| sbezverk | kfox1111: no idea, but they do test starting vm at the gate too | 00:00 |
|---|---|---|
| kfox1111 | I can believe it works. just wondering what unintended sideffects of running it without a cell1 are. | 00:00 |
| sbezverk | kfox1111: well it goes through all your basic tests | 00:01 |
| kfox1111 | basic tests not nessisarily equate to no production problems. | 00:01 |
| kfox1111 | stuff like this may only show up when upgrading from 4.x to 5.x. | 00:02 |
| sbezverk | I think it is a good start :), but agree about prod.. | 00:02 |
| kfox1111 | this is the kind of thing ops loose sleep over. | 00:02 |
| sbezverk | well, adding cell1 later will take very little time if it is really needed | 00:02 |
| kfox1111 | assuming it works though. | 00:03 |
| kfox1111 | it may munge the db in a way thats not seperable. | 00:03 |
| kfox1111 | or some cell0 cleaner task doesn't nuke all the instances cuase it doesn't expect live ones. | 00:03 |
| kfox1111 | we really shoudl confirm with the nova folks cell0 only is an ok way to operate. | 00:04 |
| sbezverk | kfox1111: sure thing, I will post question on irc tomorrow when lots of nova folks around | 00:04 |
| sbezverk | kfox1111: or you can send question to mailer | 00:05 |
| kfox1111 | yeah | 00:05 |
| kfox1111 | sbezverk: reviewed | 00:07 |
| kfox1111 | sbezverk: so, did you ever do a nova-manage cell_v2 discover_hosts ? | 00:07 |
| sbezverk | kfox1111: cool, thanks | 00:07 |
| kfox1111 | hmm... I wonder if the simple setup does end up createing a cell1 too. | 00:08 |
| kfox1111 | that coudl explain how it works. | 00:08 |
| sbezverk | kfox1111: nope, I think it is not needed since I made sure that compute nodes are up and running before creating cell0 | 00:08 |
| kfox1111 | oh. | 00:09 |
| sbezverk | kfox1111: one nova dude told me so | 00:09 |
| kfox1111 | uh.. is doing that inforced in the helm+kubernetes-entrypoint case? | 00:09 |
| kfox1111 | I think we need a way to allow it to be done after the fact... | 00:09 |
| kfox1111 | it sounds like you must rerun the process if you ever add more computes. | 00:09 |
| sbezverk | kfox1111: right dependency on nova-compute | 00:09 |
| kfox1111 | but doesn't nova compute depend on nova-api? | 00:10 |
| sbezverk | kfox1111: yes | 00:10 |
| kfox1111 | which shoudl depend on cells? | 00:10 |
| sbezverk | kfox1111: compute depends on api and cell depends on compute | 00:10 |
| sbezverk | api does not depend on cell | 00:11 |
| kfox1111 | I think the compute dep is oing to cause problems. | 00:12 |
| kfox1111 | I think we should add a nova-manage cell_v2 discover_hosts job. | 00:12 |
| kfox1111 | we will need it anyway. | 00:13 |
| sbezverk | kfox1111: your comment about nova.. it is actually how thet want | 00:14 |
| sbezverk | they want user nova had access to nova_cell0 db | 00:14 |
| sbezverk | there is no a separate user nova_cell0 | 00:14 |
| kfox1111 | sbezverk: but the delete doesnt' use the same user as the create | 00:15 |
| sbezverk | kfox1111: right right I have not noticed below config :) | 00:16 |
| sbezverk | sorry | 00:16 |
| kfox1111 | no worries. :) | 00:16 |
| kfox1111 | so, simple_setup looks to be the migration tool. | 00:16 |
| kfox1111 | so it may end up crating cell1 and moving stuff automatically. | 00:16 |
| kfox1111 | though no idea what mariadb it uses. :/ | 00:17 |
| *** masber has quit IRC | 00:21 | |
| *** goldyfruit has quit IRC | 00:26 | |
| kfox1111 | sbezverk: would be curious to see the output of "nova-manage cell_v2 list_cells --verbose" on your cells setup. | 00:27 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla master: Gate fixes: disable bifrost-base, fix ubuntu base image https://review.openstack.org/448666 | 00:27 |
| *** tonanhngo has quit IRC | 00:33 | |
| *** adrian_otto has joined #openstack-kolla | 00:33 | |
| sdake | kfox1111 sbezverk lets have a brainstorm session about HOW to do external configmaps if you have 30 minutes each plz | 00:34 |
| kfox1111 | I gotta take off actually. maybe tomorrow? | 00:35 |
| sdake | kfox1111 sure ping sbezverk and me when you get in | 00:36 |
| kfox1111 | kk. | 00:36 |
| sdake | kfox1111 i'd like to close on a design pronot :) | 00:36 |
| sdake | pronto that is | 00:36 |
| *** tonanhngo has joined #openstack-kolla | 00:41 | |
| *** zhurong has joined #openstack-kolla | 00:41 | |
| *** duonghq has joined #openstack-kolla | 00:45 | |
| duonghq | morning guys | 00:45 |
| *** tonanhngo has quit IRC | 00:45 | |
| *** tonanhngo has joined #openstack-kolla | 00:47 | |
| *** cuongnv has joined #openstack-kolla | 00:48 | |
| *** sayantani01 has quit IRC | 00:49 | |
| sdake | sup duonghq | 00:51 |
| sdake | steak tired | 00:51 |
| openstackgerrit | Joseph Robinson proposed openstack/kolla-ansible master: [deploy-guide] Moving the quick start to the deploy folder https://review.openstack.org/448509 | 00:51 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Virtualbmc and deployment of baremetal VM using ironic https://review.openstack.org/441550 | 00:52 |
| *** tonanhngo has quit IRC | 00:52 | |
| *** sayantani01 has joined #openstack-kolla | 00:52 | |
| *** tovin07_ has joined #openstack-kolla | 00:54 | |
| duonghq | sdake, have we unblocked the gate? | 00:54 |
| sdake | duonghq mnaser working on it i think | 00:54 |
| sdake | duonghq but nope its busted | 00:54 |
| mnaser | i think | 00:54 |
| mnaser | sdake | 00:54 |
| mnaser | i think i found it | 00:54 |
| mnaser | https://review.openstack.org/#/c/448666/ | 00:55 |
| mnaser | 3 fixes in total | 00:55 |
| mnaser | also now the logs are MUCH more readable for failed images ^_^ | 00:55 |
| mnaser | now we have logs/build which contain everything | 00:55 |
| mnaser | and the subunit problem doesnt poop out anymore | 00:55 |
| duonghq | mnaser, seem good | 00:56 |
| duonghq | you combined your 2 ps? | 00:56 |
| mnaser | yeah | 00:56 |
| sdake | what does /tmp/logs/build do? | 00:56 |
| sdake | the failed image logging is maddening to me | 00:57 |
| mnaser | sdake http://logs.openstack.org/66/448666/4/check/gate-kolla-dsvm-build-ubuntu-binary-ubuntu-xenial-nv/beea00a/logs/build/ :) | 00:57 |
| sdake | i'm curious how its more readable now | 00:57 |
| sdake | cool hrw suggested that I think on the ml | 00:58 |
| sdake | he will be happy to see it :) | 00:58 |
| mnaser | yeah the work was all done | 00:58 |
| mnaser | i didnt have to do much | 00:58 |
| *** adrian_otto has quit IRC | 00:59 | |
| duonghq | https://review.openstack.org/#/c/448666/3..4/tools/setup_gate.sh seem that we should wait for gate, | 01:00 |
| duonghq | http://logs.openstack.org/66/448666/3/check/gate-kolla-dsvm-build-centos-source-centos-7/e745bb1/console.html | 01:00 |
| * mnaser is patiently watching the zuul status | 01:07 | |
| mnaser | 8 jobs left on this patch, all green so far | 01:07 |
| duonghq | nice | 01:08 |
| mnaser | 7! | 01:10 |
| duonghq | mnaser, 6 | 01:10 |
| mnaser | ill just feel a bit relieved with this in | 01:12 |
| duonghq | do we provide note about disabling bifrost in doc? | 01:22 |
| *** zhurong has quit IRC | 01:23 | |
| mnaser | duonghq i think its more of a gate/testing thing, and not as much of an operational thing | 01:23 |
| *** MasterOfBugs has quit IRC | 01:23 | |
| mnaser | someone has to look into fixing it, i tried my best but im a bit busy these next few days | 01:23 |
| mnaser | just clearing out the gate took a lot into my $job work time and now i have to play catch up :( | 01:24 |
| duonghq | thank for your effort, gate is painful to debugging but must be done | 01:24 |
| duonghq | only 4 left | 01:25 |
| duonghq | hope this will be ok | 01:25 |
| mnaser | yeah i think its good | 01:27 |
| mnaser | i love the new build logs | 01:27 |
| mnaser | they're going to make life sooo much easier | 01:27 |
| mnaser | i dont know why we had it that way this whole time | 01:27 |
| duonghq | mnaser, you only add log_dir var and we have all nice log? | 01:27 |
| mnaser | yeah, it was already written | 01:27 |
| duonghq | ah, yep | 01:28 |
| duonghq | nice feature but not used in gate yet, | 01:28 |
| duonghq | also, I think it would be better if we put the line in separate commit, due to your commit has potential to be removed in future? | 01:29 |
| duonghq | how do you think sdake | 01:29 |
| *** masber has joined #openstack-kolla | 01:29 | |
| duonghq | mnaser, seem that gate is all green now | 01:30 |
| sdake | duonghq how not used in gate? | 01:31 |
| duonghq | sdake, I mean we should put logs_dir var in separate ps, due it isn't related to the bug but add nice output to our gate | 01:31 |
| mnaser | duonghq it actually breaks the gate | 01:31 |
| mnaser | if i dont add it there | 01:31 |
| duonghq | the logs_dir var? | 01:32 |
| *** japestinho has joined #openstack-kolla | 01:32 | |
| mnaser | because the logs are so huge, python subunit complaints that the length of the payload is too big | 01:32 |
| mnaser | and it fails | 01:32 |
| mnaser | that's why patch #3 failed | 01:32 |
| duonghq | hmm, got it | 01:32 |
| mnaser | if you look the log says "length too long" or something | 01:32 |
| mnaser | so i had no choice of doing it to unlock gate | 01:32 |
| duonghq | so we should note it somewhere | 01:32 |
| duonghq | not sure | 01:32 |
| duonghq | or we put it in separate commit, and rebase the fix bug commit into it | 01:33 |
| mnaser | not possible | 01:33 |
| mnaser | because gate wont merge | 01:33 |
| duonghq | why? | 01:33 |
| duonghq | if parent of fix bug commit is the logs_dir commit | 01:33 |
| mnaser | yes but | 01:34 |
| mnaser | the parent will never pass tests | 01:34 |
| duonghq | ah, yeah | 01:34 |
| mnaser | because it is logging all errors to stdout/stderr | 01:34 |
| mnaser | so it will never merge | 01:34 |
| duonghq | circular dep | 01:34 |
| mnaser | yep | 01:34 |
| mnaser | 3 circular deps | 01:34 |
| mnaser | lol | 01:34 |
| mnaser | log_dirs, netbase and bifrost disable | 01:34 |
| duonghq | painful situation | 01:35 |
| duonghq | your job is great | 01:35 |
| duonghq | one more test | 01:35 |
| duonghq | *your work | 01:36 |
| duonghq | mnaser, are you in US? | 01:37 |
| mnaser | east coast canada | 01:37 |
| *** dave-mccowan has quit IRC | 01:37 | |
| duonghq | nice | 01:37 |
| mnaser | at last | 01:39 |
| mnaser | a +1 | 01:39 |
| *** dave-mccowan has joined #openstack-kolla | 01:39 | |
| duonghq | mnaser, done | 01:39 |
| *** zhurong has joined #openstack-kolla | 01:43 | |
| *** manjeets has joined #openstack-kolla | 01:46 | |
| duonghq | sbezverk, are you there? | 01:46 |
| *** masber has quit IRC | 01:49 | |
| *** masber has joined #openstack-kolla | 01:53 | |
| *** hieulq has quit IRC | 01:56 | |
| *** fooliouno has quit IRC | 01:57 | |
| *** hieulq has joined #openstack-kolla | 01:59 | |
| *** dave-mcc_ has joined #openstack-kolla | 02:08 | |
| *** dave-mccowan has quit IRC | 02:10 | |
| *** goldyfruit has joined #openstack-kolla | 02:10 | |
| *** dave-mccowan has joined #openstack-kolla | 02:13 | |
| *** unicell has quit IRC | 02:15 | |
| *** dave-mcc_ has quit IRC | 02:16 | |
| *** eaguilar has quit IRC | 02:20 | |
| openstackgerrit | Joseph Robinson proposed openstack/kolla-ansible master: [deploy-guide] Moving the quick start to the deploy folder https://review.openstack.org/448509 | 02:28 |
| *** fooliouno has joined #openstack-kolla | 02:39 | |
| openstackgerrit | Merged openstack/kolla master: Gate fixes: disable bifrost-base, fix ubuntu base image https://review.openstack.org/448666 | 02:39 |
| *** hieulq has quit IRC | 02:39 | |
| *** hieulq has joined #openstack-kolla | 02:41 | |
| *** l4yerffeJ has joined #openstack-kolla | 02:41 | |
| *** manheim has joined #openstack-kolla | 02:45 | |
| *** unicell has joined #openstack-kolla | 02:48 | |
| *** manheim has quit IRC | 02:50 | |
| *** 7IZAAUKJJ has joined #openstack-kolla | 02:52 | |
| *** masber has quit IRC | 02:56 | |
| *** 7IZAAUKJJ has quit IRC | 03:00 | |
| openstackgerrit | Sayantani Goswami proposed openstack/kolla-ansible master: [deploy-guide] Moving the quick start to the deploy folder https://review.openstack.org/448509 | 03:03 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla master: Updated from global requirements https://review.openstack.org/446771 | 03:05 |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible master: Update the kubernetes-deployment.rst file https://review.openstack.org/448035 | 03:08 |
| *** l4yerffeJ has quit IRC | 03:10 | |
| *** eaguilar has joined #openstack-kolla | 03:11 | |
| *** kbaegis has quit IRC | 03:13 | |
| *** zhubingbing has joined #openstack-kolla | 03:17 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Virtualbmc and deployment of baremetal VM using ironic https://review.openstack.org/441550 | 03:24 |
| spsurya | hello all | 03:26 |
| duonghq | hi spsurya | 03:33 |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla master: Implement Keystone zero-downtime upgrade https://review.openstack.org/425833 | 03:33 |
| *** bot_Jeffrey4l has joined #openstack-kolla | 03:38 | |
| duonghq | sbezverk, ping | 03:41 |
| *** bot_Jeffrey4l has quit IRC | 03:42 | |
| *** bot_Jeffrey4l has joined #openstack-kolla | 03:43 | |
| *** eaguilar has quit IRC | 03:43 | |
| sbezverk | duonghq: hey what is up? | 03:46 |
| *** bot_Jeffrey4l has quit IRC | 03:47 | |
| *** bot_Jeffrey4l has joined #openstack-kolla | 03:47 | |
| *** bot_Jeffrey4l has quit IRC | 03:48 | |
| *** zhurong has quit IRC | 03:48 | |
| duonghq | sbezverk, about the ps: https://review.openstack.org/#/c/448258/8 | 03:48 |
| *** masber has joined #openstack-kolla | 03:48 | |
| duonghq | I mean tool for generate openrc file, the kube already has it? | 03:49 |
| *** dave-mccowan has quit IRC | 03:50 | |
| *** hieulq has quit IRC | 03:50 | |
| *** hieulq has joined #openstack-kolla | 03:53 | |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla master: Fix oslo_debug_helper not running for kolla/tests directory https://review.openstack.org/444398 | 03:56 |
| duonghq | sbezverk, what is vip are you using? | 03:59 |
| duonghq | (compare to nodes' ip) | 03:59 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Virtualbmc and deployment of baremetal VM using ironic https://review.openstack.org/441550 | 03:59 |
| sbezverk | duonghq: vip is public ip assigned to host running kube cluster | 04:00 |
| duonghq | sbezverk, what if I run in multinode setup? | 04:00 |
| duonghq | for local test | 04:01 |
| duonghq | and why dns_name of kolla is same as external_vip? | 04:01 |
| sbezverk | duonghq: then it could be ANY publically available ip | 04:02 |
| sbezverk | used by your cluster | 04:02 |
| duonghq | and about dns_name, the kubedns is not expose in my cluster, does it matter? | 04:03 |
| sbezverk | duonghq: you have to have kubedns running otherwise kolla-kube will not work | 04:05 |
| duonghq | sbezverk, kubedns is runnning but donot have external ip | 04:05 |
| duonghq | only cluster ip | 04:05 |
| sbezverk | that is normal | 04:06 |
| duonghq | and it's ok? | 04:06 |
| sbezverk | dsn_name has no realtion to kubedns | 04:06 |
| duonghq | I got trouble with network again | 04:06 |
| duonghq | Entrypoint WARNING: 2017/03/23 04:07:39 entrypoint.go:81: Resolving dependency for mariadb failed: Get https://10.233.0.1:443/api/v1/namespaces/kolla/endpoints/mariadb: dial tcp 10.233.0.1:443: getsockopt: no route to host | 04:07 |
| duonghq | 04:07 | |
| duonghq | it's output of one of job | 04:07 |
| sbezverk | dns_name was introduced just for the case when you have external dns mapped | 04:07 |
| sbezverk | to one of ip addresses used by kube cluster | 04:07 |
| duonghq | ah, got it | 04:07 |
| sbezverk | so normally they are the same | 04:08 |
| sbezverk | vip and dns name but could be different | 04:08 |
| sbezverk | if it is needed | 04:08 |
| sbezverk | duonghq: got to go, it is very late now.. talk to you tomorrow | 04:08 |
| duonghq | sbezverk, thank you, cya | 04:09 |
| fooliouno | duonghq: do you have a multi node kolla-k8s cluster up? | 04:09 |
| duonghq | fooliouno, yes | 04:10 |
| duonghq | ah, no | 04:10 |
| duonghq | I only have multinode k8s | 04:10 |
| *** fooliouno_ has joined #openstack-kolla | 04:10 | |
| fooliouno_ | sorry, got disconnected | 04:11 |
| fooliouno_ | do you have it up | 04:11 |
| duonghq | fooliouno_, I have only k8s cluster | 04:12 |
| duonghq | but not kolla-k8s | 04:12 |
| fooliouno_ | flannel/weave/canal? | 04:12 |
| duonghq | yup | 04:12 |
| fooliouno_ | I cant get flannel overlay to work on k8s. Cant get keystone_create_db to resolve mariadb host address | 04:13 |
| *** fooliouno has quit IRC | 04:14 | |
| fooliouno_ | did you run into any issues with resolving service names within containers | 04:14 |
| duonghq | fooliouno_, I just got it error today | 04:16 |
| duonghq | fooliouno_, how do you setup k8s? | 04:16 |
| fooliouno_ | kubeadm | 04:16 |
| fooliouno_ | hosted install | 04:16 |
| duonghq | hmm, I used kargo | 04:16 |
| duonghq | can you check kubeproxy log | 04:17 |
| fooliouno_ | what should I be looking for | 04:17 |
| *** goldyfruit has quit IRC | 04:18 | |
| duonghq | check if you have any iptables-related warning | 04:18 |
| fooliouno_ | ok .. let me take a look | 04:19 |
| *** tovin07_ has quit IRC | 04:25 | |
| fooliouno_ | duonghq: what version of docker do you have | 04:27 |
| duonghq | fooliouno_, Server Version: 1.13.1 | 04:28 |
| fooliouno_ | ok .. I have the same. Just want to rule out any differences :) | 04:28 |
| fooliouno_ | Were you able to resolve the error with mariadb | 04:29 |
| *** manheim has joined #openstack-kolla | 04:30 | |
| *** yee379 has quit IRC | 04:30 | |
| *** yee37914 has joined #openstack-kolla | 04:31 | |
| duonghq | I was able, but after some restart the VMs run k8s cluster, somehow it's broken | 04:31 |
| duonghq | I'm reinstall cluster | 04:31 |
| *** manheim has quit IRC | 04:34 | |
| *** jrobinson has quit IRC | 04:36 | |
| *** tovin07_ has joined #openstack-kolla | 04:36 | |
| *** tovin07_ has quit IRC | 04:38 | |
| *** masber has quit IRC | 04:40 | |
| duonghq | fooliouno_, | 04:41 |
| duonghq | I can createdb | 04:41 |
| fooliouno_ | keystone db? | 04:41 |
| duonghq | yes | 04:42 |
| fooliouno_ | thats good! I'm still trying to figure out whats wrong with mine :) | 04:42 |
| duonghq | can you provide kubectl get po -n kolla | 04:43 |
