*** jrobinson has joined #openstack-kolla | 00:01 | |
sbezverk | I will build new ubuntu libvirtd container to match version I have on the host and see if it helps | 00:03 |
---|---|---|
sdake | ok peeps check this out - I need help badly: http://paste.openstack.org/show/604394 | 00:03 |
kfox1111 | I wonder if thats related to the ceph volume mount issue I saw earlier too. | 00:03 |
*** shashank_t_ has joined #openstack-kolla | 00:03 | |
spsurya | morning all | 00:03 |
kfox1111 | sdake: kubectl describe pod | 00:04 |
sbezverk | kfox1111: possible, but it works on centos 100% | 00:04 |
sdake | kfox1111 are you dense ;) | 00:04 |
kfox1111 | sbezverk: yeah. I was seeing that too. | 00:04 |
sdake | kfox1111 - check out the paste again | 00:04 |
* sdake does a dance | 00:04 | |
*** signed8bit is now known as signed8bit_Zzz | 00:04 | |
kfox1111 | sdake: tahts a pod list. where's the describe pod on the failing pod? | 00:05 |
sdake | oh thats lvm | 00:05 |
sdake | thats because iscsi isnt' running or something | 00:05 |
*** shashan__ has joined #openstack-kolla | 00:05 | |
sdake | i know aout that | 00:05 |
sdake | the point is compute is running | 00:05 |
kfox1111 | k. | 00:05 |
kfox1111 | sdake: oh. I thought you were asking why it was broken. | 00:05 |
sdake | so placement 4.0.0 works | 00:05 |
kfox1111 | cool. :) | 00:05 |
sdake | i was making a joke - iguess poorly :) | 00:05 |
*** caoyuan has joined #openstack-kolla | 00:06 | |
kfox1111 | sdake: had a cold for the last few days. my brains pretty fuzzy. :/ | 00:06 |
sdake | kfox1111 i've been in a constant state of almost being sick for 9 months | 00:06 |
kfox1111 | bummer. :/ | 00:06 |
sdake | its actually a good thing | 00:07 |
sdake | that means my health is good :) | 00:07 |
sdake | considering i have two children who constantly bring around germs | 00:07 |
kfox1111 | I know, right? I think 98% of the bugs in the last few years correlate rather directly to the amount of wall licking my little boy has performed. :/ | 00:08 |
*** shashank_t_ has quit IRC | 00:08 | |
kfox1111 | one of these days he's gonna stop tasting things he shouldn't.... | 00:09 |
sdake | i asked my doctor about reading that stress wrecks yoru immune system | 00:10 |
sdake | my doctor said the right amount of stress + anxiety improves your health | 00:11 |
sdake | the problem most people have is balancing that | 00:11 |
kfox1111 | yeah. | 00:11 |
sdake | he also said it improves your general immune system response | 00:11 |
sdake | which blew me away | 00:11 |
sdake | hes a super pro | 00:11 |
sdake | "I find that answer vague and unconvincing." | 00:11 |
*** shashan__ has quit IRC | 00:12 | |
kfox1111 | somewhere between, what doesnt' kill you makes you stronger, and just chips away at your streangth until you have none left. | 00:12 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Simple configmap to raw kubectl https://review.openstack.org/449262 | 00:15 |
sdake | rwellum - the review seems to work | 00:15 |
sdake | rwellum if you know what your oding | 00:15 |
sdake | rwellum i will hopefully integrate everyones changes tonight | 00:15 |
sdake | rwellum although my mortal form grows weary :) | 00:16 |
sdake | since kubeadm is difficult to setup, i'm going to include a quick dummy guide to getting kubeadm + calico rolling | 00:16 |
*** signed8bit_Zzz is now known as signed8bit | 00:16 | |
sdake | until those upstreams make their docs/code more accessible | 00:17 |
*** schwicht has joined #openstack-kolla | 00:17 | |
sdake | following the current master docs in the various repos doesn't work | 00:17 |
sdake | (to setup kubernetes + cni calico driver0 | 00:17 |
kfox1111 | sbezverk: or tweak some scripts to automate it. | 00:17 |
sdake | rather canal driver) | 00:17 |
kfox1111 | sdake: or tweak some scripts to automate it. | 00:17 |
sdake | kfox1111 peopel can contribute scripts if they like - i think most people want to understand the whole workflow beginning to end before they make a decision | 00:18 |
kfox1111 | sdake: workflow's continuing to change. its likely that by the time anyone tries it, it breaks. :/ | 00:19 |
kfox1111 | the expierence of it being broken all the time is much worse then it being a black box. :/ | 00:19 |
kfox1111 | it being a black box, you know it works, yout just need to dig into how. | 00:19 |
sdake | kfox1111 fortunately i don't plan to relly on the gate for dev | 00:19 |
kfox1111 | if its broken, yhou just don't have any faith. | 00:19 |
sdake | kfox1111 so plan to rely on docs - just like the other 80% of devs int he world :) | 00:20 |
sdake | and maintain them as such | 00:20 |
kfox1111 | sdake: yeah, but devenv's don't work that way. you set it up once, then tend to ignore the docs. | 00:20 |
kfox1111 | only infrequently do you refer to it and find out its broken. | 00:20 |
sdake | this is really a deployment doc | 00:20 |
sdake | if people use it for dev, that is their perogative | 00:20 |
kfox1111 | I guess I'm just saying, docs that have a lot of steps dont get followed, and get broken a lot. | 00:21 |
kfox1111 | people don't read docs in general. | 00:21 |
sdake | i agree | 00:22 |
sdake | although what are we to point them to if we have no docs? | 00:22 |
sdake | "well go read the code luke" | 00:22 |
sdake | no ciesta bien | 00:22 |
kfox1111 | code. | 00:22 |
kfox1111 | "run this shell script" | 00:22 |
kfox1111 | or, "run this container" | 00:22 |
kfox1111 | or something like that. | 00:22 |
sdake | both operators and developers want to know how a thing works | 00:23 |
kfox1111 | ideally its 100% testable in a gate, or at least mostly. | 00:23 |
sdake | at some point they are willing to treat things as a black box | 00:23 |
kfox1111 | that can be done in docs. | 00:23 |
sdake | but not the front interface | 00:23 |
kfox1111 | but when the understanding isnt 100%, it doesnt' break getting a working system. | 00:23 |
kfox1111 | thats the difference. | 00:23 |
kfox1111 | allow it to be linux from scratched, but don't reocmment that as the default. | 00:24 |
kfox1111 | build it for them, and then explain how it work, | 00:24 |
kfox1111 | and if they want to linux from scratch, then and only then, they do that. | 00:24 |
sdake | well i suppose we have a difference of opinion on this point | 00:25 |
sdake | who is right - nobody knows :) | 00:25 |
kfox1111 | I'm the kind that wants to linux from scratch it. | 00:25 |
kfox1111 | but 90%+ of folks dont. | 00:25 |
sdake | people that use kolla-ansible want to understand the entire workflow | 00:25 |
kfox1111 | I agree. | 00:25 |
sdake | i ever tell you about corosync and people asking me about inane log messages and what they mean | 00:25 |
kfox1111 | but you don't ask them to paste in every ansible task one at a time in a doc. | 00:25 |
sdake | for 3-4 years i was asked daily what does "XYZ log msg mean" | 00:25 |
sdake | you knwo what I did? | 00:25 |
*** lucasxu has joined #openstack-kolla | 00:26 | |
inc0 | write a doc? | 00:26 |
sdake | no | 00:27 |
sdake | corosync was mature and had over 25k production deployments in finance /etc | 00:27 |
sdake | a doc would have created more problems | 00:27 |
sdake | that is the 90% yoru talking about that don't care kfox1111 | 00:27 |
sdake | the real production deployments | 00:27 |
sdake | early adoptors, do care | 00:27 |
sbezverk | inc0: do you know why we use 1.3.1 libvirt-bin on ubuntu | 00:28 |
sbezverk | and if it can be changed for PoC? | 00:28 |
kfox1111 | sdake: I get earlier adopters care, but its gotta be stableish too. it cant be broken all the time. | 00:29 |
sdake | that 90% figure you quoted kfox1111 - they care abouth ow a system behaves | 00:29 |
sdake | the 10% - the early adopters - care how a system works | 00:29 |
kfox1111 | how can be a bit more abstract. pictures, etc. | 00:30 |
sdake | early adoptors need to experience the software to learn it | 00:30 |
kfox1111 | not a big list of shell snippets. | 00:30 |
sdake | a big list of documented shell snippets helps people understand how a system works | 00:30 |
kfox1111 | I agree. but that can be done by poking at a working deployment more easily then standing one up from scratch. | 00:31 |
sdake | devenv = working deployment - yet nobody uses it | 00:31 |
kfox1111 | only if its not leading them down the wrong direction. | 00:31 |
kfox1111 | which happens if the docs get out of sync with reality. | 00:31 |
kfox1111 | which happens way to often. :/ | 00:31 |
sdake | docs do get out of sync | 00:31 |
kfox1111 | especially on a rapidly changing project. | 00:31 |
sdake | i fail to see how that is relevant though :) | 00:31 |
sdake | if they are maintained - they stay in sync and keep new devs coming | 00:32 |
sdake | saying "go read the gate code or the devnenv shell script" is a recipe for no contributors | 00:32 |
kfox1111 | if you share code, they are easier to maintain. thats my point. | 00:32 |
sdake | easier to maintain vs slower community growth | 00:32 |
sdake | tradeoff there | 00:32 |
kfox1111 | no. we share code between the gate, and the dev env, and the prod stuff. | 00:32 |
kfox1111 | and test it. | 00:32 |
SamYaple | sbezverk: http://packages.ubuntu.com/xenial/libvirt-bin | 00:33 |
kfox1111 | right now, its a copy pasted nightmare. | 00:33 |
sdake | inc0 how many times has kolla-ansible been tested by the quickstart guide? | 00:33 |
SamYaple | sbezverk: you can grab libvirt-bin 2.5.0 from ocata cloud-archive | 00:33 |
*** adrian_otto has quit IRC | 00:33 | |
*** Pavo has quit IRC | 00:33 | |
kfox1111 | sdake: rather, how long does it take a major brakage in the kolla-ansible docs to be noticed. | 00:34 |
kfox1111 | (and kolla-ansible isn't changing as rapidly as kolla-kubernetes) | 00:35 |
*** tonanhngo has quit IRC | 00:35 | |
sbezverk | SamYaple: thanks so it is default then.. infra ubuntu vm gets 2.5.0, I see some issue which could be related tp mixed versions | 00:35 |
sdake | kfox1111 not sure i have the answer to your question although the QSG in kolla-ansible was changing as rapidly as the code base | 00:36 |
sbezverk | SamYaple: if I need to build ubuntu source for nova-libvirt with 2.5.0, what do I need to change, I an more ver fluent in ubuntu.. | 00:36 |
sbezverk | SamYaple: I meant to say, I am NOT very fluent in Ubuntu | 00:37 |
SamYaple | sbezverk: change it to ocata cloud-archive | 00:37 |
SamYaple | sbezverk: looks like master is ocata cloud-archive | 00:38 |
SamYaple | so you should be pulling in that in the nova-libvirt container | 00:38 |
sbezverk | SamYaple: just tried kolla master and it still get 1.3.1 | 00:39 |
esharao | sdake recreated my vm and almost got all pods running except for cinder-lvm and my nova-libvirt is crashing | 00:39 |
SamYaple | sbezverk: which container are you expecting to have libvirt-bin 2.5.0 again | 00:39 |
sbezverk | nova-libvirt | 00:40 |
*** gkadam has quit IRC | 00:40 | |
SamYaple | sbezverk: can you run `apt-cache policy libvirt-bin` in the nova-libvirt container | 00:40 |
*** sayantan_ has quit IRC | 00:42 | |
sdake | esharao that is good to hear :) | 00:43 |
sdake | esharao i mean not the libvirt crashing | 00:43 |
sdake | but the almost all running | 00:43 |
sdake | esharao did you git am the 0001-* patch? | 00:43 |
sdake | i htink cinder-lvm requires iscsi | 00:43 |
esharao | nope | 00:43 |
esharao | ahh ok.. i will check it out | 00:43 |
sdake | esharao want me to walk you through a recovery? | 00:43 |
sbezverk | SamYaple: I can but I cannot get results quickly as it is a gate job | 00:44 |
sdake | ehsso you don't have to reset from scratch? | 00:44 |
sdake | esharao ^^ | 00:44 |
esharao | oh nice.. | 00:44 |
esharao | one more thing.. my compute keeps on crashing.. | 00:44 |
esharao | is that due to using ubuntu as my base os? | 00:44 |
sdake | esharao yup - we have solutiosn for those | 00:44 |
sdake | no its due to incorrect documentation | 00:44 |
sdake | esharao http://cisco.webex.com/join/stdake if its not too late for ya | 00:45 |
esharao | its a bit late.. can we do it tomorrow? | 00:45 |
eanylin | sdake: question, did you try and create VM? | 00:45 |
sdake | eanylin i have only deployed on bare metal using all of the wealth of information in the review | 00:45 |
*** tonanhngo has joined #openstack-kolla | 00:45 | |
eanylin | yeah, same here, all my containers are running | 00:46 |
sdake | eanylin grats :) | 00:46 |
sdake | esharao i am free at 12 PM MST | 00:46 |
eanylin | nova hypervisor-list shows that nova is running well | 00:46 |
eanylin | sdake: Thanks :) | 00:46 |
esharao | ok.. perfect.. will ping you then | 00:46 |
esharao | thanks | 00:46 |
eanylin | but I get no valid host found error | 00:46 |
sdake | esharao happy to help | 00:46 |
sbezverk | SamYaple: to build ubuntu source using 2.5.0, which config I need to change on kolla side? config.py? | 00:47 |
eanylin | as seen in the nova scheduler logs... | 00:47 |
SamYaple | sbezverk: im just pointing out that ocata xenial-updates which is added in base sources.list for ubuntu has 2.5.0 libvirt-bin | 00:47 |
esharao | thanks for all your help sdake | 00:47 |
esharao | tty tomorrow | 00:47 |
SamYaple | sbezverk: you shouldnt need to change anything | 00:47 |
sdake | eanylin boy that erorr is annoying - can yo urun kubectl get pods -n kube-system -> pasteservice | 00:47 |
SamYaple | sbezverk: oh.. wait.. is this arm based? | 00:47 |
sbezverk | SamYaple: hmm I am positive that currently 4.0.0 images still use 1.3.1 | 00:48 |
sdake | eanylin note i haven't spun up a vm as of yet - i was dancing when i got kubernetes rolling and then kolla-kubernetes rolling in one day thanks to all the work done in that doc review | 00:48 |
