple | put in my debug scripts! | 00:48 |
| SamYaple | did you write your own gate too? | 00:48 |
| kfox1111 | hehe. | 00:48 |
| kfox1111 | yeah. | 00:48 |
| SamYaple | ugh. why | 00:48 |
| kfox1111 | was tired of having to test everything manually on my machine. :) | 00:48 |
| SamYaple | well im not helping then | 00:49 |
| SamYaple | it works for kolla | 00:49 |
| SamYaple | go copy what i did there :P | 00:49 |
| kfox1111 | looks like maybe all I need to do is specify the loopback devices to partprobe. | 00:49 |
| kfox1111 | rather then have it rescan everything. | 00:49 |
| SamYaple | if you cant run partprobe, you are doing something wrong anyway | 00:50 |
| SamYaple | but you should include dump_info.sh from kolla repo | 00:50 |
| kfox1111 | the partprobe is failing on rax. | 00:50 |
| kfox1111 | so its a problem with rax. | 00:50 |
| kfox1111 | kolla-ansible? | 00:50 |
| SamYaple | all the repos | 00:51 |
| SamYaple | except weird ol' kolla-k8s | 00:51 |
| kfox1111 | I can add that stfuf to https://github.com/openstack/kolla-kubernetes/blob/master/tests/bin/gate_capture_logs.sh :) | 00:52 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 00:52 |
| kfox1111 | a lot o logging hapening already. :) | 00:52 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: RAX partprobe fix https://review.openstack.org/412244 | 00:55 |
| kfox1111 | ok... lets see if that'll fix it. | 00:55 |
| SamYaple | ugh why would you rewrite al lthe gate stuff | 00:56 |
| SamYaple | it was enven broken into easy functions too. you could have just grabbed a few functions | 00:56 |
| Jeffrey4l | SamYaple, yep | 00:57 |
| kfox1111 | SamYaple: no one was around that understood the other gate. | 00:57 |
| kfox1111 | at least that I found. | 00:57 |
| kfox1111 | and its mostly just setting up k8s and then driving k8s. | 00:58 |
| kfox1111 | also has multinode, which no one knew how to do. had to reverse engineer some other projects jobs. | 00:58 |
| SamYaple | kfox1111: its literally broken into functions! | 00:59 |
| SamYaple | kfox1111: im giving you a hard time. but really though, it has functions. itll save you headaches | 00:59 |
| SamYaple | as for multinode, that works in kolla gate, its mostly setup. but was waiting on zuulv3, which i believe is in place now | 01:00 |
| SamYaple | Jeffrey4l: btw. big ol' security hole | 01:00 |
| kfox1111 | SamYaple: probably should work at making them into a library more easily shared. | 01:00 |
| kfox1111 | then try to share it between all 3. | 01:00 |
| SamYaple | Jeffrey4l: https://review.openstack.org/#/c/412235/ | 01:00 |
| openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 01:00 |
| kfox1111 | SamYaple: no zuulv3 yet. infra says soon they hope. | 01:00 |
| SamYaple | oh | 01:00 |
| Jeffrey4l | SamYaple, why we need v3? devstack multnode works. | 01:00 |
| SamYaple | kfox1111: re: library, i mean, they are functions | 01:00 |
| kfox1111 | SamYaple: yeah. but I can't just include a file and then use the functions. | 01:01 |
| SamYaple | Jeffrey4l: its not a requirement, but the way we are currently testing (with networking) we would need to have zuulv3 to keep doing it that way | 01:01 |
| kfox1111 | yeah. 3+ nodes should be easier with zuulv3, or so I've been told. | 01:02 |
| kfox1111 | but they did say it can be doable now. which they didn't say a few months ago. | 01:02 |
| *** zhubingbing has joined #openstack-kolla | 01:02 | |
| Jeffrey4l | 3+ nodes is impossible. | 01:02 |
| kfox1111 | so, the dns fix seems to be holding... removing those notes. | 01:02 |
| kfox1111 | mostly down to the loopback issue... | 01:03 |
| Jeffrey4l | SamYaple, still why? devstack multi node works. why kolla not? afaik, kolla should work in 2-vm node ( network ) | 01:03 |
| kfox1111 | Jeffrey4l: kolla-kubernetes does. | 01:04 |
| Jeffrey4l | kfox1111, cool | 01:04 |
| SamYaple | Jeffrey4l: i dont want to go into it, we can make it work without zuulv3, but i dont feel like spending my time on it and no one else seems to understand it | 01:04 |
| SamYaple | with zuulv3 it will be really really easy | 01:04 |
| *** yuanying has joined #openstack-kolla | 01:04 | |
| kfox1111 | Jeffrey4l: have ceph/k8s across 2 nodes deploying a whole openstack, laucnhing vm's testing network, and moving a cinder volume from vm to vm with writes. | 01:04 |
| SamYaple | so im personally jsut waiting | 01:04 |
| Jeffrey4l | roger. i will try to continue the 2-vms gate based on this doc https://github.com/openstack-infra/devstack-gate/blob/master/multinode_setup_info.txt | 01:05 |
| kfox1111 | Jeffrey4l: if you want an example, kolla-kubernetes one has been working quite well for a while now. | 01:05 |
| Jeffrey4l | kfox1111, i will check it . thanks. | 01:06 |
| kfox1111 | np. let me know if you have any questions. :) | 01:06 |
| SamYaple | Jeffrey4l: 90% of the work is already done | 01:07 |
| Jeffrey4l | no, right now. i have some idea about the multi node test. ;) | 01:07 |
| SamYaple | it already sets up docker on multiple hosts | 01:07 |
| SamYaple | it can easily deploy to multiple hosts, but we dont have an overlay network between the two | 01:07 |
| Jeffrey4l | SamYaple, yep. the only need to do is: 1. configration docker registry on multi node 2. configure inventory for multi node 3. some network trick. | 01:08 |
| *** mattmceuen is now known as mattmceuen_away | 01:08 | |
| Jeffrey4l | SamYaple, we can use vxlan. | 01:08 |
| SamYaple | no you cant Jeffrey4l. not until zuulv3 ;) | 01:08 |
| kfox1111 | SamYaple: should be able to use vxlan's manually. | 01:09 |
| SamYaple | yea but then openvswitch and other stuff | 01:09 |
| kfox1111 | right. | 01:09 |
| SamYaple | like i said, its _doable_ but not worth my time. v3 makes it easy | 01:09 |
| SamYaple | the real thing v3 adds is the ability to reconnect | 01:09 |
| Jeffrey4l | SamYaple, we really can. i can pint the vm in one vm test. through floating-ip -> vxlan -> fixed-ip. | 01:09 |
| SamYaple | so i can move the ip from the raw interface to an openvswitch bridge | 01:09 |
| Jeffrey4l | SamYaple, we can add the ip into br-ex ;) | 01:10 |
| *** zhubingbing has quit IRC | 01:10 | |
| SamYaple | Jeffrey4l: feel free to ignore me and see it not work for you | 01:10 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 01:10 |
| SamYaple | because, you know, it wont/ because of zuulv2 | 01:10 |
| Jeffrey4l | https://review.openstack.org/402122 | 01:10 |
| Jeffrey4l | enabled tempest on gate. add a ip into br-ex directly :) | 01:11 |
| *** zhubingbing has joined #openstack-kolla | 01:11 | |
| kfox1111 | bbiab. | 01:11 |
| SamYaple | Jeffrey4l: thats not what im talking about. that isn't multinode | 01:11 |
| SamYaple | Jeffrey4l: you can't add the interface on the host into br-ex | 01:12 |
| SamYaple | anyway i dont care about this | 01:12 |
| Jeffrey4l | ok | 01:12 |
| SamYaple | when zuulv3 lands its like 5 lines of code to make this work | 01:12 |
| SamYaple | without v3 its ugly and all kinds of manual vxlan stuff i dont want to do | 01:12 |
| SamYaple | youre welcome to. but i think there are more important things to do | 01:12 |
| Jeffrey4l | roger. | 01:13 |
| SamYaple | the uid patch is coming along though Jeffrey4l. give it a look | 01:14 |
| Jeffrey4l | cool. | 01:14 |
| *** qiliang27 has joined #openstack-kolla | 01:18 | |
| *** saneax-_-|AFK is now known as saneax | 01:18 | |
| openstackgerrit | Zhigang Li proposed openstack/kolla-kubernetes: Replace basestring with six.string_types https://review.openstack.org/401144 | 01:21 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] static uid/gid in Dockerfiles https://review.openstack.org/412231 | 01:27 |
| *** duonghq has joined #openstack-kolla | 01:28 | |
| *** tovin07 has joined #openstack-kolla | 01:32 | |
| *** Pavo has quit IRC | 01:33 | |
| *** jianyi has joined #openstack-kolla | 01:36 | |
| *** yingjun has joined #openstack-kolla | 01:37 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Enable solum dashboard when enable_solum is yes https://review.openstack.org/408985 | 01:43 |
| openstackgerrit | Li Yingjun proposed openstack/kolla-kubernetes: Fix different repo issue for documentation https://review.openstack.org/405054 | 01:44 |
| duonghq | morning | 01:49 |
| duonghq | portdirect, around? | 01:50 |
| *** wcb has quit IRC | 01:53 | |
| *** williamcaban has joined #openstack-kolla | 01:55 | |
| *** jimmygc has joined #openstack-kolla | 01:56 | |
| *** Pavo has joined #openstack-kolla | 01:56 | |
| *** williamcaban has quit IRC | 01:59 | |
| *** jianyi has quit IRC | 02:01 | |
| *** Pavo has quit IRC | 02:14 | |
| *** fragatina has quit IRC | 02:17 | |
| *** Pavo has joined #openstack-kolla | 02:19 | |
| *** Pavo_ has joined #openstack-kolla | 02:23 | |
| openstackgerrit | zhangshuai proposed openstack/kolla-ansible: Add karbor ansible role https://review.openstack.org/402229 | 02:24 |
| jimmygc | Good morning, everyone. I am new to this channel. I uploaded a review on (https://review.openstack.org/#/c/411154/). It will be great if you give it a recheck or review. Thanks:) | 02:27 |
| *** yingjun has quit IRC | 02:28 | |
| Pavo_ | grrrr | 02:28 |
| Pavo_ | this cisco router is pissing me off | 02:28 |
| SamYaple | jimmygc: i gave it a review. Have you tested the sed? | 02:33 |
| *** sayantan_ has quit IRC | 02:37 | |
| *** imcsk8 has quit IRC | 02:40 | |
| *** imcsk8 has joined #openstack-kolla | 02:40 | |
| *** yingjun has joined #openstack-kolla | 02:41 | |
| *** papacz has quit IRC | 02:44 | |
| *** williamcaban has joined #openstack-kolla | 02:45 | |
| *** saneax is now known as saneax-_-|AFK | 02:45 | |
| jimmygc | yes, I tested inside docker container | 02:45 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Switch Ubuntu to overlay2 docker driver https://review.openstack.org/412255 | 02:47 |
| SamYaple | jimmygc: i posted a comment. its not the same. slightly off. please heck again | 02:47 |
| *** williamcaban has quit IRC | 02:50 | |
| jimmygc | OK, checking them now :) | 02:51 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Switch Ubuntu to overlay2 docker driver https://review.openstack.org/412255 | 02:54 |
| *** yingjun has quit IRC | 02:56 | |
| *** yuanying_ has joined #openstack-kolla | 02:57 | |
| *** saneax-_-|AFK is now known as saneax | 02:58 | |
| *** sayantan_ has joined #openstack-kolla | 02:58 | |
| *** yuanying has quit IRC | 03:00 | |
| *** rafael_sisweb has quit IRC | 03:00 | |
| *** rafael_sisweb has joined #openstack-kolla | 03:00 | |
| *** williamcaban has joined #openstack-kolla | 03:01 | |
| *** sayantan_ has quit IRC | 03:02 | |
| *** zhubingbing has quit IRC | 03:02 | |
| *** zhubingbing has joined #openstack-kolla | 03:03 | |
| *** Jeffrey4l has quit IRC | 03:05 | |
| *** Jeffrey4l has joined #openstack-kolla | 03:05 | |
| *** zhubingbing_ has joined #openstack-kolla | 03:07 | |
| zhubingbing_ | hello guys | 03:07 |
| *** zhubingbing has quit IRC | 03:08 | |
| zhubingbing_ | ;) | 03:08 |
| *** ediardo has quit IRC | 03:09 | |
| *** saneax is now known as saneax-_-|AFK | 03:09 | |
| *** yingjun has joined #openstack-kolla | 03:09 | |
| *** ediardo has joined #openstack-kolla | 03:11 | |
| *** yuanying_ has quit IRC | 03:13 | |
| *** yuanying has joined #openstack-kolla | 03:13 | |
| *** lrensing has quit IRC | 03:13 | |
| *** gomarivera has joined #openstack-kolla | 03:15 | |
| *** yuanying has quit IRC | 03:17 | |
| openstackgerrit | jimmygc proposed openstack/kolla: Readability improvement to base image https://review.openstack.org/411154 | 03:19 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Enable murano dashboard when enable_murano is yes https://review.openstack.org/406476 | 03:21 |
| *** Pavo_ has quit IRC | 03:22 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Enable murano dashboard when enable_murano is yes https://review.openstack.org/406476 | 03:24 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 03:25 |
| *** alanmeadows has quit IRC | 03:25 | |
| kfox1111 | sbezverk: still around? | 03:25 |
| kfox1111 | sbezverk: https://review.openstack.org/#/c/412238/ is ready. | 03:25 |
| *** alanmeadows has joined #openstack-kolla | 03:26 | |
| kfox1111 | hey alan | 03:26 |
| *** Pavo has quit IRC | 03:31 | |
| *** yingjun has quit IRC | 03:31 | |
| openstackgerrit | jimmygc proposed openstack/kolla: Readability improvement to base image https://review.openstack.org/411154 | 03:35 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Remove the unnecessary space https://review.openstack.org/412268 | 03:35 |
| *** williamcaban has quit IRC | 03:39 | |
| *** zhangqiankun has quit IRC | 03:43 | |
| *** zhangqiankun has joined #openstack-kolla | 03:43 | |
| jimmygc | SamYaple: Done, suggestion accepted :) | 03:43 |
| *** tovin07_ has joined #openstack-kolla | 03:48 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: Helm-entrypoint git hooks https://review.openstack.org/412238 | 03:50 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 03:52 |
| *** williamcaban has joined #openstack-kolla | 03:55 | |
| *** sdake has joined #openstack-kolla | 03:56 | |
| sdake | yo peeps | 03:57 |
| *** williamcaban has quit IRC | 03:59 | |
| *** coolsvap has joined #openstack-kolla | 03:59 | |
| duonghq | sup sdake | 04:01 |
| sdake | watching tv | 04:01 |
| sdake | putting up christmas lights | 04:01 |
| sdake | my daughter harranged me into it | 04:02 |
| duonghq | have you done with the lights? | 04:02 |
| sdake | nah we just bought em today | 04:02 |
| sdake | its dark now | 04:02 |
| duonghq | ah | 04:02 |
| sdake | she made the blueprint | 04:02 |
| sdake | and she said she wants to put em up | 04:02 |
| sdake | therefore, I'll let her climb up the ladder 25 times to do so :) | 04:03 |
| sdake | we put up a wreath too | 04:03 |
| sdake | actually atm watching curiosity tv | 04:03 |
| sdake | its pretty slid | 04:03 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 04:04 |
| kfox1111 | sbezverk: awake? | 04:07 |
| sbezverk | kfox1111: barely :-) | 04:08 |
| sbezverk | waiting for last run results and go to bed.. what is up? | 04:08 |
| kfox1111 | what do you think of https://review.openstack.org/#/c/412238/ ? | 04:09 |
| kfox1111 | should be able to plugin the glance service package you've been working on into it. | 04:09 |
| kfox1111 | and do a check experimental to test. | 04:09 |
| kfox1111 | oh. nm. you already saw it. | 04:10 |
| kfox1111 | sry. :) | 04:10 |
| sdake | sup kfox1111 | 04:10 |
| openstackgerrit | Ashish Billore proposed openstack/kolla-kubernetes: Fix Kolla-Kubernetes developer quickstart guide https://review.openstack.org/412062 | 04:10 |
| kfox1111 | sdake: hey. :) | 04:10 |
| sdake | sup sbezverk | 04:10 |
| sdake | man u guys have adopted my schedule for the last 2 years ;) | 04:11 |
| sdake | i take it easy during the holidays :) | 04:11 |
| sdake | only time to get anything done | 04:11 |
| sdake | so check this out | 04:11 |
| sdake | i took my car in for a detail yesterday | 04:11 |
| sdake | 6 hrs later, I had to go to a party | 04:11 |
| sdake | so I asked for my vehicle and it was pretty clean | 04:12 |
| sdake | but they said they were not done and to bring it back tomorrow | 04:12 |
| sdake | so I brought it in at 8am | 04:12 |
| sdake | and they were done at 1pm | 04:12 |
| sdake | so like, ANOTHER 5 hours | 04:12 |
| kfox1111 | wow. | 04:12 |
| sdake | the vehicle looked better then it did when I bought it1 | 04:12 |
| sdake | ! | 04:12 |
| sdake | that is :) | 04:12 |
| kfox1111 | you should be able to eat off of any part of it for that amount of time. :) | 04:12 |
| sdake | i am taking my panamera there next | 04:12 |
| sdake | 150 bucks | 04:13 |
| sdake | not too bad for what they did | 04:13 |
| sdake | these guys really busted ass - i mean the vechile is spotless | 04:13 |
| * kfox1111 nods. | 04:13 | |
| sdake | or virtually so | 04:13 |
| sdake | after i get my porsche serviced i'm taking it iin for the 300$ deteail | 04:13 |
| sdake | i think that is sort of what they did on my rubicon | 04:14 |
| sdake | because it took them so long - and because of holidays and weekend | 04:14 |
| sdake | they are off the hook busy | 04:14 |
| kfox1111 | :) | 04:14 |
| sdake | porsche needs new brakes, oil change, gas thing is busted, few other minor issues | 04:15 |
| sdake | after the new year - i will finally have it in the garage | 04:15 |
| sdake | my jeep is a 2011 - 28k miles | 04:15 |
| sdake | my porsche is a 2015 - 33k miles.. | 04:15 |
| sdake | lol | 04:15 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 04:15 |
| sdake | i guess i'll have some catching up to do on the review queue when i get back from pto on the 4th | 04:16 |
| sdake | i'm already starting to clear out in prep for the shutdown | 04:16 |
| sdake | kfox1111 how are things going with kolla-kubernetes? | 04:16 |
| sdake | still on target for a tag on the 4th? | 04:16 |
| kfox1111 | sdake: pretty well. | 04:17 |
| kfox1111 | got almost all of the stuff converted to microservices now. | 04:17 |
| sdake | nice! | 04:17 |
| sdake | any movement on the service layer? | 04:17 |
| kfox1111 | sbezverk and I have a couple of service packages started. | 04:18 |
| sdake | nice! | 04:18 |
| sdake | i hope some day I can catch up :) | 04:18 |
| kfox1111 | the service package layer will probably need some changes to the microservices. | 04:18 |
| kfox1111 | so I'm hoping we can get a few service packages under our belt in the next week or so, | 04:19 |
| kfox1111 | and then get the microservices tweaked as needed. | 04:19 |
| kfox1111 | then when we tag we should have a pretty stable microservices layer at that point. | 04:19 |
| kfox1111 | and then the next version can focus primarily on the serivce layer. | 04:19 |
| sdake | precisely my thinking ;) | 04:20 |
| sdake | I'm glad you guys have been following along :) | 04:20 |
| sdake | thanks for that | 04:20 |
| kfox1111 | ;) | 04:20 |
| kfox1111 | np. | 04:20 |
| sdake | this is pretty much how opensource dev works | 04:21 |
| sdake | i'll be back to do proper reviews and hopefully have some tim to write some code in the new yer | 04:21 |
| sdake | year | 04:21 |
| kfox1111 | cool. | 04:21 |
| sdake | partially depends if i am elected to the board of directors for openstack | 04:21 |
| sdake | and partially depends on internal work | 04:21 |
| kfox1111 | when will you know? | 04:21 |
| sdake | ihttps://www.openstack.org/election/2017-individual-director-election/ | 04:22 |