| fooliouno_ | Do you know of anyone who has gone through the entire kolla-k8s setup and created nova instances? | 04:43 |
| duonghq | seem that no one online now | 04:43 |
| fooliouno_ | I just did a kubeadm reset to reset the cluster and start from scratch | 04:43 |
| *** tovin07_ has joined #openstack-kolla | 04:44 | |
| *** MasterOfBugs has joined #openstack-kolla | 04:44 | |
| *** Jeffrey4l has joined #openstack-kolla | 04:44 | |
| fooliouno_ | duonghq: one other question regarding the container images. Do you build them from ubuntu source. what is your build date | 04:44 |
| duonghq | I use centos binary image | 04:45 |
| duonghq | 3 weeks ago | 04:46 |
| fooliouno_ | Ah . .ok. And do you use the kolla-kubernetes commands or helm | 04:46 |
| duonghq | I use helm | 04:47 |
| duonghq | kubectl | 04:47 |
| duonghq | due to kolla-kubernetes is not usable yet | 04:47 |
| fooliouno_ | But I am using the commands on the kolla-k8s page and they seem to bring up the containers well. Is there a link for the helm commands | 04:47 |
| duonghq | fooliouno_, some here: https://review.openstack.org/#/c/447356/ | 04:49 |
| fooliouno_ | duonghq: This is what I am using: http://pastebin.com/Yw1dEanS | 04:49 |
| sdake | fooliouno_ i have done so | 04:50 |
| fooliouno_ | sdake: baremetal, ubuntu 16.04, kubeadm, flannel is what I have. Is yours similar | 04:51 |
| duonghq | I still have trouble about keystone cannot connect to keystone-internal | 04:51 |
| sdake | fooliouno_ video of the work: https://www.youtube.com/watch?v=rHCCUP2odd8 | 04:51 |
| sdake | fooliouno_ in my setup I use centos | 04:52 |
| fooliouno_ | duonghq: I think I saw that error as well when my cluster was up. | 04:52 |
| duonghq | fooliouno_, ya | 04:52 |
| sdake | fooliouno_ we are working on a deployment guide here: https://review.openstack.org/#/c/447356/ | 04:53 |
| sdake | duonghq you know i recall that error | 04:53 |
| sdake | duonghq let me hunt in my notes and see if i have a referene to how i fixed it | 04:53 |
| fooliouno_ | sdake: I have been trying to get kolla-k8s to work using the kolla-kube commands on the quickstart guide pages for a few weeks now. Let me watch your video. Thanks! | 04:53 |
| duonghq | sdake, I resoled the resolve-conf | 04:53 |
| sdake | duonghq right - i have a different solution then that | 04:54 |
| *** Jeffrey4l has left #openstack-kolla | 04:55 | |
| sdake | fooliouno_ the quickstart guide pages are woefully out of date | 04:55 |
| sdake | fooliouno_ hence why we are working on a proper deployment guide | 04:55 |
| *** Jeffrey4l_ has quit IRC | 04:55 | |
| sdake | duonghq can you run openstack catalog list | 04:56 |
| sdake | duonghq or the command that is similar to that | 04:56 |
| duonghq | openstack user list is ok? | 04:56 |
| sdake | duonghq no, must be catalog list | 04:56 |
| fooliouno_ | sdake: unfortunately, thats what I have been struggling with. So, if I follow your new deployment guide all the way, I should be able to create nova instances as well in a multi node environment? | 04:56 |
| duonghq | yeah | 04:56 |
| sdake | duonghq yeah that works ? | 04:56 |
| *** Jeffrey4l has joined #openstack-kolla | 04:56 | |
| duonghq | http://paste.openstack.org/show/603852/ | 04:57 |
| duonghq | yes | 04:57 |
| sdake | duonghq winning - i have a solution for your problem | 04:57 |
| sdake | fooliouno_ if you can understand the guide as it stands today - it needs editing :) | 04:57 |
| duonghq | sdake, you don't say I should set dns_name? | 04:57 |
| sdake | duonghq the issue has nothign to do with dns | 04:58 |
| sdake | keystone 35357 is not externally routed | 04:58 |
| sdake | any command that requires administrative privileges will not work (such as user list, endpoint list etc) | 04:58 |
| sdake | duonghq there is a solution to this | 04:58 |
| sdake | fooliouno_ we are working on fixing the deploy guide and hope to have it published this week | 04:59 |
| sdake | (note iso standard says weeks start on sunday :) | 04:59 |
| fooliouno_ | sdake: that would be a life saver for me. Been struggling with it for too long to admit :) | 04:59 |
| sdake | fooliouno_ right -the docs are terrible | 04:59 |
| sdake | all across kolla unfortunately | 04:59 |
| sdake | fooliouno_ we don't have a professional documentation writer in kolla at this point in time that i am aware of | 05:00 |
| sdake | doesn't mean there aren't people doing that :) | 05:00 |
| sdake | fooliouno_ i feel your pain re struggling it took me 2 weeks to get kolla-kubernetes up and running and documented in a basic state in review https://review.openstack.org/#/c/447356/ | 05:01 |
| fooliouno_ | sdake: Let me start with that guide tomorrow. I'll be more engaged on this channel hereafter :) | 05:02 |
| duonghq | fooliouno_, which tz are you in? | 05:02 |
| sdake | fooliouno_ nice - tonight i'm going to fix up that deploy guide a bit so its more readable | 05:02 |
| fooliouno_ | Central | 05:02 |
| sdake | MST here | 05:02 |
| *** shashank_t_ has joined #openstack-kolla | 05:03 | |
| duonghq | sdake, [OT] how about the weather at your place these days? | 05:03 |
| sdake | duonghq i'm good with ot discussion | 05:04 |
| sdake | its actually cold atm | 05:04 |
| duonghq | hmm, nice | 05:04 |
| sdake | 17C | 05:04 |
| fooliouno_ | duonghq: sdake: Thanks for your help. I'm gonna call it a night. Will be back tomorrow. | 05:04 |
| duonghq | fooliouno_, night | 05:04 |
| sdake | duonghq although during the days its 90F which is pretyt warm | 05:04 |
| sdake | fooliouno_ enjoy - hope the video was helpful :) | 05:05 |
| sdake | fooliouno_ it can be done :) | 05:05 |
| duonghq | sdake, same here | 05:05 |
| sdake | so duonghq now that I understand what your problem is | 05:06 |
| sdake | duonghq let me sort out what was done to fix it | 05:06 |
| duonghq | roger | 05:06 |
| sdake | duonghq how did you generate your cloud.yaml? | 05:06 |
| duonghq | sdake, some script in project, then modify as common_workflow_config. | 05:08 |
| sdake | duonghq i see | 05:08 |
| duonghq | sdake, tools/build_example_yaml.py | 05:08 |
| sdake | duonghq you need to override... something | 05:09 |
| sdake | moment | 05:09 |
| *** jrobinson has joined #openstack-kolla | 05:14 | |
| duonghq | sdake, here is my cloud.yaml: http://paste.openstack.org/show/603854/ | 05:23 |
| sdake | [22:15:50] <sdake>duonghq http://paste.openstack.org/show/603853/ | 05:24 |
| sdake | [22:16:13] <sdake>duonghq you need line 35 | 05:24 |
| sdake | [22:16:30] <sdake>duonghq that defaults to false for whatever reason | 05:24 |
| sdake | [22:16:49] <sdake>you need it under the global.kolla.all namespace | 05:24 |
| *** cuongnv has quit IRC | 05:25 | |
| duonghq | sdake, I already set it to true for keystone | 05:25 |
| sdake | duonghq i dont see it in your cloud.yaml you pasted | 05:26 |
| duonghq | sdake, line 166 | 05:26 |
| sdake | duonghq it needs to be in the global.kolla.all section not in the keystone.all section | 05:27 |
| duonghq | sdake, ok | 05:27 |
| sdake | delete line 166 while your at it :) | 05:27 |
| duonghq | I only need to delete keystone chart and install again? | 05:28 |
| sdake | helm delete compute-kit --purge | 05:28 |
| sdake | helm nstall compute-kit | 05:28 |
| duonghq | can I just install keystone | 05:28 |
| sdake | you can try and report back :) | 05:28 |
| duonghq | ya | 05:28 |
| duonghq | I need a quick nap, cya | 05:29 |
| duonghq | thank you | 05:29 |
| *** cuongnv has joined #openstack-kolla | 05:32 | |
| sdake | np | 05:33 |
| spsurya | sayantani01: ping... | 05:42 |
| sayantani01 | spsurya Hi :) | 05:42 |
| spsurya | sayantani01: can we add some measures reagdring gates issue within this patch https://review.openstack.org/#/c/436462/ | 05:43 |
| spsurya | ? | 05:43 |
| spsurya | sayantani01: though all looks good | 05:44 |
| spsurya | but bifrost gate failure a big pain | 05:44 |
| sayantani01 | Since this patch deals with only documentation, wouldn't it be better to take up the gate issue in a separate patch? | 05:45 |
| *** skramaja has joined #openstack-kolla | 05:46 | |
| sayantani01 | spsurya, I am about to go to bed. I will take another look tomorrow :) | 05:46 |
| spsurya | sayantani01: sure | 05:47 |
| *** bmace has quit IRC | 05:47 | |
| *** sayantani01 has quit IRC | 05:48 | |
| *** bmace has joined #openstack-kolla | 05:48 | |
| *** zhurong has joined #openstack-kolla | 05:53 | |
| *** jascott1 has quit IRC | 05:54 | |
| *** jascott1 has joined #openstack-kolla | 05:55 | |
| *** shashank_t_ has quit IRC | 05:57 | |
| *** jascott1 has quit IRC | 05:59 | |
| zhubingbing | dough | 06:01 |
| zhubingbing | sup sdake suoghq | 06:01 |
| zhubingbing | duonghq | 06:01 |
| zhubingbing | ;) | 06:01 |
| zhubingbing | i know kolla-k8s dns problems | 06:01 |
| zhubingbing | ;) | 06:01 |
| zhubingbing | i fix it | 06:01 |
| zhubingbing | ;) | 06:01 |
| *** qiliang28 has joined #openstack-kolla | 06:04 | |
| *** haplo37_ has joined #openstack-kolla | 06:06 | |
| sdake | hey zhubingbing | 06:06 |
| zhubingbing | hi sdake | 06:06 |
| *** tonyb_ has joined #openstack-kolla | 06:06 | |
| duonghq | sdake, I'm installing compute-kit | 06:07 |
| duonghq | hope that it'll be fine | 06:07 |
| sdake | duonghq sounds good | 06:08 |
| sdake | working on docs now | 06:08 |
| sdake | need testing soon :) | 06:08 |
| duonghq | roger | 06:08 |
| *** bpetit has quit IRC | 06:08 | |
| *** bpetit has joined #openstack-kolla | 06:08 | |
| *** kbyrne_ has joined #openstack-kolla | 06:09 | |
| duonghq | hmm, got iscsi crashbackloo | 06:09 |
| duonghq | crashloopbackoff | 06:09 |
| *** unicell1 has joined #openstack-kolla | 06:10 | |
| duonghq | but what is 3.0.3-beta.1? it's from ocata? | 06:10 |
| *** unicell has quit IRC | 06:11 | |
| *** jrobinson has quit IRC | 06:11 | |
| *** tonyb has quit IRC | 06:13 | |
| *** flaper87 has quit IRC | 06:13 | |
| *** haplo37 has quit IRC | 06:13 | |
| *** kbyrne has quit IRC | 06:13 | |
| *** qiliang27 has quit IRC | 06:13 | |
| *** haplo37_ is now known as haplo37 | 06:13 | |
| *** kbyrne_ is now known as kbyrne | 06:13 | |
| *** qiliang28 is now known as qiliang27 | 06:13 | |
| duonghq | sdake, not sure why I get this potion of service list: | 06:13 |
| duonghq | keystone-admin 10.233.27.189 <none> 35357/TCP 9m | 06:13 |
| duonghq | keystone-internal 10.233.55.70 192.168.122.248 5000/TCP 9m | 06:13 |
| duonghq | keystone-public 10.233.22.176 192.168.122.248 5000/TCP 9m | 06:13 |
| sdake | duonghq its pre-ocata | 06:13 |
| sdake | duonghq its acutally 3.0.3 pre-release | 06:13 |
| duonghq | sdake, it's from dockerhub? | 06:13 |
| sdake | duonghq yes | 06:13 |
| duonghq | I'm using my local registry, hope that it will be fine | 06:14 |
| *** jrobinson has joined #openstack-kolla | 06:15 | |
| duonghq | hmm, I get keystone-manage-db-prwnp 1/1 Running 0 12m | 06:15 |
| duonghq | but keystone blames that Entrypoint WARNING: 2017/03/23 06:17:56 entrypoint.go:81: Resolving dependency for keystone-manage-db failed: Job keystone-manage-db is not completed yet | 06:15 |
| duonghq | ah still in migration process | 06:16 |
| Jeffrey4l | ls | 06:17 |
| zhubingbing | http://paste.openstack.org/show/603857/ | 06:18 |
| zhubingbing | sdake duoghq | 06:18 |
| openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: WIP: Deployment Guide Documentation https://review.openstack.org/447356 | 06:18 |
| zhubingbing | can u look it | 06:18 |
| sdake | zhubingbing i did look | 06:19 |
| sdake | zhubingbing i'm a bit winded (11pm here) - not sure what to make of it :) | 06:19 |
| zhubingbing | write a mess | 06:20 |
| zhubingbing | duonghq he know it | 06:20 |
| duonghq | zhubingbing, so we should match the kubeadm api with machine ip, | 06:21 |
| duonghq | weired, it doesn't describe in the official doc, | 06:21 |
| duonghq | iirc | 06:21 |
| *** masber has joined #openstack-kolla | 06:22 | |
| zhubingbing | ;) | 06:22 |
| sdake | zhubingbing makes up his own docs :) | 06:23 |
| duonghq | sdake, I don't know why but with admin_port_external: true the keystone-admin doesn't get external ip | 06:24 |
| sdake | duonghq paste your cloud.yaml | 06:25 |
| duonghq | http://paste.openstack.org/show/603858/ | 06:25 |
| *** tovin07_ has quit IRC | 06:26 | |
| zhubingbing | dns_name: 192.168.122.248 | 06:28 |
| zhubingbing | don;t use it | 06:28 |
| sdake | duonghq http://paste.openstack.org/show/603859/ | 06:29 |
| sdake | duonghq set that for your keystone section | 06:29 |
| zhubingbing | yeah sdake +1 | 06:30 |
| duonghq | sdake, so global.kolla.all doesn't set for kolla.keystone.all? | 06:30 |
| *** jaosorior has joined #openstack-kolla | 06:30 | |
| sdake | duonghq kolla.keystone.all overrides anything in globals.kolla.all | 06:30 |
| sdake | counterintuitive i know | 06:31 |
| duonghq | I think it's ok | 06:31 |
| duonghq | but I do not set anything in kolla.keystone.all | 06:31 |
| duonghq | or the default value of its overrides globals one? | 06:31 |
| sdake | yes you need to override the defaults | 06:31 |
| sdake | just cut and paste it at line 162 | 06:32 |
| duonghq | it'll surprising somebody | 06:32 |
| sdake | (in your cloud.ayml) | 06:32 |
| duonghq | ya, I'm deleting old compute-kti | 06:32 |
| zhubingbing | Do you think to record a video with how to deploy kolla-k8s | 06:33 |
| zhubingbing | ；） | 06:33 |
| duonghq | but even three keystone service got external ip, it still cannot contact to keystone-internal iirc | 06:33 |
| zhubingbing | https://review.openstack.org/447356 | 06:33 |
| zhubingbing | use this guide documentation | 06:33 |
| sdake | duonghq do you have a working external ip for keystone? | 06:34 |
| duonghq | for keystone-admin and keystone-public: yes, | 06:34 |
| sdake | duonghq keystone endpoint list working? | 06:35 |
| duonghq | not trying this yet | 06:35 |
| duonghq | mins, I need to bring up it again | 06:35 |
| sdake | zhubingbing are you using 101 atm? | 06:35 |
| duonghq | helm delete --purge compute-kits takes long time | 06:35 |
| sdake | duonghq about 8 mins on my gear ;) | 06:35 |
| zhubingbing | em yes sdake | 06:36 |
| sdake | zhubingbing how are you using it - i want to setup kubeadm on it and run through the documentation | 06:36 |
| sdake | zhubingbing to test it and see if it works | 06:36 |
| zhubingbing | ok | 06:36 |
| duonghq | so it will be worse on my mini PC | 06:36 |
| sdake | zhubingbing better yet, you can test it if you like | 06:36 |
| zhubingbing | no prolemes i will test in | 06:36 |
| sdake | although for some reason 102 and 103 are not coming up | 06:37 |
| openstackgerrit | Lingyong Xu proposed openstack/kolla master: Using assertIsNone() instead of assertEqual(None) https://review.openstack.org/448923 | 06:37 |
| sdake | zhubingbing i'm going to go look at 102 and 103 and see what the eal is with it | 06:37 |
| zhubingbing | ok | 06:37 |
| openstackgerrit | Lingyong Xu proposed openstack/kolla master: Using assertIsNone() instead of assertEqual(None) https://review.openstack.org/448923 | 06:39 |
| sdake | zhubingbing can you try out 101 and taint it so you can run on just 101 | 06:39 |
| *** masber has quit IRC | 06:39 | |
| openstackgerrit | jianyi proposed openstack/kolla-ansible master: Add zun ansible role https://review.openstack.org/416179 | 06:40 |
| zhubingbing | hmm i can ,i am use halcyon to set my dev env | 06:40 |
| sdake | zhubingbing i see, I want to run through the instructions in the deployment guide - on bare metal | 06:41 |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible master: Modify the permission of directory https://review.openstack.org/448932 | 06:42 |
| zhubingbing | hmm yeah | 06:42 |
| sdake | zhubingbing the review i just pushed - so I'll do that now | 06:42 |
| zhubingbing | i will fix the gate , let the same as the deployment guide | 06:43 |
| *** shashank_t_ has joined #openstack-kolla | 06:47 | |
| *** satyar has joined #openstack-kolla | 06:48 | |
| *** rwsu has joined #openstack-kolla | 06:48 | |
| duonghq | sdake, if I set the keystone to your snippet, the keystone-internal does not get external ip | 06:48 |
| sdake | duonghq what about keystone external ? | 06:49 |
| sdake | keystone admin/ | 06:49 |
| sdake | i mean | 06:49 |
| *** shashank_t_ has quit IRC | 06:50 | |
| duonghq | sdake, here | 06:50 |
| duonghq | keystone-admin 10.233.14.21 192.168.122.248 35357/TCP 2m | 06:50 |
| duonghq | keystone-internal 10.233.36.85 <none> 5000/TCP 2m | 06:50 |
| duonghq | keystone-public 10.233.57.160 192.168.122.248 5000/TCP 2m | 06:50 |
| sdake | well thats progress :) | 06:50 |
| duonghq | openstack endpoint list still blame: Failed to contact the endpoint at http://keystone-internal:5000/v3 for discovery. Fallback to using that endpoint as the base url. | 06:51 |
| sdake | still progress | 06:51 |
| duonghq | yeah, I got this error again | 06:51 |
| sdake | prior to adding the snippet did keystone-admin have an external ip? | 06:51 |
| duonghq | I got ip for keystone-admin and public in place before | 06:51 |
| duonghq | I managed to have it sometime | 06:51 |
| duonghq | but w/o the snippet, it can not have the ip, absolutely | 06:52 |
| sdake | it being which object? | 06:52 |
| duonghq | sorry? | 06:52 |
| sdake | without the snipped what does nothave the external ip? | 06:53 |
| sdake | i provided two snippets want to know which ones triggered which behavior | 06:53 |
| duonghq | hmm, the keystone snippet enable keystone-admin ip, but not keystone-internal | 06:53 |
| sdake | ok | 06:54 |
| duonghq | the other does not much thing | 06:54 |
| duonghq | (it my case) | 06:54 |
| sdake | prior to the snippet keystone-internla was enabled correct? | 06:54 |
| duonghq | iirc, yes | 06:55 |
| duonghq | http://paste.openstack.org/show/603861/ the current error | 06:55 |
| sdake | set this to true: node_port_enabled: false | 06:55 |
| duonghq | sdake, Error: release keystone failed: Service "keystone-public" is invalid: spec.ports[0].nodePort: Invalid value: 5000: provided port is not in the valid range. The range of valid ports is 30000-32767 | 06:57 |
| sdake | zhubingbing any ideas? | 06:58 |
| duonghq | got I before, so I left it disable | 06:59 |
| zhubingbing | i look it | 06:59 |
| zhubingbing | wait 1min | 06:59 |
| zhubingbing | i thinks keystone craete endpoint have some problems | 07:00 |
| *** tovin07_ has joined #openstack-kolla | 07:01 | |