SamYaple | sbezverk: i said master! | 00:48 |
SamYaple | sbezverk: newton doesnt have a new version of libvirt-bin :P | 00:48 |
sdake | eanylin my brain can only take so much success in one day :) | 00:48 |
*** tovin07_ has joined #openstack-kolla | 00:48 | |
SamYaple | sbezverk: i can walk you though how to add it i suppose... but its going to probably be a pain | 00:48 |
eanylin | sdake: http://paste.openstack.org/show/604398/ | 00:48 |
sbezverk | SamYaple: ok let me re-try, I tried master 20 minutes ago and it still got 1.3.1 | 00:49 |
eanylin | haha, I know. I was in that state 2 days ago :) | 00:49 |
sdake | eanylin kubectl get pods -n kolla -> paste | 00:49 |
eanylin | its in the same paste | 00:49 |
SamYaple | sbezverk: https://github.com/openstack/kolla/blob/master/docker/base/sources.list.ubuntu#L11 | 00:49 |
SamYaple | sbezverk: http://ubuntu-cloud.archive.canonical.com/ubuntu/dists/xenial-updates/ocata/main/binary-amd64/Packages | 00:49 |
SamYaple | sbezverk: search libvirt-bin on the second link | 00:49 |
sdake | eanylin are you anthony? | 00:49 |
eanylin | yah | 00:49 |
SamYaple | sbezverk: with those you should be pulling in 2.5.0 libvirt-bin, unless some apt pinning breaks it | 00:50 |
*** tonanhngo has quit IRC | 00:50 | |
sdake | eanylin nah - you did kube-system bck to back in that paste | 00:50 |
sdake | eanylin we have you to thank for figuring all this stuff out :) | 00:50 |
eanylin | sdake: No problem :) | 00:50 |
eanylin | sdake: I did another paste, http://paste.openstack.org/show/604399/ | 00:51 |
eanylin | sdake: Took a long time to figure out though, I was looking at the config used for the gate test, should have looked at that long time ago... | 00:52 |
*** tovin07 has joined #openstack-kolla | 00:52 | |
*** esharao has quit IRC | 00:54 | |
duonghq | I got TSP :D | 00:55 |
sbezverk | SamYaple: got it thanks | 00:58 |
sdake | eanylin did you run init-runonce? | 01:01 |
*** yangyapeng has joined #openstack-kolla | 01:02 | |
*** cuongnv has joined #openstack-kolla | 01:04 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add vSphere support to nova_compute https://review.openstack.org/449529 | 01:08 |
spsurya | duonghq: grats, you earned it | 01:09 |
duonghq | thank spsurya | 01:10 |
sdake | duonghq we pretty much ave kolla-kubernetes rolling if you want to give it a spin | 01:12 |
sdake | duonghq i know its been a long road to get to a deployment | 01:12 |
duonghq | sdake, roger, currently, which tool you uses for bring k8s cluster? | 01:12 |
sdake | duonghq kubeadm | 01:13 |
sdake | duonghq which I'd recommend you use if you are up to it | 01:13 |
sdake | duonghq i dont know how to use other tools and it took the community a week to sort out deploying with kubeadm | 01:13 |
duonghq | sdake, hmm, I'm using with, but something wrong with the network | 01:13 |
sdake | duonghq with? | 01:13 |
duonghq | pods cannot communicate with others | 01:14 |
sdake | duonghq if you ahe an hour or so i can walk you through the setup | 01:14 |
duonghq | I'm reading your ps | 01:15 |
sdake | duonghq read the comments too | 01:16 |
sdake | duonghq hwoever there is a key peice missing and that is setting up kubernetes | 01:16 |
duonghq | sdake, ya, sometime painful | 01:18 |
sdake | duonghq are you setting up in a vm or bare metal? | 01:19 |
duonghq | sdake, I'm using VMs | 01:19 |
sdake | duonghq ok - well if your up to get going let me knwo | 01:20 |
sdake | if not, i'll fix the docs up s othey are as correct as I can make them | 01:20 |
*** caoyuan_ has joined #openstack-kolla | 01:22 | |
*** signed8b_ has joined #openstack-kolla | 01:24 | |
*** qiliang27 has quit IRC | 01:25 | |
*** caoyuan has quit IRC | 01:25 | |
*** ipsecguy has quit IRC | 01:25 | |
*** qiliang27 has joined #openstack-kolla | 01:26 | |
*** ipsecguy has joined #openstack-kolla | 01:27 | |
*** signed8bit has quit IRC | 01:27 | |
*** MarginHu has joined #openstack-kolla | 01:27 | |
*** tovin07 has quit IRC | 01:28 | |
*** tovin07 has joined #openstack-kolla | 01:29 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add vSphere support to nova_compute https://review.openstack.org/449529 | 01:30 |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add VMware NSXV support to neutron https://review.openstack.org/450092 | 01:31 |
duonghq | sdake, in multi node cluster, how do you label node? | 01:32 |
masber | hi, for kolla-ansible, both neutron_external_interface and network_interface needs to have access to the internet right? | 01:32 |
*** unicell has quit IRC | 01:33 | |
sdake | duonghq i did just one node | 01:33 |
sdake | duonghq alhtough labeling multiple nodes is jus tlike labeling one node | 01:33 |
duonghq | how many node should be controller/compute? do you have any suggest? | 01:34 |
*** zhubingbing has joined #openstack-kolla | 01:34 | |
*** rhallisey has quit IRC | 01:34 | |
openstackgerrit | Lingyong Xu proposed openstack/kolla-kubernetes master: Drop py34 target in tox.ini https://review.openstack.org/449898 | 01:36 |
*** signed8b_ is now known as signed8bit_Zzz | 01:38 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Add Warning https://review.openstack.org/450498 | 01:39 |
eanylin | sdake: Yeah, so the neutron and keystone and all that is working | 01:39 |
openstackgerrit | shaofeng cheng proposed openstack/kolla-ansible master: Fix secure_proxy_ssl_header option https://review.openstack.org/449859 | 01:40 |
eanylin | sdake: I suspect we are missing the micro services for placement endpoint and service create | 01:40 |
eanylin | the cell0 database should not be empty | 01:40 |
sdake | eanylin entirel ypossible | 01:45 |
*** dixiaoli has joined #openstack-kolla | 01:46 | |
sdake | eanylin here is my start.sh script: http://paste.openstack.org/show/604401 | 01:47 |
eanylin | ok | 01:47 |
eanylin | I think mine is probably similar | 01:48 |
eanylin | you see any tables generated for nova_cell0? | 01:49 |
eanylin | if you go into your mariadb> | 01:49 |
eanylin | ? | 01:49 |
duonghq | sdake, around? | 01:49 |
sdake | duonghq yes | 01:49 |
*** signed8bit_Zzz is now known as signed8b_ | 01:50 | |
duonghq | I get error: KollaDirNotFoundException: Unable to locate None directory in search_paths | 01:50 |
sdake | duonghq you needt o cherrypick a patch and git commit it | 01:50 |
duonghq | I cherrypick zhu's ps and resolved conflict | 01:51 |
sdake | duonghq run git log and verify the patch is applied | 01:51 |
duonghq | Fix problem with kollakube res create configmap | 01:51 |
duonghq | - | 01:51 |
duonghq | yes | 01:51 |
sdake | duonghq here is my patch - moemnt | 01:52 |
duonghq | which dir are you clone kolla-* into? | 01:52 |
duonghq | I fixed it once | 01:52 |
sdake | i did a mkdir kolla-bringup | 01:52 |
sdake | cd kolla-bringup | 01:52 |
sdake | git clone both repos | 01:52 |
sdake | pip uninstalled all kolla related things | 01:52 |
*** zhurong has joined #openstack-kolla | 01:52 | |
sdake | then pip install kolla-ansible/ kolla-kubernetes/ | 01:52 |
sdake | if you leave the tailing / off, bad things happen :) | 01:52 |
duonghq | sdake, are you mentioning this? https://review.openstack.org/#/c/439740 | 01:53 |
sdake | after applying the patch you have to reinstall kolla-ansible | 01:53 |
sdake | duonghq thats the one that needs to be added | 01:53 |
duonghq | this is kolla-k8s one? | 01:53 |
sdake | there is also one that needs to be added ofr kolla-ansible | 01:53 |
sdake | perhaps i should modify the documentation to be more clear | 01:54 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Unpin Canal https://review.openstack.org/450423 | 01:54 |
kfox1111 | they may have identified the fix for canal. | 01:55 |
sdake | kfox1111 do you have a link? | 01:55 |
kfox1111 | 2 lines up. | 01:55 |
sdake | what does that do? | 01:56 |
duonghq | sdake, your pip installs packages to /usr/share? | 01:56 |
sdake | duonghq indeed | 01:56 |
kfox1111 | adds an env var. | 01:56 |
kfox1111 | if it works, I think they will submit it upstream. | 01:56 |
kfox1111 | just testing for now. | 01:56 |
duonghq | hmm, in my system, it installs to dist-package | 01:57 |
kfox1111 | duonghq: best to use venv's. | 01:58 |
kfox1111 | then you dont risk getting your base system all junked up. | 01:58 |
sdake | kfox1111 he is in a vm | 01:58 |
sdake | duonghq some files get instaleld in /usr/share some files in /usr/lib/python2.7/site-packages | 01:58 |
sdake | duonghq depends on the file | 01:58 |
duonghq | hmm, ok, it my system, it's /usr/local/share | 01:59 |
sdake | duonghq on ubuntu? | 01:59 |
duonghq | sdake, yes, VMs is ubuntu | 01:59 |
sdake | duonghq i see | 01:59 |
sdake | duonghq docs are designed around centos at present | 01:59 |
sdake | duonghq there isn't a whole lot of delta | 01:59 |
duonghq | either ubuntu or venvs, the path is differ from docs | 01:59 |
duonghq | sdake, roger | 01:59 |
sdake | you have found one of them - where share is located | 02:00 |
duonghq | it's not a big deal | 02:00 |
kfox1111 | sdake, less important then, but still a good practice to get into. | 02:00 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Fixing ironic deployment failure on Ubuntu https://review.openstack.org/449803 | 02:00 |
kfox1111 | cause you also never know who else is watching. | 02:00 |
sdake | kfox1111 everyone is watching ;) | 02:00 |
kfox1111 | exactly. :) | 02:01 |
kfox1111 | and they might not know the finer points of thowaway vm's vs bare metal for python deployment. | 02:01 |
kfox1111 | virtualevs are safe either way. | 02:01 |
sdake | virtualenvs provide no protection for kubernetes | 02:01 |
sdake | kubernetes litters the system with all kinds of crap | 02:02 |
kfox1111 | best to minimize the litter. | 02:02 |
duonghq | the interface is set in globals.yml? | 02:04 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Helm Native Values. https://review.openstack.org/449826 | 02:08 |
*** signed8b_ is now known as signed8bit_Zzz | 02:08 | |
*** dixiaoli has left #openstack-kolla | 02:11 | |
*** dixiaoli has joined #openstack-kolla | 02:11 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Helm Native Values. https://review.openstack.org/449826 | 02:12 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Helm Native Values. https://review.openstack.org/449826 | 02:13 |
duonghq | sdake, what is dns name if I used VMs? iirc, it's the IP of one of kubeapi? | 02:18 |
sdake | duonghq working on doc update - let me finish that, then you can run through the instrucitons again? | 02:19 |
*** eanylin has quit IRC | 02:19 | |
duonghq | sure | 02:19 |
duonghq | not know you're writting that | 02:19 |
*** dixiaoli has quit IRC | 02:23 | |
*** dixiaoli has joined #openstack-kolla | 02:25 | |
*** signed8bit_Zzz is now known as signed8b_ | 02:33 | |
*** eanylin has joined #openstack-kolla | 02:34 | |
zhubingbing | hi guys | 02:51 |
*** MarginHu has quit IRC | 02:53 | |
duonghq | sdake, 2 VMs out of 3 crashed by unknown reason :( | 02:53 |
duonghq | zhubingbing, hi | 02:53 |
*** duonghq has quit IRC | 02:53 | |
sdake | sup zhubingbing | 02:54 |
sdake | i got kubernetes deployed | 02:54 |
sdake | and most of kolla-kubernetes as well | 02:54 |
sdake | working on docs now | 02:54 |
zhubingbing | ji | 02:54 |
zhubingbing | ji | 02:54 |
sdake | zhubingbing we got some ansibles for you to do | 02:54 |
zhubingbing | ok | 02:54 |
sdake | zhubingbing if your up for it | 02:54 |
zhubingbing | np | 02:54 |
zhubingbing | np | 02:54 |
sdake | we are ejecting kolla-ansible as a dep from kolla-kubernetes | 02:55 |
sdake | we are doing that by copying only what is needed from kolla-ansible into kolla-kubernetes | 02:55 |
sdake | inc0 has started that work | 02:55 |
zhubingbing | i look inc0‘s commiit | 02:55 |
*** schwicht has quit IRC | 02:55 | |
sdake | zhubingbing can you make a blueprint for it | 02:55 |
sdake | i feel that is blueprint worhty work | 02:55 |
zhubingbing | all kolla-ansible project? | 02:55 |
zhubingbing | ok i know | 02:56 |
sdake | no jsut genconfig | 02:56 |
zhubingbing | emm | 02:56 |
zhubingbing | sure | 02:57 |
sdake | i think inc0 is in training for the next 3 or 4 days or something | 02:57 |
sdake | zhubingbing create a blueprint after looking at kolla-kubernets/ansible dir | 02:57 |
zhubingbing | i fixing it | 02:57 |
sdake | zhubingbing and i'll help you sort out the blueprint work items | 02:57 |
zhubingbing | hmm perfect | 02:57 |
kolla-slack | <inc0> You're writing BP for ansible config in k8s? | 03:00 |
*** duonghq has joined #openstack-kolla | 03:00 | |
sdake | inc0 feel we should | 03:00 |
sdake | inc0 as well as release notes | 03:00 |
* duonghq back | 03:00 | |
sdake | inc0 spread the work around | 03:00 |
sdake | inc0 i thought youwere out - but if your in - perhaps you can help zhubingbing with spreading hte work | 03:01 |
sdake | sup duonghq | 03:01 |
zhubingbing | https://blueprints.launchpad.net/kolla-kubernetes/+spec/move-config-to-kolla-k8s | 03:01 |
kolla-slack | <inc0> Zhubingbing if your want to help, check out my glance patch | 03:01 |
sdake | inc0 got a link? | 03:01 |
zhubingbing | yesterday i had look it | 03:01 |
kolla-slack | <inc0> I'm writing from phone | 03:02 |
zhubingbing | I've seen before | 03:02 |
zhubingbing | https://review.openstack.org/#/c/449755/ | 03:02 |
zhubingbing | sdake | 03:02 |
zhubingbing | sdake https://blueprints.launchpad.net/kolla-kubernetes/+spec/move-config-to-kolla-k8s can u help me sort out the blueprint work items | 03:03 |
zhubingbing | ;) | 03:03 |
kolla-slack | <inc0> Zhubingbing take cinder if you want | 03:05 |
zhubingbing | hmm | 03:05 |
zhubingbing | ok | 03:05 |
sdake | inc0 are you working on nova atm? | 03:06 |