| sdake | january 13th is when the election concludes | 04:22 |
| sdake | starts january 9th | 04:22 |
| kfox1111 | ah. k. | 04:22 |
| sdake | ok well off to bed | 04:26 |
| sdake | i am beat | 04:26 |
| sdake | long long day | 04:26 |
| sdake | i did manage to get all my paperwork done for 2016 ;-) | 04:26 |
| sdake | and booked the ptg hotel and travel | 04:26 |
| sdake | and got my car detialed | 04:27 |
| kfox1111 | sdake: nice. | 04:27 |
| sdake | and got everyone all their gifts along with my wife for christmas | 04:27 |
| kfox1111 | yup. I'm scheduled for the ptg too. so see you there. :) | 04:27 |
| sdake | actually still need to get 2 boxes of chocalate for my parents | 04:27 |
| sdake | from my children | 04:27 |
| sdake | nice! | 04:27 |
| *** sayantan_ has joined #openstack-kolla | 04:27 | |
| sdake | and bought christmas lights | 04:28 |
| sdake | at the last minute | 04:28 |
| sdake | groan | 04:28 |
| sdake | nothing but crap left | 04:28 |
| sdake | night - see ya tomorrow peeps | 04:28 |
| *** sdake has quit IRC | 04:28 | |
| kfox1111 | night. | 04:28 |
| openstackgerrit | Merged openstack/kolla: Remove /usr/bin/install from allowed commands https://review.openstack.org/412235 | 04:31 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 04:36 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: [wip] enable ceilometer related project in gate https://review.openstack.org/412278 | 04:37 |
| *** jianyi has joined #openstack-kolla | 04:37 | |
| *** Pavo has joined #openstack-kolla | 04:39 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 04:39 |
| *** Pavo_ has joined #openstack-kolla | 04:40 | |
| *** Pavo has quit IRC | 04:40 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: [wip] test scenario2 in gate https://review.openstack.org/412279 | 04:40 |
| *** Pavo has joined #openstack-kolla | 04:40 | |
| *** zhubingbing_ has quit IRC | 04:40 | |
| kfox1111 | sbezverk: got a min for https://review.openstack.org/#/c/412244 ? | 04:41 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: [wip] enable ceilometer related project in gate https://review.openstack.org/412278 | 04:41 |
| *** Pavo_ has quit IRC | 04:41 | |
| portdirect | duonghq: ping | 04:45 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Gate fix for helm sometimes failing https://review.openstack.org/412282 | 04:46 |
| *** Jeffrey4l has quit IRC | 04:53 | |
| *** sayantan_ has quit IRC | 04:54 | |
| *** Pavo has quit IRC | 05:03 | |
| *** Pavo has joined #openstack-kolla | 05:04 | |
| *** zhubingbing_ has joined #openstack-kolla | 05:12 | |
| *** newmember has joined #openstack-kolla | 05:13 | |
| *** newmember has quit IRC | 05:14 | |
| *** newmember has joined #openstack-kolla | 05:14 | |
| *** unicell1 has joined #openstack-kolla | 05:16 | |
| *** unicell has quit IRC | 05:16 | |
| *** senk has joined #openstack-kolla | 05:17 | |
| portdirect | anyone free to review: https://review.openstack.org/#/c/399042/ & https://review.openstack.org/#/c/412205/ ? | 05:20 |
| *** sayantan_ has joined #openstack-kolla | 05:24 | |
| *** hfu has joined #openstack-kolla | 05:24 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 05:24 |
| *** sayantan_ has quit IRC | 05:29 | |
| *** Pavo has quit IRC | 05:33 | |
| *** Pavo has joined #openstack-kolla | 05:33 | |
| *** saneax-_-|AFK is now known as saneax | 05:35 | |
| *** senk has quit IRC | 05:38 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Memcached svc https://review.openstack.org/412287 | 05:39 |
| *** gomarivera has quit IRC | 05:40 | |
| *** jianyi has quit IRC | 05:44 | |
| *** mdnadeem has joined #openstack-kolla | 05:47 | |
| *** wwriverrat_ has quit IRC | 06:00 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 06:01 |
| *** yuanying has joined #openstack-kolla | 06:14 | |
| *** yingjun has joined #openstack-kolla | 06:15 | |
| *** yuanying has quit IRC | 06:19 | |
| *** jianyi has joined #openstack-kolla | 06:20 | |
| kfox1111 | so, kolla's kubernetes-entrypoint container seems broken. | 06:30 |
| kfox1111 | I'm not seeing kubernetes-entrypoint in it at all. | 06:31 |
| *** Pavo has quit IRC | 06:31 | |
| *** enginy has joined #openstack-kolla | 06:31 | |
| *** Pavo has joined #openstack-kolla | 06:32 | |
| *** yingjun has quit IRC | 06:33 | |
| *** enginy has quit IRC | 06:34 | |
| *** Pavo has quit IRC | 06:35 | |
| *** enginy has joined #openstack-kolla | 06:35 | |
| *** enginy has quit IRC | 06:36 | |
| *** qwertyco has joined #openstack-kolla | 06:36 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Update the word choice for rst file https://review.openstack.org/392032 | 06:46 |
| *** v1k0d3n has quit IRC | 06:52 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Update the word choice for rst file https://review.openstack.org/392032 | 06:53 |
| *** chandankumar has joined #openstack-kolla | 07:11 | |
| *** jianyi has quit IRC | 07:13 | |
| *** senk has joined #openstack-kolla | 07:13 | |
| *** yingjun has joined #openstack-kolla | 07:15 | |
| *** fragatina has joined #openstack-kolla | 07:18 | |
| *** JRobinson__ has quit IRC | 07:21 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Update the word choice for rst file https://review.openstack.org/392032 | 07:23 |
| *** portdirect has quit IRC | 07:26 | |
| *** zhangqiankun has quit IRC | 07:28 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move manila precheck into its own role https://review.openstack.org/404338 | 07:33 |
| *** pcaruana has joined #openstack-kolla | 07:33 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 07:36 |
| *** senk has quit IRC | 07:39 | |
| *** zhangqiankun has joined #openstack-kolla | 07:45 | |
| bjolo | morning | 07:50 |
| *** senk has joined #openstack-kolla | 07:58 | |
| openstackgerrit | qiankun.zhang proposed openstack/kolla: Install tacker-horizon into horizon image,add Bug info,fix conflict https://review.openstack.org/407810 | 08:01 |
| *** shardy has joined #openstack-kolla | 08:02 | |
| openstackgerrit | qiankun.zhang proposed openstack/kolla: Install tacker-horizon into horizon image,add Bug info,fix conflict https://review.openstack.org/407810 | 08:06 |
| *** msimonin has joined #openstack-kolla | 08:11 | |
| *** jianyi has joined #openstack-kolla | 08:15 | |
| *** msimonin has quit IRC | 08:17 | |
| *** yingjun has quit IRC | 08:18 | |
| *** qwertyco has quit IRC | 08:24 | |
| *** jianyi has quit IRC | 08:28 | |
| *** chandankumar has quit IRC | 08:29 | |
| *** chandankumar has joined #openstack-kolla | 08:30 | |
| *** qwertyco has joined #openstack-kolla | 08:36 | |
| *** egonzalez90 has joined #openstack-kolla | 08:37 | |
| *** williamcaban has joined #openstack-kolla | 08:40 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 08:41 |
| *** zhangqiankun has quit IRC | 08:42 | |
| *** zhangqiankun has joined #openstack-kolla | 08:43 | |
| *** williamcaban has quit IRC | 08:45 | |
| *** senk has quit IRC | 08:45 | |
| openstackgerrit | Vladislav Belogrudov proposed openstack/kolla-ansible: Use default user group consistently https://review.openstack.org/411742 | 08:45 |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 08:46 |
| *** qwertyco has quit IRC | 09:06 | |
| *** portdirect has joined #openstack-kolla | 09:10 | |
| *** chandankumar has quit IRC | 09:12 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 09:13 |
| *** portdirect_ has joined #openstack-kolla | 09:17 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 09:17 |
| *** gfidente has joined #openstack-kolla | 09:18 | |
| *** gfidente has joined #openstack-kolla | 09:18 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 09:20 |
| *** senk has joined #openstack-kolla | 09:20 | |
| *** portdirect_ has quit IRC | 09:24 | |
| *** senk has quit IRC | 09:25 | |
| *** yingjun has joined #openstack-kolla | 09:26 | |
| *** senk has joined #openstack-kolla | 09:27 | |
| *** hieulq has joined #openstack-kolla | 09:29 | |
| *** rsjethani has joined #openstack-kolla | 09:31 | |
| *** portdirect_ has joined #openstack-kolla | 09:32 | |
| rsjethani | Hi all, a quick question: after 'kolla-ansible build' i can see the required image say ceph-mon present via 'docker images' but during 'kolla-ansible pull' we are again downloading the image why so? I am working on 3.0.0/newton | 09:34 |
| rsjethani | also i am using ubuntu base and binary mode | 09:35 |
| rsjethani | i the pull step before deploy necessary? | 09:36 |
| egonzalez90 | rsjethani: which tag have you in globals.yml? | 09:36 |
| egonzalez90 | maybe differs from the docker image and from there | 09:36 |
| rsjethani | 3.0.0 | 09:37 |
| *** chandankumar has joined #openstack-kolla | 09:37 | |
| rsjethani | i check the images after build phase have 3.0.0 tag only | 09:37 |
| egonzalez90 | also, in global.yml have your registry configured properly or are pointing to kolla's registry? | 09:38 |
| rsjethani | as i am following all in one guide i am mot required to have a local registry so i skipped it | 09:38 |
| egonzalez90 | yup, but by default is pointing to kolla's one | 09:39 |
| egonzalez90 | need to configure to use your tag | 09:39 |
| rsjethani | how? | 09:39 |
| egonzalez90 | docker_namespace: "egonzalez90" | 09:40 |
| egonzalez90 | egonzalez90/centos-source-mongodb | 09:40 |
| egonzalez90 | namespace should match with the image namespace, ^^ as example | 09:40 |
| rsjethani | but this is not happening for all images... | 09:41 |
| rsjethani | AFAIK all images should be downloaded again if by default we are pointing to kolla's registry | 09:43 |
| egonzalez90 | can you paste in paste.openstack.org the output from docker images and the content in /etc/kolla/globals.yml ? | 09:44 |
| rsjethani | ok | 09:44 |
| *** papacz has joined #openstack-kolla | 09:45 | |
| *** yingjun has quit IRC | 09:45 | |
| egonzalez90 | yup, if the registry is pointing to a different one all images will be pull | 09:45 |
| *** athomas has joined #openstack-kolla | 09:46 | |
| rsjethani | by the way which would be faster install from source or binary? | 09:46 |
| rsjethani | egonzalez90: docker images result: http://paste.openstack.org/show/592768/ | 09:48 |
| rsjethani | egonzalez90: globals.yml: http://paste.openstack.org/show/592769/ | 09:51 |
| *** senk has quit IRC | 09:52 | |
| *** senk has joined #openstack-kolla | 09:53 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: WIP add fluentd role https://review.openstack.org/407392 | 09:55 |
| egonzalez90 | rsjethani: that's weird, if you do kolla-ansible pull only pull ceph-mon or when finish ceph-mon it pull more images? | 09:57 |
| *** leseb has joined #openstack-kolla | 09:59 | |
| *** tyrola has quit IRC | 10:00 | |
| portdirect_ | hey leseb o/ | 10:00 |
| *** tyrola has joined #openstack-kolla | 10:01 | |
| zhubingbing_ | sup egonzalez90 | 10:01 |
| *** duonghq has quit IRC | 10:01 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Remove /usr/bin/install from allowed commands https://review.openstack.org/412377 | 10:02 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Remove /usr/bin/install from allowed commands https://review.openstack.org/412378 | 10:02 |
| openstackgerrit | zhangshuai proposed openstack/kolla-ansible: Add karbor ansible role https://review.openstack.org/402229 | 10:05 |
| SamYaple | morning | 10:06 |
| SamYaple | leseb: go to see you round these parts | 10:06 |
| zhubingbing_ | morning | 10:06 |
| *** tovin07_ has quit IRC | 10:07 | |
| portdirect_ | SamYaple where are you based again? It's gotta be pretty early for you? | 10:10 |
| leseb | SamYaple hey morning :) | 10:11 |
| *** sp_ has joined #openstack-kolla | 10:12 | |
| sp_ | portdirect: hi.. | 10:12 |
| leseb | SamYaple good to be here as well, I'm actually a bit surprised to see you here :p | 10:12 |
| portdirect | leseb, SamYaple's one of the old guard :) | 10:14 |
| *** hfu has quit IRC | 10:15 | |
| *** yuanying has joined #openstack-kolla | 10:15 | |
| rsjethani | egonzalez90: performing pull again returns quickly hence doesn't seem to download more images | 10:18 |
| bjolo | portdirect, are you the guy from scotland i met in barcelona? | 10:20 |
| egonzalez90 | rsjethani: thats weird, actually kolla-ansible pull use the same file to pull that when applying the role | 10:20 |
| portdirect | bjolo: just eating a chocolate deep fried pizza at the moment. sup dude? :) | 10:21 |
| bjolo | hehe | 10:21 |
| bjolo | whats up! :) | 10:21 |
| *** zhubingbing_ has quit IRC | 10:22 | |
| portdirect | just been playing with the kolla cats pretty much non stop since then :) getting k8s up and running - hows tricks with you? | 10:22 |
| *** portdirect_ has quit IRC | 10:22 | |
| bjolo | same here, but with kolla-ansible | 10:23 |
| bjolo | our cloud is in beta testing now | 10:23 |
| bjolo | so far so good | 10:23 |
| SamYaple | portdirect: its 5am here | 10:23 |
| bjolo | after new year, my plan is to start hacking on kolla-ansible database backup/recovery | 10:23 |
| SamYaple | leseb: work is pushing me back into kolla stuff. and im working on kolla-salt | 10:23 |
| bjolo | kolla-salt?! :) | 10:24 |
| *** Jeffrey4l has joined #openstack-kolla | 10:24 | |
| bjolo | dont know much about salt, except some friends that like it alot. | 10:24 |
| SamYaple | bjolo: https://etherpad.openstack.org/p/kolla-salt | 10:25 |
| SamYaple | with working code | 10:25 |
| bjolo | SamYaple, whats the pros&cons of using salt instead of ansible in your opinion? | 10:25 |
| SamYaple | ill do the paperwork ot make it official beginning of year probably | 10:25 |
| SamYaple | bjolo: no cons, all pros | 10:25 |
| bjolo | haha of course | 10:25 |
| SamYaple | bjolo: *maybe* slightly higher setup cost due to having queues and stuff | 10:25 |
| rsjethani | egonzalez90: let me look more closely again | 10:26 |
| SamYaple | im not really targeting salt-ssh | 10:26 |
| leseb | SamYaple oh great :) | 10:29 |
| *** senk has quit IRC | 10:30 | |
| bjolo | SamYaple, let me refrase then. What issues/problems/cons in kolla-ansible are making you go for yet another kolla orchestration tool. (not questioning why, just want to understand and learn) | 10:30 |
| SamYaple | bjolo: heh. well for starters the project i merged in to kolla to bring ansible to kolla was originally.... salt! but salt didnt have all the orchestration pieces that I needed at the time | 10:31 |
| *** Serlex has joined #openstack-kolla | 10:32 | |
| SamYaple | so i begrudginly went with ansible | 10:32 |
| *** portdirect_ has joined #openstack-kolla | 10:32 | |
| SamYaple | ansible makes complex things very hard | 10:32 |
| SamYaple | and ansible is _not_ DRY | 10:32 |
| SamYaple | this means when we do an update to some logic flow, we have to update HUNDREDS of lines, rather than one section | 10:33 |
| SamYaple | this leads to missing things, more mistakes, etc etc | 10:33 |
| *** portdirect_ has quit IRC | 10:33 | |
| SamYaple | salt, well salt we can write "tasks" in pure python should we need that flexibility | 10:33 |
| SamYaple | salt is to python as ansible is to bash | 10:34 |
| SamYaple | as far as flexibility and what not | 10:34 |
| bjolo | hmm ok sounds interesting. makes me wanna take a look at salt in general | 10:35 |
| bjolo | we use ansible today for all orchestration, so kolla-ansible makes a lot of sense to us from that perspective | 10:35 |
| SamYaple | bjolo: remember, python has a slightly higher learning curve than python, same with salt | 10:36 |
| SamYaple | python higer than bash* | 10:36 |
| bjolo | gotcha | 10:36 |
| SamYaple | and coming from ansible, youll have to forget/relearn some thing | 10:36 |
| bjolo | i have hacked quite a bit of python over the years, but not in the last one or two | 10:36 |
| *** gomarivera has joined #openstack-kolla | 10:36 | |
| *** rafael_sisweb has quit IRC | 10:37 | |
| egonzalez90 | SamYaple: when doing designate-manage pool update /etc/designate/pools.yaml, pool config is not applied in designate, any idea? | 10:37 |
| SamYaple | egonzalez90: how are you verifying its not applied? | 10:38 |
| egonzalez90 | docker exec -ti designate_pool_manager designate-manage pool show_config | 10:38 |
| SamYaple | what about looking at the database directly? | 10:39 |
| *** gomarivera has quit IRC | 10:41 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add python-gnocchiclient to ceilomter base image https://review.openstack.org/412396 | 10:42 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: [wip] enable ceilometer related project in gate https://review.openstack.org/412278 | 10:43 |
| sp_ | portdirect: Jeffrey4l pbourke please review https://review.openstack.org/411313 in reference to the bug | 10:50 |
| Jeffrey4l | sp_, roger. | 10:50 |
| sp_ | also please confirm | 10:51 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] static uid/gid in Dockerfiles https://review.openstack.org/412231 | 10:51 |
| sp_ | Jeffrey4l: thanks | 10:51 |
| * Jeffrey4l is checking. | 10:51 | |
| Jeffrey4l | sp_, how to trigger the issue: ERROR: unknown environment 'cover'? | 10:52 |
| Jeffrey4l | and who is using this env? | 10:52 |
| sp_ | Jeffrey4l: after getting the repo of kolla run the tox -ecover command | 10:53 |
| sp_ | Jeffrey4l: while in case of other projects the same execution steps work | 10:54 |
| Jeffrey4l | who use tox -ecover? the patch is doing nothing except simple create cover environment. | 10:54 |
| Jeffrey4l | sp_, check this https://github.com/openstack/nova/blob/master/tox.ini#L78 | 10:55 |
| sp_ | Jeffrey4l: yes | 10:55 |
| sp_ | Jeffrey4l: we keep cover environment creation in tox.ini | 10:56 |
| Jeffrey4l | i think you need run "coverage" command at least. | 10:56 |
| Jeffrey4l | why not add the coverage command? | 10:56 |
| Jeffrey4l | i do not think a simple cover environment is useful. | 10:56 |
| SamYaple | Jeffrey4l: https://review.openstack.org/#/c/412255/ <-- shaves about 10% time off the ubuntu gates | 10:57 |
| sp_ | Jeffrey4l: coverage command works fine. It produces index.html too which represent the %coverage of file too | 10:58 |
| Jeffrey4l | sp_, yep. but why not generate the report when running "tox -ecoverage: | 10:58 |
| portdirect | sp_: not my area of expertise :( Jeffrey4l is much better here | 10:59 |
| sp_ | portdirect: thanks | 10:59 |
| sp_ | Jeffrey4l: does "tox -ecoverage" works in other core component too | 11:00 |
| sp_ | ?? | 11:00 |
| portdirect | Jeffrey4l: if you get time it would br great if you could have a look at this https://review.openstack.org/#/c/412205/ <<-- k8s images for kolla-k8s with ceph support | 11:01 |