| *** sbezverk_ has joined #openstack-kolla | 07:01 | |
| *** targon has joined #openstack-kolla | 07:01 | |
| zhubingbing | If i can log in to the machine operation just fine | 07:02 |
| zhubingbing | duonghq | 07:02 |
| zhubingbing | ；） | 07:02 |
| duonghq | it's not possible here, | 07:02 |
| duonghq | quite sad | 07:02 |
| *** dmellado has quit IRC | 07:03 | |
| duonghq | zhubingbing, do you have network setting in systemd? | 07:03 |
| duonghq | calico... | 07:03 |
| zhubingbing | hmm | 07:03 |
| zhubingbing | Ayesha | 07:03 |
| zhubingbing | i use flnall | 07:03 |
| *** bpetit has quit IRC | 07:03 | |
| zhubingbing | flanel | 07:03 |
| zhubingbing | ;)but i think | 07:04 |
| zhubingbing | you can attach public the machetes ? | 07:04 |
| *** kfox1111 has quit IRC | 07:04 | |
| *** kfox1111 has joined #openstack-kolla | 07:04 | |
| *** sbezverk has quit IRC | 07:05 | |
| *** bpetit has joined #openstack-kolla | 07:05 | |
| duonghq | zhubingbing, sorry? | 07:05 |
| sdake | duonghq try pinging bradjones or sbezverk_ when they appear | 07:08 |
| *** dmellado has joined #openstack-kolla | 07:08 | |
| sdake | i have to hit the rack | 07:08 |
| duonghq | sdake, roger | 07:08 |
| sdake | zhubingbing duonghq if either or both of you could go throug hthe deploy guide that would be good :) | 07:08 |
| sdake | and provide a review | 07:08 |
| sdake | fix mistakes/etc | 07:08 |
| sdake | (in comments or by updating the review directly | 07:09 |
| duonghq | got it | 07:09 |
| zhubingbing | hmm | 07:09 |
| zhubingbing | ok | 07:09 |
| sdake | up since 5am - need atleast 6-7 hrs a night of sleep :) | 07:09 |
| sdake | enjoy | 07:09 |
| *** jmccarthy has joined #openstack-kolla | 07:09 | |
| zhubingbing | ;) | 07:09 |
| *** jrobinson has quit IRC | 07:11 | |
| duonghq | night sdake | 07:17 |
| jaosorior | mandre: hey, could you check this out https://review.openstack.org/#/c/446911/ ? | 07:24 |
| mandre | jaosorior: looks like you did it :) | 07:26 |
| mandre | jaosorior: what about commenting out the 'Listen 443 https' line intead of deleting it, to be more consistent with the change made to httpd.conf for port 80? | 07:29 |
| *** cuongnv has quit IRC | 07:29 | |
| jaosorior | mandre: sure | 07:29 |
| mandre | jaosorior: it's not a big deal though | 07:30 |
| jaosorior | mandre: I'll do that, it'll be in a couple of hours though, since I'm finishing up some other task | 07:32 |
| *** manheim has joined #openstack-kolla | 07:33 | |
| *** manheim has quit IRC | 07:36 | |
| *** manheim has joined #openstack-kolla | 07:36 | |
| *** mnasiadka has joined #openstack-kolla | 07:39 | |
| *** Serlex has joined #openstack-kolla | 07:52 | |
| *** matrohon has joined #openstack-kolla | 07:53 | |
| openstackgerrit | Jun Hu proposed openstack/kolla master: Fix ironic-donductor failed when build with binary https://review.openstack.org/448977 | 07:55 |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla master: Fix oslo_debug_helper not running for kolla/tests directory https://review.openstack.org/444398 | 07:55 |
| openstackgerrit | Merged openstack/kolla master: Add libosinfo package for source installation https://review.openstack.org/445701 | 07:56 |
| openstackgerrit | Jun Hu proposed openstack/kolla master: Fix ironic-donductor failed when build with binary https://review.openstack.org/448977 | 07:57 |
| *** cuongnv has joined #openstack-kolla | 07:58 | |
| spsurya | mandre: https://review.openstack.org/444398 your commnets has been incorporated and some addition information for your query | 07:59 |
| openstackgerrit | Jun Hu proposed openstack/kolla master: Fix ironic-donductor failed when build with binary https://review.openstack.org/448977 | 08:03 |
| *** rmart04 has joined #openstack-kolla | 08:03 | |
| mandre | jaosorior: I noticed we forgot to disable port 80 for a few images, I'm going to push a patch and have yours depend on it... I'll also edit your patch to comment out the line | 08:04 |
| *** shardy has joined #openstack-kolla | 08:05 | |
| jaosorior | mandre: oh, that would be great dude, thanks | 08:05 |
| *** pcaruana has joined #openstack-kolla | 08:06 | |
| *** flaper87 has joined #openstack-kolla | 08:06 | |
| *** flaper87 has quit IRC | 08:07 | |
| *** flaper87 has joined #openstack-kolla | 08:07 | |
| *** rmart04 has quit IRC | 08:07 | |
| mandre | spsurya: the thing is, I didn't see a difference with explicitely listing directories or not | 08:12 |
| spsurya | mandre: after adding the kolla/tests. We are able to run all the remaining tests | 08:13 |
| spsurya | around 30 cases are in tests and 27 are in kolla/tests | 08:14 |
| spsurya | so now we are able to run 57 cases | 08:15 |
| mandre | spsurya: "tox -e debug" runs the same number of tests for me with and without your patch, unless I'm missing something | 08:15 |
| *** mkoderer has joined #openstack-kolla | 08:17 | |
| mandre | spsurya: if I switch 'commands' to 'commands1', it does run all the 57 tests, and also much faster | 08:18 |
| mandre | spsurya: idk what commands1 stands for | 08:18 |
| *** matrohon has quit IRC | 08:20 | |
| *** Administrator_ has quit IRC | 08:20 | |
| spsurya | mandre: i have seen somewhr before that is why i change that ..... but after your comment i reverted | 08:21 |
| *** manheim has quit IRC | 08:21 | |
| *** Administrator_ has joined #openstack-kolla | 08:21 | |
| spsurya | so eaxctly not much idea but it works | 08:21 |
| *** manheim has joined #openstack-kolla | 08:21 | |
| spsurya | mandre: and yes "tox -e debug" doesn't run all the cases | 08:21 |
| spsurya | it runs only 30 If i see correclty right now | 08:22 |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla master: Fix oslo_debug_helper not running for kolla/tests directory https://review.openstack.org/444398 | 08:23 |
| spsurya | mandre: with "commands1" below is the logs where we can see the difference w.r.t "commands" only | 08:27 |
| spsurya | https://www.irccloud.com/pastebin/QYHrhMCJ/ | 08:27 |
| spsurya | https://www.irccloud.com/pastebin/dmfOTlmh/ | 08:28 |
| spsurya | mandre: please see you got something of it | 08:28 |
| *** rmart04 has joined #openstack-kolla | 08:30 | |
| *** matrohon has joined #openstack-kolla | 08:35 | |
| *** rmart04 has quit IRC | 08:35 | |
| *** egonzalez has joined #openstack-kolla | 08:44 | |
| *** bjolo has joined #openstack-kolla | 08:45 | |
| bjolo | morning | 08:45 |
| duonghq | morning bjolo | 08:45 |
| duonghq | morning eglute | 08:45 |
| duonghq | egonzalez, | 08:45 |
| bjolo | hi duonghq | 08:45 |
| duonghq | sorry eglute | 08:45 |
| bjolo | hows vietnam? | 08:46 |
| egonzalez | morning | 08:46 |
| duonghq | bjolo, weather is becoming summer, quite hot now, it'll be colder in weekend (with rain) | 08:46 |
| hawi | hi. question - is there some way to make Horizon running quicker for multiple users? as i undestand Horizon is just one container, yes? | 08:47 |
| *** mgoddard has joined #openstack-kolla | 08:47 | |
| duonghq | egonzalez, if you have spare time, can you take a look at this: https://review.openstack.org/#/c/425833/ iirc, the ubuntu-binary gate red due to fixed bug | 08:47 |
| egonzalez | duonghq, error in ubuntu in this PS is unrelated tho, rechecked to ensure | 08:48 |
| duonghq | egonzalez, ara error was fixed once | 08:49 |
| spsurya | bjolo: morning | 08:49 |
| duonghq | I'm not sure why it still appear here | 08:50 |
| spsurya | duonghq: hi... | 08:50 |
| egonzalez | hawi, for running quicker what you mean? if you have a lot of users, have to scale horizon container with more nodes | 08:50 |
| egonzalez | duonghq, the error was about mariadb not connecting. is not the same | 08:50 |
| spsurya | mandre: ^^ | 08:50 |
| *** jaosorior is now known as jaosorior_brb | 08:55 | |
| *** qiliang28 has joined #openstack-kolla | 08:56 | |
| openstackgerrit | Merged openstack/kolla master: Update to Kubernetes 1.5.4 https://review.openstack.org/447217 | 08:58 |
| *** shardy has quit IRC | 08:58 | |
| *** qiliang27 has quit IRC | 08:59 | |
| *** qiliang28 is now known as qiliang27 | 08:59 | |
| *** shardy has joined #openstack-kolla | 09:00 | |
| *** openstackgerrit has quit IRC | 09:03 | |
| *** athomas has joined #openstack-kolla | 09:03 | |
| *** openstackgerrit has joined #openstack-kolla | 09:03 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Improved documentation for central logging. https://review.openstack.org/444477 | 09:03 |
| hrw | mnaser: THANK YOU! | 09:09 |
| hrw | sdake: thanks too - would not noticed without you | 09:10 |
| *** david-lyle has quit IRC | 09:11 | |
| *** jrist has quit IRC | 09:12 | |
| *** david-lyle has joined #openstack-kolla | 09:13 | |
| *** jrist has joined #openstack-kolla | 09:13 | |
| spsurya | https://www.irccloud.com/pastebin/8MpW4q6I/ | 09:13 |
| spsurya | egonzalez: what is the exact difference between commands and commands1, if we use in tox.ini debug environment | 09:13 |
| spsurya | please take a look on above log | 09:14 |
| egonzalez | pbourke, is this a bug or is expected to have build logs from mitaka in master? https://github.com/openstack/kolla/blob/master/docker/base/Dockerfile.j2#L148 | 09:15 |
| *** cu5 has quit IRC | 09:22 | |
| *** kbaegis has joined #openstack-kolla | 09:22 | |
| egonzalez | spsurya, to be fair I don't know. Cannot find any reference for command1 in other tox files | 09:25 |
| *** david-lyle_ has joined #openstack-kolla | 09:27 | |
| egonzalez | spsurya, I think is a typo, if you look at the other changes made to the same bug, none include command1, only add the correct tests path | 09:27 |
| *** david-lyle has quit IRC | 09:27 | |
| *** cu5 has joined #openstack-kolla | 09:28 | |
| *** cu5 has quit IRC | 09:28 | |
| spsurya | egonzalez: but don't know why it is running 57 even take less time to execute :( | 09:29 |
| spsurya | yeh it is typo | 09:29 |
| egonzalez | try to execute it verbose and see what tests are made | 09:29 |
| *** fooliouno_ has quit IRC | 09:32 | |
| *** manheim has quit IRC | 09:37 | |
| *** manheim has joined #openstack-kolla | 09:37 | |
| spsurya | egonzalez: sorry | 09:42 |
| spsurya | but i have not idea how to execute in verbose | 09:42 |
| *** matrohon has quit IRC | 09:42 | |
| *** matrohon has joined #openstack-kolla | 09:43 | |
| *** rmart04 has joined #openstack-kolla | 09:43 | |
| egonzalez | spsurya, tox -e (test_to_execute) --debug | 09:44 |
| egonzalez | spsurya, nope, --debug does not exist , sorry | 09:44 |
| spsurya | egonzalez: roger | 09:45 |
| spsurya | i think only debug would be ok | 09:45 |
| spsurya | for dingle test case to run | 09:45 |
| spsurya | single* | 09:46 |
| *** rmart04 has quit IRC | 09:47 | |
| egonzalez | spsurya, can set an output json file with detailed info --result-json /tmp/result.json | 09:48 |
| spsurya | egonzalez: roger | 09:51 |
| spsurya | egonzalez: http://paste.openstack.org/show/603878/ but seems like getting nothing of it | 09:56 |
| *** duonghq has quit IRC | 09:59 | |
| *** MasterOfBugs has quit IRC | 09:59 | |
| *** satyar has quit IRC | 10:00 | |
| egonzalez | spsurya, with command1 is not running all tests http://paste.openstack.org/show/603879/ | 10:01 |
| spsurya | egonzalez: you mean commands1 not running any case | 10:03 |
| egonzalez | spsurya, yep | 10:03 |
| spsurya | egonzalez: updated on same http://paste.openstack.org/show/603880/ | 10:07 |
| spsurya | so why the output of ``tox -e debug`` showing 57 tests passed ? | 10:08 |
| spsurya | on stdout | 10:08 |
| openstackgerrit | Martin André proposed openstack/kolla master: RHEL: Add mod_ssl for services running over httpd https://review.openstack.org/446911 | 10:08 |
| openstackgerrit | Martin André proposed openstack/kolla master: Prevent apache from listening on port 80 https://review.openstack.org/449039 | 10:08 |
| openstackgerrit | Martin André proposed openstack/kolla master: Prevent apache from listening on port 80 for heat-base https://review.openstack.org/449040 | 10:08 |
| *** kbaegis has quit IRC | 10:11 | |
| egonzalez | spsurya, replied http://paste.openstack.org/show/603883/ | 10:13 |
| egonzalez | spsurya, as it is a wrong command statement, is not using the content of the command1 | 10:14 |
| *** pbourke has quit IRC | 10:14 | |
| *** pbourke has joined #openstack-kolla | 10:16 | |
| *** kbaegis has joined #openstack-kolla | 10:16 | |
| *** cuongnv has quit IRC | 10:18 | |
| *** kbaegis has left #openstack-kolla | 10:18 | |
| spsurya | egonzalez: got that command1 content is not being used as per --showconfig | 10:19 |
| egonzalez | yep | 10:19 |
| *** rmart04 has joined #openstack-kolla | 10:21 | |
| spsurya | egonzalez: sorry to say but the only thing is bugging me is ``tox -e debug`` output | 10:22 |
| spsurya | :( | 10:22 |
| spsurya | thnks for getting the command1 content | 10:23 |
| *** rmart04 has quit IRC | 10:25 | |
| hrw | omg... jenkins gave +1 to my patch... | 10:33 |
| spsurya | egonzalez: from these logs it seems it running all the slowest tests | 10:34 |
| spsurya | and skiping the third argument | 10:34 |
| *** manheim has quit IRC | 10:35 | |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla master: Fix oslo_debug_helper not running for kolla/tests directory https://review.openstack.org/444398 | 10:37 |
| *** athomas has quit IRC | 10:41 | |
| *** rabel has joined #openstack-kolla | 10:49 | |
| *** zhubingbing has quit IRC | 10:49 | |
| *** rhallisey has joined #openstack-kolla | 10:58 | |
| *** jaosorior_brb is now known as jaosorior | 11:02 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: ironic-pxe: handle non-x86 architectures https://review.openstack.org/434817 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: enable all images enabled for Ubuntu https://review.openstack.org/432787 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: Add support for non-x86 architectures (aarch64, ppc64le) https://review.openstack.org/430940 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: openstack-base/kolla-toolbox: use mariadb-devel for CentOS and !x86-64 https://review.openstack.org/444832 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: nova-libvirt: handle ppc64le https://review.openstack.org/434810 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: ceph: move ceph-fuse package to cephfs-fuse image https://review.openstack.org/440634 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: kubernetes: disable for architectures other than x86-64 https://review.openstack.org/445638 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: nova-compute: handle rtslib(-fb) for debian-binary builds too https://review.openstack.org/435941 | 11:03 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: move to stretch https://review.openstack.org/434453 | 11:03 |
| *** jistr is now known as jistr|afk | 11:06 | |
| nea1 | Hi, has anyone tested qos on ocata with ovs? - I'm setting a rule but there is nothing happening the logs look fine "Port cb3d964a-0f5f-4787-bd08-b245090643b4 updated. Details: {u'profile': {}, u'network_qos_policy_id': u'1fc76301-e0cd-477e-a2c6-4b062fdfef88'," | 11:13 |
| nea1 | but it doesn't seam to have any effect | 11:13 |
| nea1 | iptables on the node don't contain any new entries either | 11:14 |
| *** tovin07_ has quit IRC | 11:15 | |
| *** matrohon has quit IRC | 11:16 | |
| hrw | I will send a patch which will make some of centos/rhel/oraclelinux images fail. and this will be GOOD | 11:17 |
| *** tovin07 has quit IRC | 11:19 | |
| *** manheim has joined #openstack-kolla | 11:21 | |
| *** athomas has joined #openstack-kolla | 11:21 | |
| *** satyar has joined #openstack-kolla | 11:23 | |
| nea1 | hrw: that actually doesn't sound GOOD | 11:23 |
| hrw | nea1: no, it will be good | 11:23 |
| Serlex | lol? | 11:24 |
| hrw | nea1: when you tell "yum install pkg1 pkg2" it can say "there is no pkg1" and install just pkg2 and be happy. | 11:24 |
| hrw | then your image is useless as it does not contain pkg1 which is crucial to make image usable | 11:24 |
| hrw | apt in debian world in such situation says "no pkg1, cannot continue" | 11:25 |
| hrw | my patch will make yum say "no pkg1, cannot continue" as well == broken images do not build == GOOD | 11:26 |
| hrw | INFO:kolla.image.build.openstack-base:No package Percona-Server-shared-56 available. | 11:26 |
| hrw | INFO:kolla.image.build.openstack-base:Error: Not tolerating missing names on install, stopping. | 11:26 |
| hrw | INFO:kolla.image.build.openstack-base: | 11:27 |
| hrw | INFO:kolla.image.build.openstack-base:Removing intermediate container d5e506330b08 | 11:27 |
| hrw | ERROR:kolla.image.build.openstack-base:Error'd with the following message | 11:27 |
| nea1 | ok ^^ | 11:27 |
| nea1 | back to my qos test, it seams like in the netns there are iptable rules to mark the packages, but tc doesn't contain anything so the throttle never happens | 11:28 |
| jaosorior | mandre: thanks for updating the patch | 11:31 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: make yum fail on missing packages https://review.openstack.org/449086 | 11:32 |
| hrw | can you guys take a look at that patch and plus it? | 11:33 |
| nea1 | hm... no but iptables don't change if add a rule so it seams like the marks set there are not for qos | 11:33 |
| *** masber has joined #openstack-kolla | 11:33 | |
| hrw | sdake: https://review.openstack.org/449086 | 11:35 |
| * hrw off | 11:35 | |
| *** manjeets has quit IRC | 11:36 | |
| *** athomas has quit IRC | 11:43 | |
| *** vi01et03_ has joined #openstack-kolla | 11:50 | |
| *** vi01et03_ has quit IRC | 11:53 | |
| masber | hi, I am having lots of issues making br-ex to work after a clean installation of openstack using kolla-ansible 4.0.0. I have been reading for a while and I think my problem is that openvswitch is not connecting br-ex with the openstack router gateway. I know troubleshooting these things are quite painful and takes a lot of time, but I really ran out of ideas and don't know what to do next. These are some of the details http://paste.openstack.org/s | 11:55 |
| masber | how/603896/ | 11:55 |
| masber | http://paste.openstack.org/show/603896/ | 11:55 |
| *** athomas has joined #openstack-kolla | 12:00 | |
| *** shardy is now known as shardy_lunch | 12:08 | |
| *** srwilkers has joined #openstack-kolla | 12:22 | |
| *** zhurong has quit IRC | 12:22 | |
| *** jistr|afk is now known as jistr | 12:27 | |