kolla-slack | <inc0> Nova in progress, ks glance and neutron done | 03:06 |
*** schwicht has joined #openstack-kolla | 03:08 | |
sdake | merged or not merged? | 03:09 |
sdake | if not merged they are in progress :) | 03:09 |
kolla-slack | <inc0> Merged | 03:09 |
sdake | please self assign: https://blueprints.launchpad.net/kolla-kubernetes/+spec/move-config-to-kolla-k8s | 03:12 |
zhubingbing | OK | 03:12 |
zhubingbing | glance have merged | 03:13 |
duonghq | found the issue: my HDD is running out of space :( | 03:14 |
sdake | duonghq that will do it | 03:18 |
eanylin | sdake: Do let me know if you guys need my help in any way :) | 03:19 |
sdake | eanylin YES | 03:20 |
sdake | eanylin docs, code, maintaining launchpad, there are many things to be done | 03:20 |
eanylin | sdake: Ok, agree | 03:20 |
eanylin | sdake: I can help out :) | 03:21 |
*** fooliouno has joined #openstack-kolla | 03:22 | |
kolla-slack | <vhosakot> @inc0 I’d like to help contributing to https://blueprints.launchpad.net/kolla-kubernetes/+spec/move-config-to-kolla-k8s too… helps genconfig… I’ll refer https://review.openstack.org/#/c/449755/ | 03:25 |
kolla-slack | <vhosakot> @inc0 is @zhubingbing doing cinder, I’ll take neutron and horizon.. | 03:25 |
masber | hi, I have a vm with 2 virtual interfaces which are connected to same VLAN and hence to the same physical nic on the hypervisor, I tried to deploy openstack through kolla-ansible but I can't access the internet from the router namespace. Am I doing something stupid by using same VLAN for both kolla interfaces (neutron_external_interface and network_interface) even though I am using different virtual nics? | 03:26 |
*** g3ek has quit IRC | 03:27 | |
sdake | eanylin you can help with this one: https://blueprints.launchpad.net/kolla-kubernetes/+spec/deploy-guide | 03:28 |
eanylin | sdake: ok | 03:28 |
sdake | vhosakot please self assign in the blueprint | 03:28 |
sdake | eanylin are you familiar with the gerit workflow of submitting patches? | 03:28 |
kolla-slack | <vhosakot> @sdake ok, will do, thanks.. | 03:29 |
eanylin | I just started last week actually | 03:29 |
sdake | eanylin cool | 03:29 |
sdake | eanylin whats your tz | 03:29 |
kolla-slack | <vhosakot> @sdake so, did you get your single-VM up and all OpenStack pods up (Running state without restarting) | 03:29 |
eanylin | what is tz? sorry | 03:29 |
sdake | eanylin timezone | 03:29 |
sdake | eanylin i am gmt-7 (MST) | 03:29 |
eanylin | oh, I am in Central | 03:29 |
sdake | vhosakot yes i did, although vm got stuck in the "BULIDING" state | 03:30 |
sdake | vhosakot i deployed on bare metal - docs hopefully coming in next hour or so | 03:30 |
kolla-slack | <vhosakot> @sdake I’m planning to write a python script to re-use the gate scripts to automatically deploy dev-env. I did something like that for kolla almost a year ago, and it worked perfectly and I was able to re-use the gate scripts to deploy kolla’s dev env. | 03:30 |
sdake | there are alot of comments to go thorugh | 03:30 |
*** jrobinson_ has joined #openstack-kolla | 03:30 | |
sdake | vhosakot we already have something like that claled 'devenv' | 03:30 |
sdake | vhosakot here is our roadmap to 1.0.0: | 03:31 |
kolla-slack | <vhosakot> @sdake nice! you reached the point to boot a nova VM… great | 03:31 |
*** jrobinson has quit IRC | 03:31 | |
sdake | https://etherpad.openstack.org/p/kolla-pike-ptg-k8s-release-roadmap | 03:32 |
sdake | vhosakot here is our blueprints for 0.6.0: https://launchpad.net/kolla-kubernetes/+milestone/0.6.0 | 03:32 |
kolla-slack | <vhosakot> cool, thanks for the info @sdake | 03:33 |
kolla-slack | <inc0> Vhosakot neutron is done | 03:33 |
kolla-slack | <inc0> Horizon is for the taking | 03:33 |
*** jrobinson_ is now known as jrobinson | 03:33 | |
kolla-slack | <vhosakot> @inc0 cool, I’ll pick horizon then.. | 03:33 |
kolla-slack | <inc0> Thanks | 03:33 |
kolla-slack | <vhosakot> cool | 03:34 |
kolla-slack | <vhosakot> @sdake do you know if https://github.com/openstack/kolla-kubernetes/blob/master/tools/setup_dev_env.sh is used and works OTB? | 03:34 |
sdake | vhosakot it worked at one point | 03:34 |
sdake | does it work with 4.0.0 images? unkkwon | 03:34 |
sdake | bmace may be able to answer - he did that work | 03:34 |
kolla-slack | <vhosakot> cool | 03:34 |
kolla-slack | <vhosakot> @sdake I’ll give my comments for https://review.openstack.org/#/c/447356/. | 03:35 |
*** duonghq has left #openstack-kolla | 03:35 | |
kolla-slack | <vhosakot> if the user retries the steps, I think there must be a section to “cleanup the dev env”.. I’ll add this comment. | 03:36 |
kolla-slack | <vhosakot> without proper cleanup, the user needs to re-image the box to retry the steps in the review. | 03:36 |
*** g3ek has joined #openstack-kolla | 03:36 | |
eanylin | vhosakot: Would 'kubeadm reset' and mariadb cleanup suffice? | 03:37 |
*** duonghq has joined #openstack-kolla | 03:37 | |
kolla-slack | <vhosakot> @eanylin, not really :slightly_smiling_face: | 03:38 |
eanylin | ok :) | 03:38 |
fooliouno | sdake: I think you discussed the cinder-volume-lvm issue with spsurya earlier today. I see that the cinder-volume-lvm pod failed for me on Ubuntu. | 03:38 |
*** vhosakot has joined #openstack-kolla | 03:38 | |
fooliouno | sdake: Is there a workaround for it | 03:39 |
spsurya | fooliouno: did i fixed somehting related to cinder-volume-lvm ? | 03:40 |
fooliouno | spsurya: Hi! Did you get it to work? | 03:41 |
fooliouno | spsurya: Mine fails when I try helm install | 03:41 |
spsurya | fooliouno: i didnt check | 03:41 |
spsurya | fooliouno: the issue i faced during kolla-k8s setup was related to vagrant, vbox and libvert | 03:46 |
spsurya | fooliouno: not with the cinder-volume-lvm | 03:46 |
fooliouno | spsurya: ok. I mistakenly assumed that you encountered the same issue. | 03:47 |
spsurya | fooliouno: np | 03:47 |
openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Update the references link for some project https://review.openstack.org/450552 | 03:47 |
*** tovin07 has quit IRC | 03:53 | |
*** zhurong has quit IRC | 03:57 | |
*** caoyuan has joined #openstack-kolla | 03:59 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move cinder config to kolla-k8s https://review.openstack.org/450556 | 04:00 |
*** caoyuan_ has quit IRC | 04:01 | |
*** signed8b_ is now known as signed8bit_Zzz | 04:06 | |
*** signed8bit_Zzz is now known as signed8b_ | 04:07 | |
*** vhosakot has quit IRC | 04:08 | |
sdake | fooliouno i think the answer to that problem is to install iscsi | 04:08 |
sdake | fooliouno mine also fails (cinder-volume-lvm) - I am working on getting hte docs into a shape where I can rerun through them to debug the last parts | 04:09 |
sdake | there are like 50 comments on that review | 04:09 |
sdake | and i want ot unify them - i am nearing completion | 04:09 |
fooliouno | sdake: I cam across this page that states that additional steps are needed for Ubuntu: https://docs.openstack.org/developer/kolla-ansible/cinder-guide.html | 04:09 |
bmace | vhosakot: i haven't run my devenv scripts in a few weeks but last time i did they totally worked :) | 04:25 |
*** dixiaoli has quit IRC | 04:26 | |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: WIP: Deployment Guide Documentation https://review.openstack.org/447356 | 04:32 |
eanylin | fooliouno: Try labeling your node with kolla_storage as well? Not sure if that will help but I labeled my node with kolla_compute, kolla_controller and kolla_storage | 04:32 |
fooliouno | eanylin: Thanks. Let me try that. | 04:33 |
eanylin | fooliouno: No problem | 04:33 |
sbezverk | sdake: most of services you listed in service clean up blueprint are completed.. | 04:34 |
*** signed8b_ is now known as signed8bit_Zzz | 04:34 | |
*** lucasxu has quit IRC | 04:35 | |
*** signed8bit_Zzz is now known as signed8b_ | 04:36 | |
*** jrobinson has quit IRC | 04:36 | |
sdake | sbezverk cool - feel free to update - need to feed hte children ;) | 04:37 |
sdake | fooliouno eanylin try checking out that new review in a fresh env and reporting problems ;) | 04:37 |
fooliouno | sdake: will do. Thansk for updating it. | 04:37 |
fooliouno | sdake: as reported by duonghq, the paths are /usr/local/share on Ubuntu instead of /usr/share. I had to update pathfinder.py locally for it to work. Will add a comment. | 04:40 |
*** lucasxu has joined #openstack-kolla | 04:41 | |
eanylin | sdake: Reading | 04:42 |
eanylin | I still think that we are missing something for the placement keystone endpoints | 04:42 |
sdake | eanylin agreed | 04:42 |
eanylin | I think the guide for placement talks about that | 04:43 |
sdake | eanylin there were 27 comments outstanding in that last review - i think i addressed them all | 04:43 |
eanylin | yeah, think you did :) | 04:43 |
sdake | eanylin note that there are other microservices related to placement - some create endpoints and setup the db | 04:44 |
sdake | rather microcharts | 04:44 |
eanylin | sdake: I think so but it didnt really work for me I think | 04:44 |
eanylin | the endpoint one errors out | 04:44 |
eanylin | went into crash loop | 04:45 |
*** iceyao has joined #openstack-kolla | 04:45 | |
eanylin | for placement to be installed, I think we need to create placement service, user, endpoints, sync DB | 04:46 |
sdake | ya - they need to be syncrhonized | 04:47 |
sdake | hence they need to be in the nova service | 04:47 |
sdake | which means (I believe) they need some form of conditional | 04:48 |
sdake | the reason they work in the gate is the gate syncronizes every step of the way | 04:48 |
eanylin | most probably the case | 04:48 |
eanylin | I was reading on forum, it talks about config for placement_database in the nova.conf as well | 04:48 |
sdake | sbezverk do you have any suggestions for the placement api - seems like a real blocker | 04:48 |
eanylin | not sure if that is needed | 04:49 |
sdake | eanylin we do that with kolla genconfig | 04:49 |
eanylin | ok | 04:49 |
eanylin | but to enabling placement didnt seem to generate [placement_database] | 04:49 |
eanylin | or maybe I missed something in the process of doing so | 04:49 |
sbezverk | sdake: what do you mean blocker?? the gate is completely green | 04:49 |
sdake | eanylin you need a check | 04:50 |
eanylin | ok | 04:50 |
sdake | sbezverk the gate isn't representative of a real deployment | 04:50 |
sdake | sbezverk tell me how precisely you deploy the placement api without wait_pods 600 between each step? | 04:50 |
sbezverk | sdake: you should check the code sometime, wait 600 is long gone | 04:51 |
sdake | sbezverk i know its gone | 04:51 |
sdake | sbezverk you do understand my point do you not? | 04:51 |
*** yangyapeng has quit IRC | 04:51 | |
sdake | sncrhonization should happen in the service level charts | 04:51 |
*** lucasxu has quit IRC | 04:51 | |
sbezverk | sdake: frankly no. I have cluster with placement workign fine, gate also working fine.. | 04:51 |
sdake | that is the whole point of service level charts | 04:51 |
*** yangyapeng has joined #openstack-kolla | 04:52 | |
sdake | sbezverk you just run helm install on all the placement microcharts al at once and everything just comes up? | 04:52 |
sbezverk | yes | 04:52 |
*** signed8b_ has quit IRC | 04:52 | |
sdake | eanylin reports that doesn't work for him | 04:53 |
sbezverk | sdake: I do exactly what deploymnet script at the agte does | 04:53 |
sbezverk | sdake: then he has a misconfiguration | 04:53 |
*** signed8bit has joined #openstack-kolla | 04:54 | |
*** jrobinson has joined #openstack-kolla | 04:54 | |
sdake | placement api comes up for me, however, my vms stuck in build state forever | 04:54 |
*** skramaja has joined #openstack-kolla | 04:54 | |
sdake | sbezverk can you please advise best course of action to get placement apis into the service level charts | 04:55 |
eanylin | sbezverk, any particular area to look at for the configurations? | 04:55 |
sdake | placement apis/placmenet api microcharts that is | 04:56 |
*** zhurong has joined #openstack-kolla | 04:56 | |
sbezverk | eanylin: not from the top of my head.. but I strongly suggest to analize the gate job deployig it | 04:56 |
*** yangyapeng has quit IRC | 04:56 | |
fooliouno | enaylin: the dns_name in cloud.yaml ... what should it point to? the kube-dns service address? | 04:56 |
eanylin | sbezvek: ok | 04:56 |
sbezverk | guys kube-dns was intriduced just for keystone | 04:57 |
sdake | fooliouno i cna't get an answer on what dns_name shoudl be set to, although kfox1111 has indicated it hsould be set to your physical nic ip | 04:57 |
fooliouno | sdake: Thanks. Will try that | 04:57 |
sbezverk | when you have a public dns name you want to use for keystone public end point you need to use it | 04:57 |
sbezverk | for other scenarios it is not useful | 04:57 |
eanylin | fooliouno: I used eth0 IP, which is my management/OAM NIC | 04:58 |
fooliouno | eanylin: thank you | 04:58 |
eanylin | np | 04:58 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Fixing ironic deployment failure on Ubuntu https://review.openstack.org/449803 | 04:58 |
fooliouno | sdake: globals.yml has a hardcoded keystone-public address | 05:01 |
fooliouno | keystone_public_url: "http://192.168.80.232:5000/v3" | 05:01 |
sdake | fooliouno no idea what that hsould be | 05:02 |
sdake | it probably shoudln't be that though | 05:02 |
eanylin | fooliouno: I hit the same problem as you with the LVM after I scratched my environment. I updated the cloud.yaml with LVM info, seems to have got past the error | 05:02 |
fooliouno | I was wondering if keystone-public:5000/v3 would work (like the rest of the endpoints) | 05:02 |