| Jeffrey4l | SamYaple, LGTM. does overlay2 is production ready? | 11:01 |
| Jeffrey4l | sp_, i think so. check the tox -e coverage in nova project. | 11:01 |
| Jeffrey4l | portdirect, roger | 11:02 |
| SamYaple | Jeffrey4l: yea overlay2 is basically overlay with a few tweaks. its not much new code, its just not compatible with the original overlay driver so they have to make a new one | 11:03 |
| sp_ | Jeffrey4l: checking | 11:03 |
| Jeffrey4l | SamYaple, thanks. | 11:03 |
| Jeffrey4l | SamYaple, then we do not need setup_disk when using overlay2, right? | 11:04 |
| Jeffrey4l | i see. | 11:05 |
| Jeffrey4l | ignore me | 11:05 |
| portdirect | Jeffrey4l: are you are in china? | 11:06 |
| Jeffrey4l | portdirect, yep. | 11:06 |
| portdirect | can you access https://storage.googleapis.com? if not i could rework that ps to use the tarball from github? | 11:07 |
| SamYaple | Jeffrey4l: yea, i moved it to trusty only. | 11:07 |
| Jeffrey4l | i can. but need some vpn ;) | 11:07 |
| portdirect | np then :) | 11:07 |
| Jeffrey4l | tbh, it is very hard to build kolla's image in china ;(. some site are blocked. for example, aws. | 11:08 |
| Jeffrey4l | ( in kolla, some packages come from s3. ) | 11:08 |
| portdirect | :( wow didnt realise that was - hub.docker.io is ok though? | 11:09 |
| Jeffrey4l | in the pass hub.docker.io is very slow. but it is faster recently. So i think maybe hub.docker.io add china cdn now. | 11:10 |
| SamYaple | Jeffrey4l: yea we need to get infra to mirror all of our stuff | 11:11 |
| openstackgerrit | YumengBao proposed openstack/kolla: Signed-off-by: YumengBao <bao.yumeng@zte.com.cn> https://review.openstack.org/412404 | 11:11 |
| Jeffrey4l | it will be cool if they can provide mirrors for all repo. like elasticsearch, rabbitmq etc. | 11:12 |
| Jeffrey4l | on other hand, kolla do not support change all repo url now. | 11:13 |
| Jeffrey4l | like: how can i change the epel.repo's url? | 11:13 |
| Jeffrey4l | customization is a solution, but not that better. | 11:13 |
| *** rafael_sisweb has joined #openstack-kolla | 11:14 | |
| Jeffrey4l | here is an example, i used when building images in china. | 11:15 |
| Jeffrey4l | which changes the url to china site mostly. | 11:15 |
| Jeffrey4l | https://review.openstack.org/#/c/329060/19/etc/kolla/kolla-template-override.j2 | 11:15 |
| Jeffrey4l | SamYaple, ^ | 11:15 |
| openstackgerrit | YumengBao proposed openstack/kolla: Fix port error in image-building.rst https://review.openstack.org/412404 | 11:17 |
| rafael_sisweb | Hi, i'm trying to build images with kolla 4.0.0.0b2 but some images i could see some images has pointing "binary not yet available for centos" with a fix codition into Dockerfile.j2. Why that? | 11:17 |
| SamYaple | Jeffrey4l: yikes lol. | 11:17 |
| portdirect | Jeffrey4l: why not make it mergeable? seems wrong that you are at a disadvantage when building kolla :( | 11:17 |
| SamYaple | rafael_sisweb: not all services have implemented all distros | 11:17 |
| SamYaple | rafael_sisweb: rdo must not have packaged whatever you are trying to build | 11:18 |
| SamYaple | rafael_sisweb: or conversely, kolla hasnt added support for it | 11:18 |
| *** senk has joined #openstack-kolla | 11:19 | |
| Jeffrey4l | portdirect, it shouldn't be merged. it just use some customization feature. | 11:19 |
| SamYaple | Jeffrey4l: i think portdirect means we should include the ability to specify the repos to use and keys to add more dynamically | 11:21 |
| rafael_sisweb | But if i build with install type source it should work? | 11:21 |
| Jeffrey4l | hrm. may be we can. l'd like the diskimage-builder solution. when DIB_EPEL_REPO_URL exist, replace the default url with it. | 11:22 |
| SamYaple | rafael_sisweb: i dont know. what image is it? | 11:22 |
| SamYaple | rafael_sisweb: not all images work on all distros (regardless of source vs binary) | 11:22 |
| SamYaple | rafael_sisweb: i think only ubuntu source has _all_ images building | 11:22 |
| berendt | Jeffrey4l: I think it makes most sense to use the template override mechanism to provide custom apt sources files | 11:23 |
| openstackgerrit | Martin André proposed openstack/kolla: Expect monasca-base to fail build on Ubuntu binary https://review.openstack.org/412408 | 11:23 |
| openstackgerrit | Martin André proposed openstack/kolla: Sort excluded images alphabetically https://review.openstack.org/412409 | 11:23 |
| berendt | or repo files... | 11:24 |
| SamYaple | berendt: agreed | 11:24 |
| berendt | does it makes sense to open a blueprint for it? | 11:24 |
| portdirect | Jeffrey4l: I'd be up for helping with implmenting that (though would not have time untill jan), it would also help me speed up my builds, though thats just a luxury for me. | 11:24 |
| SamYaple | berendt: im pretty sure we do something in the gate for that | 11:24 |
| SamYaple | i remember writing something.... | 11:24 |
| Jeffrey4l | berendt, the issue is: the customization in only exist at the beginning and ending. in some line, kolla still access Internet. | 11:24 |
| openstackgerrit | Merged openstack/kolla: Add fluent-plugin-grok-parser in fluentd https://review.openstack.org/410089 | 11:25 |
| Jeffrey4l | SamYaple, yep. we did. but only for Ubuntu. | 11:25 |
| rafael_sisweb | Thanks for clarify that SamYaple, i got the binary not available to a lot of images when runing simple kolla-build with centos distro. manasca-base is a sample of that | 11:25 |
| berendt | SamYaple: on the gates we run some seds on the existing files | 11:25 |
| berendt | SamYaple: in setup_gate.sh | 11:25 |
| Jeffrey4l | portdirect, you can make some change of my patch. | 11:26 |
| *** dave-mccowan has joined #openstack-kolla | 11:26 | |
| berendt | Jeffrey4l: in my build environment i use apt-cacher-ng to workaround the static source files | 11:26 |
| SamYaple | berendt: ahhh thats right. This was all implemented in debian only | 11:26 |
| berendt | and I create a /etc/pip.conf file to use a local pip cache | 11:27 |
| Jeffrey4l | https://review.openstack.org/365721 | 11:27 |
| Jeffrey4l | use infra mirrors for centos ^^ | 11:28 |
| Jeffrey4l | i will fix the merge conflict ;) | 11:28 |
| Jeffrey4l | berendt, not familiar with apt. i use centos mostly. | 11:28 |
| egonzalez90 | SamYaple: yes, it was in the database, was missing --pool_id parameter in the command. Thanks | 11:29 |
| *** esmiurium has joined #openstack-kolla | 11:30 | |
| *** mnasiadka has joined #openstack-kolla | 11:33 | |
| *** portdirect is now known as portdirect_away | 11:34 | |
| SamYaple | Jeffrey4l: why did you remove ceph repo from ubuntu? | 11:34 |
| SamYaple | Jeffrey4l: now it doesnt get z releases of ceph | 11:34 |
| SamYaple | Jeffrey4l: thats the reason it was there | 11:34 |
| openstackgerrit | zhangshuai proposed openstack/kolla-ansible: Add karbor ansible role https://review.openstack.org/402229 | 11:35 |
| Jeffrey4l | hrm. there was a bug, ubuntu-cloudarchive depends on a higher ceph version, which ubuntu-cloudarchive provides it. | 11:35 |
| SamYaple | Jeffrey4l: nah thats because you didnt switch to the jewel repo, you were still on hammer | 11:36 |
| SamYaple | im going to add it back, it really shouldnt have been removed | 11:36 |
| Jeffrey4l | i do not think so. just a z stream version, iirc. | 11:36 |
| Jeffrey4l | let me check | 11:36 |
| SamYaple | either way, cloudarchive doesnt get updates like it should, and it won't get a z release | 11:37 |
| SamYaple | so im going to add ceph repo back in. ill resolve any conflict that might exist | 11:37 |
| Jeffrey4l | ok. | 11:37 |
| SamYaple | if that was the only reason it was removed, ill get it in shape again. thanks for the info | 11:37 |
| sp_ | Jeffrey4l: correct command for cover environment check is "tox -ecover" | 11:39 |
| sp_ | Jeffrey4l: same is in working nova too | 11:39 |
| sp_ | Jeffrey4l: so I think my implementation and logged bug is valid in which we will | 11:40 |
| sp_ | enable -ecover option for kolla | 11:41 |
| sp_ | Jeffrey4l: BRB | 11:41 |
| berendt | sp_: which bug report? | 11:42 |
| *** dave-mcc_ has joined #openstack-kolla | 11:45 | |
| *** dave-mccowan has quit IRC | 11:48 | |
| openstackgerrit | zhangshuai proposed openstack/kolla-ansible: Add karbor ansible role https://review.openstack.org/402229 | 11:53 |
| *** hfu has joined #openstack-kolla | 11:59 | |
| *** liyifeng has joined #openstack-kolla | 12:02 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla-ansible: Add zookeeper role https://review.openstack.org/412423 | 12:02 |
| sp_ | berendt: https://review.openstack.org/411313 this is the review link for the respective bug | 12:05 |
| sp_ | berendt: sorry i wasn't around | 12:05 |
| *** Bico_Fino has joined #openstack-kolla | 12:07 | |
| *** Jeffrey4l has quit IRC | 12:10 | |
| *** senk has quit IRC | 12:12 | |
| *** shardy is now known as shardy_lunch | 12:15 | |
| *** dave-mcc_ has quit IRC | 12:18 | |
| *** sdake has joined #openstack-kolla | 12:18 | |
| sdake | morning peeps | 12:18 |
| SamYaple | morning | 12:19 |
| bjolo | SamYaple, about https://review.openstack.org/#/c/412231/9 | 12:21 |
| SamYaple | yea | 12:21 |
| openstackgerrit | prameswar proposed openstack/kolla: Mariadb log file permission changed https://review.openstack.org/394007 | 12:21 |
| bjolo | why uid/gid under 500 as default? | 12:21 |
| bjolo | http://unix.stackexchange.com/questions/101313/what-are-the-dangers-of-creating-a-normal-user-with-uid-500 | 12:21 |
| bjolo | dont know if it is an issue or not, just asking if you had any reason for choosing the 400 range | 12:22 |
| *** egonzalez90_ has joined #openstack-kolla | 12:22 | |
| *** egonzalez90 has quit IRC | 12:22 | |
| *** sdake_ has joined #openstack-kolla | 12:22 | |
| SamYaple | bjolo: so centos family uses 500 -- 60000 as the 'pool' to pull an autogened uid from | 12:22 |
| SamYaple | that means if we use in that range we get conflicts | 12:22 |
| SamYaple | sub 500 is not pulled from for anything other than _system_ accounts. and those (by default) are the lowest available | 12:23 |
| *** sdake has quit IRC | 12:23 | |
| SamYaple | so 400-500 is our best range to ensure no conflicts arise | 12:23 |
| sp_ | sdake_: morning | 12:23 |
| SamYaple | ubuntu uses 1000-60000, but we want to be consistent across both | 12:23 |
| bjolo | good point | 12:24 |
| sdake_ | yo sp_ | 12:24 |
| bjolo | yes, im in favor of consistency | 12:24 |
| bjolo | especially between different distros | 12:24 |
| sdake_ | upgrade needs to work | 12:24 |
| sp_ | sdake_: https://bugs.launchpad.net/kolla-kubernetes/+bug/1650266 please confirm this bug | 12:25 |
| openstack | Launchpad bug 1650266 in kolla-kubernetes "Running the coverage tool gets failed" [Undecided,New] | 12:25 |
| sdake_ | sp_ i'll triage it | 12:25 |
| sdake_ | sp_ i can't confrim atm, i just woke up:) | 12:25 |
| sp_ | sdake_: np | 12:25 |
| sdake_ | sp_ it would help if you included the command to generate the error- I asume its tox -e cover? | 12:26 |
| sp_ | sdake_: but it gr8 whenever you got some movement please check :( | 12:26 |
| bjolo | i have argued for this before that in production, operators will have a mix and match between containers and versions. wrote a BP about it https://blueprints.launchpad.net/kolla/+spec/support-for-mixed-container-types | 12:26 |
| sp_ | sdake_: would be* | 12:27 |
| *** yingjun has joined #openstack-kolla | 12:27 | |
| sp_ | sdake_: yes "tox -ecover" | 12:27 |
| sp_ | is the command | 12:27 |
| SamYaple | bjolo: so *way* back when (im talking pre-liberty) this used to be more configurable | 12:28 |
| SamYaple | bjolo: it more than triples the boilerplate code for _each_ image | 12:28 |
| SamYaple | bjolo: so it was removed since what you purpose *can* be done as is | 12:28 |
| *** portdirect_away is now known as portdirect | 12:29 | |
| SamYaple | sure that one operator is goign to have a messier config, but it allows the project to have a smaller footprint, smaller footprint == less code == less bugs | 12:29 |
| *** sdake has joined #openstack-kolla | 12:29 | |
| SamYaple | or so goes that logic | 12:29 |
| bjolo | SamYaple, ive only been around since this summer so dont know the history or the code well enough do estimate code impact | 12:29 |
| SamYaple | well, very specifically what you purpose was *explicitly* removed. it used to exist | 12:29 |
| SamYaple | im not saying we cant add it back now that the project is bigger | 12:30 |
| SamYaple | just for your info | 12:30 |
| SamYaple | but for what its worth, the user can just have a one line override per image | 12:30 |
| SamYaple | i dont think thats too bad | 12:30 |
| SamYaple | nova_compute_image_full: full/image-with:tag | 12:31 |
| *** hfu has quit IRC | 12:31 | |
| SamYaple | it used to have per image registry too bjolo | 12:31 |
| SamYaple | really bloated | 12:31 |
| *** sdake_ has quit IRC | 12:32 | |
| SamYaple | bjolo: https://github.com/openstack/kolla-ansible/blob/3ac7da64d12c631def3640369998d292fd1a199f/ansible/roles/keystone/defaults/main.yml | 12:33 |
| SamYaple | thats as far back as I can find. | 12:33 |
| *** hfu has joined #openstack-kolla | 12:33 | |
| sdake | thanks for the bug report sp_ :) | 12:33 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] static uid/gid in Dockerfiles https://review.openstack.org/412231 | 12:33 |
| SamYaple | sdake: you might have an opinion on that ^ | 12:34 |
| SamYaple | centralizes the uid creation to base and uses static numbers | 12:34 |
| sdake | SamYaple so here is what I wanted last cycle | 12:34 |
| sp_ | sdake: ok | 12:34 |
| sdake | I wanted in the base image a list of UUID and GUIDs that were standardized for kolla | 12:34 |
| sdake | independent of distro | 12:35 |
| SamYaple | thats what this is | 12:35 |
| sdake | we just didn't have time for it | 12:35 |
| SamYaple | yea thats what this patch is | 12:35 |
| sdake | SamYaple sorry i was dcing and stuff to connect to my home network to validate verify sp_ 's bug report, so I missed most of your message | 12:35 |
| sdake | SamYaple further I wanted it to work with binary and source | 12:36 |
| SamYaple | sdake: im not sure what you missed, but everything you just said, thats what this patch is. standardized kolla uid/gid | 12:36 |
| SamYaple | well youll have to talk to rdo about that one | 12:36 |
| SamYaple | they hardcode uid for nova (and cinder? maybe a few others) | 12:36 |
| sdake | SamYaple - rpm will use an existing uid/gid | 12:36 |
| SamYaple | nope | 12:36 |
| SamYaple | well yea | 12:36 |
| SamYaple | youre right. it will use an _existing_ | 12:37 |
| sdake | it is definately yes- they do hardcode, but it doesn't matter | 12:37 |
| SamYaple | yea youre right. i thought you meant something different at first | 12:37 |
| sdake | it will try to useradd, and fail, and then rpm will adopt what is already there | 12:37 |
| sdake | ok | 12:37 |
| SamYaple | this works with all of our distros, _but_ it doesnt match rdo hardcode | 12:37 |
| sdake | right that is where the third requirement comes in | 12:37 |
| sdake | want it to work with upgrades ;) | 12:37 |
| SamYaple | if the user already exists, rdo wont chance uid | 12:38 |
| SamYaple | upgrade is a kolla-* problem | 12:38 |
| sdake | agreed | 12:38 |
| SamYaple | but it could be done with config.json stuff | 12:38 |
| sdake | still needs to work | 12:38 |
| SamYaple | the code already exists for that | 12:38 |
| sdake | right - we didn't have config.json prior | 12:38 |
| sdake | that is what we did to solve the upgrade problem in general last cycle because we discovered it was a problem about t-2 weeks to the final 3.0.0 tag | 12:39 |
| SamYaple | yea | 12:39 |
| sdake | actually jeffrey4l did it iirc | 12:39 |
| SamYaple | this removes the nova-libvirt -> nova-base dep | 12:39 |
| sdake | we did not have time to standaridze the uid/gids however | 12:39 |
| sdake | as that was a big job | 12:39 |
| SamYaple | so right now, i do homedir and shell creation _with_ user and group creation | 12:40 |
| *** hfu has quit IRC | 12:40 | |
| SamYaple | that can be split out into individual dockerfiles | 12:40 |
| SamYaple | and jsut have the user/group created in the base | 12:40 |
| SamYaple | im ok with that | 12:40 |
| SamYaple | that way we dont have a bunch of /var/lib/<service> directories in each container | 12:41 |
| SamYaple | but beyond that, i dont have much more to do one this patch | 12:41 |
| *** senk has joined #openstack-kolla | 12:42 | |
| *** portdirect is now known as portdirect_away | 12:43 | |
| *** Jeffrey4l has joined #openstack-kolla | 12:43 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla-ansible: Add zookeeper role https://review.openstack.org/412423 | 12:44 |
| *** flaper87 has joined #openstack-kolla | 12:45 | |
| sdake | SamYaple when i wake up - i'll provide a review | 12:46 |
| sdake | SamYaple i hate reviewing right when i wake up - not particularly sharp | 12:46 |
| SamYaple | sdake: cool. I wish I could break it up into smaller patches, but most of this patch was written with regex stuff and I just dont have the time to pick through it manually | 12:47 |
| sdake | SamYaple i'll hunt it down in the kolla review queue - i'd like jeffrey to review as well to make sure it actually works for kolla-ansible since we don't have cross gating atm | 12:47 |
| sdake | cross-repo gating | 12:47 |
| SamYaple | sdake: no but we do deploy in the kolla gate still | 12:47 |
| SamYaple | its just non-voting | 12:47 |
| SamYaple | so the results are still coming up green | 12:47 |
| sdake | you mean kolla-ansible gate? | 12:47 |
| SamYaple | no | 12:47 |
| sdake | maybe cross-repo gating is in! | 12:47 |
| SamYaple | no | 12:47 |
| sdake | i need to see that | 12:48 |
| SamYaple | its not | 12:48 |
| sdake | got a link? | 12:48 |
| SamYaple | sdake: https://review.openstack.org/#/c/412255/ | 12:48 |
| SamYaple | thats the kolla repo, but we still have deploy jobs | 12:48 |
| *** Bico_Fino has quit IRC | 12:48 | |
| SamYaple | you are talking about kolla-ansible doing a depends-on test, right? | 12:48 |
| sdake | cross repo gating is where kolla gates using all kolla-ansible and kolla-kubernetes jobs | 12:48 |
| sdake | and visa-versa | 12:48 |
| SamYaple | my point is, the ansible (commited) code is still being exersized | 12:48 |
| SamYaple | at least currently | 12:49 |
| sdake | yup- cross repo gating is in! | 12:49 |
| SamYaple | because the deploy gates are there and they git cloen kolla-ansible | 12:49 |