| *** matrohon has joined #openstack-kolla | 12:38 | |
| *** qiliang27 has quit IRC | 12:41 | |
| *** mkoderer has quit IRC | 12:47 | |
| sdake | zhenguo any luck? | 12:48 |
| sdake | zhenguo disregard plz | 12:49 |
| mnaser | o/ | 12:49 |
| mnaser | morning | 12:49 |
| sdake | sup folks | 12:49 |
| srwilkers | morning sdake | 12:49 |
| sdake | sup srwilkers | 12:49 |
| sbezverk_ | good morning | 12:49 |
| mnaser | we dont have a broken gate | 12:49 |
| mnaser | so thats a thing | 12:49 |
| sbezverk_ | Jeffrey4l: ping | 12:49 |
| mnaser | does recheck automatically rebase | 12:50 |
| mnaser | i think this discussion was had yesterday | 12:50 |
| *** athomas has quit IRC | 12:50 | |
| sdake | mnaser yes | 12:51 |
| sdake | mnaser it doesn't rebsae the gerrit commit in gerrit | 12:51 |
| sdake | mnaser it does rebase on master before running the gate with zuul cloner | 12:52 |
| *** schwicht has joined #openstack-kolla | 12:52 | |
| srwilkers | not much sdake, just enjoying the quiet office | 12:52 |
| * portdirect throws things at srwilkers | 12:53 | |
| mnaser | im going to recheck the stuff that is blocked by gate and got a +2 already | 12:53 |
| *** dave-mccowan has joined #openstack-kolla | 12:54 | |
| sdake | mnaser here is evidence: https://review.openstack.org/#/c/448043/ | 12:55 |
| sdake | mnaser this patch was jenkins -1 on the check jobs prior to the recheck I did | 12:56 |
| sdake | mnaser no rebase involved | 12:56 |
| *** jascott1 has joined #openstack-kolla | 12:56 | |
| sbezverk_ | sdake: srwilkers: appreciate your review of this one https://review.openstack.org/441550 !!! do not workflow it I need kfox1111 ok too. | 12:56 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: make yum fail on missing packages https://review.openstack.org/449086 | 12:56 |
| hrw | bug opened and mentioned in commit | 12:57 |
| sdake | sbezverk_ notice there is a gate failure | 12:57 |
| sdake | sbezverk_ any thoughts on that? | 12:57 |
| *** shardy_lunch is now known as shardy | 12:57 | |
| hrw | mnaser: great work on gate! | 12:57 |
| sdake | oh right - this patch breaks subuntu | 12:57 |
| sdake | on 2.0.2 it appears only? | 12:57 |
| sbezverk_ | sdake: which one? | 12:58 |
| sdake | sbezverk_ the one you linked | 12:58 |
| mnaser | thank you hrw | 12:58 |
| sdake | https://review.openstack.org/#/c/441550/ | 12:58 |
| sbezverk_ | sdake: that job has NO any change, so it is not related | 12:58 |
| sbezverk_ | sdake: if you check the log, it got stuck on ceph cluster thing, and ceph is not my thing ;) | 12:59 |
| sdake | sbezverk_ i did look at the log | 13:00 |
| *** mnasiadka has quit IRC | 13:00 | |
| sdake | i dontsee stuck on ceph - still learning how to understand these gates output | 13:00 |
| *** jascott1 has quit IRC | 13:00 | |
| hrw | mnaser: now debuging why jenkins gave -1 took me minute | 13:01 |
| mnaser | hrw the per-image build folder is so nice right? | 13:01 |
| hrw | mnaser: yes, it is | 13:01 |
| mnaser | i dont know whoever implemented it but much love | 13:01 |
| mnaser | all i did was enable it | 13:01 |
| hrw | mnaser: once I found it I fall in love | 13:01 |
| mnaser | no more 5 mb files opening in browser :> | 13:02 |
| sbezverk_ | sdake: http://logs.openstack.org/50/441550/119/check/gate-kolla-kubernetes-deploy-ubuntu-binary-2-ceph-nv/4cb4600/logs/openstack/kolla-ceph-osd0.txt | 13:02 |
| mnaser | we can thank harlowja for adding log_dir which makes gates clean today - https://bugs.launchpad.net/kolla/+bug/1586478 | 13:03 |
| openstack | Launchpad bug 1586478 in kolla "Tech-debt: use image specific loggers" [Wishlist,Fix released] - Assigned to Joshua Harlow (harlowja) | 13:03 |
| sdake | sbezverk_ last night we had a question as to how to export the keystone-admin endpoint and keystone-internal endpoint - we have been through this before (me and you) and I couldn't recall what advice to give to duounghq - moment on logs: | 13:03 |
| sdake | http://eavesdrop.openstack.org/irclogs/%23openstack-kolla/%23openstack-kolla.2017-03-23.log.html#t2017-03-23T04:51:46 | 13:04 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: openstack-base/kolla-toolbox: use mariadb-devel for CentOS and !x86-64 https://review.openstack.org/444832 | 13:04 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: ceph: move ceph-fuse package to cephfs-fuse image https://review.openstack.org/440634 | 13:04 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: nova-compute: handle rtslib(-fb) for debian-binary builds too https://review.openstack.org/435941 | 13:04 |
| sbezverk_ | sdake: tools/build_local_admin_keystonerc.sh | 13:05 |
| hrw | mnaser: I still open that 5MB file to find error but do not have to dig in it | 13:05 |
| mnaser | hrw it shouldnt have any logs at all | 13:05 |
| mnaser | console.log should be pretty much empty | 13:06 |
| sdake | sbezverk_ yes i know that command builds the adminrc | 13:06 |
| mnaser | unless there's a traceback or something | 13:06 |
| sbezverk_ | sdake: for external consumtion, you need to run it with "ext" parameter | 13:06 |
| hrw | mnaser: ah, right - its 113K now | 13:06 |
| sdake | sbezverk_ read the log for our diag | 13:06 |
| mnaser | :D | 13:06 |
| sdake | dialog | 13:06 |
| hrw | mnaser: fast network connection and I did not noticed as new what I want to look for | 13:06 |
| hrw | mnaser: magic phrase usually is "Expected image" | 13:07 |
| mnaser | yeah exactly | 13:07 |
| mnaser | i should probably make a little write up about that at some point | 13:07 |
| sdake | sbezverk_ looking at that code, that sounds fine | 13:07 |
| *** rwallner has joined #openstack-kolla | 13:07 | |
| sdake | sbezverk_ the problem is admin privs are often needed to do many operations | 13:08 |
| sdake | sbezverk_ and when we went through debug (adding my cloud.yaml snippets) there was no longer an external ip for keystone | 13:08 |
| *** rwallner has quit IRC | 13:08 | |
| sdake | https://github.com/openstack/kolla-kubernetes/blob/master/tools/build_local_admin_keystonerc.sh#L10 | 13:08 |
| sbezverk_ | sdake: please ifyou see someone tries to resolve local problem by modifying common_workflow_config, stop them right away! they do something wrong | 13:08 |
| *** rwallner has joined #openstack-kolla | 13:09 | |
| hrw | 3 more jenkins jobs and my patchset should be V:+1 :d | 13:09 |
| sdake | sbezverk_ well people gonna do what people gonna do when there are no docs ;) | 13:09 |
| *** zhubingbing_ has joined #openstack-kolla | 13:09 | |
| sdake | sbezverk_ its the only way to experience the code and understand the problems | 13:10 |
| sbezverk_ | sdake: well, then do not complain ;) | 13:10 |
| sdake | sbezverk_ I wasn't ocmplaining | 13:10 |
| hrw | found new bug with that yum change ;D | 13:10 |
| sdake | sbezverk_ I was asking a question - what cloud.yaml snippets are needed to get external keystone to work properly | 13:10 |
| sbezverk_ | sdake: I was not refering to you, but to duonghq | 13:10 |
| sdake | sbezverk_ i understand | 13:10 |
| sbezverk_ | just two lines are needed. | 13:11 |
| sdake | sbezverk_ the defaults do not work correctly for both the admin (35357) and non-admin case (port 5000) | 13:11 |
| *** haplo37 has quit IRC | 13:11 | |
| hrw | ironic-conductor was broken for centos/source builds and no one noticed | 13:12 |
| sdake | hrw you noticed :) | 13:12 |
| hrw | sdake: jenkins did | 13:13 |
| sdake | hrw openstack doesn't use jenkins | 13:13 |
| sdake | hrw openstack instead uses zuul | 13:13 |
| sdake | jenkins didn't scale to the 2kjobs/hour that openstack requires | 13:13 |
| hrw | sdake: whatever. mails are signed "Jenkins" :D | 13:14 |
| sbezverk_ | sdake: here is my cloud.yaml and I highlighted what you need http://paste.openstack.org/show/603915/ | 13:14 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: ironic-conductor: add missing comma for centos/source build https://review.openstack.org/449135 | 13:14 |
| sdake | sbezverk_ note i am reviewing your patch but struggling asi justwoke up - brain booting | 13:15 |
| sdake | sbezverk_ there is this vague comment here on line 42: https://review.openstack.org/#/c/447356/5/doc/source/deployment-guide.rst | 13:16 |
| openstackgerrit | Christian Berendt proposed openstack/kolla master: Fix name of the mistral-dashboard horizon plugin https://review.openstack.org/449136 | 13:16 |
| sdake | sbezverk_ it says to prune cloud.yaml | 13:17 |
| sdake | sbezverk_ I spoke with kevin on irc about it, he said to use generate_cloud_yaml_default (iirc) to generate just the absolute changes from the defaults that are needed | 13:17 |
| sdake | sbezverk_ i am struggling with this activity | 13:17 |
| hrw | so if someone has spare 2 minutes then please look at https://review.openstack.org/449135 one | 13:17 |
| sbezverk_ | sdake: first time I hear about generate_cloud_yaml_default | 13:19 |
| sdake | # TODO(sdake): This file is generated by | 13:20 |
| sdake | # /opt/kolla-kubernetes/tools/build_example_yaml.py | 13:20 |
| sdake | # https://etherpad.openstack.org/p/kolla-cloudyaml | 13:20 |
| sdake | # The example should be pruned to just the changes needed to run, rather | 13:20 |
| sdake | # than force all defaults to their default value. Its hard to see what | 13:20 |
| sdake | # actually changed this way. | 13:20 |
| sdake | buidl_exaple_yaml.py rather | 13:20 |
| *** haplo37 has joined #openstack-kolla | 13:20 | |
| sdake | how does one remove tainting from the master node? | 13:21 |
| sdake | zhubingbing_ you awake? | 13:21 |
| zhubingbing_ | hear | 13:21 |
| zhubingbing_ | sbezverk_ | 13:22 |
| mnaser | wow good catch hrw | 13:22 |
| zhubingbing_ | we should change something in gate ;) | 13:22 |
| mnaser | im trying to see how we can avoid having that happen again | 13:22 |
| sbezverk_ | sdake: whatever you do, please make sure you leave build_example_yaml.py as is | 13:23 |
| sdake | sbezverk_ i wasn't going to modify it | 13:23 |
| sdake | sbezverk_ i was going to include it in th edocumentation hoepfully | 13:23 |
| sdake | sbezverk_ what i'd really like is to understand how to generate a cloud.yaml file without having a phd in kolla-kubernetes | 13:23 |
| sbezverk_ | sdake: ah ok, because for my case it does, is making perfect sense | 13:24 |
| sdake | sbezverk_ huh? | 13:24 |
| sbezverk_ | zhubingbing_: hey, you pinged me? | 13:24 |
| sdake | sbezverk_ we are all attempting to sort out how to deploy kolla-kubernetes outside a dev env and outside the gate | 13:25 |
| sdake | sbezverk_ all the pings re the same thing | 13:25 |
| sbezverk_ | sdake: sorry, I mean build_example_yaml.py is making perfect sense | 13:25 |
| sdake | sbezverk_ ok well it may make perfect sense in your head, but if its not documented it doesn't make sense at all | 13:25 |
| sdake | can you help me with wording for that TODO above | 13:25 |
| sbezverk_ | sdake: cloud.yaml I pasted you is from my local kube cluster | 13:25 |
| sbezverk_ | it has been tested inside out | 13:26 |
| sbezverk_ | can you use it as a base? | 13:26 |
| sdake | sbezverk_ right - kfox1111 had indicated instead I should trim that to the bare minimum | 13:26 |
| sdake | he siad to use build_example_yaml.py to do that | 13:26 |
| sdake | trim that = trim cloud.yaml | 13:26 |
| sdake | sbezverk_ i could use your cloud.yaml as a base, although it has a bunch of defaults in it | 13:27 |
| sdake | sbezverk_ kfox1111 's position was there is no point in defining the defaults twice | 13:27 |
| sdake | in the documentation | 13:27 |
| sdake | which sounds reasonable | 13:27 |
| sdake | how to do that - I dont know | 13:27 |
| sdake | I need a trimmed cloud.yaml | 13:27 |
| sbezverk_ | sdake: ok I understand what you are saying, I do not understand a reason, because that cloud.yaml is primary configuration tool for a person who deploy kolla-kube | 13:27 |
| sdake | and a technique to do so | 13:27 |
| sbezverk_ | if you trim default options then that person becomes unaware what can be tweaked | 13:28 |
| sdake | sbezverk_ the reason given in the review was that the operator should be able to see what the minimum stuf fis that is needed to get going | 13:28 |
| sdake | sbezverk_ versus the "here are the 100 options needed" | 13:28 |
| sdake | sbezverk_ i can document both types of workflow, however, i see merit in understadning just the deltas | 13:29 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: ironic-conductor: add missing comma for centos/source build https://review.openstack.org/449135 | 13:29 |
| hrw | bug filled, commmit updated | 13:29 |
| sbezverk_ | sdake: I come from a different world, I need to see all my options and I change only what I need | 13:29 |
| sdake | sbezverk_ right - thats cool - i can document that too | 13:29 |
| sdake | sbezverk_ i personally would like to understand the deltas and I think other operators do too | 13:29 |
| *** Serlex has quit IRC | 13:29 | |
| sdake | sbezverk_ the problem is we have hundreds of options and none of them are documented | 13:30 |
| sdake | sbezverk_ so changing just what you need involves looking deeply into helm charts | 13:30 |
| sdake | sbezverk_ and understanding what that means | 13:30 |
| sdake | sbezverk_ this is a QSG for deploy not a PHD dissertation on deployment :) | 13:30 |
| sbezverk_ | sdake: what do you need from me then? | 13:30 |
| sdake | we can get to documenting all the options later | 13:30 |
| sdake | sbezverk_ a trimmed down delta of just the minimal otpions needing configuration in cloud.yaml | 13:31 |
| sdake | sbezverk_ as in a minimal cloud.yaml | 13:31 |
| sdake | sbezverk_ or a technique to generate one | 13:31 |
| sdake | or and a techniue to generate one | 13:32 |
| *** mbruzek has joined #openstack-kolla | 13:32 | |
| *** mbruzek has joined #openstack-kolla | 13:32 | |
| sdake | i mean I could put yours in the repo as the cononical example, and then diff against whatever is generated in cloud yaml creator or whatever its called | 13:32 |
| sdake | and then document the diff | 13:32 |
| sdake | woudl that be the correct approach? | 13:32 |
| sdake | I dont need you to do the work, I need to know how o do it :) | 13:33 |
| sbezverk_ | sdake: since I have never thought about it, so I do not know how.. the thing is, that cloud.yaml is very highly customized. even things like image tags, which would not be the case if we did not default to 2.0.2 images etc | 13:34 |
| sdake | ok - well action to update the defaults from 2.0.2 to somethign more recent | 13:35 |
| rwellum | we need to start with a solid and clear description of each important option and how they inter-relate imo. Some of them are not obvious to me.. 'external_vip' for example. | 13:35 |
| sbezverk_ | sdake: I am afraid when you run diff you will get 80%+ differences | 13:35 |
| sdake | there is zero reason the defaults should be 3 cycles bck | 13:35 |
| sdake | rwellum ++++++ | 13:35 |
| sdake | sbezverk_ rwellum hit it on the head -some options don't matter | 13:36 |
| sdake | sbezverk_ some do | 13:36 |
| sdake | sbezverk_ i'd liek to focus on the ones that do | 13:36 |
| sbezverk_ | rwellum: some names used were due to the baggage we got from kolla-ansible | 13:36 |
| sdake | sbezverk_ thats fine - nobody cares about why names were used | 13:37 |
| sdake | we care about what the names mean | 13:37 |
| sbezverk_ | rwellum: I would call it differently, but it is how it/was in kolla ansible | 13:37 |
| sdake | names being variables in this case | 13:37 |
| sdake | for example in kolla-ansible docs here: | 13:38 |
| sdake | https://docs.openstack.org/developer/kolla-ansible/quickstart.html | 13:38 |
| rwellum | sbezverk_: ack | 13:38 |
| sdake | http://paste.openstack.org/show/603926/ | 13:39 |
| sdake | need that same sort of thing for kolla-kubernetes | 13:39 |
| sdake | what options must one configure to get from defaults to a working system | 13:39 |
| sdake | fwiw the defaults should yield a working system :) | 13:40 |
| sdake | they dont today - which is fine - so lets document how what gets a working system | 13:40 |
| sbezverk_ | sdake: absolutely NOT | 13:40 |
| *** jtriley has joined #openstack-kolla | 13:40 | |
| sdake | sbezverk_ i said 5 different things there which one do you object to | 13:40 |
| sbezverk_ | sdake: common, how default can get you a working system, if you need your personal IP to attach to>?!?!?!? | 13:40 |
| sdake | sbezverk_ right -t here is always some minimal level of configuration needed | 13:41 |
| sdake | what is that minimal level of configuraiton | 13:41 |
| sdake | see past eabove | 13:41 |
| sdake | this is an exmaple from the kolla-ansible quickstart | 13:41 |
| *** skramaja has quit IRC | 13:41 | |
| sdake | in kolla-ansible ther eare like 5 config value sthat need to be configured to obtain a functional system | 13:41 |
| sbezverk_ | sdake: setting you mentioned are not "deployment" setting, at best for dev env. | 13:42 |
| sdake | sbezverk_ listen dude Im super proud of th ework the community has done - I merely want to document it so others can enjoy in my proudness :) | 13:42 |
| sdake | sbezverk_ ok - well then best for dev env - gotta start somewhere :) | 13:43 |
| sdake | sbezverk_ atm we are blocked on debating whether we should even document cloud.yaml or not | 13:43 |
| sdake | right now cloud.yaml is a black box | 13:43 |
| sbezverk_ | sdake: since you plan to document it, I do not see what stopping you just do it :) I am not twisting your arms or something | 13:43 |
| *** krtaylor has joined #openstack-kolla | 13:44 | |
| sdake | sbezverk_ the faster i learn how cloud.yaml works and the faster rwellum learn's how cloud yaml works and he faster zhubingbing_ learns how cloud.yaml works, the faster we learn together | 13:44 |
| sdake | sbezverk_ I dont want to document the whole thing | 13:44 |
| sdake | sbezverk_ just the basics | 13:44 |
| sdake | sbezverk_ I don't know what the basics are | 13:44 |
| zhubingbing_ | ok | 13:44 |
| sdake | sbezverk_ hence I need some help here | 13:44 |
| sdake | sbezverk_ you mentioned the ip to attach to | 13:44 |
| sdake | that is a basic | 13:44 |
| sdake | so how about something like this: | 13:45 |
| sdake | step 1. generate cloud yaml | 13:45 |
| *** Margin has joined #openstack-kolla | 13:45 | |
| sdake | step 2. configure IP (how) | 13:45 |