eanylin | fooliouno: http://paste.openstack.org/show/604416/ | 05:02 |
eanylin | fooliouno: You can set it to keystone-public | 05:03 |
eanylin | it should pick up the VIP | 05:03 |
sdake | eanylin line 25 is unnecessary | 05:03 |
eanylin | ok | 05:03 |
sdake | line 34 will override line 25 always | 05:03 |
eanylin | we can remove that line | 05:03 |
sdake | actually 25 will override line 34 | 05:03 |
fooliouno | eanylin: I'll try the keystone-public as well as the lvm update in cloud.yml. | 05:04 |
sdake | but since they are the same, they are redundant | 05:04 |
eanylin | fooliouno: Cool | 05:04 |
eanylin | sdake: Agree | 05:04 |
*** shashank_t_ has joined #openstack-kolla | 05:04 | |
sdake | but you need one in all: or newton is used iirc | 05:04 |
sdake | globals.allt hat is | 05:04 |
eanylin | sdake: Oh, ok | 05:05 |
*** shashank_t_ has quit IRC | 05:08 | |
*** iceyao has quit IRC | 05:09 | |
*** lucasxu has joined #openstack-kolla | 05:12 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Fixing ironic deployment failure on Ubuntu https://review.openstack.org/449803 | 05:13 |
sdake | Jeffrey4l you around? | 05:16 |
*** lamt has joined #openstack-kolla | 05:19 | |
Jeffrey4l | sdake, yep. | 05:21 |
sdake | Jeffrey4l sbezverk had question re the rationale behind 1.3 of libvirt in ubuntu containers | 05:22 |
Jeffrey4l | sdake, it is resolved already sbezverk | 05:22 |
sdake | Jeffrey4l i recall there was some last minute changes | 05:22 |
sdake | Jeffrey4l apparently it breaks kolla-kubernetes | 05:22 |
Jeffrey4l | sbezverk, mixed up something. container is OK. | 05:22 |
sdake | Jeffrey4l or thats the speculation - could you explain why | 05:22 |
sdake | Jeffrey4l thanks | 05:23 |
Jeffrey4l | np | 05:24 |
*** dixiaoli has joined #openstack-kolla | 05:26 | |
*** dixiaoli has quit IRC | 05:32 | |
*** dixiaoli has joined #openstack-kolla | 05:32 | |
*** targon has joined #openstack-kolla | 05:36 | |
*** tvignaud has quit IRC | 05:40 | |
*** lucasxu has quit IRC | 05:42 | |
openstackgerrit | Gao Zexu proposed openstack/kolla-ansible master: Fix some blank space issues https://review.openstack.org/447348 | 05:45 |
*** tvignaud has joined #openstack-kolla | 05:53 | |
*** imcsk8_ has joined #openstack-kolla | 06:01 | |
*** imcsk8 has quit IRC | 06:01 | |
*** lamt has quit IRC | 06:01 | |
*** lamt has joined #openstack-kolla | 06:05 | |
*** jascott1_ has quit IRC | 06:08 | |
*** caowei has joined #openstack-kolla | 06:08 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add vSphere support to nova_compute https://review.openstack.org/449529 | 06:09 |
*** satyar has joined #openstack-kolla | 06:09 | |
*** jascott1_ has joined #openstack-kolla | 06:10 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add vSphere support to nova_compute https://review.openstack.org/449529 | 06:11 |
*** kfox1111 has quit IRC | 06:12 | |
*** jascott1_ has quit IRC | 06:15 | |
sdake | qwang you about? | 06:18 |
*** lamt has quit IRC | 06:19 | |
*** dixiaoli has quit IRC | 06:22 | |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Deployment Guide Documentation https://review.openstack.org/447356 | 06:23 |
*** dixiaoli has joined #openstack-kolla | 06:31 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move rabbitmq config to kolla-k8s https://review.openstack.org/450581 | 06:32 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move cinder config to kolla-k8s https://review.openstack.org/450556 | 06:36 |
*** bjolo has joined #openstack-kolla | 06:36 | |
*** zhuzeyu has joined #openstack-kolla | 06:40 | |
*** manheim has joined #openstack-kolla | 06:41 | |
*** lamt has joined #openstack-kolla | 06:42 | |
*** manheim has quit IRC | 06:43 | |
*** manheim has joined #openstack-kolla | 06:43 | |
*** lamt has quit IRC | 06:46 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move memcached config to kolla-k8s https://review.openstack.org/450556 | 06:48 |
*** sambetts|afk has quit IRC | 06:48 | |
*** lamt has joined #openstack-kolla | 06:51 | |
*** tovin07 has joined #openstack-kolla | 06:52 | |
*** sambetts_ has joined #openstack-kolla | 06:52 | |
*** haplo37 has quit IRC | 06:55 | |
*** matrohon has joined #openstack-kolla | 07:00 | |
*** haplo37 has joined #openstack-kolla | 07:04 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move heat config to kolla-k8s https://review.openstack.org/450589 | 07:07 |
hrw | is anyone able to build bifrost images? centos, debian, ubuntu ones failed for me (both binary and source) | 07:07 |
*** lamt has quit IRC | 07:10 | |
openstackgerrit | Bertrand Lallau proposed openstack/kolla-ansible master: Remove ml2_conf.ini file from some Neutron agents https://review.openstack.org/447992 | 07:12 |
satyar | Hi pbourke | 07:12 |
satyar | pbourke: how are you | 07:12 |
*** pcaruana has joined #openstack-kolla | 07:13 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move heat config to kolla-k8s https://review.openstack.org/450589 | 07:15 |
bjolo | morning | 07:17 |
duonghq | morning bjolo | 07:17 |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add vSphere support to nova_compute https://review.openstack.org/449529 | 07:18 |
*** jrobinson has quit IRC | 07:21 | |
*** dixiaoli has quit IRC | 07:22 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move cinder config to kolla-k8s https://review.openstack.org/450556 | 07:24 |
*** dixiaoli has joined #openstack-kolla | 07:26 | |
*** jaosorior has joined #openstack-kolla | 07:29 | |
*** bdek has joined #openstack-kolla | 07:30 | |
*** hrito has joined #openstack-kolla | 07:33 | |
*** Mech422 has quit IRC | 07:33 | |
*** Mech422 has joined #openstack-kolla | 07:33 | |
*** hrito has quit IRC | 07:35 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move memcached config to kolla-k8s https://review.openstack.org/450556 | 07:37 |
*** dixiaoli has quit IRC | 07:38 | |
*** manheim has quit IRC | 07:39 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Split globals.yml into mutiple part by service https://review.openstack.org/450106 | 07:40 |
*** manheim has joined #openstack-kolla | 07:41 | |
*** dixiaoli has joined #openstack-kolla | 07:44 | |
*** egonzalez has joined #openstack-kolla | 07:50 | |
*** jmccarthy has joined #openstack-kolla | 07:51 | |
hrw | time for morning's refresh of patchset | 07:51 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move cinder config to kolla-k8s https://review.openstack.org/450556 | 07:52 |
*** Mech422 has quit IRC | 07:54 | |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: iscsid: enable Debian support https://review.openstack.org/450612 | 07:55 |
*** Mech422 has joined #openstack-kolla | 07:55 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add vSphere support to nova_compute https://review.openstack.org/449529 | 07:55 |
*** furlongm has joined #openstack-kolla | 07:57 | |
hrw | I hate adding reviewers in gerrit | 07:58 |
hrw | normally it is a matter of typing 'firstname lastname' but with openstack devlist it just does not work ;( | 07:58 |
hrw | do you have a better way for it? | 08:03 |
*** iceyao has joined #openstack-kolla | 08:03 | |
jaosorior | egonzalez: thanks for the review | 08:05 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move memcached config to kolla-k8s https://review.openstack.org/450620 | 08:19 |
*** mgoddard has joined #openstack-kolla | 08:20 | |
*** nathharp has joined #openstack-kolla | 08:23 | |
zhubingbing | egonzalez around ? | 08:24 |
egonzalez | zhubingbing, yep | 08:24 |
*** dgonzalez has quit IRC | 08:24 | |
nathharp | Hi all - hoping for some guidance. I have a multi node deployment via kolla-ansible. Everything deploys properly. However, attempting to assign a floating IP gives an error via horizon: | 08:29 |
nathharp | Error: Request Failed: internal server error while processing your request. Neutron server returns request_ids: ['req-b1a2a917-c163-4dbb-b298-61ed592957b3'] | 08:29 |
*** iceyao has quit IRC | 08:29 | |
*** dgonzalez has joined #openstack-kolla | 08:30 | |
nathharp | Neutron server log gives a backtrace finishing with: neutron.api.v2.resource AttributeError: 'NoneType' object has no attribute 'get' | 08:30 |
nathharp | this is using stable/ocata and 4.0.0 dockerhub images | 08:31 |
duonghq | zhubingbing, around? | 08:32 |
zhubingbing | hi | 08:32 |
zhubingbing | i am hear | 08:32 |
duonghq | can you kubectl describe po mariadb-0 -n kolla | 08:33 |
duonghq | and show me the mount section | 08:33 |
zhubingbing | i destroy my dev enviroment | 08:33 |
zhubingbing | ;( | 08:33 |
duonghq | for unknown reason, in my system, resolve-conf doesn't mount in keystone-create-db job | 08:33 |
duonghq | do you get it before? | 08:34 |
zhubingbing | yeah | 08:34 |
egonzalez | nathharp, can you share full trace of the issue and also check logs from neutron containers | 08:34 |
duonghq | zhubingbing, how do you solve the issue? | 08:34 |
zhubingbing | you must let k8s-dns is can resolve ip address | 08:35 |
duonghq | but the resolve.conf does not point to this cluster-ip | 08:36 |
*** iceyao has joined #openstack-kolla | 08:36 | |
nathharp | egonzalez - http://paste.openstack.org/show/604437/ - full trace | 08:36 |
nathharp | egonzalez - that’s the only error I found | 08:37 |
duonghq | zhubingbing, in other words, it got wrong dns server | 08:37 |
egonzalez | nathharp, any service enabled out of default services? | 08:37 |
zhubingbing | hmm can u pause your command | 08:37 |
nathharp | egonzalez - on the networking side - designate | 08:38 |
nathharp | egonzalez - plus heat, murano and external ceph | 08:38 |
duonghq | zhubingbing, sorry? | 08:40 |
duonghq | zhubingbing, I exec ping mariadb in memcached pod for test | 08:40 |
egonzalez | nathharp, have you followed designate guide? https://github.com/openstack/kolla-ansible/blob/master/doc/designate-guide.rst | 08:40 |
duonghq | also, keystone-createdb job cannot connect to mariadb | 08:40 |
*** iceyao has quit IRC | 08:41 | |
nathharp | egonzalez - aha, I’d not seen that. I’ve only recenly enabled designate at the request of one of my devs | 08:41 |
*** iceyao has joined #openstack-kolla | 08:42 | |
nathharp | egonzalez - first things first, I’ll disable and see if the problem goes away | 08:42 |
egonzalez | nathharp, designate is different in the deploy method, as for integrate with nova/neutron need a domain_id in their files configured, but until the service is fully operational cannot be created (having to do a reconfigure after deploy) | 08:43 |
*** targon has quit IRC | 08:46 | |
*** g3ek has quit IRC | 08:48 | |
*** haplo37 has quit IRC | 08:49 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Add vSphere support to nova_compute https://review.openstack.org/449529 | 08:49 |
*** haplo37 has joined #openstack-kolla | 08:50 | |
*** g3ek has joined #openstack-kolla | 08:51 | |
jaosorior | Bertrand Lallau are you around? | 08:52 |
*** pbandark has joined #openstack-kolla | 08:53 | |
*** tovin07 has quit IRC | 08:55 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Split globals.yml into mutiple part by service https://review.openstack.org/450106 | 08:56 |
*** gfidente has joined #openstack-kolla | 08:58 | |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Split globals.yml into mutiple part by service https://review.openstack.org/450106 | 09:04 |
*** iceyao has quit IRC | 09:09 | |
*** shashank_t_ has joined #openstack-kolla | 09:10 | |
openstackgerrit | Zeyu Zhu proposed openstack/kolla-kubernetes master: Remove useless moudles https://review.openstack.org/450662 | 09:10 |
*** bmace has quit IRC | 09:14 | |
openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible master: Modify the permission of directory https://review.openstack.org/448932 | 09:14 |
*** shashank_t_ has quit IRC | 09:14 | |
*** nathharp has quit IRC | 09:16 | |
*** Serlex has joined #openstack-kolla | 09:16 | |
manheim | docker no longer uses cgconfig.conf? cannot find it anywhere in a running env I have | 09:23 |
manheim | i want to see something related to cgroups | 09:24 |
*** bmace has joined #openstack-kolla | 09:26 | |
openstackgerrit | Zeyu Zhu proposed openstack/kolla-kubernetes master: Remove useless moudles https://review.openstack.org/450662 | 09:26 |
*** nathharp has joined #openstack-kolla | 09:26 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move cinder config to kolla-k8s https://review.openstack.org/450556 | 09:31 |
openstackgerrit | jimmygc proposed openstack/kolla-ansible master: Split globals.yml into mutiple part by service https://review.openstack.org/450106 | 09:47 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move rabbitmq config to kolla-k8s https://review.openstack.org/450581 | 09:48 |
*** shardy has joined #openstack-kolla | 09:51 | |
*** gfidente has quit IRC | 09:52 | |
*** bdek has quit IRC | 09:53 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move heat config to kolla-k8s https://review.openstack.org/450589 | 09:58 |
*** caoyuan has quit IRC | 09:59 | |
*** nhlfr has quit IRC | 10:04 | |
*** retr0h has quit IRC | 10:04 | |
*** vcn[m] has quit IRC | 10:04 | |
*** bachp has quit IRC | 10:04 | |
*** vcn[m] has joined #openstack-kolla | 10:08 | |
*** athomas has quit IRC | 10:10 | |
*** tovin07_ has quit IRC | 10:10 | |
*** athomas has joined #openstack-kolla | 10:11 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move ironic config to kolla-k8s https://review.openstack.org/450701 | 10:15 |
*** manheim has quit IRC | 10:15 | |
*** athomas has quit IRC | 10:16 | |
*** gfidente has joined #openstack-kolla | 10:19 | |
*** zhurong has quit IRC | 10:20 | |
*** nhlfr has joined #openstack-kolla | 10:22 | |
*** bachp has joined #openstack-kolla | 10:22 | |
*** retr0h has joined #openstack-kolla | 10:22 | |
*** pbourke has quit IRC | 10:25 | |
*** satyar has quit IRC | 10:26 | |
*** pbourke has joined #openstack-kolla | 10:27 | |
*** zhurong has joined #openstack-kolla | 10:28 | |
*** athomas has joined #openstack-kolla | 10:41 | |
*** shardy has quit IRC | 10:44 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move rabbitmq config to kolla-k8s https://review.openstack.org/450581 | 10:44 |
*** shardy has joined #openstack-kolla | 10:44 | |
*** manheim has joined #openstack-kolla | 10:48 | |
openstackgerrit | Zeyu Zhu proposed openstack/kolla-kubernetes master: Remove useless modules https://review.openstack.org/450662 | 10:49 |