| sdake | although our gate jobs dont' test upgrade | 12:49 |
| *** Bico_Fino has joined #openstack-kolla | 12:49 | |
| sdake | they test upgrade is a minimal sense | 12:49 |
| sdake | as in "does the upgrade command work" | 12:49 |
| sdake | not does it upgrade and run properly afterwards | 12:50 |
| SamYaple | oh yea. for sure. but upgrade in the gate is worthless unless we can pull old images | 12:51 |
| SamYaple | this one will upgrade all day long, because static uid's | 12:51 |
| sdake | SamYaple we can build old images theoretically :) | 12:52 |
| sdake | although I think that is broken atm with rdo | 12:52 |
| sdake | not quite sure | 12:52 |
| sdake | i read libvirt and centos7.3 are busted atm | 12:52 |
| SamYaple | eh idk. but its goign to be hard to test building old images i the gate, time constrained and all | 12:52 |
| sdake | yup | 12:52 |
| SamYaple | maybe ubuntu since it can finish in under 20m | 12:53 |
| sdake | if we had mirrors of epel centos would be much faster | 12:53 |
| sdake | i have submitted mirror jobs, but to no avail | 12:53 |
| sdake | they dont get reviewed unfortunately | 12:53 |
| sdake | only a few people know how to get mirroring working in infra and those few are super busy with other things | 12:53 |
| *** senk has quit IRC | 12:55 | |
| sdake | SamYaple are you on 3rd shift or whatever it was again? | 12:56 |
| SamYaple | nah. im just around | 12:56 |
| sdake | ya - i typically wake up around this time of day | 12:56 |
| sdake | 5-6am | 12:56 |
| sdake | if i go to bed at 9-10 | 12:56 |
| SamYaple | yea my sleep is rotating atm | 12:56 |
| sdake | so busy with holidays barely have time to do much work | 12:57 |
| *** egonzalez90_ has quit IRC | 12:57 | |
| SamYaple | btw sdake kolla-salt is coming along. want to make it official deliverable. dont know what kollas latest process is on that | 12:58 |
| sdake | SamYaple to do that, would have to talk to inc0, we would vote on it as a community - just as we have in the past | 12:58 |
| sdake | SamYaple the voting is a little fuzzy to me, as we have multiple core teams now | 12:58 |
| sdake | SamYaple in other words we don't really ahve a policy for that | 12:59 |
| srwilkers_zzzz | morning sdake SamYaple | 12:59 |
| sdake | SamYaple i guess what makes most sense is all core reviewer teams would vote | 12:59 |
| SamYaple | morning srwilkers_zzzz still tired? | 12:59 |
| sdake | SamYaple we would probably first need to have a vote to sort that problem out | 12:59 |
| SamYaple | sdake: idk about all core teams. why should ansible or k8s (or salt) teams vote on a... puppet repo? | 12:59 |
| sdake | something like "For the addition of deliverables, all core teams must vote" | 12:59 |
| sdake | it creates strain on everyone to add a deliverables | 13:00 |
| sdake | deliverable | 13:00 |
| SamYaple | it creates a strain on the kolla-core team | 13:00 |
| *** skramaja_ has joined #openstack-kolla | 13:00 | |
| sdake | SamYaple that team has been split into 3 separate core reviewer teams - check out gerrit | 13:00 |
| SamYaple | i dont know if an -ansible only core would care what k8s does | 13:00 |
| *** skramaja has quit IRC | 13:00 | |
| sdake | right, we dont care | 13:00 |
| sdake | however, we do care when adding deliverables | 13:01 |
| SamYaple | sdake: yea the kolla-core team still exists, no? | 13:01 |
| SamYaple | there is also kolla-ansible and kolla-k8s | 13:01 |
| sdake | right kolla-core in the future will only be responsible for the kolla deliverable not all of kolla | 13:01 |
| *** srwilkers_zzzz is now known as srwilkers | 13:01 | |
| sdake | sup srwilkers | 13:01 |
| srwilkers | not much sdake | 13:02 |
| sdake | morning :) | 13:02 |
| sbezverk | good morning | 13:02 |
| sdake | sup sbezverk | 13:02 |
| srwilkers | SamYaple, always tired | 13:02 |
| srwilkers | morning sbezverk :) | 13:02 |
| *** srwilkers has quit IRC | 13:02 | |
| sdake | SamYaple does that make sense? | 13:03 |
| *** chandankumar has quit IRC | 13:03 | |
| SamYaple | sdake: right, but i feel only the kolla-core team will be impacted by a new kolla-* delieverable realistically | 13:03 |
| sdake | I think the best course of action (not being the ptl) is | 13:03 |
| sdake | 1. vote on _who_ can vote on the addition of deliverables | 13:03 |
| SamYaple | yea im not doing anything with the repo side until new year | 13:03 |
| sdake | 2. once 1 is settled, vote on the addition of kolla-salt | 13:03 |
| SamYaple | im working out of my repo until then so i dont realyl care | 13:03 |
| sdake | nah it does effect everyone | 13:03 |
| SamYaple | no i mean im not _blocked_ at the moment | 13:03 |
| sdake | right | 13:03 |
| SamYaple | by not being in kolla namespace | 13:04 |
| SamYaple | and thats all i care about at the moment | 13:04 |
| sdake | inc0 is our ptl, so he should have the ability to get this rolling :) | 13:04 |
| SamYaple | yep. new years is fine with me | 13:04 |
| sdake | I can do so as well - anyone in the community can submit policy votes | 13:04 |
| SamYaple | by that time, i should have all the core services implemented | 13:04 |
| sdake | inc0 has lost alot of state about the project progress in the last 3 weeks during his cross country move | 13:05 |
| sdake | or 2 weeks | 13:05 |
| sdake | or whatever it is | 13:05 |
| sdake | plus the holiday shutdown | 13:05 |
| *** jimmygc has quit IRC | 13:05 | |
| sdake | so he will need to catch up :) | 13:05 |
| sdake | he did complain about not making policy changes while he was on the road via vote, and it is hte holidays as well - some people not checking their email | 13:06 |
| sdake | i think waiting until the 4th-ish to get things rolling is prudent | 13:06 |
| SamYaple | yea im not worried. to be honest, official repo (with gerrit) would really slow me down at the moment | 13:07 |
| SamYaple | you know how initial projects are | 13:07 |
| sdake | ya - need reviewers | 13:07 |
| sdake | which makes it harder | 13:07 |
| SamYaple | need reviewers, changes happen quickly | 13:07 |
| SamYaple | not worth it until i have something concrete in my opinion | 13:07 |
| sdake | that is where we were with kolla-kubernetes about 6 months ago | 13:07 |
| sdake | had no reviewers | 13:07 |
| sdake | now we have a cor eteam that can actually process reviews in a timely manner :) | 13:08 |
| SamYaple | well if i can show a working deploy, that goes along way to get people testing/reviewing | 13:08 |
| sdake | SamYaple there is a risk with that if you write all the code nobody will want to play ;) | 13:08 |
| *** hfu has joined #openstack-kolla | 13:08 | |
| *** senk has joined #openstack-kolla | 13:08 | |
| SamYaple | highly doubt that ;) | 13:08 |
| berendt | I want to play with it.. :) | 13:08 |
| SamYaple | wrote all the ansible code first too | 13:08 |
| SamYaple | people seemed to like that | 13:08 |
| sdake | SamYaple ya - however, it went in throug hthe normal review process | 13:08 |
| sdake | and required alot of rework iirc ;) | 13:08 |
| portdirect_away | hay sdake o/ | 13:09 |
| sdake | and some jedi magic on my end to get the core reviewt eam on board | 13:09 |
| SamYaple | nah the initial code was copy-pasta | 13:09 |
| sdake | sup portdirect_away | 13:09 |
| portdirect_away | hey | 13:09 |
| *** portdirect_away is now known as portdirect | 13:09 | |
| SamYaple | and people blind reviewed lol | 13:09 |
| sdake | i dont think it was totally blind | 13:09 |
| SamYaple | nah not always | 13:09 |
| *** senk has quit IRC | 13:09 | |
| SamYaple | sometimes though | 13:09 |
| SamYaple | at least initially | 13:09 |
| sdake | but now the core review team is like "WTF why didn't we do that earlier" :) | 13:09 |
| SamYaple | once i got it working, reviewers were way better | 13:09 |
| sdake | i wasn't blind reviewing, can't speak for others | 13:09 |
| sdake | although I was learning ansible | 13:10 |
| sdake | so my reviews may have been imperfect | 13:10 |
| SamYaple | HAH! there was more than one review where you specifically said "i dont understand this +2+1" | 13:10 |
| sdake | yup - to move things along | 13:10 |
| SamYaple | well thats what i mean by blind reviewing | 13:10 |
| sdake | when our core team was minimal | 13:10 |
| sdake | oh - atleast I admitted it :) | 13:10 |
| SamYaple | nah im not putting anyone down | 13:10 |
| SamYaple | im just saying I know people will want kolla-salt. im not worried | 13:11 |
| sdake | i dont do that now, if i don't understand it totally and completely i +1 it | 13:11 |
| sdake | sometimes i do that for kolla-kubernetes to move the reviews along tho :) | 13:11 |
| sdake | kolla and kolla-ansible deliverables have vast review team | 13:12 |
| *** lamt has quit IRC | 13:12 | |
| sdake | well atleast large enough to get the job done :) | 13:12 |
| portdirect | how often do we push images to docker hub? | 13:13 |
| SamYaple | portdirect: sometime between never and no | 13:13 |
| portdirect | SamYaple: you really know how to upset me :( | 13:14 |
| sdake | portdirect ya - dockerhub is not super well maintained | 13:14 |
| sdake | portdirect its a shame really - need more minerals! | 13:14 |
| SamYaple | portdirect: if it makes you feel better, my plan was to fix the docker images _first_ and then push :) | 13:14 |
| SamYaple | portdirect: so thats two cycles away | 13:14 |
| portdirect | SamYaple: can we move it to some time between never and sometimes? :) | 13:14 |
| SamYaple | portdirect: go for it! | 13:14 |
| sdake | portdirect what specific issue do you face | 13:15 |
| sdake | portdirect howdy btw :) | 13:15 |
| sdake | i'd be more interested in getting our security analysis done | 13:15 |
| portdirect | A few images are broken (kolla-toolbox) and would like to get them sorted if poss | 13:15 |
| sdake | so we can be VMT managed | 13:16 |
| sdake | portdirect when did that timer start on the core reviewer team vote | 13:16 |
| portdirect | wed | 13:16 |
| sdake | portdirect no vetos atm, and there is a majority | 13:16 |
| bjolo | portdirect, whats broken in kolla-toolbox? | 13:16 |
| sdake | portdirect i can add you wed assuming there are no vetos | 13:17 |
| sdake | which i assume | 13:17 |
| portdirect | thats for kolla-k8s though, not kolla? | 13:17 |
| bjolo | ah ok | 13:17 |
| sdake | portdirect note i am hesistent to push things to the dockerhub at this time - because the rdo upstream is busted in some way | 13:17 |
| sdake | if you were to push, only push the toolbox | 13:17 |
| portdirect | bjolo: openstack cli, shouldnt affect I think. | 13:17 |
| sdake | portdirect which toolbox code are you talking about | 13:17 |
| sdake | oh you need a new version of openstack cli? | 13:18 |
| portdirect | yup | 13:18 |
| sdake | why? | 13:18 |
| sdake | for the 3.0.2 deliverable? | 13:18 |
| sdake | rather milestone | 13:18 |
| sdake | 3.0.3 should be out shortly | 13:18 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Split config list into lines https://review.openstack.org/412451 | 13:18 |
| sdake | its been tagged, needs to be pushed to the hub | 13:19 |
| openstackgerrit | Martin André proposed openstack/kolla: Kafka image is expected to fail on Ubuntu binary https://review.openstack.org/412452 | 13:19 |
| SamYaple | sdake: https://review.openstack.org/#/c/412377/ https://review.openstack.org/#/c/412378/ | 13:19 |
| SamYaple | sdake: those definetely need to merge before tag | 13:19 |
| portdirect | sdake: oh cool - it thats in progrees I'll hang on, we are doing this at the moment: https://github.com/openstack/kolla-kubernetes/blob/master/helm/kolla-common/templates/_common_create_keystone_endpoint.yaml#L28 | 13:19 |
| SamYaple | sdake: prepare to be upset | 13:19 |
| *** msimonin has joined #openstack-kolla | 13:19 | |
| portdirect | which works short tem but makes me sad :( | 13:20 |
| sdake | SamYaple jeffrey4l is our release liason | 13:20 |
| sdake | SamYaple the tag was done on the 15th | 13:20 |
| SamYaple | esh | 13:20 |
| sdake | SamYaple oh you mean the z streams | 13:20 |
| SamYaple | yea | 13:20 |
| *** senk has joined #openstack-kolla | 13:20 | |
| sdake | SamYaple not sure if jeffrey4l tagged the zstream - let me check the repos | 13:20 |
| Jeffrey4l | np. now. | 13:21 |
| Jeffrey4l | no* | 13:21 |
| SamYaple | if he has, this warrants a new z tag | 13:21 |
| Jeffrey4l | should we test the stable branch before making new tag? | 13:21 |
| Jeffrey4l | sdake SamYaple ^^ | 13:21 |
| sdake | https://github.com/openstack/releases/blob/master/deliverables/newton/kolla.yaml#L40 | 13:22 |
| sdake | SamYaple if you check ou thte blame, i was the last person to tag a zstream for 3.0.1 | 13:22 |
| sdake | SamYaple therefore, a zstream tag may still be done for 3.0.2 | 13:22 |
| sdake | with those reviews | 13:23 |
| SamYaple | what? | 13:23 |
| SamYaple | no sdake i just found a big security bug yesterday | 13:23 |
| SamYaple | introduced in...... kolla-mesos! | 13:23 |
| SamYaple | ugh | 13:23 |
| sdake | GAH | 13:23 |
| SamYaple | i know right? | 13:23 |
| sdake | how did /usr/bin/install end up in the sudoers file | 13:24 |
| *** senk has quit IRC | 13:24 | |
| SamYaple | sdake: https://review.openstack.org/#/c/241438/ | 13:24 |
| SamYaple | fix merged in master | 13:24 |
| SamYaple | big 'ol mesos patch | 13:24 |
| sdake | hmm hidden away | 13:25 |
| sdake | oh well | 13:25 |
| SamYaple | basically drop-root hasnt been a thing since implementation lol | 13:25 |
| sdake | Jeffrey4l please ack https://review.openstack.org/#/c/412377/1 | 13:25 |
| SamYaple | sdake: there are two of them | 13:25 |
| Jeffrey4l | done | 13:25 |
| SamYaple | https://review.openstack.org/#/c/412377/1 https://review.openstack.org/#/c/412378/1 | 13:25 |
| SamYaple | these are backports | 13:25 |
| *** zhubingbing_ has joined #openstack-kolla | 13:26 | |
| sdake | Jeffrey4l is our gate broken: https://review.openstack.org/#/c/412378/1 for mitaka? | 13:26 |
| Jeffrey4l | yes. | 13:27 |
| sdake | ok | 13:27 |
| openstackgerrit | Pete Birley proposed openstack/kolla: Add Kubernetes Images for Kolla-Kubernetes https://review.openstack.org/412205 | 13:27 |
| sdake | Jeffrey4l is this because of the rdo breakage upstream? | 13:27 |
| SamYaple | sdake: looks like it deploys fully | 13:28 |
| SamYaple | ah nvm | 13:28 |
| Jeffrey4l | no idea. nova-libvirt failed with some issue. | 13:28 |
| SamYaple | it breaks on novaclient | 13:28 |
| sdake | ya - novalibvirt is busted | 13:28 |
| openstackgerrit | Merged openstack/kolla-kubernetes: RAX partprobe fix https://review.openstack.org/412244 | 13:28 |
| sdake | dmsimard posted on the mialing list about it | 13:28 |
| Jeffrey4l | http://logs.openstack.org/93/410593/1/check/gate-kolla-dsvm-deploy-centos-binary-centos-7-nv/23cb517/logs/kolla/libvirt/libvirtd.log | 13:28 |
| SamYaple | sdake: http://logs.openstack.org/78/412378/1/check/gate-kolla-dsvm-deploy-centos-binary-centos-7-nv/3cae562/console.html#_2016-12-19_10_17_39_258914 | 13:28 |
| sdake | moment let me see if i can find it | 13:28 |
| sdake | maybe there is a workaround or something | 13:28 |
| openstackgerrit | Pete Birley proposed openstack/kolla: Add Kubernetes Images for Kolla-Kubernetes https://review.openstack.org/412205 | 13:28 |
| SamYaple | sdake: no look again, its novaclient thats busted | 13:28 |
| SamYaple | sdake: it successfully launches instance and breaks on polling | 13:29 |
| sdake | SamYaple ya - returning an error from libvirtd | 13:29 |
| Jeffrey4l | SamYaple, do not think so. check this "No valid host was found. There are not enough hosts available." | 13:29 |
| SamYaple | im not seeing that | 13:29 |
| Jeffrey4l | check this http://logs.openstack.org/93/410593/1/check/gate-kolla-dsvm-deploy-centos-binary-centos-7-nv/23cb517/logs/kolla/nova/nova-compute.log | 13:29 |
| *** ipsecguy_ is now known as ipsecguy | 13:29 | |
| SamYaple | oh jeez yea | 13:29 |
| sdake | SamYaple its a centos virtualization sig problem only from what i understand | 13:29 |
| SamYaple | we need clearer logs | 13:29 |
| SamYaple | thanks for the link | 13:30 |
| sdake | SamYaple n odoubt - have to be a rocket scientist to detect what is wrong with the gate :) | 13:30 |
| SamYaple | lol yea right | 13:30 |
| sdake | ok - libvirt email moment | 13:30 |
| sdake | maybe that will give people something to work with that don't have a holiday shutdown | 13:30 |
| SamYaple | btw Jeffrey4l. right now for the static uid patch, I can move user homedir creation and setting shell into individual containers | 13:31 |
| openstackgerrit | Merged openstack/kolla: Switch Ubuntu to overlay2 docker driver https://review.openstack.org/412255 | 13:31 |
| SamYaple | Jeffrey4l: if thats desired. so we dont autocreate any home directory | 13:31 |
| sdake | subject is: [openstack-dev] [all] Known issue with CentOS 7.3 and qemu-kvm(-ev)2.6.0 | 13:31 |
| Jeffrey4l | how about the user creation? SamYaple | 13:32 |
| SamYaple | Jeffrey4l: what about user creation? | 13:32 |
| sdake | Jeffrey4l note - we have it in mitaka | 13:32 |
| sdake | Jeffrey4l i'm surprised master gate is working | 13:32 |
| * Jeffrey4l is checking the mail. | 13:32 | |
| sdake | Jeffrey4l lets focus on the qemu-kvm thing for a moment ;) | 13:32 |
| Jeffrey4l | the issue exist before centos 7.3 release. | 13:32 |
| sdake | i see | 13:32 |
| sdake | mitaka gate broke when? | 13:33 |
| Jeffrey4l | i am not sure. | 13:33 |
| sdake | roger | 13:33 |
| sdake | wonderif infra has a way to get that info :) | 13:33 |
| SamYaple | sdake: looks like master just borke too https://review.openstack.org/#/c/412231/ | 13:33 |
| sdake | ya- everything is probably busted | 13:33 |
| sdake | this is why voting gates are not optimal for the deploy jobs | 13:33 |
| sdake | unless we have mirroring :) | 13:34 |
| Jeffrey4l | SamYaple, where the user creation will be? base or each containers | 13:34 |
| SamYaple | Jeffrey4l: user creation is in base, but we can _configure_ the user (shell, homedir) in individual containers. | 13:34 |
| Jeffrey4l | SamYaple, that PS is different. | 13:34 |
| SamYaple | we dont have to, but it might be nice | 13:34 |
| sdake | our gate jobs use qemu hypervisor Jeffrey4l | 13:34 |
| Jeffrey4l | but the nova-libvirt issue happens now and then in master. | 13:34 |
| *** v1k0d3n has joined #openstack-kolla | 13:35 | |
| SamYaple | Jeffrey4l: ok ill work on it | 13:35 |
| sdake | Jeffrey4l alhtough pavo was mentioining pip install wasn't working very well recently either | 13:35 |
| sdake | software updates ftw ;) | 13:35 |