| sdake | step 3. configure keystone for external access (how) | 13:45 |
| *** manheim has quit IRC | 13:45 | |
| sdake | step 4. set image tags since the defaults are borked (how) | 13:45 |
| sdake | i need to know how to do these things | 13:45 |
| sbezverk_ | sdake: and who is going to take care of how part? | 13:45 |
| sdake | if you explain it to me, i will writ ethe docs | 13:45 |
| zhubingbing_ | yeah | 13:46 |
| sdake | sbezverk_ i was hoping you would tell me | 13:46 |
| sdake | sbezverk_ or someone would tellm e that knows | 13:46 |
| sdake | sbezverk_ this is the #1 blocker in my mind for producinga deployment guide | 13:46 |
| sdake | or whatever doc we are producing here :) | 13:46 |
| *** manheim has joined #openstack-kolla | 13:46 | |
| sbezverk_ | sdake: I am sorry but I do not have cycles to do it. You know it is not my main job and I have other things on going.. I can | 13:46 |
| sbezverk_ | helm answering "some" questions | 13:47 |
| sbezverk_ | s/helm/help/ | 13:47 |
| rwellum | As a minimum, 'external_vip': your external ip address, 'ext_interface_name', 'ext_bridge_name' and 'tunnel_interface' - is what I had to change | 13:47 |
| rwellum | FYI | 13:47 |
| sdake | rwellum thanks - this is the kind of info I am looking for sbezverk_ | 13:47 |
| sdake | rwellum did you manage to bring up kolla-kubernetes? | 13:47 |
| sdake | so lets start with these 4 variables | 13:48 |
| rwellum | sdake: I got stuck at this step ;) | 13:48 |
| sdake | since rwellum had to chang eem | 13:48 |
| sdake | sbezverk_ ^^ everyone gets stuck at this step including me | 13:48 |
| sdake | we know we need to change these 4 variables - what should they be changed to? | 13:48 |
| rwellum | There was a mistake in the file that someone pointed out in the review. But like sbezverk_ I got pulled off into other things - I'm ready to try again tho. | 13:48 |
| sdake | rwellum good to hear :) | 13:49 |
| sdake | rwellum I was travelling for 2 weeks so feel your pain :( | 13:49 |
| rwellum | Also inc0 was trying the last I saw I was sitting back and watching the experts figure it out :) | 13:49 |
| sdake | ok - well waiting for someone else to solve ap roblem doesn't solve a problem ;) | 13:49 |
| rwellum | truism | 13:50 |
| sdake | we need to solve it together | 13:50 |
| sbezverk_ | external_vip: ip address assigned to your host and kubernetes api server listen on | 13:50 |
| sdake | sbezverk_ ++ :) | 13:50 |
| sbezverk_ | sdake: funny thing, I do not know how many times I have already mentioned it, something tells me it will not be the last one | 13:50 |
| *** fguillot has joined #openstack-kolla | 13:51 | |
| sdake | sbezverk_ it is being written down | 13:51 |
| sdake | sbezverk_ and published on docs.oo | 13:51 |
| sbezverk_ | sdake: you wrote it down numerous time, and here we are again | 13:51 |
| sdake | i am writing your comments in this review as you type them | 13:51 |
| sdake | https://review.openstack.org/#/c/447356/6/doc/source/deployment-guide.rst | 13:51 |
| sdake | that guide will be published on docs.oo | 13:52 |
| rwellum | sbezverk_: you've told me numerous times too - sorry - some I wrote down and some I didn't mainly due to lack of knowledge | 13:52 |
| sdake | so this question will be asked once and answered | 13:52 |
| sbezverk_ | ext_interface_name: name of interface or bridge which will be used by neutron's provider interface | 13:53 |
| hrw | bye | 13:53 |
| sdake | later hrw | 13:53 |
| sbezverk_ | ext_bridge_name: name of bridge you want neutron to use as external bridge, by default it shold be br-ex, but you can change it by using this option | 13:54 |
| sbezverk_ | tunnel_interface: interface name used for connectivity between nodes in kubernetes cluster, in most of cases it matches the name you kubernetes host management interface | 13:55 |
| sbezverk_ | well, here is the problem, if you do not look in the code for these things, then how you plan to learn the code and then develop?? | 13:56 |
| sbezverk_ | hopefully it will unblock you ;) | 13:57 |
| sdake | note comments in https://review.openstack.org/#/c/447356/6 on line 24 | 13:57 |
| sdake | sbezverk_ there are hundreds of variables - looking at code for all of them is not scalable from a developer POV - from a deployer/operator POV, they often don't care about the code or hae time to decode it | 13:58 |
| sdake | sbezverk_ can you think of other variables we need to set? | 13:58 |
| sdake | sbezverk_ we identified 4 that needed to be changed always | 13:58 |
| sbezverk_ | sdake: storage options | 13:59 |
| *** schwicht has quit IRC | 13:59 | |
| sdake | sbezverk_ anything else? | 14:00 |
| *** lucker has joined #openstack-kolla | 14:01 | |
| sbezverk_ | sdake: I think the rest is not extremely essential.. | 14:02 |
| sdake | good to hear :) | 14:02 |
| sdake | rwellum what tz are you in? | 14:03 |
| rwellum | I'm in RTP sdake | 14:04 |
| rwellum | EST | 14:04 |
| sdake | rwellum roger est then | 14:04 |
| *** Serlex has joined #openstack-kolla | 14:04 | |
| sdake | rwellum i am MST | 14:04 |
| rwellum | cool | 14:05 |
| rwellum | sbezverk_: this is really useful - thanks | 14:05 |
| *** rabel_b1 has joined #openstack-kolla | 14:05 | |
| sdake | rwellum up for some collaboration on producing a deploy today via webex? | 14:05 |
| sdake | rwellum i'm struggling to write this document - and could use a second set of eyeballs | 14:06 |
| sdake | zhubingbing_ you will probablyb easleep by the time I have open slots on my calendar | 14:06 |
| sdake | zhubingbing_ if you have any heavy wizardry experience to add, can you add it to the review? | 14:07 |
| *** rabel has quit IRC | 14:08 | |
| *** lucasxu has joined #openstack-kolla | 14:08 | |
| *** Margin has quit IRC | 14:09 | |
| sdake | sbezverk_ re the storage config options | 14:09 |
| sdake | sbezverk_ mind sketching those out for us? | 14:09 |
| *** caoyuan has joined #openstack-kolla | 14:10 | |
| zhubingbing_ | yeah | 14:13 |
| zhubingbing_ | please add me to the revierwer | 14:13 |
| caoyuan | hi | 14:14 |
| sbezverk_ | sdake: sorry very low on available time, since I am on that training program. you ask specific questions, I will try to answer | 14:15 |
| *** yangyape_ has joined #openstack-kolla | 14:15 | |
| yangyape_ | caoyuan: ping | 14:15 |
| sdake | sbezverk_ oh sitll in training? | 14:15 |
| sdake | sbezverk_ sounds good re answering specific Q&A | 14:15 |
| yangyape_ | caoyuan: hi | 14:16 |
| sbezverk_ | sdake: until first week of june, when we need to preent our project | 14:16 |
| sdake | summit is 5 weeks out | 14:19 |
| spsurya | sdake: yep | 14:19 |
| spsurya | sdake: many of the contributors must be facing US VISA approval issue :( | 14:21 |
| sdake | spsurya possible | 14:22 |
| sdake | spsurya thankfully next summit is in canada which is more lenient | 14:22 |
| *** goldyfruit has joined #openstack-kolla | 14:23 | |
| Serlex | not Sydney? | 14:23 |
| sdake | sorry sydey then canda | 14:23 |
| sdake | (vancouver) | 14:23 |
| Serlex | that would be nice | 14:23 |
| sdake | not sure what sydney's visa process is like - US typically has visa-free entry to most countries | 14:23 |
| sdake | the only place I have ever applied for a visa before is china | 14:23 |
| spsurya | sdake: yes | 14:24 |
| spsurya | its Sydney | 14:24 |
| sdake | ya summit after is vancouver | 14:24 |
| sdake | foundation indicated via superuser they were having all summits in 2018 outside the US | 14:25 |
| *** fguillot has quit IRC | 14:25 | |
| sdake | superuser article thatis | 14:25 |
| Serlex | bummer | 14:26 |
| sdake | Serlex bummre why? | 14:27 |
| *** shashank_t_ has joined #openstack-kolla | 14:28 | |
| spsurya | sdake: roger | 14:28 |
| Serlex | I only say that because how popular Openstack is in US | 14:30 |
| rwellum | sdake: I am fine to webex together whenever you need, except 2-4pm EST. | 14:32 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla master: Updated from global requirements https://review.openstack.org/446771 | 14:33 |
| sdake | rwellum i have to get the kids to school from now for about 1 hr, and then have two hours of meetings (which ends at 2pm EST if I did the TZ math correctly) | 14:33 |
| sdake | if 4pm EST onwards ok for you? | 14:34 |
| rwellum | sdake yup | 14:34 |
| sdake | rwellum i'm farily convinced if we can get some reasonable docs written on how this stuff works - we will be able to make forward progress as a community | 14:35 |
| sdake | rwellum cool - ill ping you as soon as my torture sessions end for the day :) | 14:35 |
| sdake | rwellum actually since yours are after mine, perhaps you should ping me | 14:35 |
| *** shashank_t_ has quit IRC | 14:36 | |
| rwellum | sdake: agree and sounds good I will. | 14:36 |
| *** shashank_t_ has joined #openstack-kolla | 14:36 | |
| rwellum | Your environment or mine? | 14:37 |
| sdake | rwellum both ? :) | 14:37 |
| sdake | rwellum whatever works best for you - we can use yours | 14:37 |
| sdake | rwellum two of my machines in my home lab seem to have gone out to lunch | 14:37 |
| *** dciabrin has quit IRC | 14:38 | |
| sdake | so that may add more complication to getting us going as I only have one bare metal node to work with | 14:38 |
| *** dciabrin has joined #openstack-kolla | 14:38 | |
| rwellum | Yes same in that I have to use VM's for the most part. | 14:38 |
| dmsimard | Serlex: political instability/uncertainty is my guess | 14:39 |
| *** yangyape_ has quit IRC | 14:45 | |
| *** yangyapeng has joined #openstack-kolla | 14:47 | |
| *** fooliouno has joined #openstack-kolla | 14:48 | |
| *** shashank_t_ has quit IRC | 14:49 | |
| *** athomas has joined #openstack-kolla | 14:49 | |
| *** schwicht has joined #openstack-kolla | 14:51 | |
| *** yangyapeng has quit IRC | 14:52 | |
| *** manheim has quit IRC | 14:54 | |
| mnaser | http://grafana.openstack.org/dashboard/db/kolla-failure-rate?from=1490194436373&to=1490280836373 | 14:55 |
| mnaser | weee lets try to get that as low as possible | 14:55 |
| mnaser | i still wonder why rhel-based is around 10% higher on failures usually | 14:55 |
| *** caoyuan has quit IRC | 14:58 | |
| spsurya | night all | 14:59 |
| sdake | mnaser how did you set that up? | 14:59 |
| mnaser | sdake https://review.openstack.org/#/c/446697/ | 15:01 |
| mnaser | anyone can add grafana graphs | 15:01 |
| *** adrian_otto has joined #openstack-kolla | 15:03 | |
| *** zhubingbing_ has quit IRC | 15:04 | |
| *** sayantani01 has joined #openstack-kolla | 15:06 | |
| *** zhubingbing_ has joined #openstack-kolla | 15:13 | |
| *** matrohon has quit IRC | 15:14 | |
| *** matrohon has joined #openstack-kolla | 15:15 | |
| *** new2Kolla has joined #openstack-kolla | 15:15 | |
| openstackgerrit | Merged openstack/kolla master: Speed up builds by increasing threads for builds https://review.openstack.org/446701 | 15:16 |
| sdake | hey fooliouno | 15:17 |
| kfox1111 | morning. | 15:17 |
| rwellum | Source build failure rates at 70%? Was that just a blip - seems crazy high? Be cool if the graph had a weekly, monthly view.... mnaser ? | 15:17 |
| sdake | fooliouno we had some discussion this morning around the docs | 15:17 |
| mnaser | rwellum you can click in the top right and select timelines | 15:17 |
| sdake | rwellum click "zoom out" | 15:17 |
| mnaser | past N hours ,etc | 15:17 |
| sdake | rwellum ya our gate was wedged | 15:18 |
| mnaser | and yeah, gate was blocked till this was done - https://review.openstack.org/#/c/448666/ | 15:18 |
| sdake | ironic upstream was broken | 15:18 |
| sdake | and possibly something else :) | 15:18 |
| mnaser | im mostly happy about the log-file-per-image | 15:18 |
| mnaser | finding image build problems is so much easier now | 15:18 |
| sdake | rather biforst upstream | 15:18 |
| *** jtriley has quit IRC | 15:18 | |
| kfox1111 | ah. nice to see openstack adopting grafana | 15:18 |
| mnaser | kfox1111 you can add kube graphs in there too! | 15:19 |
| kfox1111 | nice. :) | 15:19 |
| mnaser | i just dont know much about how the ci works there to muck around it but yeah it's a useful thing to have | 15:19 |
| mnaser | https://git.openstack.org/cgit/openstack-infra/project-config/tree/grafana/kolla.yaml | 15:19 |
| *** shashank_t_ has joined #openstack-kolla | 15:19 | |
| mnaser | hopefully once its at reasonable #s we can make kolla-ansible deploy gates vote | 15:19 |
| kfox1111 | that link mentions grafyaml and grafana in project-config. | 15:20 |
| mnaser | yeah, you can just propose a change review just like any other openstack change | 15:21 |
| rwellum | Ah that's what zoom-out means. :) | 15:21 |
| mnaser | also guys cool idea | 15:21 |
| kfox1111 | looks pretty much like a yaml dump of the grafana dashboard. | 15:21 |
| mnaser | https://github.com/openstack-dev/devstack/tree/master/files/rpms | 15:21 |
| mnaser | what if we look into adopting that | 15:21 |
| mnaser | https://github.com/openstack-dev/devstack/tree/master/files/debs | 15:21 |
| mnaser | that way we dont have to sit and maintain which packages we need | 15:21 |
| mnaser | its all in one centralized spot.. | 15:21 |
| sdake | mnaser i dont follow | 15:22 |
| mnaser | for example, the other day source deploys were broken because haproxy become a requirement | 15:22 |
| openstackgerrit | Merged openstack/kolla master: Remove py34 settings from gating https://review.openstack.org/447247 | 15:22 |
| mnaser | we didnt have it installed so all deploys failed. | 15:22 |
| mnaser | say if we relied on something like that, it would have been added here, and we'd have picked up the change - https://github.com/openstack-dev/devstack/blob/master/files/debs/neutron#L5 | 15:22 |
| mnaser | and the effort of maintaining which project needs which packages is simplified and unified across devstack and kolla | 15:23 |
| sdake | are you proposing to reimplement EPEL/RDO/UCA? :) | 15:23 |
| mnaser | source builds dont install dependencies | 15:23 |
| mnaser | we install all these dependencies manually | 15:23 |
| sdake | yes, and they come from the RDO and UCA repos | 15:23 |
| mnaser | what im saying is | 15:24 |
| sbezverk_ | kfox1111: good morning, have you seen my reply about nova cells? | 15:24 |
| mnaser | instead of us maintaining a list of packages and devstack maintaining (effectively the same list) | 15:24 |
| kfox1111 | sbezverk_: no. in the scroll back? | 15:24 |
| mnaser | https://github.com/openstack-dev/devstack/blob/master/files/debs/neutron vs https://github.com/openstack/kolla/blob/1d9b41f3d8ef05ea1b9516a6e853eb7f84c67e18/docker/neutron/neutron-base/Dockerfile.j2#L66-L79 | 15:24 |
| *** manheim has joined #openstack-kolla | 15:25 | |
| sdake | instead of maintaing separately maintain where? | 15:25 |
| kfox1111 | sbezverk_: a huge amount of scroll back this morning. could you summerize? | 15:25 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla master: Updated from global requirements https://review.openstack.org/446771 | 15:26 |
| mnaser | sdake either use what devstack uses, or propose to share it somehow. | 15:26 |
| EmilienM | can someone approve https://review.openstack.org/#/c/448043/ ? it has 3 °2 | 15:26 |
| EmilienM | +2 | 15:26 |
| sdake | EmilienM we are waiting on the gate to go green which it should | 15:27 |
| *** jtriley has joined #openstack-kolla | 15:28 | |
| sdake | EmilienM usually this can be achieved thorugh rechecks - although the implementation may be off slightly | 15:28 |
| sdake | looking at the problem now | 15:28 |
| mnaser | sdake its all green, except it looks like a non voting deploy job failed | 15:28 |
| mnaser | which should be ok | 15:28 |
| mnaser | considering it seems like mariadb bootstrap failed | 15:28 |
| EmilienM | sdake: it's green now. | 15:28 |
| EmilienM | ? | 15:28 |
| sdake | just because its non-voting doesn't mean we should have a regression in the gate jobs | 15:29 |
| mnaser | but gate doesnt even run deploy jobs | 15:29 |
| sdake | EmilienM this gat ejob fails: http://logs.openstack.org/43/448043/1/check/gate-kolla-dsvm-deploy-oraclelinux-binary-centos-7-nv/a74a093/ | 15:29 |
| EmilienM | not sure it's related to my patch | 15:29 |
| mnaser | thats a check job, and its non voting because its unreliable, and it seems pretty clear that the cause is not that | 15:29 |
| EmilienM | my patch is about RDO repos | 15:29 |
| sdake | EmilienM agreed that is why we recheck | 15:29 |
| sdake | EmilienM oraclelinux uses RDO repos | 15:30 |
| EmilienM | well, I'm switching http to https | 15:30 |
| sbezverk_ | kfox1111: it is in review reply | 15:30 |
| EmilienM | not sure it's related but ok... | 15:30 |
| sbezverk_ | should be in your mail box | 15:30 |
| kfox1111 | sbezverk_: ah ok. sec | 15:30 |
| sdake | EmilienM yup - our gate is a little flakey | 15:30 |
| sdake | EmilienM we sitll dont want to have a regression so we treat the gate as voting even though it doesn't vote | 15:30 |
| *** hrw has quit IRC | 15:30 | |
| sdake | EmilienM we are moving the gating to voting as soon as feasible | 15:31 |
| EmilienM | well, I work on RDO every day and I can tell that using https instead of http will help you to be more stable | 15:31 |
| EmilienM | we saw some cases where redirect doesn't work | 15:31 |
| kfox1111 | sbezverk_: ah. ok. that explains it, and makes me feel much better. :) | 15:31 |
| sdake | EmilienM all good - i am looking at the specific failure now hang tight | 15:31 |
| kfox1111 | I guess simple setup does that. | 15:31 |
| sdake | mnaser changed how the gate logs re ecorded so I need to learn that now :) | 15:31 |
| mnaser | sdake only for image builds | 15:32 |
| kfox1111 | thanks for getting verification. | 15:32 |
| mnaser | this was a deploy | 15:32 |
| mnaser | so everything is still in html | 15:32 |
| mnaser | console.html | 15:32 |
| mnaser | i checked it and mariadb failed to bootstrap | 15:32 |
| mnaser | http://logs.openstack.org/43/448043/1/check/gate-kolla-dsvm-deploy-oraclelinux-binary-centos-7-nv/a74a093/logs/kolla/mariadb/mariadb.txt.gz | 15:33 |
| mnaser | ERROR 2003 (HY000): Can't connect to MySQL server on '10.31.184.114' (111 "Connection refused") | 15:33 |
| kfox1111 | sbezverk_: I think for the iptables stuff, since your using ovs now, all you may need to do is add some iptables INPUT rules for allowing traffic from your tenant bridge. | 15:34 |
| mnaser | actually that change has zero impact on that job | 15:34 |