openstackgerrit | Merged openstack/kolla master: Add ironic sudoers file to binary packages as well https://review.openstack.org/450274 | 10:50 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move ironic config to kolla-k8s https://review.openstack.org/450701 | 10:50 |
*** manheim has quit IRC | 10:53 | |
openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Remove deprecated options designate https://review.openstack.org/450715 | 10:55 |
*** dixiaoli has quit IRC | 10:56 | |
hrw | https://pastebin.com/RhaE2qDJ - is it worth submitting? Would get rid of that 'supposed to be debug' output at the end of kolla job | 11:01 |
*** manheim has joined #openstack-kolla | 11:02 | |
*** cuongnv has quit IRC | 11:02 | |
*** sambetts_ is now known as sambetts | 11:04 | |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: move dump of results into debug level https://review.openstack.org/450716 | 11:05 |
hrw | sent anyway | 11:05 |
*** manheim has quit IRC | 11:06 | |
*** manheim has joined #openstack-kolla | 11:07 | |
*** athomas has quit IRC | 11:09 | |
*** athomas has joined #openstack-kolla | 11:17 | |
*** furlongm has quit IRC | 11:19 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move horizon config to kolla-k8s https://review.openstack.org/450725 | 11:24 |
*** zhurong has quit IRC | 11:25 | |
*** furlongm has joined #openstack-kolla | 11:31 | |
*** caoyuan has joined #openstack-kolla | 11:32 | |
*** openstackgerrit has quit IRC | 11:33 | |
*** erlon has joined #openstack-kolla | 11:38 | |
*** zhubingbing has quit IRC | 11:38 | |
*** rwallner has joined #openstack-kolla | 11:45 | |
*** manheim has quit IRC | 11:50 | |
mliima | morning guys | 11:52 |
*** zhuzeyu has quit IRC | 11:53 | |
*** pbandark is now known as pbandark`brb | 11:57 | |
*** rhallisey has joined #openstack-kolla | 11:57 | |
*** manheim has joined #openstack-kolla | 11:57 | |
*** openstackgerrit has joined #openstack-kolla | 11:57 | |
openstackgerrit | Merged openstack/kolla master: Remove gnocchi-carbonara package https://review.openstack.org/450171 | 11:57 |
*** iceyao has joined #openstack-kolla | 11:58 | |
nathharp | egonzalez - quick update - Designate was the problem here. Disabled and the problem went away. onwards and upwards! | 12:01 |
nathharp | egonzalez - thanks | 12:02 |
*** caoyuan_ has joined #openstack-kolla | 12:03 | |
*** caoyuan has quit IRC | 12:06 | |
*** Pavo has joined #openstack-kolla | 12:07 | |
hrw | gnocchi in, yum change next | 12:11 |
*** harbie has quit IRC | 12:13 | |
*** harbie has joined #openstack-kolla | 12:15 | |
*** esharao has joined #openstack-kolla | 12:18 | |
*** esharao has quit IRC | 12:22 | |
*** eaguilar has joined #openstack-kolla | 12:23 | |
*** schwicht has quit IRC | 12:23 | |
*** dave-mccowan has joined #openstack-kolla | 12:25 | |
*** zhubingbing_ has joined #openstack-kolla | 12:26 | |
jaosorior | Any +As for this https://review.openstack.org/#/q/topic:bug/1675308+branch:master+status:open ? :D | 12:26 |
*** satyar has joined #openstack-kolla | 12:29 | |
rwallner | can I run enable_haproxy: "no" in multinode mode? I have a use case where hosts have a single lo interface for bgp routing at the host and I'd like to expose the control plane over a virtual network separately | 12:33 |
rwallner | morning btw :) | 12:33 |
Daviey | rwallner: I think you'd need to try it and see if it works :) .. I thought about using a HW loadbalancer, which would require the same thing.. but not tested | 12:34 |
rwallner | fair enough, currently working on a small test env. | 12:35 |
Daviey | jaosorior: Comment added to https://review.openstack.org/#/c/446911/9 | 12:38 |
jaosorior | Daviey: that file comes from the package that is just installed there. It's not something that folks had before. | 12:39 |
jaosorior | Daviey: There's details about that in the commit message | 12:39 |
Daviey | jaosorior: SSL worked before right? Doesn't this break it? | 12:39 |
jaosorior | Daviey: it didn't work | 12:40 |
Daviey | jaosorior: Are you sure? | 12:40 |
jaosorior | Daviey: the package just wasn't avialable, so one couldn't configure SSL at all in those containers | 12:40 |
jaosorior | Daviey: yes, I tried, that's why I submitted that | 12:40 |
*** eaguilar_ has joined #openstack-kolla | 12:40 | |
*** eaguila__ has joined #openstack-kolla | 12:41 | |
jaosorior | Daviey: I'm talking about SSL in the internal network. Regular SSL cases are not handled by these containers, those are handled by HAProxy which uses it's own OpenSSL-based backend | 12:41 |
Daviey | jaosorior: Ohhhh... this enables SSL on the backend... not just for haproxy ? | 12:42 |
jaosorior | Daviey: this doesn't touch haproxy | 12:42 |
Daviey | I SEE! Sorry, i am blind | 12:42 |
*** caowei has quit IRC | 12:42 | |
Daviey | jaosorior: Can you just summarise that on gerrit as it wasn't clear to me, and i'll +2/A | 12:42 |
jaosorior | Daviey: done | 12:43 |
sdake | morning | 12:43 |
*** pbourke_ has joined #openstack-kolla | 12:44 | |
*** eaguilar has quit IRC | 12:44 | |
*** zhubingbing_ has quit IRC | 12:44 | |
*** eaguilar_ has quit IRC | 12:45 | |
Daviey | jaosorior: Done | 12:45 |
jaosorior | Daviey: thanks | 12:45 |
*** rwallner has quit IRC | 12:50 | |
*** manheim has quit IRC | 12:53 | |
*** rwallner has joined #openstack-kolla | 12:54 | |
Daviey | rwallner: Hey... before you disappear - i'd love to hear the results of your test | 12:54 |
*** rhallisey has quit IRC | 12:54 | |
Daviey | Gah | 12:54 |
*** manheim has joined #openstack-kolla | 12:55 | |
Daviey | Oh.. different people :) | 12:55 |
rwallner | ill let you know :) | 12:55 |
*** schwicht has joined #openstack-kolla | 12:55 | |
*** sbezverk has quit IRC | 12:55 | |
rwallner | right now we have some hardware setup to do still, but im testing the theory out in a test environment buyt turning ha proxy /vrrp off and using weave to expose the control plane over cumulus switch | 12:55 |
rwallner | the theory is we would have a custom routing cluster running haproxy but talking overy the overlay to the control nodes | 12:56 |
*** iceyao has quit IRC | 12:57 | |
rwallner | we're running newton in prod today as a pilot and changing up the architecture from flat/vlan to bgp hence the changes and moving to ocata | 12:57 |
Daviey | rwallner: What is the benefit of using weave here? Doesn't it add more complexity ? | 12:57 |
rwallner | well, in this case its experimental. meaning out bgp routes are only internal to the spine and leaf so we still need to bridge out to the vlan world. The idea is to use an overlay (whether its weave or something esle) to bridge the control plane | 12:58 |
rwallner | the provider networks is a separate concern as well | 12:59 |
*** caoyuan has joined #openstack-kolla | 13:00 | |
Daviey | rwallner: It sounds interesting.. Would you mind doing a write up of your experiment? Blog / Mailing list or something.. I'm sure others would appreciate it | 13:00 |
*** caoyuan_ has quit IRC | 13:01 | |
*** goldyfruit has joined #openstack-kolla | 13:02 | |
*** yangyapeng has joined #openstack-kolla | 13:02 | |
*** caoyuan has quit IRC | 13:03 | |
*** schwicht has quit IRC | 13:03 | |
*** eaguila__ has quit IRC | 13:03 | |
*** signed8bit has quit IRC | 13:03 | |
rwallner | Daviey, sure, would love to. will probably be a few weeks, just kicked off a sprint yesterday to experiment | 13:04 |
Daviey | neat | 13:04 |
*** schwicht has joined #openstack-kolla | 13:04 | |
*** eaguilar has joined #openstack-kolla | 13:05 | |
sdake | rwellum thanks for the review | 13:07 |
sdake | rwellum were you able to get a functional system with those docs? | 13:07 |
*** skramaja has quit IRC | 13:08 | |
*** caoyuan has joined #openstack-kolla | 13:08 | |
rwellum | I will try in about 2 hours sdake and let you know of course. | 13:08 |
*** eaguilar has quit IRC | 13:09 | |
sdake | rwellum cool - i need to run through them again to make sure they are still valid | 13:09 |
sdake | rwellum i may have done that last night - not sure :) | 13:09 |
rwellum | I see a couple of steps I missed due to noise of the original doc. | 13:09 |
sdake | rwellum right- the original doc was super noisy | 13:09 |
sdake | also you can render this one with tox -e docs | 13:09 |
sdake | it renders pretty well | 13:09 |
sdake | when i run al lthis i get a vm creation in the "BUILD" state | 13:09 |
sdake | something not quite right | 13:10 |
sdake | i think some helm microcharts are not launched properly | 13:10 |
sdake | although its close | 13:10 |
sdake | rwellum i'd recommend rendering the docs - makes it so much easier to C&P | 13:10 |
rwellum | Yes - I was getting similar yesterday. Everything ran but some failures. Will take a look this morning and redo. For me it's just a VM so no issue reverting to stable state. | 13:11 |
rwellum | Ok will render ty. Didn't think of that. | 13:11 |
sdake | docs rendered in doc/build/html i think | 13:12 |
sdake | its actually very easy for me to reset to a clean slate as well | 13:12 |
sdake | rwellum that doc includes a cleanup script - which speeds up dev as the charts don't haveto be reuilt every time | 13:12 |
rwellum | Yeah if you have a host where others don't yell at you if you 'accidentally' bring it down... | 13:13 |
sdake | ya my machine is dedicated to a few people | 13:13 |
*** esharao has joined #openstack-kolla | 13:14 | |
*** lamt has joined #openstack-kolla | 13:14 | |
rwellum | You're up early? | 13:15 |
sdake | always up by 6am at latest | 13:15 |
sdake | usually up by 5 | 13:15 |
sdake | kids go to school at 6:30 | 13:15 |
sdake | or I have to get ready for them to go to school atleast | 13:16 |
rwellum | Similar but he drives himself. :) | 13:16 |
sdake | must be nice | 13:16 |
sdake | my daughter is 14 and she is almost self sufficient | 13:16 |
sdake | my son is 12, he needs alot of attention ot get to schol | 13:16 |
rwellum | Very similar here - but my 13 yr old bus comes 7:30 | 13:17 |
sdake | wow thats early | 13:17 |
sdake | my son starts school at 8:45, daughter at 7:30 | 13:17 |
rwellum | Makes more sense to me. | 13:18 |
rwellum | It's great when they can drive themselves around, but the insurance+worry jumps exponentially... | 13:18 |
*** signed8bit has joined #openstack-kolla | 13:18 | |
*** rhallisey has joined #openstack-kolla | 13:21 | |
*** signed8b_ has joined #openstack-kolla | 13:23 | |
*** schwicht has quit IRC | 13:25 | |
*** signed8bit has quit IRC | 13:25 | |
*** manheim has quit IRC | 13:30 | |
*** eanylin has quit IRC | 13:32 | |
openstackgerrit | Merged openstack/kolla-ansible master: Fix some blank space issues https://review.openstack.org/447348 | 13:33 |
*** manheim has joined #openstack-kolla | 13:33 | |
openstackgerrit | Merged openstack/kolla master: Fix source install for vmtp https://review.openstack.org/448515 | 13:39 |
*** hachi has joined #openstack-kolla | 13:40 | |
*** sbezverk has joined #openstack-kolla | 13:40 | |
*** srwilkers has joined #openstack-kolla | 13:41 | |
*** bjolo has quit IRC | 13:42 | |
*** eanylin has joined #openstack-kolla | 13:45 | |
*** schwicht has joined #openstack-kolla | 13:46 | |
*** iceyao has joined #openstack-kolla | 13:47 | |
*** iceyao has quit IRC | 13:51 | |
esharao | GM guys | 13:51 |
esharao | sdake i have some comments on the doc based on using ubuntu as base os | 13:51 |
*** cuongnv has joined #openstack-kolla | 13:52 | |
*** lamt has quit IRC | 13:52 | |
*** manheim has quit IRC | 13:54 | |
*** manheim has joined #openstack-kolla | 13:55 | |
*** manheim has quit IRC | 13:55 | |
*** blallau has joined #openstack-kolla | 13:55 | |
*** krtaylor has quit IRC | 13:56 | |
openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible master: Pg_num should be configurable https://review.openstack.org/450797 | 13:56 |
openstackgerrit | Merged openstack/kolla master: Prevent apache from listening on port 80 for heat-base https://review.openstack.org/449040 | 13:56 |
openstackgerrit | Merged openstack/kolla master: RHEL: Add mod_ssl for services running over httpd https://review.openstack.org/446911 | 13:57 |
*** krtaylor has joined #openstack-kolla | 14:00 | |
eanylin | rwellum esharao sdake: I added some comments for Cinder LVM as well | 14:00 |