| Jeffrey4l | sdake, any logs for this? i do not see it. sdake | 13:36 |
| *** skramaja has joined #openstack-kolla | 13:36 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 13:36 |
| sdake | Jeffrey4l what logs do you need? | 13:36 |
| sdake | morning sbezverk :) | 13:36 |
| *** skramaja_ has quit IRC | 13:36 | |
| Jeffrey4l | "alhtough pavo was mentioining pip install wasn't working very well recently either" sdake | 13:36 |
| sdake | Jeffrey4l dunno feew days ago | 13:36 |
| Jeffrey4l | ok. | 13:36 |
| sbezverk | sdake: morning | 13:36 |
| sdake | Jeffrey4l can't be everywhere all the time :) | 13:37 |
| sdake | Jeffrey4l lets fix gate first | 13:37 |
| sdake | i think thatsolution is easy | 13:37 |
| Jeffrey4l | eyp | 13:37 |
| Jeffrey4l | yep | 13:37 |
| sdake | cpu_model = host_type - is that being sset? | 13:38 |
| *** portdirect has quit IRC | 13:38 | |
| sdake | (is it hte default for libvirt_type=qemu? | 13:39 |
| Jeffrey4l | no idea. | 13:39 |
| *** jheroux has joined #openstack-kolla | 13:39 | |
| sdake | Jeffrey4l there is a way to find out right - via looking at the startup of the container in qemu mode | 13:39 |
| sdake | or looking at code | 13:39 |
| sdake | moment let me look at the code see if I can decipher the mysteries of nova | 13:40 |
| sdake | i know we do not set cpu_mode | 13:40 |
| sdake | Jeffrey4l did you pull of dmsimard 's email? | 13:40 |
| Bico_Fino | Jeffrey4l: Hi! Tried that patch for ironic-inspector, updated the info on bugs page: 1650345 | 13:40 |
| Jeffrey4l | start up of container in qemu mode? or start up vm in qemu mode? | 13:40 |
| sdake | Jeffrey4l if you start up kolla with virt_type=qemu what is host_mode? | 13:41 |
| Jeffrey4l | /me is checking. | 13:41 |
| sdake | thanks :) | 13:41 |
| Jeffrey4l | Bico_Fino, no :( no time for that. | 13:41 |
| dmsimard | sdake: it's not host_mode, it's cpu_mode | 13:41 |
| Bico_Fino | Jeffrey4l: np. :) | 13:41 |
| sdake | Bico_Fino link to me plz - i'll take a look | 13:41 |
| sdake | dmsimard sorry cpu_mode | 13:41 |
| Bico_Fino | sdake: https://bugs.launchpad.net/kolla/+bug/1650345 | 13:41 |
| openstack | Launchpad bug 1650345 in kolla "Ironic Inspector DBAPIError" [High,Confirmed] - Assigned to Jeffrey Zhang (jeffrey4l) | 13:41 |
| sdake | dmsimard right, so we don't set that at all, which means we take the nova default - which may be cpu_mode=host-mode | 13:42 |
| dmsimard | host-model* yes, which with qemu triggers the libvirt issue. | 13:42 |
| sdake | ya sorry host_model = 6am - brain still rebooting :) | 13:42 |
| dmsimard | cpu_mode needs to be set to none | 13:42 |
| sdake | dmsimard yup i read that email, I'm wondering what the default is which Jeffrey4l is checking | 13:43 |
| sdake | dmsimard might put [rdo] tags in email in the future so people can filter on them | 13:43 |
| sdake | dmsimard I don't think anyone would find that offensive - but rather helpful | 13:43 |
| dmsimard | sdake: well this is not an RDO issue, it's a CentOS 7.3 issue | 13:43 |
| dmsimard | I did post it to rdo-list as well, though | 13:44 |
| sdake | dmsimard yup - not everyone on that list althoug htye shoudl be ;) | 13:44 |
| Jeffrey4l | tl;dr; but at least in master, the gate works ( may red now and then ). so we may not meet this issue. | 13:44 |
| dmsimard | If you install from source on CentOS 7.3 you'd have the issue | 13:44 |
| Bico_Fino | When 3.0.2 will be released? | 13:44 |
| sdake | Bico_Fino - probably shortly after the holidays | 13:44 |
| sdake | Bico_Fino just a guess tho | 13:45 |
| Bico_Fino | sdake: thanks. ;) | 13:45 |
| Jeffrey4l | in the gate configuration, we use virt_type=qemu and no cpu_mode. | 13:45 |
| sdake | Jeffrey4l right - which means cpu_mode is taking a default | 13:45 |
| sdake | what is that default? | 13:45 |
| *** erhudy has joined #openstack-kolla | 13:45 | |
| sdake | dmsimard i recognize its a virtualization sig problem, but affects rdo in a negative way - would help raise visibility that rdo is affected by rdo transitive dependencies :) | 13:46 |
| Jeffrey4l | If virt_type="kvm|qemu", it will default to "host-model", otherwise it will | 13:46 |
| Jeffrey4l | default to "none". | 13:46 |
| Jeffrey4l | hrm, default is host-model. | 13:46 |
| sdake | right | 13:46 |
| sdake | so that needs to be hard coded to none | 13:46 |
| sdake | thanks dmsimard :) | 13:46 |
| dmsimard | sdake: it's not a virtualization sig issue, it's a qemu-kvm problem, even upstream RHEL has the issue | 13:46 |
| sdake | (in the gate) | 13:46 |
| dmsimard | just setting facts straight :p | 13:47 |
| Jeffrey4l | anyway, i do not think it is the same issue. i will push a test patch to collect more logs to check. | 13:47 |
| dmsimard | there are various bugs upstream that tracks the issue | 13:47 |
| sdake | dmsimard cool - well whoever's issue it is, a tag in the mailing list would help :) | 13:47 |
| dmsimard | I'll tag RDO next time. | 13:47 |
| sdake | dmsimard thanks :) | 13:47 |
| * sdake goes to setup an rdo filter tag | 13:47 | |
| *** mdnadeem has quit IRC | 13:49 | |
| *** srwilkers has joined #openstack-kolla | 13:50 | |
| *** mattmceuen_away is now known as mattmceuen | 13:50 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: [wip] enable ceilometer related project in gate https://review.openstack.org/412278 | 13:53 |
| sdake | Jeffrey4l are you working on testing that gate workaround that dmsimard brought up? | 13:54 |
| openstackgerrit | Merged openstack/kolla: Remove /usr/bin/install from allowed commands https://review.openstack.org/412377 | 13:55 |
| openstackgerrit | Merged openstack/kolla: Remove /usr/bin/install from allowed commands https://review.openstack.org/412378 | 13:55 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: NOMERGE: DEBUG MITAKA BRANCH https://review.openstack.org/412459 | 13:55 |
| sdake | Jeffrey4l or shall I | 13:55 |
| sdake | I think its as simple as copying a file to /etc/kolla/config/nova.conf | 13:55 |
| Jeffrey4l | sdake, yep. i tried push a patch in mitaka to collect logs. just a cherry-pick, tbh. | 13:55 |
| sdake | we already do this to override libvirt | 13:55 |
| sdake | we can override one more value for the moment | 13:55 |
| Jeffrey4l | sdake, i am just collect logs. no workaround. | 13:56 |
| Jeffrey4l | hrm, i still do think it is cause by cpu_model ;( | 13:57 |
| Jeffrey4l | do not think* | 13:57 |
| Jeffrey4l | btw, could u ack this review? https://review.openstack.org/411403 it collect some useful logs in gate. and i prefer to backport it into stable branch sdake | 13:59 |
| srwilkers | hey mattmceuen o/ | 14:03 |
| mattmceuen | GM srwilkers! | 14:03 |
| *** chandankumar has joined #openstack-kolla | 14:04 | |
| *** hfu has quit IRC | 14:06 | |
| sdake | Jeffrey4l reviewing - give me 10-15 - lots of bits in here | 14:08 |
| Jeffrey4l | roger. thanks. | 14:08 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 14:10 |
| sdake | Jeffrey4l done | 14:10 |
| sdake | Jeffrey4l looks good dude - nice work there ! | 14:10 |
| Jeffrey4l | thanks. | 14:10 |
| sdake | Jeffrey4l note i did leave one comment if you want to keep it in your backlog | 14:11 |
| sdake | ok back on gate jobs :) | 14:11 |
| *** senk has joined #openstack-kolla | 14:11 | |
| sdake | Jeffrey4l if dmsimard says its broken and red hat says its broken, its broken | 14:11 |
| Jeffrey4l | sdake, may be i need use a better name. collect_info? | 14:11 |
| sdake | Jeffrey4l since its a gate job, lets just fix it | 14:11 |
| sdake | Jeffrey4l well that patch merged - what I was thinking is in the future we may have a need to add more sudoers entries | 14:12 |
| *** lamt has joined #openstack-kolla | 14:12 | |
| sdake | and and finding that particular sudoers in the gate scripts sort of like needle in haystack ;) | 14:12 |
| *** senk has quit IRC | 14:13 | |
| Jeffrey4l | hrm, it is host OS sudoers not container. we only added one item ( jenkins ) . and i do no think we will add others in the future. | 14:13 |
| Jeffrey4l | but i do not see the error message: "qemu-kvm: CPU feature arat not found" | 14:14 |
| Jeffrey4l | sdake, ^^ | 14:14 |
| sdake | Jeffrey4l i understand that | 14:15 |
| sdake | Jeffrey4l lets just try it atleast :) | 14:15 |
| sdake | Jeffrey4l ok - well in future if there are, you know where to look :) | 14:15 |
| Jeffrey4l | sdake, ok. i will push that workaround into mitaka branch. | 14:15 |
| Jeffrey4l | roger. | 14:15 |
| sdake | so we do not see hte cpu feature not found in the gate scrtips | 14:15 |
| sdake | but it may be that it manifests in some different way across the 10 different cloud providers used in openstack | 14:16 |
| *** shardy_lunch is now known as shardy | 14:16 | |
| sdake | they are all subtly different | 14:16 |
| sdake | sup shardy | 14:16 |
| sdake | ok - have meeting in 10 mins | 14:16 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: NO_MERGE: TRY CPU_MODEL WORKAROUND https://review.openstack.org/412470 | 14:20 |
| *** sdake_ has joined #openstack-kolla | 14:20 | |
| *** sdake has quit IRC | 14:23 | |
| *** srwilkers is now known as srwilkers_coffee | 14:28 | |
| *** fguillot has joined #openstack-kolla | 14:29 | |
| *** liyifeng has quit IRC | 14:38 | |
| *** gomarivera has joined #openstack-kolla | 14:38 | |
| *** mnasiadka has quit IRC | 14:38 | |
| SamYaple | i need someone running centos-binary that has build all the images to run a command for me | 14:40 |
| SamYaple | anyone around that has all the images built? | 14:40 |
| SamYaple | they can be old images | 14:41 |
| *** gomarivera has quit IRC | 14:42 | |
| *** yingjun has quit IRC | 14:44 | |
| *** TxGirlGeek has joined #openstack-kolla | 14:45 | |
| openstackgerrit | Merged openstack/kolla: Collect logs at the end of gate https://review.openstack.org/411403 | 14:45 |
| *** jtriley has joined #openstack-kolla | 14:50 | |
| *** v1k0d3n has quit IRC | 14:51 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 14:56 |
| *** mdnadeem has joined #openstack-kolla | 14:57 | |
| *** msimonin has quit IRC | 14:58 | |
| *** msimonin has joined #openstack-kolla | 14:58 | |
| *** msimonin has quit IRC | 14:58 | |
| *** adrian_otto has joined #openstack-kolla | 14:58 | |
| *** srwilkers_coffee is now known as srwilkers | 15:02 | |
| *** adrian_otto has quit IRC | 15:03 | |
| *** msimonin has joined #openstack-kolla | 15:06 | |
| *** Pavo has joined #openstack-kolla | 15:09 | |
| *** tovin07_ has joined #openstack-kolla | 15:10 | |
| *** msimonin has quit IRC | 15:12 | |
| *** msimonin has joined #openstack-kolla | 15:12 | |
| *** msimonin has quit IRC | 15:13 | |
| *** Pavo has quit IRC | 15:14 | |
| *** msimonin has joined #openstack-kolla | 15:17 | |
| *** eaguilar has joined #openstack-kolla | 15:17 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 15:27 |
| SamYaple | sbezverk: is kolla-k8s going to do mullinode maraidb? | 15:29 |
| sbezverk | SamYaple: not atm.. I think rhallisey is playing with galera | 15:29 |
| SamYaple | ok. id be curious how youd pull that off in k8s to be honest | 15:30 |
| sbezverk | SamYaple: att folks already run galera on 3 nodes without any issues afaik | 15:31 |
| SamYaple | sbezverk: i mean the orchestration of setting up 3 node galera | 15:33 |
| *** chandankumar has quit IRC | 15:33 | |
| SamYaple | corner cases too, like supporting single node and multinode, growing from single to multi etc | 15:33 |
| sbezverk | SamYaple: I see, I think these aspects are not yet covered, but in near future we will have to.. | 15:34 |
| *** mattmceuen is now known as mattmceuen_away | 15:36 | |
| *** msimonin has quit IRC | 15:38 | |
| *** zhubingbing_ has quit IRC | 15:41 | |
| *** saneax is now known as saneax-_-|AFK | 15:43 | |
| *** lrensing has joined #openstack-kolla | 15:47 | |
| sdake_ | kfox1111 ping | 15:49 |
| *** lbeliveau has quit IRC | 15:50 | |
| *** lbeliveau has joined #openstack-kolla | 15:51 | |
| sbezverk | sdake_: mariadb service if running :-) | 15:51 |
| sdake_ | NICE | 15:51 |
| sdake_ | IF/is :) | 15:51 |
| sbezverk | is | 15:52 |
| srwilkers | sbezverk, thats awesome man | 15:53 |
| sdake_ | sbezverk is it in a gerrit review - or not quite yet? | 15:54 |
| srwilkers | sdake_, https://review.openstack.org/#/c/412237/14 | 15:54 |
| *** lrensing has quit IRC | 15:54 | |
| sbezverk | sdake_: yes but it is not supposed to be merged atm we need to decide how it is going to be tested | 15:54 |
| sbezverk | sdake_ kfox1111 built nice infra, we just need to put it in a probably separate job | 15:56 |
| sdake_ | sbezverk yup i hear ya | 15:56 |
| *** lrensing has joined #openstack-kolla | 15:56 | |
| *** chandankumar has joined #openstack-kolla | 16:00 | |
| *** sp__ has joined #openstack-kolla | 16:07 | |
| *** tovin07_ has quit IRC | 16:07 | |
| sp__ | sdake_: ping..... | 16:08 |
| openstackgerrit | Martin André proposed openstack/kolla: Download kafka binaries from within the container https://review.openstack.org/412452 | 16:08 |
| sdake_ | sp__ whats up? | 16:11 |
| rsjethani | hi can anyone help me with cinder with lvm issue? | 16:11 |
| sp__ | sdake_: today my US visa application got disapproved | 16:12 |
| rsjethani | i am gettinh config name lvm-1 uninitialzed error in cinder-volume logs | 16:12 |
| sdake_ | sp_ sadface :( | 16:13 |
| sp__ | sdake_: is there any plan for virtual PTG | 16:13 |
| sp__ | ? | 16:13 |
| sdake_ | sp__ unknown - i think the best person to ask on that topic is inc0 when he returns today hopefully :) | 16:13 |
| sp__ | sdake_: thanks :( | 16:13 |
| *** manjeets has quit IRC | 16:14 | |
| *** pcaruana has quit IRC | 16:15 | |
| *** chandankumar has quit IRC | 16:16 | |
| *** manjeets has joined #openstack-kolla | 16:17 | |
| rsjethani | sdake_ can you help me with the above issue? | 16:17 |
| *** mattmceuen_away is now known as mattmceuen | 16:21 | |
| *** msimonin has joined #openstack-kolla | 16:21 | |
| *** sayantan_ has joined #openstack-kolla | 16:24 | |
| *** bmace has quit IRC | 16:25 | |
| kfox1111 | morning. | 16:25 |
| *** williamcaban has joined #openstack-kolla | 16:25 | |
| *** bmace has joined #openstack-kolla | 16:25 | |
| *** eaguilar has quit IRC | 16:25 | |
| kfox1111 | sbezverk: sdake_: ping | 16:27 |
| srwilkers | morning kfox1111 :) | 16:28 |
| sbezverk | kfox1111: hey good mornig | 16:28 |
| kfox1111 | srwilkers: sbezverk: morning. :) | 16:28 |
| sbezverk | kfox1111: so are we going to have separate gate job for service level testing? | 16:29 |
| kfox1111 | sbezverk: docker run -it --rm docker.io/kolla/centos-binary-kubernetes-entrypoint:3.0.1 /bin/bash | 16:29 |
| kfox1111 | kolla-entrypoint | 16:29 |
| kfox1111 | bash: kolla-entrypoint: command not found | 16:29 |
| kfox1111 | we seem to have a broken container. | 16:29 |
| kfox1111 | sbezverk: yeah. there's an experimental one already. | 16:29 |
| sbezverk | hm, it is working for me | 16:29 |
| sbezverk | [root@8f348489fe55 /]# | 16:30 |
| kfox1111 | kubernetes-entrypoint doesn't seem to be installed. | 16:30 |
| kfox1111 | the docker file says it should be in /usr/local/bin/ but nothing is there. :/ | 16:31 |
| kfox1111 | I tried to hook it into the neutron service package, and it was acting like entrypoint was just letting everything through. is really weird. | 16:31 |
| *** sayantan_ has quit IRC | 16:31 | |
| *** sayantan_ has joined #openstack-kolla | 16:32 | |
| kfox1111 | just tried pulling the ubuntu varient, and there isn't one. :/ | 16:32 |
| sbezverk | kfox1111: I tested entrypoint at one point and I saw it was holding on the resource when it was not available, for openvswitch | 16:32 |
| kfox1111 | did you build the container yourself? | 16:33 |
| sbezverk | that I do not rememeber, I might have | 16:33 |
| *** Serlex has quit IRC | 16:34 | |
| kfox1111 | know who built the one on the hub? | 16:35 |
| sbezverk | kfox1111: not sure, I am rebuilding on my local machine | 16:37 |
| sbezverk | to see if it is working or not | 16:37 |
| *** sp__ has quit IRC | 16:38 | |
| sbezverk | kfox1111: now I rememeber , it was inc0 | 16:38 |
| sbezverk | I asked him tu push one for us | 16:38 |
| kfox1111 | sbezverk: looking at the dockerfile, I thin kit should work. but the files listed just arn't in the one on the hub for some reason. | 16:39 |
| *** msimonin has quit IRC | 16:39 | |
| kfox1111 | ok. we should see if he can push an updated one. | 16:39 |
| kfox1111 | or, I guess I could try too. | 16:39 |
| sbezverk | yep it gets build fine | 16:41 |
| sbezverk | "/usr/local/bin/kubernetes-entrypoint" | 16:41 |
| sbezverk | I can pus it to docker if you want | 16:41 |
| kfox1111 | ok. I'll try to build too and push if it works. | 16:41 |
| kfox1111 | oh. ok. cool. :) | 16:41 |
| *** cinerama has joined #openstack-kolla | 16:43 | |
| sbezverk | kfox1111: should be at docker hub in 30 seconds, pushing | 16:43 |
| kfox1111 | awesome. thanks. :) | 16:44 |
| sbezverk | kfox1111: how do I extend experimental to test mariadb service, I got it running in all gate jobs | 16:44 |
| *** cinerama has quit IRC | 16:44 | |
| sbezverk | kfox1111: please try now | 16:44 |
| openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 16:44 |
| *** cinerama has joined #openstack-kolla | 16:44 | |
| kfox1111 | yup. looks better. thanks. :) | 16:48 |
| kfox1111 | rechecking in the gate. | 16:48 |
| kfox1111 | ok. so, for your gate question. | 16:48 |
| kfox1111 | I put hooks in for bypassing the openstack services so they can have service packages put in place. | 16:49 |
| kfox1111 | doing mariadb/memcached/rabbit/keystone will be trickier though. :/ | 16:49 |
| sbezverk | kfox1111: what if we have a completely separate job for services? | 16:50 |
| kfox1111 | since we're not quite merging yet, you could probably just launch the mariadb package right before the pv's get created and exit, and then we just don't merge it till we're ready to split out the ceph_workflow into multiple files? | 16:50 |
| *** msimonin has joined #openstack-kolla | 16:51 | |
| kfox1111 | yeah. I thin kwe should do that long term, but having it like it is lets us test the openstack service packages with their deps in place. | 16:51 |
| kfox1111 | like: | 16:51 |
| kfox1111 | https://review.openstack.org/#/c/412239/9/tests/bin/ceph_workflow.sh | 16:51 |
| sbezverk | kfox1111: well the only concern I have is if we keep these services longs enough it will be nightmare to go back and resolve merging conflicts | 16:52 |
| sbezverk | lots of unnecessary extra work | 16:53 |