| mnaser | because the deploy job downloads images instead of building them | 15:35 |
| *** hrw has joined #openstack-kolla | 15:36 | |
| sbezverk_ | kfox1111: it seems it is not required as traffic goes in/out fine.. | 15:36 |
| *** jascott1 has joined #openstack-kolla | 15:37 | |
| sbezverk_ | adding complexity when it is not absolutely requiered is not great approach | 15:37 |
| *** manheim has quit IRC | 15:38 | |
| kfox1111 | sbezverk_: could be. I think I rewrote the iptables rules to mostly only apply to the external interface. | 15:38 |
| *** schwicht has quit IRC | 15:39 | |
| sbezverk_ | kfox1111: it would have been broken if it were really required | 15:40 |
| sdake | mnaser the job is not non-voting because it is unreliable | 15:40 |
| kfox1111 | true | 15:40 |
| mnaser | sdake im aware, but if we just recheck it till it goes green. we're not really solving the issue | 15:40 |
| mnaser | and we're using up resources for no reason | 15:40 |
| kfox1111 | sbezverk_: initramf_url: is the f a typo? | 15:40 |
| mnaser | i'm all for making jobs vote again, but by looking into the issues and resolving them | 15:40 |
| kfox1111 | or a missing 's'? | 15:41 |
| sdake | mnaser wfm - jeffrey4l and inc0 are leading the gating work | 15:41 |
| sdake | (for kolla-ansible) | 15:41 |
| sdake | egonzales is helping as well | 15:41 |
| sdake | i'm not really involved | 15:41 |
| sdake | just following the process we have in place | 15:41 |
| sbezverk_ | kfox1111: yep, itshould be initramfs | 15:41 |
| kfox1111 | sbezverk_: k. I'll mark it. | 15:42 |
| sbezverk_ | kfox1111: but it is missspelled consistently ;) | 15:44 |
| jaosorior | mandre: will you file the LP bug that was requested here https://review.openstack.org/#/c/446911/ ? | 15:45 |
| kfox1111 | sbezverk_: yeah. :) | 15:45 |
| *** new2kolla_ has joined #openstack-kolla | 15:47 | |
| *** new2Kolla has quit IRC | 15:47 | |
| *** shasha_t_ has joined #openstack-kolla | 15:48 | |
| *** shashank_t_ has quit IRC | 15:52 | |
| *** shashank_t_ has joined #openstack-kolla | 15:54 | |
| *** lamt has joined #openstack-kolla | 15:56 | |
| *** david-lyle_ is now known as david-lyle | 15:58 | |
| *** mliima has joined #openstack-kolla | 15:59 | |
| mliima | hello! | 16:00 |
| *** jtriley has quit IRC | 16:00 | |
| *** targon has quit IRC | 16:01 | |
| *** matrohon has quit IRC | 16:06 | |
| *** unicell has joined #openstack-kolla | 16:07 | |
| *** unicell1 has quit IRC | 16:09 | |
| *** jascott1_ has joined #openstack-kolla | 16:09 | |
| *** lucker has quit IRC | 16:09 | |
| *** jtriley has joined #openstack-kolla | 16:11 | |
| mandre | jaosorior: do you mind filing that bug? sorry, I'm a bit short on time | 16:14 |
| sbezverk_ | kfox1111: how do you suggest we proceed with that ps? I change initramfs and then we can merge it? | 16:14 |
| jaosorior | mandre: we all are. No biggie; I'll file it. | 16:15 |
| jaosorior | mandre: thanks again for updating the patch. | 16:15 |
| openstackgerrit | Merged openstack/kolla-ansible master: Update the "virt_driver" to "virt_type" https://review.openstack.org/447336 | 16:15 |
| *** manheim has joined #openstack-kolla | 16:16 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Fix mongodb cluster using wrong network interface https://review.openstack.org/442999 | 16:19 |
| openstackgerrit | Merged openstack/kolla-ansible master: Add an extra network node for HA https://review.openstack.org/448494 | 16:20 |
| inc0 | good morning | 16:24 |
| mandre | jaosorior: I'm thinking we'll need a release note for your SSL patch | 16:25 |
| jaosorior | mandre: no biggie man, I'll add that once I add the LP bug. | 16:25 |
| *** matrohon has joined #openstack-kolla | 16:26 | |
| mandre | jaosorior: perfect, thanks | 16:26 |
| *** jascott1_ has quit IRC | 16:30 | |
| *** lucasxu has quit IRC | 16:32 | |
| hrw | re | 16:32 |
| hrw | feel strange. all patches got +1 from CI... and one which did not failed because of already-present-but-somehow-missed bug | 16:33 |
| *** mbruzek has quit IRC | 16:33 | |
| hrw | hm... that --threads patch which got recently merged will be fun on ppc64le machines | 16:36 |
| hrw | ah. tht's tests only | 16:36 |
| sbezverk_ | inc0: good morning, qq can you change config gen ps to write config files in a different from /etc/kolla location? | 16:39 |
| hrw | can someone give +2 to https://review.openstack.org/449135 so we can get https://review.openstack.org/449086 verified and then +2 and merged? | 16:41 |
| *** rhallisey has quit IRC | 16:44 | |
| *** adrian_otto1 has joined #openstack-kolla | 16:46 | |
| inc0 | sbezverk_: check out my patchset - you have CONFIG_DIR=/etc/kolla in ansible-playbook call | 16:47 |
| inc0 | you can change this to whatever you want | 16:47 |
| egonzalez | hrw, after fixing ironic build you will find another error and you know it ;) | 16:48 |
| sbezverk_ | inc0: it is not exactly what I had in mind :) | 16:48 |
| inc0 | then what do you have in mind?:) | 16:49 |
| sbezverk_ | would be breat to be able to write configs in helm/microservice/{microservice-name}-configmap/data ;) | 16:49 |
| *** adrian_otto has quit IRC | 16:49 | |
| hrw | egonzalez: sure, but ironic-conductor bug keeps yum from entering. then I will have to deal with lot of centos failures but that's excepted anyway. and with yum fix being in I will have errors in proper places == less debugging == more work done | 16:49 |
| inc0 | well, doable, I can make this happen...hell you can use my other patchset to write rendered configmap | 16:50 |
| inc0 | and template that | 16:50 |
| inc0 | but copying isn't big deal...templating is | 16:51 |
| sbezverk_ | inc0: I do not give up the idea to use helm | 16:51 |
| inc0 | neither do I | 16:51 |
| *** rhallisey has joined #openstack-kolla | 16:51 | |
| inc0 | what I'm saying is, when we do it, we should do it correctly | 16:51 |
| sbezverk_ | inc0: placing config files into that folder would allows us to do it very very easily | 16:51 |
| inc0 | well it's really simple playbook | 16:52 |
| inc0 | like this one https://review.openstack.org/#/c/447733/ | 16:52 |
| inc0 | I can change it to output either fully rendered configmap or just internal data | 16:53 |
| inc0 | but I don't really like this approach tbh | 16:54 |
| inc0 | if we're getting rid of ansible, let's get rid of it alltogethe | 16:54 |
| inc0 | r | 16:54 |
| inc0 | and that means moving templating of configmaps all the way to helm | 16:54 |
| sbezverk_ | inc0: atm there is a limitation in helm :( it can access files only located inside of chart folder | 16:55 |
| inc0 | sbezverk_: what I'm saying is, I'm -1 to adding files dynamically to tree | 16:56 |
| inc0 | either we render them all in helm | 16:57 |
| inc0 | (including templates) | 16:57 |
| inc0 | or we add them to k8s externally | 16:57 |
| inc0 | reason I don't want to change things in tree is git | 16:57 |
| inc0 | and stuff like that | 16:57 |
| inc0 | code tree should not be touched | 16:57 |
| *** egonzalez has quit IRC | 16:58 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-kubernetes master: WIP Move config generation to kolla-k8s https://review.openstack.org/448258 | 17:00 |
| *** new2kolla_ has quit IRC | 17:01 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-kubernetes master: WIP: Playbook to add configmaps and secrets https://review.openstack.org/447733 | 17:03 |
| inc0 | sbezverk_: I addressed you comment | 17:03 |
| *** sean-k-mooney2 has joined #openstack-kolla | 17:05 | |
| *** sean-k-mooney2 has quit IRC | 17:05 | |
| *** schwicht has joined #openstack-kolla | 17:06 | |
| *** unicell has quit IRC | 17:06 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/kolla master: Prevent apache from listening on port 80 https://review.openstack.org/449039 | 17:08 |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/kolla master: RHEL: Add mod_ssl for services running over httpd https://review.openstack.org/446911 | 17:08 |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/kolla master: Prevent apache from listening on port 80 for heat-base https://review.openstack.org/449040 | 17:08 |
| jaosorior | mandre: done ^^ | 17:09 |
| jaosorior | mandre: added you as a co-author of the patch that adds mod_ssl | 17:09 |
| *** pbandark has joined #openstack-kolla | 17:14 | |
| *** pcaruana has quit IRC | 17:14 | |
| openstackgerrit | Merged openstack/kolla master: Implement Keystone zero-downtime upgrade https://review.openstack.org/425833 | 17:17 |
| sdake | inc0 the key is making defaults via helm config maps and making non-defaults via helm config maps as well | 17:25 |
| sdake | kfox1111 inc0 sbezverk_ you around to ddiscuss that now? | 17:25 |
| *** manheim has quit IRC | 17:27 | |
| *** jmccarthy has quit IRC | 17:29 | |
| *** jaosorior has quit IRC | 17:30 | |
| *** tonanhngo has joined #openstack-kolla | 17:30 | |
| mnaser | hrw got me worried that my threads would have broke stuff :D | 17:31 |
| inc0 | sdake: I'd like to have single way to do it | 17:32 |
| *** lucasxu has joined #openstack-kolla | 17:32 | |
| inc0 | if we do ansible, we do ansible | 17:32 |
| inc0 | if we do helm, we go for it | 17:32 |
| sdake | you said you woudln't be married to ansible... | 17:32 |
| inc0 | I'm not married to ansible, I'll do it as quick fix | 17:32 |
| inc0 | but what I don't want to do is some middle ground prothesis just to say "we do helm" even if we don't really | 17:33 |
| sdake | right - so ansible eventuallyg et s ejected | 17:33 |
| sdake | if it can't be done it can't be done | 17:33 |
| inc0 | we need to move it to helm, that's true, but when we do it we do it properly | 17:33 |
| hrw | mnaser: that was before I noticed that it tests. | 17:33 |
| sdake | i think it can be done | 17:33 |
| hrw | mnaser: I was running kolla on 80 cpu system | 17:34 |
| kfox1111 | sdake: yeah. | 17:34 |
| inc0 | https://review.openstack.org/#/c/448258/ <- this took 2 days of work | 17:34 |
| inc0 | that's why I'd rather go with ansible | 17:35 |
| inc0 | now | 17:35 |
| inc0 | we *might* go with ansible for 1.0, depends how we do | 17:35 |
| inc0 | but moving it all to helm is priority | 17:35 |
| hrw | bye | 17:35 |
| inc0 | good night hrw | 17:36 |
| openstackgerrit | Merged openstack/kolla master: Updated from global requirements https://review.openstack.org/446771 | 17:36 |
| kfox1111 | inc0: +1 for just copying from kolla-ansible to kolla-kubernetes for now, just to decouple us, prevent further breakages, and simplify the deployment. thats a lot of benifit for very little initial work. | 17:37 |
| inc0 | exactly | 17:37 |
| *** jtriley has quit IRC | 17:39 | |
| *** krtaylor has quit IRC | 17:42 | |
| *** jtriley has joined #openstack-kolla | 17:44 | |
| *** Serlex has quit IRC | 17:45 | |
| openstackgerrit | Merged openstack/kolla master: Switch trunk/cbs/buildlogs to use https https://review.openstack.org/448043 | 17:45 |
| mnaser | good and simple +W for any cores https://review.openstack.org/#/c/449135/ | 17:47 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Virtualbmc and deployment of baremetal VM using ironic https://review.openstack.org/441550 | 17:48 |
| *** deadnull has joined #openstack-kolla | 17:51 | |
| sbezverk_ | kfox1111: answered you latest comments | 17:51 |
| *** adrian_otto1 has quit IRC | 17:52 | |
| sdake | inc0 i think we all agree ejecting kolla-ansible as a dep is mandatory - and the approach you took to do so is solid | 17:52 |
| sdake | inc0 i think we may disagree that ansible should be a dependency at all - i'm not sure | 17:52 |
| sdake | I don't think it is mandatory | 17:52 |
| sdake | you indicated for 1.0.0 it is mandatory | 17:53 |
| *** rhallisey has quit IRC | 17:53 | |
| sdake | i think this is open for debate if someone comes along and implements external config maps - because this removes most of the need for any type of ansible in the codebase | 17:53 |
| sdake | inc0 what we really want to avoid as a team is a workflow using ansible | 17:53 |
| sdake | inc0 if we were to do workflow with ansible we would be permanently locked into ansible as a dependency | 17:54 |
| inc0 | sdake: workflow using ansible is still better than alternative we have now | 17:54 |
| *** rhallisey has joined #openstack-kolla | 17:54 | |
| sdake | inc0 workflows shipping with 1.0.0 = cli = permanent thing | 17:54 |
| sdake | that needs to be maintained and dicctates the permanent architecture of the system | 17:54 |
| kfox1111 | sbezverk_: looks good to me. :) | 17:54 |
| sdake | that is why I asked you if you would be married to ansible | 17:55 |
| sdake | you said no | 17:55 |
| sdake | and i hope you stick to that ;) | 17:55 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-kubernetes master: Move keystone configuration to kolla-k8s https://review.openstack.org/449254 | 17:55 |
| *** mbruzek has joined #openstack-kolla | 17:56 | |
| *** mbruzek has quit IRC | 17:56 | |
| *** mbruzek has joined #openstack-kolla | 17:56 | |
| inc0 | sdake: I hope you already saw that I'm just trying to do what makes sense | 17:56 |
| inc0 | if something else is better, I'll drop my approach happily | 17:56 |
| *** adrian_otto has joined #openstack-kolla | 17:56 | |
| sdake | inc0 right - pragmatism is good, dictating a design of workflow in ansible is not something we want to do - we already had broad consensus on this point | 17:57 |
| kfox1111 | inc0: I'd be happy to get the kolla-ansible templates -> kolla-kubernetes thign done asap. that should be pretty quick to do. | 17:57 |
| kfox1111 | the other one, replacing kollakube is going to be a lot of work. | 17:57 |
| inc0 | kfox1111: I just submitted another service;) 30min of work | 17:57 |
| kfox1111 | realitive to the first one. | 17:57 |
| kfox1111 | inc0: cool. :) | 17:57 |
| inc0 | I don't believe it's going to be *that* much of work | 17:58 |
| inc0 | I mean kollakube does 2 things really -create configmaps and create secrets | 17:58 |
| kfox1111 | inc and theres some services that still need transition. | 17:58 |
| *** jtriley has quit IRC | 17:58 | |
| inc0 | like what? | 17:58 |
| sdake | I dont care about replacing kollakube i care about not being locked into ansible permanently | 17:58 |
| kfox1111 | keepalived, half of ceph, and openvswitch-set-ip | 17:59 |
| kfox1111 | sdake: +1. | 17:59 |
| kfox1111 | kollakube isn't too bad. depending on genconfig's pretty painful. | 17:59 |
| inc0 | approach I'm taking is - as long as files lands in /etc/kolla | 17:59 |
| inc0 | I don't care who creates there | 17:59 |
| kfox1111 | inc0: +1. | 17:59 |
| inc0 | them* | 17:59 |
| inc0 | so no, we won't be locked to ansible ever | 17:59 |
| sdake | right - you have sent messages on irc that indicate you want to do workflow in ansible because it is pragmatic | 18:00 |
| inc0 | because it is pragmatic | 18:00 |
| sdake | if we ship 1.0.0 with a dep on ansible we are locked into ansible forever | 18:00 |
| inc0 | but workflow in ansible != being locked to it | 18:00 |
| sdake | (if the workflow is designed in ansible) | 18:00 |
| inc0 | no... | 18:00 |
| *** zhubingbing_ has quit IRC | 18:00 | |
| kfox1111 | sdake: inc0 this is where we get into workflow=religious, and I wanted microservices so that people can pick workflow that they want. | 18:00 |
| inc0 | I just create files | 18:00 |
| kfox1111 | if someone wants ansible workflow, great. heat workflow, great. pure helm, great. | 18:01 |
| inc0 | we can provide one | 18:01 |
| inc0 | on our own | 18:01 |
| sdake | i agree we should provide one | 18:01 |
| inc0 | as long as it's optional and pragmatic | 18:01 |
| sdake | and i disagree it should be ansible :) | 18:01 |
| kfox1111 | thats the issue. are we going to all be able to agree on just one? | 18:01 |
| inc0 | kfox1111: I'd stick to one in repo | 18:02 |
| inc0 | (just one) | 18:02 |
| inc0 | supplying multiple will make it night impossible to work with | 18:02 |
| kfox1111 | inc0: or 0. | 18:02 |
| *** MasterOfBugs has joined #openstack-kolla | 18:02 | |
| sdake | our architecture spec indicates a helm + operators apprroach | 18:02 |
| kfox1111 | inc0: yeah, but agreeing to one might be impossible. | 18:02 |
| inc0 | no, we need quick start | 18:02 |
| inc0 | kfox1111: we agreed on helm right?;) | 18:02 |
| kfox1111 | thus requiring, a kolla like repo split. | 18:02 |
| kfox1111 | kolla-kubernetes-helm, kolla-kubernetes-ansible, etc. | 18:03 |
| inc0 | negative, requiring someone to take the job of translating ansible to helm | 18:03 |
| sdake | kfox1111 sounds good to me :) | 18:03 |
| inc0 | then we drop ansible | 18:03 |
| kfox1111 | and kolla-kubernetes is all the shared stuff. | 18:03 |
| inc0 | it's as simple as that | 18:03 |
| sdake | ansible is fire and forget, hence it doesn't work well for managing openstack clouds | 18:03 |
| inc0 | if we have full helm, I don't see value on keeping ansible | 18:04 |
| sdake | it works well for deployment upgrade and other forms of management, I'd like to take that a step further | 18:04 |
| inc0 | I'm working with ansible now because it takes 30min to create new role | 18:04 |
| inc0 | and port it from kolla-ansible | 18:04 |
| inc0 | and I'm lazy person | 18:04 |
| inc0 | so I'm all for quick wins;) | 18:04 |
| kfox1111 | ok, so the agreed thing to do for now, is | 18:04 |
| *** eaguilar has joined #openstack-kolla | 18:05 | |
| kfox1111 | copy templates from kolla-ansible -> kolla-kubernetes to reduce deps, and allow us to set new, saner for kolla-kubernetes defaults. | 18:05 |
| *** jtriley has joined #openstack-kolla | 18:05 | |
| kfox1111 | step 2, convert config to helm. | 18:05 |
| kfox1111 | ? | 18:05 |
| inc0 | and simpler workflow to start with kolla-k8s | 18:05 |
| inc0 | yes | 18:05 |
| *** bjolo has quit IRC | 18:05 | |
| kfox1111 | what workflow? | 18:05 |
| inc0 | only question is, who and when will work on step 2 | 18:05 |
| inc0 | well, quickstart | 18:05 |
| inc0 | git clone kolla-kubernetes, run this command, run that command, you have openstack | 18:06 |
| rwellum | workflow = deployment or development inc0 ? (or both). | 18:06 |
| inc0 | rwellum: both, development as testing our stuff | 18:07 |
| kfox1111 | so, there's kinf of 4 workflow tools either being discussed seriously, or already exists. | 18:07 |
| kfox1111 | 1. shell scripts | 18:07 |
| kfox1111 | 2. helm+kubernetes-entrypoint | 18:07 |