*** tovin07 has joined #openstack-kolla | 14:01 | |
*** lucasxu has joined #openstack-kolla | 14:04 | |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: ironic-pxe: handle non-x86 architectures https://review.openstack.org/434817 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: enable all images enabled for Ubuntu https://review.openstack.org/432787 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: Add support for non-x86 architectures (aarch64, ppc64le) https://review.openstack.org/430940 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: make yum fail on missing packages https://review.openstack.org/449086 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: use mariadb-devel for CentOS and !x86-64 https://review.openstack.org/444832 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: nova-libvirt: handle ppc64le https://review.openstack.org/434810 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: ceph: move ceph-fuse package to cephfs-fuse image https://review.openstack.org/440634 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: openstack-base: Percona-Server is x86-64 only https://review.openstack.org/449965 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: kubernetes: disable for architectures other than x86-64 https://review.openstack.org/445638 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: nova-compute: handle rtslib(-fb) for debian-binary builds too https://review.openstack.org/435941 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: move to stretch https://review.openstack.org/434453 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: mariadb: handle lack of external repos on non x86 https://review.openstack.org/450111 | 14:08 |
openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: nova-compute: enable Debian/source build https://review.openstack.org/450805 | 14:08 |
hrw | hate those merge conflicts ;( | 14:10 |
*** jaosorior has quit IRC | 14:12 | |
*** iceyao has joined #openstack-kolla | 14:13 | |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Deployment Guide Documentation https://review.openstack.org/447356 | 14:18 |
Serlex | :-D | 14:22 |
*** manheim has joined #openstack-kolla | 14:29 | |
*** shashank_t_ has joined #openstack-kolla | 14:29 | |
*** mnasiadka has joined #openstack-kolla | 14:30 | |
*** duonghq has quit IRC | 14:30 | |
*** duonghq has joined #openstack-kolla | 14:30 | |
duonghq | evening guys | 14:31 |
mnasiadka | Hi | 14:31 |
*** duonghq_ has joined #openstack-kolla | 14:31 | |
mnasiadka | Is horizon somewhat broken in ubuntu/binary (stable/ocata)? | 14:31 |
mnasiadka | It's in restart loop just after deployment... | 14:31 |
*** manheim has quit IRC | 14:33 | |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Deployment Guide Documentation https://review.openstack.org/447356 | 14:34 |
*** manheim has joined #openstack-kolla | 14:34 | |
*** fooliouno has quit IRC | 14:35 | |
sdake | sup peeps | 14:37 |
duonghq | hi sdake | 14:37 |
sdake | eanylin esharao rwellum thanks for all the comments - esharao i htink i haven't integrated yours yet however plan to do so shortly | 14:37 |
sdake | sup duonghq | 14:37 |
duonghq | I'm trying deploy kolla-k8s again at home | 14:38 |
sdake | duonghq cool - the latest docs hopefully should be very close | 14:38 |
duonghq | got trouble with keystone password | 14:38 |
duonghq | somehow, keystone-managedb failed to connect to mariadb due to wrong password | 14:38 |
duonghq | but keystone-createdb is ok | 14:38 |
spsurya | sdake: sup | 14:39 |
sdake | duonghq rwellum eanylin esharao when running through the docs - you mgiht find it helpful to tox -e docs and open up build/html/index.html in chrome or your favoriate brower | 14:39 |
duonghq | sdake, roger | 14:39 |
*** manheim has quit IRC | 14:39 | |
rwellum | sdake: ack | 14:40 |
sdake | esharao its the kid hour atm - so i'll integrate your comments after i get my son off to schoool | 14:40 |
*** manheim has joined #openstack-kolla | 14:42 | |
esharao | sdake.. anytime.. is good | 14:47 |
*** lamt has joined #openstack-kolla | 14:47 | |
*** shashank_t_ has quit IRC | 14:49 | |
sdake | sbezverk if you have the time today - can you throw your wealth of knowledge and wisdom about kolla-kubernetes into this semi-final review- goal is to get it merged by ednesday: https://review.openstack.org/#/c/447356/ | 14:58 |
sdake | qwang jascott1 ^^ | 15:00 |
*** mnasiadka has quit IRC | 15:01 | |
sdake | dims at one point you had offered - the docs are actually semi-working - a fresh eye would be helpful ^ | 15:01 |
*** magicboiz has quit IRC | 15:02 | |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Deployment Guide Documentation https://review.openstack.org/447356 | 15:06 |
*** adrian_otto has joined #openstack-kolla | 15:07 | |
sdake | eanylin did you set up an LVM backing store for kolla-kubernetes in that review? | 15:13 |
openstackgerrit | Merged openstack/kolla master: Using assertIsNone() instead of assertEqual(None) https://review.openstack.org/448923 | 15:14 |
openstackgerrit | Merged openstack/kolla stable/newton: Fix hardcoded path for custom configuration of Neutron VPNaaS https://review.openstack.org/449975 | 15:14 |
eanylin | sdake: I havent really tested cinder part | 15:15 |
mnaser | hm | 15:15 |
mnaser | is there a gate issue by any chance | 15:15 |
*** shashank_t_ has joined #openstack-kolla | 15:15 | |
mnaser | for stable/newton | 15:15 |
eanylin | sdake: But I think it is created as part of that chart since the initial error was that it cannot mount | 15:15 |
eanylin | and once we put that in, it stops complaining | 15:15 |
mnaser | wait maybe we did backport the sudo stuff back to stable/newton | 15:16 |
eanylin | which is telling me that it probably created it as part of the chart | 15:16 |
mnaser | yup | 15:16 |
mnaser | stable/newton deploy gates broken because of a change that wasnt backported welp my abd | 15:16 |
*** lamt has quit IRC | 15:17 | |
eanylin | sdake: Hmm, seems like its looking for Cinder v3 enpoints which doesnt exists | 15:18 |
eanylin | [root@kube03 ~]# cinder list | 15:19 |
eanylin | ERROR: publicURL endpoint for volumev3 service not found | 15:19 |
*** sbezverk has quit IRC | 15:20 | |
*** lamt has joined #openstack-kolla | 15:21 | |
*** zhubingbing has joined #openstack-kolla | 15:21 | |
hawi | hi. question - why is that i cant enter index name or pattern for Kibana? | 15:23 |
openstackgerrit | Merged openstack/kolla-ansible master: Fix Telegraf role reconfigure action https://review.openstack.org/450314 | 15:23 |
egonzalez | mnaser, what gates? kolla, kolla-ansible, kolla-kubernetes? | 15:23 |
mnaser | egonzalez nevermind, found out the issue, 3.0.3 was released, now images are being built with 3.0.4, all.yml points to 3.0.3 in stable/newton | 15:23 |
mnaser | fixing right now to bump | 15:23 |
mnaser | similar to this change https://github.com/openstack/kolla/commit/be846040251ccd193135d7782cb9e04def85c828 | 15:23 |
*** tovin07 has left #openstack-kolla | 15:24 | |
sdake | hawi i belieeve its log-* | 15:24 |
openstackgerrit | Mohammed Naser proposed openstack/kolla stable/newton: Bump openstack_release to 3.0.4 in group_vars https://review.openstack.org/450849 | 15:24 |
mnaser | ^ to unbreak stable/newton deploy jobs | 15:25 |
openstackgerrit | Sayantani Goswami proposed openstack/kolla-ansible master: [deploy-guide] Moving the quick start to the deploy folder https://review.openstack.org/448509 | 15:25 |
hawi | sdake: no, it doesnt work :( | 15:25 |
sdake | hawi not sure then - someone on the channel asked the same quesiton monday - try checking the logs (in the subject) | 15:26 |
*** Guest52040 has joined #openstack-kolla | 15:26 | |
*** alop has joined #openstack-kolla | 15:27 | |
*** sayantan_ has joined #openstack-kolla | 15:28 | |
*** skramaja has joined #openstack-kolla | 15:29 | |
dims | sdake : yep, i had. apologies, juggling etcd and some go/containers at the moment. | 15:29 |
sdake | dims its all good - pretty much eveyrone dirnking from multiple firehoses :) | 15:29 |
sdake | dims the instructions take about 1 hour to go through assumin they work | 15:30 |
sdake | dims so don't necessarily ened a week of your time :) | 15:30 |
*** matrohon has quit IRC | 15:31 | |
*** manheim has quit IRC | 15:33 | |
duonghq | sdake, do I need to add 2 virtual NIC to my VM? | 15:33 |
duonghq | for kolla-k8s | 15:33 |
duonghq | one for network_interface, one for neutron_external_interface | 15:34 |
*** manheim has joined #openstack-kolla | 15:34 | |
zhubingbing | i thinks need 2 virtual NI in kolla-k8s | 15:34 |
esharao | sdake.. i used eanylin's solution and my cinder volume is up | 15:34 |
sdake | @duonghq i don't know the answer to that question - I do know that I setup my network with 2 interfaces | 15:34 |
*** hrw has quit IRC | 15:34 | |
duonghq | zhubingbing, me too | 15:34 |
esharao | i just need help with nova compute on the ubuntu and i shouldbe good to go | 15:35 |
sdake | @esharao did you create an lvm volume? | 15:35 |
*** Serlex has quit IRC | 15:35 | |
esharao | hmm no.. pod is running let me try | 15:35 |
sdake | @esharao no i mean did you create one to bootstrap the thing | 15:35 |
sdake | @esharao or was one already there? | 15:35 |
sdake | @esharao in other words, are you doing anything outside of the instructions - I would think that lvm section should reference an lvm backing store | 15:36 |
esharao | sdake no | 15:36 |
*** hrw has joined #openstack-kolla | 15:36 | |
esharao | just added to cloud.yaml the config provided by eanylin | 15:36 |
esharao | did a helm delete cinder-volume --purge and did helm install | 15:37 |
openstackgerrit | Bertrand Lallau proposed openstack/kolla master: Rabbitmq: refactor Dockerfile.j2 file https://review.openstack.org/450858 | 15:37 |
*** sbezverk has joined #openstack-kolla | 15:37 | |
sdake | esharao cool thanks :) | 15:37 |
*** klindgren_ has joined #openstack-kolla | 15:37 | |
sdake | i think we may need an lvm storage volume - although I really dont know for sure | 15:38 |
sdake | (to get cinder to work) | 15:38 |
sdake | it may start, but may not operate correctly | 15:38 |
esharao | hmm also what about the version issue reported by eanylin? | 15:38 |
sdake | bbiaf need to ru nkid to shcool | 15:38 |
*** klindgren has quit IRC | 15:38 | |
*** lamt has quit IRC | 15:41 | |
*** blallau has quit IRC | 15:43 | |
*** shashank_t_ has quit IRC | 15:44 | |
openstackgerrit | Eduardo Gonzalez proposed openstack/kolla master: Alphabetize packages https://review.openstack.org/450150 | 15:45 |
*** fooliouno has joined #openstack-kolla | 15:50 | |
*** tonanhngo has joined #openstack-kolla | 15:51 | |
sdake | egonzalez did you have a solution for logstash search filter? | 15:52 |
inc0 | good morning | 15:53 |
* inc0 semi-afk but here in general | 15:53 | |
sdake | inc0 qq do you know what logstash filter to use? | 15:54 |
inc0 | sdake: negative, akwasnie do you?;) | 15:54 |
egonzalez | sdake, hawi is flog-* tho https://review.openstack.org/#/c/449384/ | 15:54 |
sdake | egonzalez thanks :) | 15:54 |
sdake | egonzalez thats wierd, that should be in the docs | 15:55 |
*** manheim has quit IRC | 15:55 | |
sdake | hawi check out the docs here: https://docs.openstack.org/developer/kolla-ansible/central-logging-guide.html | 15:55 |
*** lamt has joined #openstack-kolla | 15:56 | |
*** iceyao has quit IRC | 15:58 | |
akwasnie | inc0, hawi, sdake: it was changed after replacing Heka with fluentd - with Heka it was log (so for Mitaka it is log), now it is flog as egonzalez said | 15:59 |
*** iceyao has joined #openstack-kolla | 16:00 | |
*** sbezverk has quit IRC | 16:01 | |
*** vhosakot has joined #openstack-kolla | 16:03 | |
*** sbezverk has joined #openstack-kolla | 16:04 | |
*** egonzalez has quit IRC | 16:05 | |
*** erlon has quit IRC | 16:05 | |
*** sbezverk has quit IRC | 16:07 | |
*** cuongnv has quit IRC | 16:09 | |
*** kfox1111 has joined #openstack-kolla | 16:09 | |
sdake | kfox1111 mind adding your wealth of knowledge and wisdom to review: https://review.openstack.org/#/c/447356/ | 16:14 |
*** iceyao has quit IRC | 16:14 | |
*** rwallner has quit IRC | 16:14 | |
sdake | kfox1111 specifically i am after how things behave and why they behave that way - if that makes any sense | 16:14 |
*** skramaja has quit IRC | 16:14 | |
sdake | kfox1111 goal is to get this initial doc merged wednesday with followups as needed | 16:15 |
*** zhubingbing has quit IRC | 16:15 | |
*** zhubingbing__ has joined #openstack-kolla | 16:15 | |
kfox1111 | sdake: I'll try and go through it tonight. | 16:16 |
sdake | kfox1111 cool thanks - would appreciate anything that is missing as well :) | 16:16 |
sdake | kfox1111 goal is to merge by wed so it shows up on docs.o.o | 16:17 |
*** shashank_t_ has joined #openstack-kolla | 16:18 | |
*** unicell has joined #openstack-kolla | 16:20 | |
kfox1111 | k. | 16:20 |
kfox1111 | I'll see what I can do. | 16:20 |
kfox1111 | depends on what else decides to explode on me today. :/ | 16:20 |
*** Reepicheep has joined #openstack-kolla | 16:21 | |
vhosakot | inc0: I see zhubingbing__ already did horizon :) I was going to pick it today - https://review.openstack.org/#/c/450725/ | 16:25 |
zhubingbing__ | ok | 16:25 |
inc0 | haha | 16:25 |
*** pcaruana has quit IRC | 16:26 | |
kfox1111 | inc0: https://review.openstack.org/#/c/449262/ | 16:26 |
inc0 | vhosakot: ironic is ripe for taking:) | 16:26 |
vhosakot | inc0: np.. zhubingbing__ is anything left for me? https://review.openstack.org/#/q/status:open+project:openstack/kolla-kubernetes+branch:master+topic:bp/move-config-to-kolla-k8s | 16:26 |
zhubingbing__ | can u help me look rabbitmq? | 16:26 |
kfox1111 | inc0: a large chunk of kollakube gone. | 16:26 |
inc0 | kfox1111: I don't have time for reviews now:( listening to fast speaking golang tutor | 16:27 |
kfox1111 | the rest of the stuff is all things kollakube actually does some heavy lifting, and should be ported to helm. | 16:27 |
kfox1111 | k. | 16:27 |
sdake | @kfox1111 roger- well even if you do have emergencies would appreciate a +2 on that doc even if it is imperfect so we can get on with improving it | 16:28 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Helm Native Values. https://review.openstack.org/449826 | 16:28 |
zhubingbing__ | vhosakot can u help me look it ? https://review.openstack.org/#/c/450725/ | 16:29 |