| kfox1111 | I guess mariadb's one of the very first things though, so if you want to do the split in your review, we can rebase the service packages on the mariadb one as they get done, and eventually all will be over. | 16:53 |
| kfox1111 | yeah. I get that... | 16:53 |
| kfox1111 | well, lets see... | 16:53 |
| sbezverk | kfox1111: so you talk about splitting cepf workflow right? | 16:54 |
| kfox1111 | I was kind of hoping to get the rest of the microservices in before forking the file. | 16:54 |
| sbezverk | kfox1111: I see that make sense too | 16:54 |
| kfox1111 | thought it might be a bit easier. | 16:54 |
| kfox1111 | but... actually, there is so little left to do, and we have to touch it all anyway to add the service packages, maybe it doesn't matter at this point. | 16:54 |
| *** v1k0d3n has joined #openstack-kolla | 16:55 | |
| sbezverk | kfox1111: ok so to split or not to split (now) is the question :-) | 16:55 |
| v1k0d3n | morning all | 16:55 |
| v1k0d3n | cores, question... | 16:56 |
| kfox1111 | sbezverk: yeah. hmm.. | 16:56 |
| kfox1111 | v1k0d3n: morning | 16:56 |
| *** eaguilar has joined #openstack-kolla | 16:56 | |
| sbezverk | v1k0d3n: morning | 16:56 |
| v1k0d3n | who wants to attend a call with deis to discuss future, and gaps. | 16:56 |
| v1k0d3n | kevin you're not accepting messages. | 16:56 |
| v1k0d3n | but i need emails | 16:56 |
| v1k0d3n | then will coordinate cores from each of the projects. | 16:56 |
| sbezverk | v1k0d3n: sbezverk@cisco.com | 16:56 |
| *** rsjethani has quit IRC | 16:56 | |
| v1k0d3n | thanks sbezverk :) | 16:57 |
| kfox1111 | v1k0d3n: weird... not sure why not. I don't know irssi very well though. :/ | 16:57 |
| kfox1111 | v1k0d3n: Kevin.Fox@pnnl.gov | 16:57 |
| srwilkers | v1k0d3n, wilkers.steve@gmail.com | 16:58 |
| v1k0d3n | steve, probably use your att email? | 16:58 |
| srwilkers | not a core, but still interested | 16:58 |
| kfox1111 | v1k0d3n: thanks for orchestrating. :) | 16:58 |
| v1k0d3n | (that was my plan) | 16:58 |
| srwilkers | that'd work too i suppose | 16:58 |
| v1k0d3n | srwilkers: thought... | 16:58 |
| sdake_ | v1k0d3n stdake@cisco.com | 16:58 |
| v1k0d3n | sorry man. i can include both, just in case :) | 16:58 |
| sbezverk | kfox1111: so if we copy ceph workflow into ceph_workflow_service and then keep it in sync with new added microservices and gradually replacing with services | 16:59 |
| sdake_ | v1k0d3n i don't think it should be limited to core reviewers fwiw | 16:59 |
| v1k0d3n | kfox1111: absolutely. we all have something to gain from this meeting. | 16:59 |
| sdake_ | v1k0d3n all devs make contributions not just core reviewers | 16:59 |
| v1k0d3n | ok, sdake_ you want everyone from all communities to come? | 16:59 |
| *** msimonin has quit IRC | 17:00 | |
| sbezverk | it would just require an additional gate job to run it at least for now | 17:00 |
| sdake_ | v1k0d3n well we don't want a circus on our hands :) | 17:00 |
| v1k0d3n | i don't really care, but this is just to identify core contribs and architecture goals. | 17:00 |
| v1k0d3n | sdake_: that was the point of inviting cores. | 17:00 |
| sdake_ | v1k0d3n core reviewers :) | 17:00 |
| v1k0d3n | this doesn't "exclude" anyone. | 17:00 |
| kfox1111 | sbezverk: I'm ok with that plan. the gate job just needs to be tweaked very slightly to make that work. | 17:00 |
| sdake_ | i'm not quite sure how deis' governance model works, i guess they are called cores? | 17:01 |
| v1k0d3n | sdake_: well...you tell me what you want. | 17:01 |
| srwilkers | sbezverk, kfox1111: i think that plans good | 17:01 |
| v1k0d3n | since you brought up :) | 17:01 |
| *** sayantan_ has quit IRC | 17:01 | |
| srwilkers | sorry, tabbed back over and had to read through all the scroll back | 17:01 |
| sdake_ | v1k0d3n i'd say send an email to the openstack-dev ml [kolla] tag and ask for email addresses of anyone that wants to attend a deis conference | 17:01 |
| sdake_ | v1k0d3n sorry not conference conference call | 17:01 |
| sbezverk | kfox1111: let me start it but I will need your help with gate tweaking, since you are the main gate tweaked and keeper ;-) | 17:02 |
| kfox1111 | sbezverk: have a look at setup_gate.sh at the very bottom. | 17:02 |
| v1k0d3n | sdake_: i can let you arrange then. | 17:02 |
| kfox1111 | sbezverk: there's ceph_workflow, then an if for helm-entrypoint. | 17:02 |
| v1k0d3n | sorry this is community on the deis side. | 17:02 |
| sdake_ | v1k0d3n i can't arrange who attends from deis | 17:02 |
| v1k0d3n | so you're welcome to join. | 17:02 |
| sbezverk | kfox1111: will do thanks for pointer | 17:02 |
| kfox1111 | sbezverk: probably just need to rework that slightly to call the fork in an else and move the ceph workflow into the if. | 17:02 |
| sdake_ | v1k0d3n ok - so not being a jerk, struggling to get my kid to school atm, he is not feeling well | 17:03 |
| sdake_ | and i'm a little tired and hangry :) | 17:03 |
| srwilkers | sdake_, all good man. do what you gotta do :) | 17:03 |
| *** sayantan_ has joined #openstack-kolla | 17:03 | |
| sdake_ | v1k0d3n i can arrange a list of emails of people that want to attend from the kolla community | 17:03 |
| sdake_ | v1k0d3n and hand that list off to you | 17:04 |
| *** msimonin has joined #openstack-kolla | 17:04 | |
| sdake_ | v1k0d3n and you can arrange the whoel thing after that? :) | 17:04 |
| v1k0d3n | sdake_: i'm just a bit confused. | 17:05 |
| v1k0d3n | this is what i can do... | 17:06 |
| v1k0d3n | and how it will work | 17:06 |
| v1k0d3n | we are starting a formal engagement with Helm core developers to discuss gaps, plans/future, and engage with all of the projects using helm (openstack related; core develoers....architects who have the visions for the project). | 17:07 |
| v1k0d3n | i am going to give a few email addresses to lachlan, and he will email if you are interested. | 17:08 |
| v1k0d3n | to be clear, this is a zoom, like the other meetings held in the kubernetes community. | 17:08 |
| v1k0d3n | you are welcome to join...or not. | 17:08 |
| v1k0d3n | and anyone can attend. it's a community call. | 17:08 |
| v1k0d3n | but the cores can take back to this group. | 17:08 |
| *** mattmceuen is now known as mattmceuen_away | 17:08 | |
| v1k0d3n | i do not want to be responsible for that. | 17:08 |
| v1k0d3n | so i have three/four email addresses...i will use that, and figure that those folks can tell anyone else. | 17:09 |
| v1k0d3n | but in community fashion, this is to give equal time to all projects, equal voice, etc. the gaps will be similar in many ways. | 17:09 |
| kfox1111 | sbezverk: seems to be doing something different now. :) http://logs.openstack.org/39/412239/9/experimental/gate-kolla-kubernetes-deploy-centos-binary-helm-entrypoint-nv/a02bbf3/console.html | 17:11 |
| *** msimonin has quit IRC | 17:11 | |
| v1k0d3n | gtg guys. thanks kfox1111 sbezverk and srwilkers | 17:12 |
| v1k0d3n | be on the lookout soon | 17:12 |
| kfox1111 | v1k0d3n: np. thanks. l8r. | 17:12 |
| srwilkers | v1k0d3n, thanks man \o | 17:14 |
| sdake_ | v1k0d3n when is the meeting | 17:14 |
| v1k0d3n | sdake_: hasn't been decided. needs to be coordinated but this week. | 17:14 |
| v1k0d3n | i don't want it blocked because someone can't attend. as long as core folks can come...perfect. | 17:14 |
| kfox1111 | hmm.... does kubernetes-entrypoint do long polling or periodic polling? | 17:15 |
| sdake_ | v1k0d3n ok - need portdirect there too i'd assume :) | 17:15 |
| v1k0d3n | it's a community event. so... | 17:15 |
| sdake_ | right | 17:15 |
| v1k0d3n | but this week | 17:15 |
| *** sdake_ is now known as sdake | 17:15 | |
| sdake | like not christmas, right and not christmas eve? | 17:15 |
| sdake | as in tue or wed? | 17:15 |
| v1k0d3n | :) of course not. | 17:15 |
| *** sayantan_ has quit IRC | 17:15 | |
| sdake | ok | 17:15 |
| sdake | wfm then | 17:15 |
| sdake | moment let me hunt up pete's email | 17:15 |
| sdake | v1k0d3n pete@port.direct | 17:16 |
| sdake | not sure what dounghq's email is | 17:16 |
| sdake | moment i can probably find it | 17:16 |
| sdake | v1k0d3n i'll have to email you duounghq's email when he appears later today | 17:18 |
| *** sayantan_ has joined #openstack-kolla | 17:18 | |
| kfox1111 | so, to answer my own question, it polls at 2 second intervals. | 17:20 |
| kfox1111 | so... it shouldn't be taking so long... hmm.. | 17:21 |
| kfox1111 | ah. keystone svc is keystone-internal | 17:26 |
| SamYaple | kfox1111: | 17:27 |
| SamYaple | oops | 17:27 |
| SamYaple | disregard | 17:27 |
| kfox1111 | k. :) | 17:27 |
| SamYaple | or dont. im not your boss | 17:28 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 17:28 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 17:29 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 17:29 |
| *** msimonin has joined #openstack-kolla | 17:30 | |
| *** msimonin has quit IRC | 17:30 | |
| kfox1111 | sbezverk: looks good. | 17:31 |
| sbezverk | kfox1111: thanks.. I think with infra you built adding services will be super fun ;-) | 17:32 |
| *** papacz has quit IRC | 17:32 | |
| kfox1111 | +1. :) | 17:33 |
| kfox1111 | the neutron one's staring to come along. | 17:33 |
| kfox1111 | I'm really hoping the helm folks wil lschedule 2.1.1 soon. | 17:33 |
| kfox1111 | the global stuff looks like it will reallyl affect the microservice packages. so would be really nince to get that out of the way. | 17:34 |
| kfox1111 | we may just need to build helm from a specific commit temporarily? | 17:34 |
| *** papacz has joined #openstack-kolla | 17:34 | |
| *** x220_ has joined #openstack-kolla | 17:35 | |
| SamYaple | kfox1111: is this static uid thing going to affect k8s? | 17:35 |
| sbezverk | kfox1111: we could use canary at the gate | 17:35 |
| kfox1111 | SamYaple: yeah. | 17:35 |
| kfox1111 | sbezverk: I'm a little worried about doing that. they don't garantee not to break things. | 17:36 |
| kfox1111 | if we had somewhere to stash things, we could stash a tested canary though. | 17:36 |
| kfox1111 | but would probably hve to stash both the client and the server container. | 17:36 |
| sbezverk | kfox1111: understood, I am asking when it will get into the mainline | 17:36 |
| kfox1111 | its in trunk now. | 17:37 |
| *** mdnadeem has quit IRC | 17:37 | |
| kfox1111 | just not a release yet, and nothings scheduled. | 17:37 |
| kfox1111 | somewhere between now and feb. | 17:37 |
| *** newmember has quit IRC | 17:38 | |
| sbezverk | kfox1111: I do not know, we could ask as a favor to build "engineering special" 2.1.0 + 1 global fix | 17:38 |
| sbezverk | at cisco we sometime do it :-), not sure if helm has similar practice | 17:39 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Gate fix for helm sometimes failing https://review.openstack.org/412282 | 17:39 |
| *** msimonin has joined #openstack-kolla | 17:39 | |
| kfox1111 | sbezverk: yeah. that may be an option. | 17:40 |
| SamYaple | kfox1111: how will it affect you | 17:40 |
| kfox1111 | SamYaple: if someone upgrades from x to x+1 they may get wrong permissions and things break. | 17:40 |
| SamYaple | kfox1111: no i mean right now | 17:41 |
| *** papacz has quit IRC | 17:41 | |
| SamYaple | kfox1111: this static stuff should merge before end of cycle | 17:41 |
| SamYaple | wait | 17:41 |
| SamYaple | does kolla-k8s have a stable release yet? | 17:41 |
| kfox1111 | no stable release yet. | 17:41 |
| *** papacz has joined #openstack-kolla | 17:42 | |
| SamYaple | ok then it shouldnt affect _how_ you do things | 17:42 |
| *** msimonin has quit IRC | 17:42 | |
| kfox1111 | but doesn't really matter. containers are the thing that shoudl be stable, and not changing. | 17:42 |
| *** eaguilar has quit IRC | 17:42 | |
| *** eaguilar has joined #openstack-kolla | 17:43 | |
| sdake | kfox1111 a static mapping of UIDs wil not effect kolla-kubernetes | 17:45 |
| sdake | bbl | 17:45 |
| kfox1111 | sdake: if someone is useing persistent storage, it will. | 17:45 |
| sdake | there is no stable release | 17:45 |
| sdake | there is no upgrade needed | 17:46 |
| sdake | i dont understand the persistent storage thing | 17:46 |
| kfox1111 | sdake: I plan on trying to use 2.0.2 and upgrade to 3.0.1. | 17:48 |
| kfox1111 | so I'm not upgrading from non kolla-kubernetes to kola-kubernetes and mitaka to newton at the same time. | 17:48 |
| sdake | i see - well that does work if you use the json API | 17:49 |
| sdake | does/will :) | 17:49 |
| sdake | if the json api isn't used, that wont work | 17:49 |
| kfox1111 | you mean the config.json thing? | 17:49 |
| sdake | yup | 17:49 |
| kfox1111 | will uid/gid specification be added to it? | 17:49 |
| sdake | the json includes code to chown to the correct user | 17:50 |
| SamYaple | kfox1111: there is built ins into set_config.py that change permissions | 17:50 |
| SamYaple | kfox1111: im assuming thats what kolla-ansible will use | 17:50 |
| kfox1111 | ok. cool. yeah, that should work. | 17:50 |
| SamYaple | kolla-k8s doesnt use config.json? | 17:50 |
| kfox1111 | SamYaple: it does use config.json. right from kolla-ansible genconfig at the moment. | 17:50 |
| sdake | SamYaple it didn't and now it does :) | 17:51 |
| kfox1111 | so we should pick up the change automatically. | 17:51 |
| SamYaple | oh gotcha. yea thats a really solid entry point into the container | 17:51 |
| SamYaple | even if you dont use the permissions or config files stuff | 17:51 |
| kfox1111 | yup. | 17:51 |
| *** vhosakot has joined #openstack-kolla | 17:54 | |
| *** msimonin has joined #openstack-kolla | 17:54 | |
| sbezverk | kfox1111: there might be a chance helm rolls out 2.1.1 just for this fix for us :-) | 17:55 |
| *** msimonin has quit IRC | 17:57 | |
| *** msimonin has joined #openstack-kolla | 17:57 | |
| *** msimonin has quit IRC | 17:57 | |
| qwang | do we have a dependency diagram for all services? | 17:57 |
| pbourke | qwang: you can generate one with kolla-build | 17:58 |
| *** eaguilar has quit IRC | 17:59 | |
| SamYaple | hey pbourke. fancy seeing you around here | 17:59 |
| qwang | pbourke: that's great. thanks | 17:59 |
| kfox1111 | sbezverk: that would be very nice. | 18:00 |
| pbourke | SamYaple: fixed that migration bug | 18:00 |
| pbourke | SamYaple: kinda | 18:00 |
| *** unicell1 has quit IRC | 18:00 | |
| kfox1111 | qwang: no. but it is implied somewhat in the ordering of ceph_workflow. | 18:00 |
| SamYaple | pbourke: color me interested | 18:01 |
| kfox1111 | the wait's are bariors are there to ensure deps are handled. | 18:01 |
| pbourke | SamYaple: it seems with the old version of libvirt we're stuck on only native migration will work | 18:01 |
| qwang | kfox1111: so kolla-build would generate all related components for a service but we still can't tell the dependency among them. do i understand correctly? | 18:01 |
| pbourke | SamYaple: so had to hack the nova libvirt driver a little to make it use that API call rather than the URI type one | 18:01 |
| pbourke | SamYaple: Im still not 100% on the various forms of migration it can do | 18:02 |
| *** tonanhngo has joined #openstack-kolla | 18:02 | |
| pbourke | SamYaple: e.g. whats the difference between native and direct? | 18:02 |
| kfox1111 | qwang: we're adding the deps to the helm microservices package values. like: | 18:02 |
| kfox1111 | qwang: https://review.openstack.org/#/c/412239/10/helm/microservice/neutron-create-keystone-endpoint-admin/values.yaml | 18:03 |
| qwang | kfox1111: that's the reason I am asking, how can I figure out the dependencies as a developer | 18:04 |
| kfox1111 | dont know. :/ | 18:04 |
| kfox1111 | I've been using openstack long enough, I kind of know what each service depends on. | 18:05 |
| SamYaple | pbourke: native is qemu->libvirt->libvirt->qemu. direct is qemu->qemu | 18:05 |
| kfox1111 | maybe the openstack docs have some dependenncy info in them? | 18:05 |
| SamYaple | pbourke: i *think*. sometimes these words get reused, so it can depend on context | 18:05 |
| sbezverk | kfox1111: did you notice that openvswitch is failing for different jobs? | 18:05 |
| kfox1111 | sbezverk: got a link? | 18:06 |
| *** pbourke has quit IRC | 18:06 | |
| klindgren_ | am I m,issing a hidden config option or does kolla-ansible not provide a way to adjust the docker bridge ip? | 18:06 |
| SamYaple | klindgren_: kolla doesnt use the docker bridge | 18:07 |
| SamYaple | klindgren_: so thats all on you the operator if you want to adjust it | 18:07 |
| sbezverk | kfox1111: your neutron job and now mariadb same type of failure: http://logs.openstack.org/37/412237/16/check/gate-kolla-kubernetes-deploy-centos-binary-ceph-multi-nv/6a80ad1/console.html#_2016-12-19_17_55_10_999325 | 18:07 |
| *** Jeffrey4l has quit IRC | 18:07 | |
| klindgren_ | but it installs docker in the bootstrap phase? | 18:07 |
| SamYaple | klindgren_: it *can* install docker. sure. bootstrap isnt required | 18:08 |
| SamYaple | but either way, the bridge ip has nothing to do with kolla | 18:08 |
| kfox1111 | hmm. strange. | 18:08 |
| *** shardy has quit IRC | 18:08 | |
| SamYaple | klindgren_: you are welcome to adjust it with systemd overrides if you want | 18:08 |
| klindgren_ | well, in my case, not having that set causes huge issues | 18:08 |
| klindgren_ | because the odkcer bip overlaps with our vpn ip's | 18:09 |
| klindgren_ | which bascially means as soon a box gets docker isntalled it on | 18:09 |
| klindgren_ | its impposible to reach over vpn | 18:09 |
| kfox1111 | sbezverk: "Back-off 10s restarting failed container=dependencies pod=openvswitch-vswitchd-network-3ktgk_kolla(b537b55e-c613-11e6-83e3-fa163e2b40b7)" | 18:09 |
| SamYaple | klindgren_: the bootstrap is an optional playbook. you are welcome to use it or not. There was a bit of a fight to not even _include_ the bootstrap playbook because of things like you are complaining about | 18:09 |
| kfox1111 | sbezverk: cause its trying to run a dependency container for some reason. | 18:09 |
| SamYaple | klindgren_: kolla isnt trying to be the "docker deploy tool", there are other ansible roles that can do that for you | 18:09 |
| kfox1111 | probably was running but always suceeding because of the broken container. | 18:10 |
| sbezverk | kfox1111: there was a check for a socket for vswitchd | 18:10 |
| kfox1111 | sbezverk: yeah. probably not quite right. :/ | 18:11 |
| sbezverk | ok I will look into it right away | 18:11 |
| klindgren_ | one more question - what is the kolla-toolbox docker image actually used for? | 18:11 |
| *** rsjethani has joined #openstack-kolla | 18:11 | |