| kfox1111 | 3. kubernetes-operators | 18:07 |
| kfox1111 | 4. ansible | 18:07 |
| kfox1111 | 1 & 2 already exist. | 18:08 |
| inc0 | yeah but shell is not great tool for the job | 18:08 |
| kfox1111 | 3 is being prototyped. | 18:08 |
| kfox1111 | do we really want 4? | 18:08 |
| inc0 | and 2 is not complete - lacks config | 18:08 |
| *** matrohon has quit IRC | 18:08 | |
| kfox1111 | I agree with shell not being good for end users. | 18:08 |
| kfox1111 | that 2 or 3 is mroe for that. | 18:08 |
| inc0 | so prototype - we'll see how it goes | 18:09 |
| kfox1111 | well, we're talkign about dealing with config. does that need to influence everythign else though? | 18:09 |
| inc0 | entrypoint + helm native configmaps is best case scenerio I see now | 18:09 |
| kfox1111 | once we get helm packages for configmap, 2 & 3 can work natively with configmaps. | 18:09 |
| sbezverk_ | kfox1111: why would we need 4?? | 18:09 |
| kfox1111 | sbezverk_: I'm arguing we dont. | 18:09 |
| kfox1111 | sbezverk_: inc0 is talking about having it, at least termporarily? | 18:09 |
| inc0 | ok, nobody wants that really | 18:09 |
| sbezverk_ | kfox1111: +1 we do not | 18:09 |
| inc0 | no | 18:09 |
| inc0 | ansible roles end when configmaps are created imho | 18:10 |
| kfox1111 | inc0: ah. ok. | 18:10 |
| inc0 | and this is really really simple job to do for ansible | 18:10 |
| kfox1111 | I'm ok with that then. | 18:10 |
| *** jascott1_ has joined #openstack-kolla | 18:10 | |
| kfox1111 | well, let me ask something first... | 18:10 |
| kfox1111 | ansible takes template, spits out etc/kolla/... | 18:10 |
| kfox1111 | not sure we need etc/kolla/.... -> configmap to k8s | 18:11 |
| kfox1111 | kubectl create configmap can do that. | 18:11 |
| inc0 | well, it can, true | 18:11 |
| kfox1111 | so we talking about tjust the first part, or all the parts of configmap jen? | 18:11 |
| inc0 | but we need shell to iterate over the thing;) | 18:11 |
| inc0 | so we need something anyway | 18:11 |
| kfox1111 | yeah. but shell is trivial to understand in that. | 18:11 |
| kfox1111 | its like a 3 line bash script. | 18:12 |
| sdake | kfox1111 for you shell is trivial for inc0 ansible is trivial | 18:12 |
| inc0 | if that's true, why kollakube even exists? | 18:12 |
| kfox1111 | loop over dirs, kubectl create configmap ..... ? | 18:12 |
| openstackgerrit | Marcus Williams proposed openstack/kolla master: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 18:12 |
| kfox1111 | inc0: because it did literally EVERYTHING before helm. | 18:12 |
| sdake | kfox1111 what i'm getting at is eveyrone has a different POV on triviality | 18:12 |
| kfox1111 | sdake: I garantee the shell script will be at least 10 times smaller then the eqiv ansible. | 18:12 |
| sbezverk_ | sdake: replacing shell by ansible just to act as bash has no value | 18:13 |
| inc0 | kfox1111: alternatively instead of generating config files we can go straight to kube resources | 18:13 |
| sdake | sbezverk_ i didnt propose it no reason yell at me :) | 18:13 |
| kfox1111 | inc0: +1 on that. as thats most of the work in getting things converted to helm. | 18:13 |
| sbezverk_ | sdake: appologies, it was directed to inc0 | 18:13 |
| kfox1111 | though a lot of work. | 18:13 |
| inc0 | yeah | 18:14 |
| inc0 | so let's not dig too deep into philosophical space of how bad is ansible vs shell | 18:14 |
| inc0 | and just write damn thing;) | 18:14 |
| sbezverk_ | inc0: we have been doing it almost 8 month, hav enot you noticed ;) | 18:15 |
| inc0 | zuul is dead, long live the zuul :( | 18:15 |
| *** sayantani01 has quit IRC | 18:15 | |
| jascott1_ | there is only zuul.... | 18:15 |
| *** pbandark has quit IRC | 18:16 | |
| *** adrian_otto has quit IRC | 18:16 | |
| kfox1111 | inc0: yeah, but still, the forloop in the gate, can be replaced with a forloop with kubectl create configmap with much less effort then adding ansible into the mix there. | 18:16 |
| *** sayantani01 has joined #openstack-kolla | 18:16 | |
| *** adrian_otto has joined #openstack-kolla | 18:16 | |
| sbezverk_ | kfox1111: if we can have config files in chart folder, then we can do it using helm charts | 18:17 |
| sbezverk_ | then we have consistent helm implementation for all aspects | 18:17 |
| sdake | one thing we want to avoid is modifying charts to change configuration | 18:18 |
| sdake | charts for config specifically are fine to modify | 18:18 |
| sdake | charts for the implementaiton not ok | 18:18 |
| kfox1111 | sbezverk_: +1 for helm charts eventaully. just saying, if the goal is to get rid of kollakube, we can do that by just switching configmap uploading to kubectl creat econfigmap trivially, | 18:19 |
| sbezverk_ | sdake: I get this point, then we can push for a feature in helm --with-files | 18:19 |
| sdake | config and implementation should be separate | 18:19 |
| *** athomas has quit IRC | 18:19 | |
| kfox1111 | as thats all what kollakube does in that case. | 18:19 |
| jascott1_ | sbezverk_ did someone say new helm feature ? :) | 18:19 |
| sdake | jascott1_ go forth and conquer :) | 18:19 |
| sbezverk_ | jascott1_: yep | 18:20 |
| jascott1_ | its been conquering me lately but I am whooping it as we speak | 18:20 |
| kfox1111 | jascott1_: yeah. really looking forward to the parent chart feature. should really help. :) | 18:20 |
| kfox1111 | hoping they merge it soon. :) | 18:20 |
| sbezverk_ | jascott1_: currently chart rendering happens in tiller which has no visibility to clients file system, hence cannot include files located not in the chart's folder | 18:20 |
| openstackgerrit | Marcus Williams proposed openstack/kolla master: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 18:21 |
| jascott1_ | kfox1111 i had to fix the parent wins thing (which i did) but should have it all ready today.. release will be April fools day so not long to wait :) | 18:22 |
| sbezverk_ | jascott1_: there was a proposal sometime ago to add --with-files key to attach other files | 18:22 |
| jascott1_ | sbezverk_ i will hunt it down and check it | 18:22 |
| sbezverk_ | jascott1_: I think having this feature could simplify a lot config/configmap generation thing | 18:23 |
| sbezverk_ | jascott1_: cool, appreciate it | 18:23 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Simple configmap to raw kubectl https://review.openstack.org/449262 | 18:25 |
| jascott1_ | sbezverk_ this one? https://github.com/kubernetes/helm/issues/1754 | 18:25 |
| kfox1111 | jascott1_: awesome. :) | 18:25 |
| kfox1111 | inc0: I think that review ---^ will do most of the remaining work of getting rid of kollakube. | 18:26 |
| *** sayantani01 has quit IRC | 18:26 | |
| sbezverk_ | jascott1_: yep | 18:26 |
| openstackgerrit | Marcus Williams proposed openstack/kolla master: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 18:26 |
| kfox1111 | no ansible needed for it. | 18:26 |
| jascott1_ | cool, bookmarked | 18:26 |
| *** sayantani01 has joined #openstack-kolla | 18:26 | |
| *** jmccarthy has joined #openstack-kolla | 18:28 | |
| *** yangyapeng has joined #openstack-kolla | 18:29 | |
| *** yangyapeng has quit IRC | 18:34 | |
| *** jmccarthy has quit IRC | 18:34 | |
| *** manheim has joined #openstack-kolla | 18:36 | |
| *** manheim has quit IRC | 18:36 | |
| *** lucasxu has quit IRC | 18:38 | |
| kfox1111 | is zuul broked? | 18:40 |
| inc0 | zuul is dead | 18:40 |
| inc0 | long live the zuul | 18:40 |
| *** yangyapeng has joined #openstack-kolla | 18:42 | |
| *** jascott1_ has quit IRC | 18:42 | |
| sean-k-mooney | quick dumb question. is there an apache contiener? or does kolla install apache on the host | 18:42 |
| kfox1111 | apache is in the containers that use it. | 18:43 |
| *** unicell has joined #openstack-kolla | 18:43 | |
| sean-k-mooney | kfox1111: so horizon and keysone have seperate apache instances? | 18:44 |
| *** jtriley has quit IRC | 18:45 | |
| kfox1111 | sean-k-mooney: yes. | 18:45 |
| kfox1111 | greatly simplifies the config, and allows you to restart one without effeting the other. | 18:46 |
| *** yangyapeng has quit IRC | 18:47 | |
| sean-k-mooney | one of my deployments was scanned by our intenal vulnerbility system and has the sweet32 vlunerblity in the ssl implnetaion of the service running on port 443 | 18:48 |
| sean-k-mooney | im trying to figure out where i need to fix it | 18:48 |
| *** jtriley has joined #openstack-kolla | 18:48 | |
| kfox1111 | not sure how its done in kolla-ansible, but I'm guessing its ssl terminated at an haproxy? | 18:50 |
| openstackgerrit | Marcus Williams proposed openstack/kolla master: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 18:51 |
| sean-k-mooney | this is a sandbox of our ci so i dont know if they installed haproxy but ill check both | 18:52 |
| sean-k-mooney | hum no ha proxy and for some reason they have also instlled apach2 on the host. ssl is disabled for kolla on the sandbox it seams | 18:55 |
| *** egonzalez has joined #openstack-kolla | 18:55 | |
| mnaser | kolla-core: im going over common centos and oraclelinux failure and they seem to be related to a bug in the old docker releases | 18:56 |
| inc0 | how old mnaser ? | 18:56 |
| mnaser | i proposed this a while back and it passed everything... could we get some traction? https://review.openstack.org/#/c/446158/ | 18:56 |
| mnaser | https://github.com/docker/docker/issues/28336 | 18:57 |
| mnaser | http://logs.openstack.org/71/446771/7/check/gate-kolla-dsvm-deploy-oraclelinux-source-centos-7-nv/3ddff5e/logs/kolla/nova/nova-scheduler.txt.gz#_2017-03-23_16_16_36_556 | 18:57 |
| mnaser | http://logs.openstack.org/71/446771/7/check/gate-kolla-dsvm-deploy-oraclelinux-source-centos-7-nv/3ddff5e/logs/docker.txt.gz#_Mar_23_16_16_36 | 18:57 |
| mnaser | look at the time corrolation there | 18:57 |
| inc0 | mnaser: replied to change | 18:58 |
| mnaser | inc0 good call, also fyi - http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22has%20become%20an%20orphan%2C%20killing%20it%5C%22 | 18:59 |
| mnaser | so a lot of containers get killed in kolla jobs bc of this | 18:59 |
| *** shashank_t_ has quit IRC | 18:59 | |
| mnaser | inc0 actually shouldnt that be done in kolla-ansible re: your comment | 19:00 |
| inc0 | mnaser: why? | 19:01 |
| mnaser | because no roles in the kolla repo? | 19:01 |
| inc0 | ahh | 19:01 |
| inc0 | right... | 19:01 |
| mnaser | ill do that regardless tho | 19:01 |
| mnaser | as another review | 19:01 |
| inc0 | yeah, I'm so newton | 19:01 |
| mnaser | :P | 19:01 |
| inc0 | right, ofc you're right | 19:01 |
| mnaser | same | 19:01 |
| mnaser | i was with you too | 19:01 |
| mnaser | but i was like "wait a second why cant i find this... search isnt returning anything.." | 19:02 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla-ansible master: Update Docker to 1.13.1 https://review.openstack.org/449285 | 19:04 |
| mnaser | btw, im working off of this to try and identify more gate failures, just search this in logstash: | 19:05 |
| mnaser | build_status:"FAILURE" AND build_name:gate-kolla-dsvm* AND message:"Job complete" | 19:05 |
| mnaser | if anyone finds that useful | 19:05 |
| egonzalez | mnaser the new docker version is not having issues with libvirt-selinux? | 19:05 |
| mnaser | egonzalez i think as part of that review, i ran setenforce 0 in the gate | 19:06 |
| mnaser | the justification was 1) kolla doesnt support selinux in the first place - docs say so | 19:06 |
| mnaser | 2) that only means it worked "by luck" before | 19:06 |
| mnaser | there is a comment there about it | 19:07 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Fernet Token Support https://review.openstack.org/446274 | 19:08 |
| *** deadnull has quit IRC | 19:09 | |
| sdake | kfox1111 sweet i can abandon my review with the kollakube res fix solution | 19:09 |
| *** shashank_t_ has joined #openstack-kolla | 19:09 | |
| kfox1111 | sdake: which review? | 19:10 |
| sdake | kfox1111 i did look at that review - and it seems wrong in some way | 19:10 |
| egonzalez | mnaser added a comment in kolla-ansible PS | 19:10 |
| kfox1111 | sdake: which review? | 19:10 |
| sdake | kfox1111 https://review.openstack.org/#/c/449262/1 | 19:10 |
| sean-k-mooney | inc0: did you ever get a chance to follow up on the osic apt repo issue? | 19:11 |
| kfox1111 | sdake: wrong how so? | 19:11 |
| sean-k-mooney | inc0: im currently patch my teams systems and just taught of it since im runing apt like 50 times in the last hour | 19:11 |
| sdake | there is still kollakube still used ? | 19:11 |
| kfox1111 | sdake: like I keep telling inc0, kollakube still does more then genconfig->configmap stuff. | 19:12 |
| kfox1111 | not all things have been helmized yet. | 19:12 |
| sdake | kfox1111 i see | 19:12 |
| kfox1111 | thats why getting completely out of kollakube is nontrivial. | 19:12 |
| egonzalez | inc0 mnaser not sure if pin point to docker-engine version, maybe is better remove the pin point? | 19:13 |
| kfox1111 | some ceph stuff needs converting. some configmap stuff, keepliave and openvswitch-set-ip, and at least a secret I think. | 19:13 |
| sdake | kfox1111 the thing that triggered all this tlak of ansible is this review: https://review.openstack.org/#/c/439740/ | 19:13 |
| *** mgoddard has quit IRC | 19:13 | |
| mnaser | egonzalez cool thanks | 19:13 |
| sdake | kfox1111 kollakube is broken today | 19:13 |
| mnaser | im not sure, i mean.. i think its better to pin things just in case docker breaks thingss | 19:14 |
| mnaser | so our gate can remain relatively healthy | 19:14 |
| kfox1111 | sdake: how does it not work today? its owrking in the gate. | 19:14 |
| mnaser | but i intend to poke this up to the 17.03 release soon if things are good | 19:14 |
| sean-k-mooney | mnaser: docker has broken us in the past so pinning is good provided we also test with new versions and update promptly when required | 19:15 |
| *** krtaylor has joined #openstack-kolla | 19:15 | |
| kfox1111 | sdake: and the ps you put up breaks everything. so how does it work? | 19:15 |
| sdake | kfox1111 it breaks the gate, however works when run locally | 19:15 |
| kfox1111 | I think its to try and elminnate the need for a symlink, but it doesn't seem to work? | 19:15 |
| mnaser | sean-k-mooney i agree with that. we can slowly but surely poke it up and try to keep up to date | 19:15 |
| sdake | kfox1111 if you run kollakube withoutsymlinks in place, it fials | 19:15 |
| sdake | read the patch you will underwtaand why | 19:15 |
| kfox1111 | sdake: gate's shouldn't be different then local. so why does it break? | 19:16 |
| sdake | i dont use symlinks locally | 19:16 |
| sdake | i pip install | 19:16 |
| kfox1111 | sdake: I did. but its not clear why it breaks things to me. | 19:16 |
| kfox1111 | do we not pip install in the gate? | 19:16 |
| kfox1111 | I thought we did. | 19:16 |
| openstackgerrit | Mohammed Naser proposed openstack/kolla-ansible master: Update Docker to 1.13.1 https://review.openstack.org/449285 | 19:17 |
| sdake | pip install -e | 19:17 |
| sdake | i pip install | 19:17 |
| sdake | pip install -e only works with symlinks | 19:17 |
| mnaser | egonzalez ty for that catch ^ | 19:17 |
| sdake | pip install works in any condition | 19:17 |
| sdake | although kollakube is busted | 19:17 |
| mnaser | i think eventually the setup code in kolla/ should just rely on kolla-ansible for the deploy jobs | 19:18 |
| mnaser | but thats for another patch | 19:18 |
| kfox1111 | hmm.... | 19:18 |
| kfox1111 | so, we pip install kolla-kubernetes, but not kolla-ansible in the gate it looks like. | 19:18 |
| *** shardy is now known as shardy_afk | 19:18 | |
| kfox1111 | sdake: so, in your ps, if you remove the added symlink, and add a pip install in here: https://github.com/openstack/kolla-kubernetes/blob/master/tools/setup_gate_common.sh#L80 | 19:18 |
| kfox1111 | I think it would fix the gate and fix your concern? | 19:18 |
| egonzalez | mnaser yep, makes sense, but is not as easy because deploy gates in kolla also build the images before deploy | 19:19 |
| sdake | possible | 19:20 |
| *** shashank_t_ has quit IRC | 19:20 | |
| sdake | kfox1111 may require more gate work | 19:20 |
| kfox1111 | sdake: possibly. | 19:20 |
| kfox1111 | but we can't commit somethign that breaks the gate. | 19:20 |
| kfox1111 | so either we recommend people just add the symlink, | 19:20 |
| kfox1111 | or we fix the gate. | 19:20 |
| sdake | i agree | 19:20 |
| *** shashank_t_ has joined #openstack-kolla | 19:20 | |
| sdake | i'm not asking for it to be committed with a red gate | 19:21 |
| sdake | i'm pointing out what triggered the ansible discussion | 19:21 |
| kfox1111 | k. yeah. I'm just saying what I think might fix the review. | 19:21 |
| sdake | yup - atm working on https://review.openstack.org/#/c/447356/ - as i committed to have this ready to go by friday | 19:23 |
| *** shashank_t_ has quit IRC | 19:23 | |
| kfox1111 | ah. ok. cool. | 19:23 |
| *** adrian_otto has quit IRC | 19:24 | |
| sbezverk_ | sdake: the recheck you pushed for my ps came back all green.. | 19:24 |
| sbezverk_ | sdake: please take a look at ps when you have a chance | 19:25 |
| sdake | gerrit pokey atm | 19:26 |
| sdake | can't do anytthing with it really | 19:26 |
| *** adrian_otto has joined #openstack-kolla | 19:26 | |
| *** vhosakot has joined #openstack-kolla | 19:27 | |
| *** matrohon has joined #openstack-kolla | 19:32 | |
| sbezverk_ | sdake: ok did not know that.. | 19:34 |
| *** adrian_otto has quit IRC | 19:34 | |
| kfox1111 | gerrits getting worse. :/ | 19:41 |
| *** Margin has joined #openstack-kolla | 19:41 | |
| *** Margin has joined #openstack-kolla | 19:42 | |
| Margin | hello | 19:43 |
| *** g3ek has quit IRC | 19:43 | |
| *** shashank_t_ has joined #openstack-kolla | 19:51 | |
| *** g3ek has joined #openstack-kolla | 19:52 | |
| *** vhosakot has quit IRC | 19:54 | |
| *** vhosakot_ has joined #openstack-kolla | 19:55 | |
| *** rabel_b1 has quit IRC | 20:04 | |
| *** eaguilar has quit IRC | 20:06 | |
| *** mliima has quit IRC | 20:10 | |
| *** manheim has joined #openstack-kolla | 20:10 | |
| inc0 | can I get reviews plz? https://review.openstack.org/#/c/448258/ | 20:11 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Simple configmap to raw kubectl https://review.openstack.org/449262 | 20:12 |
| kfox1111 | inc0: so, you want to merge it in seperate chunks? | 20:13 |