sdake | kfox1111 if its in rendered format on docs.oo its alot easier for people to follow it then in a review with a bajllion comments | 16:29 |
kfox1111 | k. | 16:29 |
sdake | sup zhubingbing__ | 16:29 |
sdake | zhubingbing__ sleep is good for ya ;) | 16:29 |
kfox1111 | yeah. totally agree. | 16:29 |
sdake | thanks kfox1111 :) | 16:29 |
zhubingbing__ | ;) | 16:29 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Unpin Canal https://review.openstack.org/450423 | 16:30 |
kfox1111 | looks like they may have fixed canal. | 16:30 |
sdake | nice htanks kfox1111 - will test it out | 16:30 |
kfox1111 | the test I did last night worked, and they committed the fix. | 16:30 |
kfox1111 | lets see if it holds up. | 16:30 |
sdake | i saw your review worked in the gate | 16:30 |
*** lucasxu has quit IRC | 16:31 | |
kfox1111 | https://review.openstack.org/#/c/447031/ is making good progress | 16:31 |
kfox1111 | sdake does the board weigh in on that, or is that strictly a tc thing? | 16:31 |
duonghq | sdake, I almost bring keystone up with your guide, but get one issue, somehow it try to connect to 10.10.10.254:35357 but not 10.3.168.173 (its cluster ip) | 16:32 |
sdake | kfox1111 the board doesn't interfere with tc matters, the tc is responsible for that repository | 16:33 |
sdake | kfox1111 atleast I think that is the answer - not really sure :) | 16:33 |
zhubingbing__ | https://review.openstack.org/#/c/450556/7 kfox1111 can u help me look it ? | 16:33 |
vhosakot | zhubingbing__: yeah, I saw the gate errors of https://review.openstack.org/#/c/450725/ in http://logs.openstack.org/25/450725/1/check/gate-kolla-kubernetes-deploy-centos-binary-2-ceph-multi-nv/e8df508/console.html#_2017-03-28_11_32_28_202440. looks like the gate failed in the task "Copying over local_settings" since memcahced was not found. I'll take a look.. thanks. | 16:33 |
sdake | kfox1111 the best person to ask would be a tc member - dims is in channel and can probabaly answer your question more accurately | 16:34 |
kfox1111 | sdake: ok. thanks. | 16:34 |
sdake | duonghq define "it try" do you mean your keystone client? | 16:35 |
kfox1111 | zhubingbing__: not sure the right answer there. maybe just hardcode it to 3? | 16:35 |
zhubingbing__ | if we don't use inventory/all-in-one i don;t know how fix "groups['glance-api']" | 16:35 |
dims | kfox1111 : that review has my support | 16:36 |
*** Pavo has quit IRC | 16:36 | |
dims | kfox1111 : something in there you like or dont like? | 16:36 |
zhubingbing__ | vhosakot thank you very much | 16:36 |
vhosakot | zhubingbing__: np :) | 16:36 |
vhosakot | sdake: how is the single-vm deploy going? I'll add my comments to your review. | 16:37 |
duonghq | sdake, hmm. forgot set internal vip, but even when I set it, it cannot connect to that external-ip of the keystone-admin service | 16:38 |
sdake | vhosakot the review is shaping up - recommend rendering it with tox -e docs | 16:38 |
vhosakot | sdake: cool | 16:38 |
sdake | vhosakot compute kit seems to start - although when I ran through the docs this morning kubernetes didn'tdeploy correctly | 16:39 |
vhosakot | inc0: why are you learning go btw? :) | 16:39 |
sdake | vhosakot so wiping out my system and giving it another go | 16:39 |
inc0 | yeah | 16:39 |
*** lucasxu has joined #openstack-kolla | 16:40 | |
inc0 | I mean I know a bit of it already, what I'm trying to get to is to stop thinking "this would be sooo much easier if I'd only use python" | 16:40 |
vhosakot | sdake: yes, sanitizing/validating the empty base k8s cluster is a big thing before installing OpenStack pods on it.. | 16:41 |
inc0 | so far, besides small webservers/rest apis I struggle in finding thing that golang is better than py at | 16:41 |
sdake | vhosakot i did have kubernetes and koll-kubernetes operational last night | 16:41 |
inc0 | besides being full-hype maybe | 16:41 |
sdake | vhosakot started a vm after init-runonce - it entered the "BUILDING" state and stuck there | 16:41 |
sdake | inc0 must be nice to have time to learn a new language ;) | 16:42 |
vhosakot | sdake: having worked on OpenStack so many years, I treat OpenStack "operational" when we can ping www.google.com from inside a nova VM.. that way, we validate both neutorn(both L2 and l3) and nova and underlying routing/networking... | 16:43 |
sdake | vhosakot agreed | 16:43 |
inc0 | sdake: well, someone has to understand k8s changes in our community | 16:43 |
vhosakot | sdake: vm going to "BUILDING" state is great progress compared to where I am :) | 16:43 |
sdake | vhosakot if you havea base setup of kubernetes, run through the docs | 16:43 |
sdake | it should get you to that state | 16:43 |
vhosakot | inc0: yeah, go has a lot of hype.. | 16:43 |
vhosakot | sdake: all-in-one? | 16:44 |
sdake | vhosakot although your netowrking pod and service cidrs) may be fubared in your base env | 16:44 |
sdake | vhosakot which will cause kolla-kubernetes to not enter the running state | 16:44 |
sdake | vhosakot indeed AIO | 16:44 |
sdake | two nics | 16:44 |
vhosakot | sdake: cool, I'll try it later today.. cool | 16:44 |
sdake | bare metal | 16:44 |
vhosakot | sdake: cool, thanks for the info | 16:45 |
sdake | 4 fingers on mac = winning | 16:46 |
vhosakot | sdake: inc0: once we have a good deploy mechanism and docs (that always works), I promise to write a blog and tweet about it so we get more contributors to kolla-k8s.. | 16:46 |
inc0 | yeah, we need that | 16:47 |
inc0 | good thinking | 16:47 |
sdake | inc0 its not just vhosakot tweeting about it - its his entire network :) | 16:47 |
inc0 | a bit of marketing:) | 16:47 |
sdake | its less marketing and more technical "here is how to get kolla-kubernetes running" | 16:47 |
vhosakot | sdake: yes.. a new dev/contributor spends few hrs eval'ing a dev env... plus, we cannot learn the technology if we use the gate as the dev env.. :) | 16:48 |
vhosakot | inc0: yes, marketing is needed to attract devs... many many many ops want to deploy OpenStack on k8s... kolla-kubernetes is the only one doing it :) | 16:49 |
vhosakot | may be tectonic as well | 16:49 |
sdake | vhosakot there are about 6+ independent efforts to deploy openstack on kubernetes | 16:50 |
srwilkers | vhosakot, what sdake said. more than just kolla-kubernetes | 16:50 |
vhosakot | sdake: open-source under the OpenStack big tent? | 16:50 |
inc0 | but we're only openstack one:) | 16:50 |
vhosakot | ^^^ | 16:50 |
sdake | vhosakot nothing in big tent but kolla-kubernetes | 16:50 |
vhosakot | that's what I meant ^^^ | 16:50 |
vhosakot | :) | 16:50 |
inc0 | so technically nothing that can call itself "OpenStack" really | 16:50 |
inc0 | but whatever, let's make sure that we're best one for our audience | 16:52 |
sdake | tend to agree with you vhosakot re gate being only way to deploy kolla-kubernetes being a problem | 16:52 |
*** nathharp has quit IRC | 16:52 | |
*** lamt has quit IRC | 16:52 | |
*** gfidente is now known as gfidente|afk | 16:52 | |
srwilkers | sdake, +1 | 16:52 |
sdake | vhosakot that is why it is bieing fixed :) | 16:52 |
vhosakot | sdake: +1 | 16:52 |
srwilkers | inc0, that should be your focus. not splitting hairs over what's technically correct | 16:52 |
vhosakot | inc0 is busy going :) | 16:52 |
inc0 | srwilkers: technical correctnes (as in correct technical decisions) is important | 16:53 |
inc0 | but agree, we're solving this problem | 16:53 |
srwilkers | thats vastly different than what your comment implied | 16:53 |
inc0 | I just stated fact, nothing to do with priorities | 16:53 |
inc0 | being only one in big tent is not enough to be good project | 16:53 |
inc0 | it usually gives us credit of trust, but broken trust is really bad for project | 16:54 |
duonghq | sdake, I follow your guide but I don't know why keystone-admin cannot connect to keystone-internal endpoint but when I try to nslookup, the dns works fine | 16:55 |
kfox1111 | dims: no, It has my support too. just wondering if sdake should vote on it too, or if its purely a tc thing. | 16:55 |
inc0 | so it's really just promise that we'll hold ourselves to high standards | 16:55 |
duonghq | also, keystone-internal doesn't get external-ip, but I think it doesn't matter | 16:55 |
dims | kfox1111 : everyone can vote | 16:55 |
dims | can and should vote :) | 16:55 |
inc0 | dims: what vote? | 16:55 |
inc0 | (missing contect) | 16:55 |
kfox1111 | dims: ok. :) | 16:55 |
inc0 | context | 16:55 |
kfox1111 | inc0: https://review.openstack.org/#/c/447031/ | 16:56 |
dims | inc0 : https://review.openstack.org/#/c/447031/ | 16:56 |
*** fooliouno has quit IRC | 16:56 | |
dims | kfox1111 : if you see there's a Rollcall-Vote in there to count the tc members' vote | 16:56 |
kfox1111 | dims: ah. ok. I saw the rollcall but wasn't sure how it worked. | 16:56 |
kfox1111 | thanks for the clarification. :) | 16:56 |
inc0 | ok, Zane's writting, I'll need more time for it than I have now;) | 16:56 |
dims | inc0 : it's pretty easy read :) | 16:56 |
inc0 | well, it's long read tho and I'm on training now | 16:57 |
inc0 | starred a review tho | 16:57 |
*** lamt has joined #openstack-kolla | 16:58 | |
*** inc0 has quit IRC | 16:59 | |
*** inc0 has joined #openstack-kolla | 16:59 | |
inc0 | dims: speaking of rollcall votes https://review.openstack.org/#/c/346455/ <- | 16:59 |
zhubingbing__ | around inc0 | 17:00 |
dims | ack inc0 | 17:00 |
zhubingbing__ | https://review.openstack.org/#/c/450556/7/ansible/roles/cinder/templates/cinder.conf.j2 can u help me look it ? | 17:01 |
*** rwallner has joined #openstack-kolla | 17:01 | |
sdake | kfox1111 with purely my kolla hat on - that sentence 77-80 is a doozy :) | 17:01 |
sdake | kfox1111 i'm not even sure what that means tbh | 17:02 |
duonghq | sdake, also, seem that dns_name should be keystone-admin endpoint? | 17:02 |
*** mgoddard has quit IRC | 17:02 | |
sdake | inc0 to kick off a rollcall vote, send an email to openstack-dev with [tc] asking for a rollcall vote of your review | 17:02 |
sdake | inc0 atleast that was the general protocol i followed when i was steering the ship ;) | 17:03 |
*** caoyuan has quit IRC | 17:03 | |
sdake | inc0 otherwise tc members may not see the review and it may not be prioritized in the agenda | 17:03 |
*** bswartz has joined #openstack-kolla | 17:04 | |
sdake | duonghq no clue - kfox1111 any idea on duonghq 's question? | 17:04 |
duonghq | log of keystone issue (not related to dns_name): http://paste.openstack.org/show/604522/, can you take a look? | 17:05 |
*** zhubingbing_ has joined #openstack-kolla | 17:05 | |
*** zhubingbing__ has quit IRC | 17:08 | |
duonghq | sdake, seem that we set it in keystone-create-endpoints-job, but just set to $userDnsName var | 17:08 |
*** lucasxu has quit IRC | 17:09 | |
satyar | hi sdake | 17:10 |
*** pbandark`brb has quit IRC | 17:10 | |
*** pbandark has joined #openstack-kolla | 17:10 | |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Deployment Guide Documentation https://review.openstack.org/447356 | 17:11 |
sdake | duonghq there is one thing added to that last review | 17:11 |
sdake | duonghq https://review.openstack.org/#/c/447356/11..12 | 17:12 |
duonghq | sdake, seen | 17:12 |
*** deadnull has joined #openstack-kolla | 17:12 | |
inc0 | dims: speaking of rollcall votes https://review.openstack.org/#/c/346455/ <- | 17:13 |
inc0 | sorry | 17:13 |
inc0 | disregard ^ | 17:13 |
inc0 | zhubingbing_: I'm here | 17:13 |
openstackgerrit | Marcus Williams proposed openstack/kolla-ansible master: Split Openvswitch into own role https://review.openstack.org/416793 | 17:14 |
*** MasterOfBugs has joined #openstack-kolla | 17:14 | |
zhubingbing_ | https://review.openstack.org/#/c/450581/ | 17:14 |
*** pramodrj07 has joined #openstack-kolla | 17:14 | |
zhubingbing_ | inc0 i have some problems ;) | 17:15 |
*** lucasxu has joined #openstack-kolla | 17:15 | |
zhubingbing_ | i don't know how to fix group['xxx'] in kolla-kubernetes | 17:15 |
inc0 | zhubingbing_: you remove it;) | 17:15 |
*** lucasxu has quit IRC | 17:15 | |
inc0 | in kolla-k8s there is no group required | 17:15 |
zhubingbing_ | {% for host in groups['rabbitmq'] %} | 17:16 |
zhubingbing_ | {'rabbit@{{ hostvars[host]['ansible_hostname'] }}', disc} | 17:16 |
zhubingbing_ | {%- if not loop.last -%},{%- endif %} | 17:16 |
zhubingbing_ | {% endfor % | 17:16 |
*** lucasxu has joined #openstack-kolla | 17:16 | |
inc0 | zhubingbing_: currently we didn't solve clustering of rabbitmq | 17:16 |
satyar | does kolla support nova ssl? | 17:16 |
inc0 | so just remove this whole clustering thing | 17:16 |
inc0 | satyar: nova ssl? as in api ssl or what? | 17:17 |
satyar | yes | 17:17 |
satyar | nova ssl ceert | 17:17 |
satyar | cert | 17:17 |
zhubingbing_ | {'rabbit@{{ hostname }}', disc} ? | 17:17 |
inc0 | we terminate ssl on haproxy | 17:17 |
mnaser | kolla cores, stable/newton gates are broken since release.. super quick and straighforwrd change - https://review.openstack.org/#/c/450849/ | 17:17 |
inc0 | zhubingbing_: you need to find out how config looks in kolla-k8s today | 17:17 |
inc0 | kfox1111: can you helm Zhubing here? | 17:18 |
*** lucasxu has quit IRC | 17:18 | |
mnaser | thanks duonghq , one more +2 so we can break gate if someone doesnt mind ^ | 17:18 |
duonghq | mnaser, nice work | 17:19 |
*** lucasxu has joined #openstack-kolla | 17:19 | |
duonghq | sdake, no luck here, keystone-internal still cannot catch up with other microservice | 17:19 |
satyar | inc0: basically i am looking for novnc ssl | 17:19 |
*** lucasxu has quit IRC | 17:20 | |
*** sbezverk has joined #openstack-kolla | 17:21 | |
*** yangyape_ has joined #openstack-kolla | 17:21 | |
mnaser | thanks sdake | 17:21 |