| SamYaple | klindgren_: when ansible runs, some tasks require deps like all of the python clients (or shade). instead of deploying those deps on the host, we put them into a container | 18:12 |
| SamYaple | we then execute those tasks in that cnotainer | 18:12 |
| klindgren_ | ok - so I should probably build that guy :-D | 18:13 |
| SamYaple | klindgren_: lol probably :) | 18:13 |
| sbezverk | kfox1111: strange in one scenario it does not fail.. gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv SUCCESS | 18:15 |
| *** athomas has quit IRC | 18:17 | |
| kfox1111 | sbezverk: cause its not using the openvswitch container? | 18:17 |
| kfox1111 | sbezverk: I think dns broke again. infra says centos 7.3 is rolling out. :/ | 18:18 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: PS Do not merge Intermittent vswitchd crash https://review.openstack.org/412563 | 18:18 |
| *** sayanta__ has joined #openstack-kolla | 18:22 | |
| *** gfidente is now known as gfidente|afk | 18:23 | |
| jascott1 | o/ | 18:23 |
| sbezverk | v1k0d3n: ping | 18:23 |
| v1k0d3n | hello | 18:23 |
| sbezverk | v1k0d3n: do you know if entrypoint has a keyword to turn on debugging? | 18:24 |
| dmsimard | building images is failing consistently on percona repository setup -- trying a yum install of the percona rpm looks to work for me.. anyone else experiencing that issue ? http://pastebin.centos.org/60041/raw/ | 18:24 |
| *** sayantan_ has quit IRC | 18:24 | |
| sbezverk | v1k0d3n: I see it exist wuth code 1, but I wold like to get a bit more background.. | 18:25 |
| *** eaguilar has joined #openstack-kolla | 18:25 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: DNS fix https://review.openstack.org/412569 | 18:26 |
| v1k0d3n | are you talking about the kubernetes-entrypoint? | 18:27 |
| v1k0d3n | container | 18:27 |
| kfox1111 | v1k0d3n: yeah. | 18:28 |
| srwilkers | hey jascott1 o/ | 18:29 |
| sbezverk | v1k0d3n: right.. all I see in the log that it exited with code 1. I would like to get more debgging info, especially since I know that the socket it was supposed to watch existed | 18:30 |
| kfox1111 | sbezverk: it prints out a bunch of logs but its in container so if its exiting 1 you wont see it. | 18:30 |
| kfox1111 | sbezverk: will probably have to do a while true, and put entrypoint in it. | 18:30 |
| kfox1111 | then when it eventually times out, it should be in the logs dir. | 18:31 |
| sbezverk | kfox1111: do you mean exit command for entrypoin while true, and timeout on the gate will kill it right? | 18:32 |
| kfox1111 | sbezverk: either way. | 18:32 |
| kfox1111 | so long as it keps the container alive. | 18:33 |
| sbezverk | kfox1111: ok | 18:33 |
| *** unicell has joined #openstack-kolla | 18:34 | |
| sbezverk | kfox1111: but debug flag for it wold be really useful now | 18:34 |
| kfox1111 | sbezverk: yeah. :/ | 18:35 |
| kfox1111 | file a feature request? | 18:36 |
| *** matrohon has joined #openstack-kolla | 18:36 | |
| *** pc_m has quit IRC | 18:37 | |
| sbezverk | kfox1111: do you have link handy to do that? | 18:38 |
| kfox1111 | https://github.com/stackanetes/kubernetes-entrypoint/issues | 18:38 |
| *** pc_m has joined #openstack-kolla | 18:39 | |
| jascott1 | srwilkers whas up? | 18:41 |
| sbezverk | kfox1111: https://github.com/stackanetes/kubernetes-entrypoint/issues/14 | 18:43 |
| *** williamcaban has quit IRC | 18:43 | |
| sbezverk | please add your opinion.. | 18:44 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: DNS fix https://review.openstack.org/412569 | 18:44 |
| *** williamcaban has joined #openstack-kolla | 18:44 | |
| kfox1111 | sbezverk: commented. | 18:45 |
| kfox1111 | bbiab | 18:45 |
| sdake | sup jascott1 | 18:49 |
| sdake | wow lots of spaces | 18:49 |
| jascott1 | sdake is all spaced out | 18:49 |
| jascott1 | :) | 18:49 |
| sdake | jascott1 after htis morning - lemme tell ya | 18:49 |
| jascott1 | haha | 18:49 |
| jascott1 | i just found a nice surprise in the cabinet in my cube - K&R C book printed in '78 | 18:50 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 18:58 |
| *** williamcaban has quit IRC | 18:59 | |
| kfox1111 | back. | 19:01 |
| kfox1111 | jascott1: nice. :) | 19:03 |
| kfox1111 | been a while since I've seen K&R c... | 19:03 |
| kfox1111 | but less time then you'd think. :) | 19:03 |
| *** britthouser has quit IRC | 19:13 | |
| *** britthouser has joined #openstack-kolla | 19:13 | |
| *** fragatina has quit IRC | 19:13 | |
| *** david-lyle_ has joined #openstack-kolla | 19:13 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: DNS fix https://review.openstack.org/412569 | 19:13 |
| *** openstackstatus has quit IRC | 19:13 | |
| *** Bico_Fino has quit IRC | 19:13 | |
| *** david-lyle has quit IRC | 19:13 | |
| *** openstack has joined #openstack-kolla | 19:17 | |
| *** x220_ has quit IRC | 19:19 | |
| *** matrohon has quit IRC | 19:20 | |
| *** ipsecguy_ has joined #openstack-kolla | 19:30 | |
| *** msimonin has joined #openstack-kolla | 19:33 | |
| sbezverk | kfox1111: found an issue in entrypoint or probably the way we build it | 19:33 |
| sbezverk | check this out: http://paste.openstack.org/show/592835/ | 19:34 |
| *** ipsecguy has quit IRC | 19:35 | |
| sdake | jascott1 nice :) | 19:38 |
| openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 19:38 |
| sdake | jascott1 my kids each have a copy - if I could unplug them from overwatch long enough to train em :) | 19:38 |
| jascott1 | haha one of mine is cracked out on OverHype as well | 19:38 |
| sdake | I asked wife what my son wanted for dinner | 19:39 |
| jascott1 | i dont get, but am probably not supposed to get it as I am "old" | 19:39 |
| sdake | and she wa slike "He wants overwatch ..." | 19:39 |
| jascott1 | lol | 19:39 |
| sdake | jascott1 games like overwatch super addictive | 19:39 |
| jascott1 | sounds like mine as well | 19:39 |
| kfox1111 | sbezverk: oh... yeah. exits probably a bash built in. | 19:39 |
| sdake | just like these damn phone games | 19:39 |
| sdake | get off my lawn!! | 19:40 |
| kfox1111 | sbezverk: try /bin/true | 19:40 |
| jascott1 | sbin/init! | 19:40 |
| sbezverk | kfox1111: how about just echo "Done" | 19:40 |
| kfox1111 | /bin/echo should work too. | 19:41 |
| jascott1 | sorry just shouting random things I guess | 19:41 |
| kfox1111 | but you should never really see the output of it. | 19:42 |
| *** jmccarthy has quit IRC | 19:42 | |
| *** x220__ has quit IRC | 19:43 | |
| sbezverk | kfox1111: I reproduced the issue in my local test bed that is how I managed to get logs from the container | 19:43 |
| *** jmccarthy has joined #openstack-kolla | 19:43 | |
| kfox1111 | ah. cool. | 19:45 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: PS Do not merge Intermittent vswitchd crash https://review.openstack.org/412563 | 19:48 |
| sbezverk | kfox1111: pushed fix if it goes through we will need to merge asap :-) | 19:49 |
| *** portdirect has joined #openstack-kolla | 19:50 | |
| sayanta__ | Hi guys, it will be great if you can review the documentation work whenever you find time :) https://review.openstack.org/#/c/404993/ | 19:51 |
| kfox1111 | sbezverk: +1 | 19:51 |
| *** gfidente|afk has quit IRC | 19:51 | |
| portdirect | o/ | 19:52 |
| jascott1 | o/ sayanta__ | 19:52 |
| sayanta__ | Once, the content is approved I can go ahead and work with the openstack-manual guys to divide it by operating systems. | 19:52 |
| portdirect | reading scrollback | 19:52 |
| jascott1 | whatup portdirect! | 19:53 |
| sayanta__ | Hey jascott1 | 19:53 |
| portdirect | sbezverk / kfox1111, re kube-entrypoint: try not setting a command at all for entrypoint. They added this specifily for us based on the init-container discussion - though I dont know if its in the container we are using | 19:54 |
| portdirect | hey jascott1! | 19:54 |
| SamYaple | portdirect: in those cases set the command 'sleep infinity' | 19:54 |
| portdirect | SamYaple: lol | 19:56 |
| kfox1111 | portdirect: that shoudl be easily testable. | 19:56 |
| SamYaple | portdirect: that isnt a joke. thats convention | 19:56 |
| sbezverk | portdirect: so we specify the condition no command at all? | 19:56 |
| portdirect | https://github.com/stackanetes/kubernetes-entrypoint/pull/13/files | 19:56 |
| kfox1111 | have they had a release since then? | 19:57 |
| kfox1111 | Entrypoint WARNING: 2016/12/19 19:58:17 kubernetes-entrypoint.go:35: COMMAND env is empty | 19:58 |
| kfox1111 | echo $? | 19:58 |
| kfox1111 | 0 | 19:58 |
| kfox1111 | we're good. | 19:58 |
| SamYaple | that is a question. how does this differ from stackenetes? just using kolla images? | 19:58 |
| portdirect | sweet :) | 19:59 |
| SamYaple | is there any differing goals? | 19:59 |
| kfox1111 | SamYaple: they are embedeing kubernetes-entrypoint in every container. | 19:59 |
| kfox1111 | and using it as the container init. | 19:59 |
| portdirect | SamYaple: stackinetes uses kolla images - but with the entrypoint binary we are talking about in the containers | 19:59 |
| sbezverk | so we stay with echo in command | 19:59 |
| kfox1111 | we're using init containers so blocking the main containers from starting, then exiting when ready. | 19:59 |
| kfox1111 | sbezverk: so we can drop the command altogether and it will work. | 19:59 |
| SamYaple | portdirect: wait... stackates uses kolla images? | 20:00 |
| portdirect | SamYaple: yes | 20:00 |
| kfox1111 | SamYaple: kind of. | 20:00 |
| SamYaple | portdirect: what the heck! they are using code i wrote. crazy. | 20:00 |
| kfox1111 | I thin kthey have modified them to have kubernetes-entrypoint in them too. | 20:00 |
| SamYaple | why are they not kolla-kubernetes? | 20:00 |
| kfox1111 | SamYaple: its pretty common for most openstack on k8s projects to reuse kolla containers. | 20:01 |
| kfox1111 | they are very flexible. | 20:01 |
| kfox1111 | SamYaple: because kolla-kubernetes wasn't a thing when they started. | 20:01 |
| SamYaple | i know why didnt they start it | 20:01 |
| SamYaple | and why arent we merging with them | 20:01 |
| portdirect | SamYaple: i think (though dont quote me: https://github.com/DTadrzak/kolla/tree/dtadrzak/stackanetes) | 20:01 |
| kfox1111 | SamYaple: thats a work in progress. | 20:02 |
| SamYaple | sorry. ive just heard about stackanetes for while | 20:02 |
| *** senk has joined #openstack-kolla | 20:02 | |
| SamYaple | never looked into them. seeing they use kolla is... wierd. | 20:02 |
| *** sayanta__ has quit IRC | 20:02 | |
| kfox1111 | SamYaple: not so weird. most of the work is in the containers, not in the plubming to stand them up. ;) | 20:03 |
| portdirect | SamYaple: Yeah I think my project was one of the very few that does not use kolla images in some shape of form - even though i use all of the original kolla build tooling.. | 20:03 |
| SamYaple | kfox1111: its weird that ive talked to people about stackanetes and didnt know it was kolla :D | 20:04 |
| kfox1111 | SamYaple: ah. yeah. :) | 20:04 |
| SamYaple | portdirect: my first project built images with ansible wrapping docker | 20:04 |
| SamYaple | portdirect: and _THEN_ deploying :P | 20:04 |
| SamYaple | portdirect: i just ported the ansible bits | 20:04 |
| SamYaple | not the building bits | 20:04 |
| *** williamcaban has joined #openstack-kolla | 20:04 | |
| kfox1111 | SamYaple: I'd be willing to bet if you talked to jnust about anyone doing openstack on k8s they are using kolla containers of some form or another. fuel not withstanding. | 20:04 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Fixing vswitchd crash https://review.openstack.org/412563 | 20:04 |
| SamYaple | kfox1111: dont worry, everyone who works on fuel wishes it was kolla | 20:05 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: DNS fix https://review.openstack.org/412569 | 20:06 |
| kfox1111 | SamYaple: I still would really like to see fuel use kolla-kubernetes. | 20:06 |
| kfox1111 | their ui is awesome. | 20:06 |
| SamYaple | kfox1111: i dont disagree on the final point | 20:07 |
| kfox1111 | they can provide the ui, we can provide the plumbing. | 20:07 |
| kfox1111 | really a good match in the end. | 20:07 |
| kfox1111 | would like to see tripleo do the same. | 20:07 |
| *** n0isyn0ise has quit IRC | 20:07 | |
| SamYaple | tripleo. ugh | 20:07 |
| SamYaple | talk about a project only deveolpers would think is a good idea | 20:07 |
| *** TxGirlGeek has quit IRC | 20:08 | |
| kfox1111 | heh. | 20:08 |
| kfox1111 | I think its a great idea. I think openstack's governence has been totally broken to actually make it work though. :/ | 20:08 |
| kfox1111 | the idea is really really good. use openstack to deploy a non trivial application. | 20:08 |
| kfox1111 | (dog fooding) | 20:08 |
| *** vhosakot has quit IRC | 20:08 | |
| SamYaple | well what a suprise. a developer thinks tripleo is a good idea | 20:08 |
| SamYaple | you disgust me | 20:08 |
| kfox1111 | openstack sucks though at laumnching non trivial apps. | 20:08 |
| kfox1111 | tripleo could have identified the pain points, | 20:09 |
| kfox1111 | and fed that back to the projects, and it could have gotten fixed. | 20:09 |
| SamYaple | i actually dont know you well enough to make that joke. but that was a joke kfox1111 | 20:09 |
| kfox1111 | but the projects have their heads screwed on wrong and wont listen to feedback. :/ | 20:09 |
| kfox1111 | SamYaple: its all good. :) | 20:09 |
| kfox1111 | k8s, in contrast has done the right thing. | 20:09 |
| kfox1111 | using k8s to deploy itself. | 20:09 |
| SamYaple | what you have to realize is tripleo lives in openstack. it doesnt live in an idealized version of openstack | 20:09 |
| kfox1111 | it works quite well. | 20:09 |
| SamYaple | and therein lies the problem | 20:10 |
| *** TxGirlGeek has joined #openstack-kolla | 20:10 | |
| kfox1111 | the problem is nova won't listen. heat barely listens. | 20:10 |
| kfox1111 | it took 2 years to get conditionals into heat. :/ | 20:10 |
| SamYaple | kolla identifed that _really_ early. other projects are going to do thier thing. we need to adapt | 20:10 |
| SamYaple | tripleo is stuck on "nauhhhh you gotta do it the right way" | 20:10 |
| SamYaple | also over/under cloud? really? | 20:10 |
| SamYaple | single point of failure dhcp? | 20:11 |
| SamYaple | ugh | 20:11 |
| kfox1111 | doing it the right way, would have worked, if the other projects listended. the problem is,they didnt. | 20:11 |
| SamYaple | so poorly thought out architecturely | 20:11 |
| kfox1111 | oh, undercloud/overcloud is a great idea. | 20:11 |
| kfox1111 | i really really want it. | 20:11 |
| kfox1111 | the idea being this... | 20:11 |
| SamYaple | its a good concept implemented awfully | 20:11 |
| portdirect | kfox1111: i kinda vew kolla-k8s as that | 20:11 |
| kfox1111 | I want to hae a pool of bare hardware. | 20:11 |
| kfox1111 | I want to be able to have some of that hardware made availale to users. | 20:11 |
| kfox1111 | I also want some of that hardware availabe to user as vm's. | 20:12 |
| kfox1111 | so if you run a multitenant undercloud, | 20:12 |
| kfox1111 | and use a tenant in that undercloud to deploy a multitenant overcloud, | 20:12 |
| kfox1111 | and federate them together, | 20:12 |
| SamYaple | I agree. turns out thats not how tripleo works at all! | 20:12 |
| kfox1111 | you get the best of both worlds. :) | 20:12 |
| kfox1111 | yeah. :/ | 20:12 |
| SamYaple | if you take the concept of "over/under cloud" and run with it, then yea | 20:13 |
| SamYaple | if you say tripleo, expect it to be an awful experince | 20:13 |
| kfox1111 | portdirect: yeah. I think long term, kolla-kubernetes deplys ironic. ironic deploys k8s. kolla-kubernetes on top. win. :) | 20:13 |
| SamYaple | i dont think thats awful kfox1111 | 20:13 |
| SamYaple | it could even be kolla-salt/kolla-ansbiel at that point | 20:13 |
| kfox1111 | can go from seed to full blown over/undercloud pretty easy I think. | 20:13 |
| SamYaple | like i said, good concept | 20:13 |
| * kfox1111 nods | 20:13 | |
| kfox1111 | devils in the details. :) | 20:14 |
| SamYaple | tripleo is not the embodyment of that concept. lets say that | 20:14 |
| kfox1111 | SamYaple: they were trying. heat/nova were immature, to say the least. :/ | 20:14 |
| kfox1111 | composable roles go a long wayh. but I'm convinced now that | 20:15 |
| kfox1111 | the container plylosophy is much better then the, lets cram way too much on one os aproach. | 20:15 |
| SamYaple | i like full control of the host os | 20:15 |
| kfox1111 | tripleo could pull it off if they used heat just to laucnch k8s on bare metal and kick off kolla-kubernetes. | 20:16 |
| SamYaple | i was never fully behind the kolla-ansible "dont touch the host" thing | 20:16 |
| SamYaple | but now that its split.... i can make the project i want! | 20:16 |
| kfox1111 | never touch is proably a bit much. but as little as possible seems good. | 20:16 |
| kfox1111 | of cource ubuntu calls it /etc/unbound/unbound.conf.d/ | 20:18 |
| kfox1111 | and centos /etc/unbound/conf.d | 20:19 |
| kfox1111 | :/ | 20:19 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: DNS fix https://review.openstack.org/412569 | 20:19 |
| *** n0isyn0ise has joined #openstack-kolla | 20:20 | |
| SamYaple | kfox1111: kolla-salt will be mirroring host users with the static containers | 20:20 |
| SamYaple | kfox1111: i like that kind of integration | 20:21 |
| SamYaple | i really want a single thing ofr the operator "ssh here and manage things like you normally would, forget about 'containers'" | 20:21 |
| SamYaple | debugging will need container knowlege, sure, but i think you can get an operator by without touching "contaienrs" | 20:22 |
| * kfox1111 shrugs. I see them as apliance images. if built properly, you shouldn't have to enter them. | 20:22 | |
| kfox1111 | they are "staticly linked binaries" | 20:22 |
| SamYaple | kfox1111: what if you need to run `rabbitmqctl` | 20:22 |
| kfox1111 | just happened to be assebled out of multiple. | 20:23 |
| kfox1111 | SamYaple: there should be a job for that. | 20:23 |
| SamYaple | absolutely not. what. operators. thats operators territory | 20:23 |
| SamYaple | you cant automate operators away | 20:23 |
| SamYaple | try as people might tahts not happening | 20:23 |
| kfox1111 | no, but you can provide them the tools they need in such a way as they hardly have to think about the containers. | 20:23 |
| v1k0d3n | SamYaple: "absolutely not. what. operators. thats operators territory" << my thoughts. | 20:24 |