| inc0 | yes, please | 20:13 |
| inc0 | one role at the time | 20:13 |
| inc0 | lol it took 3hrs for gates to pass | 20:13 |
| kfox1111 | how long do you think it will take to remove the need for genconfig entirely? | 20:15 |
| inc0 | how many patchsets? # of services | 20:16 |
| inc0 | 7 or 8? | 20:16 |
| inc0 | and then last patch to remove genconfig from gates and docs | 20:17 |
| kfox1111 | no, like, wall clock? | 20:17 |
| inc0 | took me 30min to get ks up | 20:17 |
| inc0 | so if I'm alone, entirely possible to do it next week fully | 20:17 |
| inc0 | with help, less | 20:17 |
| kfox1111 | cause while its enabled in the gate, we have a mismatch between what the gate is testing, and what usuers will probably do. | 20:17 |
| kfox1111 | I guess if its just for a week, then its not that risky, and gets ut there quicker? | 20:19 |
| inc0 | quicker than what? | 20:19 |
| kfox1111 | quicker then waiting until all ps's are ready and merging all at once? | 20:20 |
| inc0 | please let's nt do that | 20:20 |
| kfox1111 | It feels a little dirty having the gate half and half for how it gen's config. | 20:20 |
| kfox1111 | though we did the same with helm. but that was really messy. :/ | 20:20 |
| inc0 | map-reduce;) | 20:21 |
| inc0 | that allows others to work with me | 20:21 |
| kfox1111 | I'm on the fence. but just slightly on the side of lets just get it done quickly and get it overwith. | 20:21 |
| inc0 | quickest way - one role at time | 20:21 |
| kfox1111 | +2'ed. | 20:21 |
| inc0 | because again, allows cooperation of multiple people | 20:21 |
| inc0 | thank you | 20:21 |
| *** egonzalez has quit IRC | 20:21 | |
| *** satyar has quit IRC | 20:21 | |
| kfox1111 | well, with the current aproach, it will cause merge conflicts | 20:22 |
| inc0 | keystone is almost ready too, zuul is finishing last gate | 20:22 |
| inc0 | all other green | 20:22 |
| kfox1111 | cause they will race on adding rm -rf /etc/kolla/xxxx's. | 20:22 |
| kfox1111 | until they are all done with genconfig. | 20:22 |
| inc0 | it's easy rebase tho | 20:22 |
| inc0 | if needed at all | 20:22 |
| kfox1111 | yeah. | 20:22 |
| kfox1111 | not a major problem. | 20:23 |
| kfox1111 | more of a problem if it wasn't like 8 things. | 20:23 |
| *** sayantani01 has quit IRC | 20:24 | |
| kfox1111 | inc0: https://review.openstack.org/#/c/441550/ | 20:24 |
| kfox1111 | inc0: ironic totally working, and tested in gate. sbezverk did a huge amount of work there. great stuff. :) | 20:24 |
| sdake | yup pretty impressive sbezverk_ ^^ | 20:25 |
| inc0 | 120 patchsets :O | 20:27 |
| inc0 | that might be record | 20:27 |
| sdake | ironic is really difficult to implement | 20:28 |
| inc0 | I know | 20:28 |
| sdake | sbezverk_ should get a gold medal in the triathalon for getting that working and better yet getting it gating | 20:28 |
| inc0 | kfox1111: https://review.openstack.org/#/c/449254/ | 20:29 |
| kfox1111 | gota head out for a little while. | 20:29 |
| inc0 | keystone genconfig | 20:29 |
| inc0 | doing nova now | 20:29 |
| inc0 | sbezverk_: mind taking a look? | 20:29 |
| inc0 | https://review.openstack.org/#/c/448258/ | 20:29 |
| kfox1111 | inc0: don't have time to look yet, | 20:29 |
| kfox1111 | but just cursory, | 20:29 |
| inc0 | np | 20:29 |
| kfox1111 | we don't need the fernet/ssh stuff. | 20:30 |
| inc0 | kk | 20:30 |
| kfox1111 | doing it a much different way. | 20:30 |
| inc0 | I removed | 20:30 |
| kfox1111 | k8s native. | 20:30 |
| inc0 | fernet stuff | 20:30 |
| kfox1111 | ok. gotta go. | 20:30 |
| inc0 | have fun | 20:30 |
| rwellum | sdake: sorry the free time I thought I'd have at 4 EST never happened - stuck doing something else. | 20:32 |
| sdake | rwellum roger - understood | 20:32 |
| sdake | rwellum perhaps tomorrow | 20:32 |
| rwellum | yes | 20:32 |
| *** sayantani01 has joined #openstack-kolla | 20:36 | |
| sbezverk_ | inc0: done, good start !! | 20:42 |
| inc0 | https://review.openstack.org/#/c/449254/ <- this is ks | 20:43 |
| inc0 | green at first try;) | 20:43 |
| *** jrobinson has joined #openstack-kolla | 20:44 | |
| sbezverk_ | inc0: I think kfox1111 doing fernet separately. Maybe we need to remove it from this ps? | 20:45 |
| inc0 | sbezverk_: yeah, mind adding a comment? | 20:46 |
| sbezverk_ | inc0: sure thing | 20:46 |
| krtaylor | inc0, sdake - (different topic) at altanta, I agreed to go look for help with the ansible extensible strategy plugin, of course, I then got pulled into 10 other things | 20:46 |
| krtaylor | inc0, sdake - but I did talk to our openstack-ansible guru | 20:47 |
| inc0 | and?:) | 20:47 |
| *** rwallner has quit IRC | 20:47 | |
| krtaylor | inc0, sdake - he said "After looking at this a bit…it doesn’t seem like Kolla *needs* to put this change directly into Ansible. Ansible is pluggable and Kolla could simply configure Ansible to use this strategy plugin versus putting it up in Ansible proper." | 20:48 |
| inc0 | btw krtaylor https://github.com/inc0/ansible/commit/beba49876974d05c38d98ec2ec855bf59eb480e2 <- this is my prototype | 20:48 |
| inc0 | krtaylor: yes and no | 20:48 |
| inc0 | yes, we can put it there | 20:48 |
| inc0 | but this file has to be GPLv3 licensed | 20:48 |
| inc0 | and if part of project is GPLv3 licensed...you know the rest | 20:49 |
| krtaylor | ah, I knew there was a catch | 20:49 |
| inc0 | it's legal not techincal | 20:49 |
| krtaylor | I couldn't remember why | 20:49 |
| openstackgerrit | Merged openstack/kolla-kubernetes master: WIP Move config generation to kolla-k8s https://review.openstack.org/448258 | 20:49 |
| inc0 | now I see that commit msg wasn't ideal | 20:50 |
| inc0 | well...what's done is doen | 20:50 |
| krtaylor | well, the bad news is I couldn't find an ansible person to drive it | 20:50 |
| krtaylor | sorry it took so long to follow up | 20:50 |
| inc0 | this might be documentation problem and create external project, licensed by gplv3 | 20:50 |
| inc0 | that will have literally one file | 20:51 |
| krtaylor | heh, brute force approach, was there any objection to including it in ansible proper? | 20:52 |
| inc0 | nothing much, but also not much was discussed/looked at | 20:52 |
| inc0 | it's really fully optional and adds value | 20:52 |
| inc0 | all is needed is some love and care | 20:53 |
| *** manheim has quit IRC | 20:53 | |
| krtaylor | I wish I had the bandwidth to take it on right now, but also not sure about if I can with the GPL | 20:54 |
| kfox1111 | inc0: the plugin can be gplv3 compatable, and only gets covered in process. | 20:54 |
| kfox1111 | and I believe the apache lisense is very explicitly v3 compatible. | 20:54 |
| kfox1111 | so, should be fine? | 20:54 |
| krtaylor | mingling code would make it viral I believe (not a legal expert) | 20:55 |
| inc0 | well...ML didnt give me clear answer | 20:55 |
| kfox1111 | viral in process. | 20:55 |
| kfox1111 | gpl's a distribution license. | 20:55 |
| inc0 | I'm positive that this file has to be gpl | 20:55 |
| kfox1111 | I'm pretty sure it doesn't. | 20:56 |
| inc0 | not sure about other files in same repo | 20:56 |
| inc0 | it does, it imports gpl code | 20:56 |
| inc0 | from very internals of ansible | 20:56 |
| inc0 | that's the problme | 20:56 |
| sdake | krtaylor thanks for getting back to me | 20:57 |
| krtaylor | I wish I had better news | 20:57 |
| kfox1111 | stil wouldn't be dirivitive until it is in process. | 20:57 |
| inc0 | kfox1111: I don't know and I'm not comfortable to navigate this field without guidance:/ | 20:57 |
| inc0 | too much gray and too big of a risk | 20:57 |
| kfox1111 | inc0: fair enough. | 20:57 |
| kfox1111 | I'm not a layer. but I have read through a lot of open source licenses. in detail. | 20:58 |
| *** adrian_otto has joined #openstack-kolla | 20:58 | |
| inc0 | kfox1111: with a bit of pressure we can make you a layer | 20:59 |
| kfox1111 | -1 | 20:59 |
| kfox1111 | :) | 20:59 |
| kfox1111 | I can parse machine code really well. I don't want to parse human code for a living. | 20:59 |
| kfox1111 | I only do becausse I need to do so I can write machine code. :) | 20:59 |
| inc0 | that's exactly my thing | 20:59 |
| krtaylor | is there openstack legal that could assist? | 21:00 |
| kfox1111 | yeah, shoudl be able to drop legal mailing list an email. | 21:00 |
| inc0 | I think I did | 21:01 |
| *** jascott1_ has joined #openstack-kolla | 21:01 | |
| kfox1111 | sbezverk_: don't think a recheck will clar that. | 21:01 |
| kfox1111 | I think we need to -workflow + workflow. | 21:01 |
| kfox1111 | cause it was in the merge check, not the normal check. | 21:02 |
| sbezverk_ | kfox1111: what then :( ?? | 21:07 |
| sbezverk_ | I see | 21:07 |
| jascott1_ | kfox1111 hit a little snag, how important is it to have the ability to merge values from two different modules into one parent key? | 21:08 |
| jascott1_ | modules=charts | 21:08 |
| *** krtaylor has quit IRC | 21:08 | |
| *** lamt has quit IRC | 21:10 | |
| *** lamt has joined #openstack-kolla | 21:12 | |
| jemcevoy | I am having an issue with a new install at a customer site... When I run "pip install kolla" I get version 4.0.0 and there is no globals.yml file in the /usr/share/kolla/etc_examples/kolla ??? What am I do incorectly? | 21:12 |
| jascott1_ | hi jemcevoy! might try checkout from git? | 21:15 |
| *** kolla-slack1 has joined #openstack-kolla | 21:18 | |
| *** jascott1_ has quit IRC | 21:20 | |
| *** jascott1_ has joined #openstack-kolla | 21:20 | |
| *** kolla-slack1 has quit IRC | 21:21 | |
| *** jtriley has quit IRC | 21:22 | |
| *** kolla-slack has joined #openstack-kolla | 21:22 | |
| *** kolla-slack has quit IRC | 21:22 | |
| *** kolla-slack1 has joined #openstack-kolla | 21:23 | |
| *** kolla-slack has joined #openstack-kolla | 21:24 | |
| sdake | jemcevoy the config files are in kolla-ansible deliverable | 21:25 |
| *** jascott1_ has quit IRC | 21:25 | |
| *** jascott1_ has joined #openstack-kolla | 21:25 | |
| *** kolla-slack has quit IRC | 21:26 | |
| *** kolla-slack has joined #openstack-kolla | 21:26 | |
| *** eaguilar has joined #openstack-kolla | 21:26 | |
| *** matrohon has quit IRC | 21:27 | |
| sdake | slackbot do something | 21:28 |
| *** lamt has quit IRC | 21:28 | |
| sdake | slackbot do something | 21:29 |
| kolla-slack | <sdake> sweet | 21:29 |
| *** jascott1_ has quit IRC | 21:30 | |
| *** jascott1_ has joined #openstack-kolla | 21:32 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-kubernetes master: Move keystone configuration to kolla-k8s https://review.openstack.org/449254 | 21:33 |
| *** pbandark has joined #openstack-kolla | 21:34 | |
| *** jascott1_ has quit IRC | 21:35 | |
| jemcevoy | sdake: What is the pip command to get kolla-ansible deliverable ( I am not a pythonista ) | 21:35 |
| inc0 | jemcevoy: pip install kolla-ansible :) | 21:36 |
| jemcevoy | Do I need both kolla and kolla-ansible packages? | 21:37 |
| inc0 | not really | 21:37 |
| sdake | kolla is for buildling | 21:37 |
| sdake | kolla-ansible is for deploying | 21:37 |
| sdake | ify ou wantot build your own or custom images you need kolla package | 21:37 |
| sdake | if your happyto pull them form dockerhub that works too | 21:37 |
| *** lucasxu has joined #openstack-kolla | 21:40 | |
| jemcevoy | I fill local repository with the binary downloads. But I think it build the docker images to fill the repo right? | 21:40 |
| *** jascott1_ has joined #openstack-kolla | 21:42 | |
| jemcevoy | sdake: This customer has one of those man-in-the-middle http/https proxies so the internet is kinda slow. So I will need both to build the local repo for the cluster install I assume. | 21:43 |
| inc0 | jemcevoy: well, not on same node tho;) | 21:44 |
| inc0 | like, no need to be on same node | 21:44 |
| kolla-slack | <jascott1> that koala is naked! | 21:45 |
| kolla-slack | <jascott1> neeked! | 21:45 |
| sdake | indeed jascott1 | 21:45 |
| kolla-slack | <jascott1> wow the integration seems to work | 21:45 |
| *** adrian_otto has quit IRC | 21:47 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-kubernetes master: Move keystone configuration to kolla-k8s https://review.openstack.org/449254 | 21:47 |
| jemcevoy | inc0: Not following... I have a new 12 node cluster and one ops server on a older reclaimed system where I plan to run the kolla-ansible from to do the Newton install. This ops server will also have cobbler to kickstart the new cluster... | 21:48 |
| *** yangyapeng has joined #openstack-kolla | 21:48 | |
| jemcevoy | do I install both kolla and kolla-ansible on the ops machine | 21:48 |
| inc0 | what I'm saying, there is no need for them to be on same machine | 21:48 |
| inc0 | it's entirely up to you | 21:48 |
| inc0 | kolla will provide private registry of images, kolla-ansible will know how to deploy from it | 21:49 |
| jemcevoy | I installed both with pip... Guess I am good to go then I see that the globals.yml is there... I had to have them generate a "real" x509 cert for my repository since the self signed one did not work in their "secure" environment... | 21:51 |
| *** eaguilar has quit IRC | 21:52 | |
| *** jascott1_ has quit IRC | 21:52 | |
| *** yangyapeng has quit IRC | 21:52 | |
| *** jascott1_ has joined #openstack-kolla | 21:53 | |
| *** eaguilar has joined #openstack-kolla | 21:56 | |
| *** jascott1_ has quit IRC | 21:57 | |
| kolla-slack | <sdake> @jascott1 ya its pretty cool - surprised it was so easy to setup | 21:58 |
| *** shashank_t_ has quit IRC | 22:00 | |
| *** jascott1_ has joined #openstack-kolla | 22:06 | |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Virtualbmc and deployment of baremetal VM using ironic https://review.openstack.org/441550 | 22:06 |
| dmsimard | inc0, sdake: you nice people see anything out of the ordinary for RDO repos recently ? yum errors etc ? | 22:18 |
| sdake | dmsimard gate seems a little flakey but not certain | 22:19 |
| fooliouno | kfox1111: you there? | 22:19 |
| sdake | dmsimard our gate was broken by other things | 22:19 |
| sdake | dmsimard not sure if it rdo was a problem or not | 22:19 |
| kolla-slack | <inc0> Oh great now I'll get ping on phone from irc | 22:19 |
| kolla-slack | <inc0> Just what I dreamt of | 22:19 |
| sdake | inc0 use the mute feature | 22:19 |
| sdake | "mute the channel" in slack | 22:20 |
| kolla-slack | <inc0> I know :) | 22:20 |
| kolla-slack | <inc0> No that's actually good | 22:20 |
| fooliouno | sdake: I see that you are setting the kolla_controller and kolla-compute to the same node. Will it work if I apply them to two separate nodes. Has it been tested :) | 22:20 |
| sdake | many irc clients have various noises they can emit if you set it up | 22:20 |
| sdake | fooliouno tested no idea | 22:20 |
| sdake | fooliouno with ceph I believe so | 22:21 |
| sdake | fooliouno but not certain | 22:21 |
| sdake | fooliouno those docs you are referencing have not been tested however | 22:21 |
| sdake | fooliouno so YMMV | 22:21 |
| sdake | fooliouno please leave comments in them | 22:21 |
| sdake | fooliouno so they can be improved | 22:21 |
| fooliouno | sdake: I will add comments as I run through them. If I use "host" configuration, then I cannot separate the two functions? | 22:22 |
| *** kolla-slack1 has quit IRC | 22:23 | |
| sdake | fooliouno ya its not ideal | 22:24 |
| fooliouno | ok .. thanks | 22:24 |
| *** kolla-slack has quit IRC | 22:24 | |
| *** kolla-slack has joined #openstack-kolla | 22:25 | |
| *** mbruzek has quit IRC | 22:29 | |
| *** fooliouno has quit IRC | 22:31 | |
| *** eaguilar has quit IRC | 22:32 | |
| *** dave-mcc_ has joined #openstack-kolla | 22:32 | |
| *** dave-mccowan has quit IRC | 22:36 | |
| *** dave-mccowan has joined #openstack-kolla | 22:38 | |
| *** lucasxu has quit IRC | 22:39 | |
| openstackgerrit | Marcus Williams proposed openstack/kolla master: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 22:40 |
| *** dave-mcc_ has quit IRC | 22:40 | |
| *** manheim has joined #openstack-kolla | 22:43 | |
| *** unicell1 has joined #openstack-kolla | 22:46 | |
| *** unicell has quit IRC | 22:46 | |
| *** vhosakot_ has quit IRC | 22:50 | |
| *** tonanhngo has quit IRC | 22:51 | |
| sdake | v1k0d3n re your personal attack in the loci review, i have a right to review anything in the community in any way as I see fit. My commentary was valid and is likely still valid. Please don't attack other people personally - that doesn't fly in OpenStack. TIA | 22:54 |
| openstackgerrit | Marcus Williams proposed openstack/kolla master: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 22:57 |
| *** tonanhngo has joined #openstack-kolla | 22:58 | |
| *** srwilkers has quit IRC | 22:58 | |
| *** g3ek has quit IRC | 23:00 | |
| *** tonanhngo has quit IRC | 23:02 | |
| *** Pavo has joined #openstack-kolla | 23:08 | |
| *** g3ek has joined #openstack-kolla | 23:09 | |
| *** Pavo has quit IRC | 23:11 | |
| v1k0d3n | sdake please stop. nobody was attacking anyone personally at all. | 23:12 |
| *** Margin has quit IRC | 23:15 | |
| *** fooliouno has joined #openstack-kolla | 23:15 | |
| *** Margin has joined #openstack-kolla | 23:16 | |
| kolla-slack | <vhosakot> nice… hellor kolla friends in IRC :slightly_smiling_face: | 23:17 |
| kolla-slack | <vhosakot> I’m sending this from k8s slack :slightly_smiling_face: | 23:17 |
| sdake | vhosakot we see that ;) | 23:18 |
| kolla-slack | <vhosakot> awesome! :slightly_smiling_face: | 23:18 |
| *** manheim has quit IRC | 23:18 | |
| spsurya | morning all | 23:19 |
| *** MasterOfBugs has quit IRC | 23:19 | |
| sdake | sup spsurya | 23:19 |
| *** Margin has quit IRC | 23:20 | |
| *** pbandark has quit IRC | 23:31 | |
| *** adrian_otto has joined #openstack-kolla | 23:42 | |
| jemcevoy | inc0: I have a new issue. kolla-build is failing on the customer site because docker build does not have the http/https man-in-the-middle proxy configured inside the centos7 containers so the rpm --import https://<blah> fails because it needs access to the ca.crt file for the proxy. Any ideas on how to get around this? | 23:45 |
| *** goldyfruit has quit IRC | 23:45 | |
| *** egonzalez has joined #openstack-kolla | 23:48 | |
| openstackgerrit | Mohammed Naser proposed openstack/kolla-ansible master: Update Docker to 1.13.1 https://review.openstack.org/449285 | 23:52 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!