| SamYaple | kfox1111: well i do, i alias rabbitmq='docker exec -it rabbitmq rabbitmqctl' | 20:24 |
| SamYaple | and there you go. you 'enter' the container, but not relaly | 20:24 |
| v1k0d3n | or if not, an operator could be specific to what you're trying to achieve. my operator could look different than zxy-awesome-company's operator. | 20:24 |
| kfox1111 | I've had to patch neutron migrate-db and rerun manually. so I do get it. but shoudl be quite rare and normally not needed. | 20:24 |
| SamYaple | i mean yea but not really | 20:24 |
| * v1k0d3n ducks out again :) | 20:25 | |
| kfox1111 | rather, provide them a k8s job for running migrate-db. then they don't have to know to enter a container and manually do something. | 20:25 |
| SamYaple | kfox1111: yea i agree. i think aliases can solve all of that | 20:25 |
| portdirect | SamYaple: thats the same approach i take - gets a little funky when you neeed to deal with commands that need to work on a file but otherwise is pretty solid. | 20:25 |
| SamYaple | portdirect: agreed, it may require a bit of distributed "knowlege" to use it successfully, butfor the most part its awesome | 20:25 |
| SamYaple | kolla-salt will be dropping aliases out the wazoo | 20:26 |
| SamYaple | i want the operator to not know the difference between bare metal or contaienr deploys | 20:26 |
| SamYaple | i want them to _dig_ for a difference | 20:26 |
| portdirect | ^^^ YES! | 20:26 |
| portdirect | (still wont use though :P ) | 20:26 |
| portdirect | but love the approach you are taking :) | 20:26 |
| SamYaple | right. but thats key, its not for every situtation or team | 20:27 |
| v1k0d3n | SamYaple: +1 | 20:27 |
| v1k0d3n | i'm with you and port direct | 20:27 |
| SamYaple | i can say with 100% certainty, its for _my_ team. they'll love it | 20:27 |
| SamYaple | and ill be running it at home | 20:27 |
| SamYaple | i think rax does something like 30% of _total_ openstack deploys too. | 20:28 |
| SamYaple | so thats a large part of opentack taht will benefit from it :) | 20:28 |
| kfox1111 | SamYaple: the best tools are those that operators write for themselves. :) | 20:28 |
| SamYaple | kfox1111: and im an operator writing this tool | 20:29 |
| kfox1111 | SamYaple: me too. :) | 20:29 |
| SamYaple | i do not 'identify (lol) as an operator | 20:29 |
| SamYaple | errr developer | 20:29 |
| SamYaple | im an operator at heart | 20:29 |
| SamYaple | theory is awesome, but it has its place | 20:29 |
| kfox1111 | I try to be about 50%/50%. | 20:29 |
| kfox1111 | I am a better dev knowing operator issues. | 20:29 |
| kfox1111 | I 'm a better operator knowing what's possible from a dev perspective. | 20:30 |
| SamYaple | i end up being more 80/20 (leaning toward developer). but I dont forget what my 20% wants | 20:30 |
| SamYaple | its just how its worked out for me at least | 20:30 |
| *** newmember has joined #openstack-kolla | 20:30 | |
| kfox1111 | my mix goes back and forth over the years. but I try and keep it balanced. | 20:31 |
| SamYaple | haha i actualyl hate operations. but i realize they are the only thing that _really_ matters. so my goal in life is to make operations easier | 20:31 |
| SamYaple | abstraction is rarely the path to that goal | 20:31 |
| kfox1111 | I get that. | 20:32 |
| kfox1111 | arg... | 20:33 |
| kfox1111 | yaml2json's proving non trivial on this k8s file.... :/ | 20:33 |
| kfox1111 | stupid dates... | 20:33 |
| SamYaple | yaml1.2 is a true superset of json | 20:35 |
| SamYaple | turns out yaml1.1 is not | 20:35 |
| *** TxGirlGeek has quit IRC | 20:38 | |
| sbezverk | kfox1111: is something wrong with gates? I have not see any results for a couple of hours | 20:39 |
| kfox1111 | sbezverk: not sure... looking | 20:39 |
| *** TxGirlGeek has joined #openstack-kolla | 20:39 | |
| kfox1111 | looks a little slow, but stuff is going through I think. | 20:40 |
| sbezverk | kfox1111: ok | 20:42 |
| *** sayantan_ has joined #openstack-kolla | 20:45 | |
| kfox1111 | arg... the init containers really need to get out of beta/alpha. | 20:48 |
| kfox1111 | they are really hard to parse back out of the json. :/ | 20:49 |
| *** sayantan_ has quit IRC | 20:49 | |
| *** sayantan_ has joined #openstack-kolla | 20:49 | |
| *** sayantan_ has quit IRC | 20:50 | |
| kfox1111 | jq -r '.items[].metadata.annotations."pod.alpha.kubernetes.io/init-containers"' /tmp/pods.json | grep -v '^null$' | jq -r .[].name | 20:51 |
| kfox1111 | seems to do the trick though... | 20:51 |
| *** n0isyn0ise has quit IRC | 20:52 | |
| openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for cinder keystone v2 endpoints https://review.openstack.org/412615 | 20:54 |
| *** senk has quit IRC | 20:55 | |
| kfox1111 | srwilkers: nice. | 20:56 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 20:56 |
| *** sayantan_ has joined #openstack-kolla | 21:00 | |
| sbezverk | kfox1111: please check https://review.openstack.org/412563 | 21:02 |
| kfox1111 | sbezverk: yup. already workflolwed it. :) | 21:02 |
| sbezverk | oops :-) missed it | 21:02 |
| sbezverk | thank you | 21:02 |
| kfox1111 | its holding up https://review.openstack.org/#/c/412569/ one too. | 21:02 |
| kfox1111 | so as soon as it merges, we should try andn get that one in too. | 21:03 |
| *** Bico_Fino has joined #openstack-kolla | 21:03 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: Fixing vswitchd crash https://review.openstack.org/412563 | 21:04 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: DNS fix https://review.openstack.org/412569 | 21:05 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 21:06 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Gate fix for helm sometimes failing https://review.openstack.org/412282 | 21:08 |
| openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for cinder keystone v2 endpoints https://review.openstack.org/412615 | 21:09 |
| sayantan_ | SamYaple you there? | 21:10 |
| *** n0isyn0ise has joined #openstack-kolla | 21:10 | |
| *** eaguilar has quit IRC | 21:10 | |
| *** Bico_Fino has quit IRC | 21:14 | |
| sbezverk | kfox1111: will keep an eye on it | 21:14 |
| kfox1111 | sbezverk: thx. | 21:16 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 21:16 |
| *** Bico_Fino has joined #openstack-kolla | 21:18 | |
| srwilkers | sbezverk, thanks for the reminder about release. :) | 21:20 |
| sbezverk | srwilkers: any time ;-) | 21:21 |
| srwilkers | sbezverk, also seems i missed the templates folder in those. oops :/ brains not running 100% today it seems | 21:22 |
| *** n0isyn0ise has quit IRC | 21:24 | |
| openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for cinder keystone v2 endpoints https://review.openstack.org/412615 | 21:25 |
| openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for nova keystone service and endpoints https://review.openstack.org/412620 | 21:27 |
| *** n0isyn0ise has joined #openstack-kolla | 21:31 | |
| *** Bico_Fino_ has joined #openstack-kolla | 21:32 | |
| *** l4yerffeJ_ has joined #openstack-kolla | 21:33 | |
| *** Bico_Fino has quit IRC | 21:33 | |
| *** Bico_Fino_ is now known as Bico_Fino | 21:33 | |
| *** l4yerffeJ has quit IRC | 21:35 | |
| *** zhangshuai has quit IRC | 21:35 | |
| *** ssurana has joined #openstack-kolla | 21:36 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 21:39 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Gate fix for helm sometimes failing https://review.openstack.org/412282 | 21:47 |
| *** fragatina has joined #openstack-kolla | 21:48 | |
| *** TxGirlGeek has quit IRC | 21:51 | |
| *** saneax-_-|AFK is now known as saneax | 21:51 | |
| *** TxGirlGeek has joined #openstack-kolla | 21:51 | |
| *** fragatina has quit IRC | 21:54 | |
| *** fragatina has joined #openstack-kolla | 21:55 | |
| *** fguillot has quit IRC | 21:56 | |
| *** TxGirlGeek has quit IRC | 21:56 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 21:58 |
| *** fragatina has quit IRC | 21:58 | |
| *** fragatina has joined #openstack-kolla | 21:59 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Fixing dependencies container https://review.openstack.org/412627 | 22:02 |
| *** dave-mccowan has joined #openstack-kolla | 22:02 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:05 | |
| *** fragatina has quit IRC | 22:06 | |
| *** dave-mccowan has quit IRC | 22:07 | |
| kfox1111 | sbezverk: there's a bug there. | 22:08 |
| openstackgerrit | Merged openstack/kolla-kubernetes: DNS fix https://review.openstack.org/412569 | 22:10 |
| kfox1111 | srwilkers: I'd recommend rebasing your two reviews to trunk. | 22:11 |
| kfox1111 | there's been a few fixes to the gate merged that should make it check more relyably. | 22:11 |
| *** dave-mccowan has joined #openstack-kolla | 22:11 | |
| srwilkers | will do, thanks kfox1111 | 22:12 |
| openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for cinder keystone v2 endpoints https://review.openstack.org/412615 | 22:13 |
| openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for nova keystone service and endpoints https://review.openstack.org/412620 | 22:13 |
| *** TxGirlGeek has quit IRC | 22:13 | |
| *** jtriley_ has joined #openstack-kolla | 22:15 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Fixing dependencies container https://review.openstack.org/412627 | 22:16 |
| *** rafael_sisweb has quit IRC | 22:17 | |
| sbezverk | I need to step out for a couple of hours. please ack dependencies fix.. | 22:17 |
| sbezverk | if it is all green | 22:17 |
| *** jtriley has quit IRC | 22:18 | |
| *** jtriley_ has quit IRC | 22:20 | |
| portdirect | kfox1111, sbezverk: why are we listing deps in uppercase? makes the values.yaml pretty inconsistent | 22:23 |
| *** Bico_Fino has quit IRC | 22:23 | |
| kfox1111 | portdirect: its a direct env passthrough. | 22:25 |
| *** srwilkers has quit IRC | 22:25 | |
| sbezverk | portdirect: your comment is not relevant to this PS, we need the fix to unblock service testing.. | 22:27 |
| *** Bico_Fino has joined #openstack-kolla | 22:28 | |
| *** lrensing has quit IRC | 22:28 | |
| SamYaple | sayantan_: im around | 22:29 |
| portdirect | sbezverk: ok, just my concern is that: https://review.openstack.org/#/c/412239/13/helm/all_values.yaml will end up causing confusion down the road, but I'll stay quite :) | 22:29 |
| SamYaple | good portdirect. know your place! | 22:29 |
| * SamYaple has no context | 22:29 | |
| kfox1111 | portdirect: should it be lowercased and then upper cased in the for loop? | 22:30 |
| kfox1111 | portdirect: it would prevent any env vars ever being lowercased. but maybe that won't matter for kubernetes-entrypoint. | 22:30 |
| portdirect | kfox1111: i think that it wont matter for k8s-entrypoint? | 22:31 |
| kfox1111 | probably not. | 22:32 |
| kfox1111 | its not the only all caps we have in places. but if you want to propose the patch, I'd be ok with it. | 22:33 |
| jascott1 | portdirect I hear ya on the consistency but at the same time entrypoint uses all caps so it doesnt seem worth it to add complexity to me | 22:33 |
| kfox1111 | codes in helm/kolla-common/templates/_common_dependency_container.yaml | 22:33 |
| kfox1111 | would just be a | uppercase or something like that in one place. | 22:34 |
| kfox1111 | hmm... for some reason deps aren't quite working... | 22:34 |
| portdirect | kfox1111: thats exactly the soultion i was thinking of | 22:34 |
| portdirect | jascott1: what I'm proposing is https://review.openstack.org/#/c/412239/13/helm/kolla-common/templates/_common_dependency_container.yaml | 22:38 |
| *** jheroux has quit IRC | 22:39 | |
| *** williamcaban has quit IRC | 22:41 | |
| jascott1 | portdirect seems legit | 22:42 |
| *** williamcaban has joined #openstack-kolla | 22:43 | |
| jascott1 | uniformed user could see the use of lower in the values file and then assume they can use lower in their own init containers | 22:44 |
| jascott1 | but I dont think its a worry for us | 22:44 |
| *** williamcaban has quit IRC | 22:46 | |
| *** williamcaban has joined #openstack-kolla | 22:46 | |
| kfox1111 | yeah. and we use it alittle differently then they do anyway. | 22:47 |
| kfox1111 | we're taking a list, while they are just taking a string. | 22:47 |
| kfox1111 | I don't think it really matters too much either way. | 22:47 |
| portdirect | kfox1111: who is they? | 22:49 |
| *** williamcaban has quit IRC | 22:51 | |
| kfox1111 | portdirect: kubernetes-entrypoint | 22:53 |
| *** sayantan_ has quit IRC | 22:55 | |
| *** sayantan_ has joined #openstack-kolla | 22:56 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:57 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 22:58 |
| openstackgerrit | Pete Birley proposed openstack/kolla-kubernetes: Change common dependency container to convert env var names to uppercase https://review.openstack.org/412640 | 23:01 |
| sayantan_ | Hi SamYaple | 23:01 |
| portdirect | kfox1111: got it, I was off checking other implementations :) | 23:02 |
| *** sdake has quit IRC | 23:02 | |
| kfox1111 | portdirect: looks good to me. | 23:03 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Memcached svc https://review.openstack.org/412287 | 23:11 |
| openstackgerrit | Pete Birley proposed openstack/kolla: Add Kubernetes Images for Kolla-Kubernetes https://review.openstack.org/412205 | 23:14 |
| *** huikang has joined #openstack-kolla | 23:14 | |
| *** eaguilar has joined #openstack-kolla | 23:17 | |
| kfox1111 | portdirect: oh. was going to mention. | 23:20 |
| kfox1111 | we might want to do the kube images differently from the rest in one other aspect. | 23:20 |
| kfox1111 | we probably shoudl version them after the k8s versions. | 23:20 |
| kfox1111 | rather then the kolla version. | 23:20 |
| portdirect | hmm | 23:21 |
| kfox1111 | they release so much more often then kolla. | 23:21 |
| portdirect | i hear/like what you are saying i think - but how to achive that? thats a tagging issue isnt it? | 23:22 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Fixing dependencies container https://review.openstack.org/412627 | 23:22 |
| kfox1111 | yeah. but the version of the binaries are hardcoded. maybe the version should be too? | 23:23 |
| portdirect | as in kolla version Label in the docker image? not so sure about that - unless we just label it 'rolling' or similar? | 23:24 |
| portdirect | seems a bit of an anti pattern to label it against the k8s binary version to me when it's being layered over a version of kolla-base? but totally open to being swayed on that. | 23:25 |
| *** huikang has quit IRC | 23:26 | |
| kfox1111 | the issue I think is, user has kube 1.5.4 and 1.5.5 comes out. | 23:27 |
| kfox1111 | what does the user do? | 23:27 |
| kfox1111 | we almst kind of need a kolla-other repo or something for stuff like this. | 23:28 |
| portdirect | +1 to that | 23:28 |
| kfox1111 | something that isn't tied so tightly to the openstack release cycle. | 23:28 |
| portdirect | the current kolla-build cannot cope with this use case | 23:28 |
| kfox1111 | hmm. k. | 23:28 |
| kfox1111 | well, we can always just update the link inside and then retag after built. | 23:29 |
| kfox1111 | just a manual bit of pain. | 23:29 |
| kfox1111 | at least until something better can be thought up. | 23:29 |
| portdirect | yeah thats what i was intending | 23:29 |
| *** sayantan_ has quit IRC | 23:29 | |
| *** TxGirlGeek has quit IRC | 23:29 | |
| *** sayantan_ has joined #openstack-kolla | 23:29 | |
| *** lamt has quit IRC | 23:32 | |
| portdirect | the primary issue here is that vitualy every k8s deployment too expects the images tag to match the k8s version, so they cannot be coupled to the kolla version as kubeadmet all would expect foo:3.0.2 to be for that version of k8s, so they need to be tagged foo:v1.5.1 to be drop in replacements (also the name's that kolla produce will not work) so for now documentation an manual retagging is the only opti | 23:35 |
| portdirect | on. I added these more for our gates than anything else, as I'd really like to see rdb support as standard... | 23:35 |
| kfox1111 | yeah. it would be nice if kolla's build scripts could take a tag override, then we could hard code it to match what k8s expects. | 23:36 |
| kfox1111 | +1 for gate support too. batteries included. some folks wont want to build rbd support just for this. | 23:36 |
| *** JRobinson__ has joined #openstack-kolla | 23:37 | |
| portdirect | should i add -amd64 to the end of each image? - that way there is a chance that some k8s tools would work ootb with only the version tag being changed? | 23:37 |
| kfox1111 | sounds reasonable. | 23:39 |
| *** Bico_Fino has quit IRC | 23:39 | |
| kfox1111 | this is soo weird... http://logs.openstack.org/39/412239/13/experimental/gate-kolla-kubernetes-deploy-centos-binary-helm-entrypoint-nv/eaa9811/logs/openstack/kolla-neutron-server-4130061660-8hmd5.txt | 23:40 |
| kfox1111 | it really looks like entrypoint is letting go too soon. | 23:40 |
| portdirect | can we get logs from: http://logs.openstack.org/39/412239/13/experimental/gate-kolla-kubernetes-deploy-centos-binary-helm-entrypoint-nv/eaa9811/logs/openstack/kolla-neutron-manage-db-gdlq4.txt? | 23:41 |
| portdirect | be interesting to look at timestaps, but yeah thats weird | 23:42 |
| kfox1111 | hmm... | 23:42 |
| kfox1111 | http://logs.openstack.org/39/412239/13/experimental/gate-kolla-kubernetes-deploy-centos-binary-helm-entrypoint-nv/eaa9811/logs/pods/kolla-neutron-server-4130061660-8hmd5-dependencies.txt | 23:42 |
| kfox1111 | looks like its just waiting for keystone. | 23:42 |
| kfox1111 | but not the jobs.... | 23:42 |
| kfox1111 | something may be wrong with the dep loop. | 23:42 |
| kfox1111 | http://logs.openstack.org/39/412239/13/experimental/gate-kolla-kubernetes-deploy-centos-binary-helm-entrypoint-nv/eaa9811/logs/pods/kolla-neutron-server-4130061660-l1840.txt looks like the deps are there.... | 23:44 |
| openstackgerrit | Pete Birley proposed openstack/kolla: Add Kubernetes Images for Kolla-Kubernetes https://review.openstack.org/412205 | 23:44 |
| portdirect | kfox1111: needs to be dependency_jobs not job | 23:45 |
| kfox1111 | oh. good catch. thanks. | 23:45 |
| portdirect | np | 23:45 |
| *** msimonin has quit IRC | 23:49 | |
| *** dave-mccowan has quit IRC | 23:50 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 23:50 |
| kfox1111 | sbezverk: https://review.openstack.org/#/c/412287/ is ready | 23:55 |
| kfox1111 | sbezverk: I think once portdirect & srwilkers are made cores, we should go back to 2 x +2 ? | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!