satyar | I want to access openstack instances vnc console using SSL/TLS | 17:21 |
sdake | mnaser sure | 17:21 |
sdake | mnaser i was looking at it earlier and got distracted by kids + school + morning chaos = winning | 17:21 |
mnaser | sdake no worries, one of my other stable/newton backports failed deploys so thats how i found out about this | 17:22 |
inc0 | satyar: I'd need to check, not sure | 17:22 |
*** fooliouno has joined #openstack-kolla | 17:23 | |
satyar | i dont see any code in the https://github.com/openstack/kolla-ansible/blob/2f34d4bd031cc1385ca852159017e86a2ca4df36/ansible/roles/nova/templates/nova-novncproxy.json.j2 | 17:23 |
*** yangyapeng has quit IRC | 17:23 | |
satyar | for ssl cert peice | 17:23 |
duonghq | sdake, dns still works fine, but keystone-internal cannot get external-ip (sure, due to keystone-public already bound to that port), but seem that keystone-admin cannoct contact with keystone-internal | 17:25 |
*** lamt has quit IRC | 17:27 | |
sdake | duonghq i am in the midst of redeploying a fresh env now - i'll let you know what i find from the rendered docs (e.g. if they still work for me) | 17:27 |
inc0 | satyar: there won't be ssl cert in nova | 17:28 |
duonghq | sdake, the resolve-conf is supposed to be mounted or not? in keystone pod | 17:28 |
*** rwallner has quit IRC | 17:28 | |
satyar | for accessing novnc we should be able to configure nova to use ssl | 17:29 |
satyar | example https://bugs.launchpad.net/nova/+bug/1243614 | 17:29 |
openstack | Launchpad bug 1243614 in OpenStack Compute (nova) "with ssl/tls configured novnc nova doesn't provide a https link without editing the config file" [Undecided,Expired] | 17:29 |
*** pbandark has quit IRC | 17:29 | |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move rabbitmq config to kolla-k8s https://review.openstack.org/450581 | 17:30 |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Deployment Guide Documentation https://review.openstack.org/447356 | 17:30 |
duonghq | sdake, where do you get keystone_public_url? | 17:31 |
sdake | duonghq no clue - it was a todo in the last review - this was the contents of my /etc/kolla/globals.yaml | 17:32 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move cinder config to kolla-k8s https://review.openstack.org/450556 | 17:32 |
duonghq | ya, I read your comment | 17:32 |
sdake | duonghq i thought it might help you since it seems in the area | 17:32 |
sdake | in the area you are having trouble (keystone) | 17:32 |
duonghq | what is ip of your box? | 17:32 |
duonghq | I set it to my box's ip | 17:33 |
sdake | sbezverk can you look at https://review.openstack.org/#/c/447356/12..13/doc/source/deployment-guide.rst and tell us what keystone_public_url should be set to? | 17:33 |
sdake | or kfox1111 - should take less then 2 minutes ;) | 17:33 |
sdake | duonghq I left it as is from a previous review from eanylin | 17:33 |
duonghq | ah, fine | 17:33 |
sdake | duonghq i dont think leaving it as is is correct however | 17:34 |
duonghq | seem that it will be external vip? | 17:34 |
sbezverk | we do not set it, it comes to us from ansible | 17:35 |
sbezverk | actually scratch that | 17:35 |
*** jmccarthy has quit IRC | 17:35 | |
*** unicell has quit IRC | 17:35 | |
sbezverk | it should be set to: tests/conf/iscsi-all-in-one/kolla_config:keystone_public_url: "http://keystone-public:5000/v3" | 17:36 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move ironic config to kolla-k8s https://review.openstack.org/450701 | 17:36 |
sdake | sbezverk thanks :) | 17:36 |
duonghq | sdake, sbezverk thanks | 17:37 |
*** adrian_otto has quit IRC | 17:37 | |
openstackgerrit | Steven Dake proposed openstack/kolla-kubernetes master: Deployment Guide Documentation https://review.openstack.org/447356 | 17:37 |
*** Guest52040 has quit IRC | 17:38 | |
sbezverk | duonghq: if you have questions about certain variables, just do grep -r {var} in kolla-kubernetes repo | 17:39 |
*** adrian_otto has joined #openstack-kolla | 17:39 | |
sbezverk | you will see where this var gets setup in the gate scritps, so you get the idea | 17:39 |
satyar | inc0: not sure how i am getting this http://paste.openstack.org/show/604544/ | 17:39 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move horizon config to kolla-k8s https://review.openstack.org/450725 | 17:39 |
sdake | sbezverk thanks - i'll shceck that file out | 17:40 |
duonghq | sbezverk, roger, I could not find why it's set to the specific IP :P | 17:40 |
sdake | duonghq that came from a prevous review | 17:40 |
duonghq | sdake, sbezverk in my setup, keystone-internal still does not get external-ip | 17:41 |
zhubingbing_ | ;) | 17:42 |
openstackgerrit | Marcus Williams proposed openstack/kolla-ansible master: Split Openvswitch into own role https://review.openstack.org/416793 | 17:42 |
*** tonanhngo has quit IRC | 17:43 | |
deadnull | What is everyone using to bootstrap their metal with CentOS to prep for a kolla deployment? TripleO? Foreman? | 17:45 |
sbezverk | duonghq: http://paste.openstack.org/show/604545/ | 17:46 |
*** mgoddard has joined #openstack-kolla | 17:46 | |
sbezverk | duonghq: here is portion of my cloud.yaml related to keystone, it gives me complete access to keystone from inside as well as outside of cluster | 17:47 |
sdake | deadnull kickstart :) | 17:47 |
sbezverk | keystone-admin 10.57.188.241 192.168.80.232 35357/TCP 8d | 17:47 |
sbezverk | keystone-internal 10.57.168.139 192.168.80.232 5000/TCP 8d | 17:47 |
sbezverk | keystone-public 10.57.186.117 192.168.80.232 5000/TCP 8d | 17:47 |
duonghq | sbezverk, so the dns_name matches external-ip? | 17:47 |
sbezverk | duonghq: it can | 17:47 |
sbezverk | duonghq: but does not have to | 17:48 |
sdake | sbezverk under what circumstances would it not match the external ip? | 17:48 |
sbezverk | duonghq: dns_name might be used if you have external registered dns name you want to use for accessign keystone from outside | 17:48 |
sbezverk | example keystone.sdake.io | 17:49 |
duonghq | sbezverk, understood | 17:49 |
deadnull | sdake thx, yeah im leaning towards just rolling a simple kickstart setup for this purpose, dont want/need something heavyweight | 17:49 |
sbezverk | duonghq: since in my case I do not have this luxury, it matches my cluster external ip | 17:49 |
sdake | deadnull in a diferrent lab, I use bifrost to bootstrap | 17:50 |
sdake | although bifrost usag eis a little complex to setup networking wise | 17:50 |
*** adrian_otto has quit IRC | 17:50 | |
duonghq | sbezverk, last try today and I need sleep :( | 17:51 |
duonghq | thanks | 17:51 |
zhubingbing_ | good night | 17:52 |
zhubingbing_ | i will go to sleep to | 17:52 |
zhubingbing_ | ;) | 17:52 |
sdake | zhubingbing_ enjoy :) | 17:52 |
duonghq | night zhubingbing_ | 17:52 |
zhubingbing_ | bye guys | 17:53 |
duonghq | sbezverk, iirc, I should rebuild helm package after changing global.yaml? | 17:53 |
*** rwallner has joined #openstack-kolla | 17:53 | |
sbezverk | duonghq: you should NOT need to change globals.yaml, actually we do not even use it. All config is supplied by cloud.yaml you feed to helm | 17:54 |
deadnull | yeah i have tripleO rocking now, just bootstrapping the undercloud myself. but I want to be able to turn this over quickly during the initial testing phase. -- now that tripleO stated they're moving towards kolla im digging deeper into what our deployment strategy is going to look like | 17:54 |
duonghq | sbezverk, do we need anything from kolla-kubernetes.yaml? | 17:55 |
duonghq | (ip stuff) | 17:55 |
*** jmontleon has quit IRC | 17:55 | |
deadnull | i guess before i dive into metal I should get this working as an all in one in VMware fusion | 17:55 |
openstackgerrit | zhubingbing proposed openstack/kolla-kubernetes master: Move rabbitmq config to kolla-k8s https://review.openstack.org/450581 | 17:56 |
*** rwallner_ has joined #openstack-kolla | 17:56 | |
sdake | deadnull do you have a public reference for tripleo moving towards kolla? | 17:57 |
sbezverk | duonghq: almost nothing | 17:57 |
*** jmontleon has joined #openstack-kolla | 17:57 | |
*** rwallner has quit IRC | 17:57 | |
sbezverk | duonghq: once final jinja2 bits are gone, then we will not need it any longer | 17:57 |
deadnull | you will have an option as i understand it | 17:58 |
deadnull | https://specs.openstack.org/openstack/tripleo-specs/specs/ocata/containerize-tripleo-overcloud.html | 17:58 |
sdake | sbezverk you man the jinaj2 bits in genconfig? | 17:58 |
deadnull | its proposed | 17:58 |
duonghq | sbezverk, got it | 17:58 |
deadnull | also https://docs.openstack.org/developer/tripleo-docs/containers_deployment/architecture.html | 17:58 |
sbezverk | there are still a couple of modules not converted to helm | 17:58 |
sbezverk | duonghq: basically forget any ANY yaml, other than your cloud.yaml :) | 17:59 |
duonghq | sbezverk, if I understand correctly, external_vip doesn't need to in service-cidr? | 17:59 |
sbezverk | duonghq: nope, service cidr is need only for internal services | 17:59 |
*** kbyrne has quit IRC | 18:00 | |
sdake | deadnull interestingly enough I submitted that review in 2014 - although it has undergone significant change since :) | 18:00 |
*** pbourke_ has quit IRC | 18:00 | |
*** rwallner_ has quit IRC | 18:01 | |
deadnull | sdake no kidding, cool! | 18:01 |
*** zhubingbing_ has quit IRC | 18:01 | |
*** rwallner has joined #openstack-kolla | 18:02 | |
*** athomas has quit IRC | 18:02 | |
openstackgerrit | Marcus Williams proposed openstack/kolla-ansible master: Split Openvswitch into own role https://review.openstack.org/416793 | 18:02 |
*** shardy has quit IRC | 18:03 | |
*** adrian_otto has joined #openstack-kolla | 18:05 | |
openstackgerrit | Merged openstack/kolla stable/newton: Bump openstack_release to 3.0.4 in group_vars https://review.openstack.org/450849 | 18:05 |
*** unicell has joined #openstack-kolla | 18:06 | |
mnaser | is it a big deal if a backported patch is not cherry picked? | 18:08 |
*** pbandark has joined #openstack-kolla | 18:08 | |
mnaser | oh i see hm | 18:08 |
mnaser | https://review.openstack.org/#/c/435687/ | 18:08 |
mnaser | what can i do to help get this merged? | 18:08 |
duonghq | sbezverk, keystone-internal still doesn't work for me, I'll try it later, thanks :) at least, all three keystone-* got external ip | 18:09 |
mnaser | i just got hit by this :( | 18:09 |
*** kbyrne has joined #openstack-kolla | 18:09 | |
duonghq | night guys | 18:13 |
*** duonghq has quit IRC | 18:13 | |
sbezverk | duonghq: check your source file | 18:13 |
sbezverk | by default the script generate url pointing to kyestone-internal | 18:14 |
sbezverk | you need to run that script with ext keyword | 18:14 |
sbezverk | then it will get external ip and build correct authentication url | 18:15 |
*** rwallner has quit IRC | 18:18 | |
mnaser | it looks like it was cherry picked | 18:19 |
mnaser | pbourke do you mind removing your -1 here? https://review.openstack.org/#/c/435687/ | 18:19 |
mnaser | barbican is broken in stable/newton, this was backported to stable/ocata and master. | 18:20 |
*** lamt has joined #openstack-kolla | 18:20 | |
openstackgerrit | Marcus Williams proposed openstack/kolla-ansible master: Split Openvswitch into own role https://review.openstack.org/416793 | 18:21 |
*** adrian_otto has quit IRC | 18:21 | |
*** shashank_t_ has quit IRC | 18:22 | |
*** signed8b_ is now known as signed8bit_Zzz | 18:23 | |
*** rwallner has joined #openstack-kolla | 18:28 | |
*** srwilkers has quit IRC | 18:32 | |
*** caoyuan has joined #openstack-kolla | 18:33 | |
*** pbandark has quit IRC | 18:34 | |
*** srwilkers has joined #openstack-kolla | 18:37 | |
*** caoyuan has quit IRC | 18:37 | |
*** srwilkers has quit IRC | 18:41 | |
*** new2kolla has joined #openstack-kolla | 18:46 | |
*** deadnull_ has joined #openstack-kolla | 18:48 | |
*** deadnull has quit IRC | 18:49 | |
*** srwilkers has joined #openstack-kolla | 18:49 | |
fooliouno | sdake: how were you able to solve your nova-compute crashloop. All my pods are fine except for nova-compute | 18:49 |
*** signed8bit_Zzz is now known as signed8b_ | 18:49 | |
*** lamt has quit IRC | 18:49 | |
*** kbyrne has quit IRC | 18:50 | |
*** rwallner has quit IRC | 18:51 | |
*** kbyrne has joined #openstack-kolla | 18:52 | |
*** rwallner has joined #openstack-kolla | 18:52 | |
*** lamt has joined #openstack-kolla | 18:55 | |
kfox1111 | nice. jascott1's parent values patch set works with nova-api-deployment switched over. :) https://review.openstack.org/#/c/449826/ | 18:56 |
kfox1111 | so, once it merges and 2.3 is cut, we shoudl be good to go. :) | 18:57 |
kfox1111 | https://review.openstack.org/#/c/450423/3 shoudl be ready | 19:02 |
sdake | fooliouno lindeed | 19:04 |
sdake | fooliouno current docs get things out of a crashloop | 19:04 |
sdake | kfox1111 gate wasn't green - i had lready +2'ed it prior to the gate runs - although i just ran a recheck | 19:06 |
esharao | sdake do we need to clean up terminated containers? | 19:08 |
sdake | esharao would be nice if kubernetes had GC | 19:09 |
esharao | :) | 19:09 |
sdake | esharao but no you dont need ot clean up terminated containers if you don't mind your etcd storage growing large | 19:09 |
kfox1111 | it shoudl gc. | 19:09 |
esharao | its just im on the dashboard (maybe we need to add as optional in the doc) and it shows a lot of terminated containers | 19:09 |
sdake | esharao i'm not sure in this case who should be responsible for gc - helm or kubernetes | 19:10 |
esharao | *pods | 19:10 |
kfox1111 | it sometimes keeps references to some things though in unexpected ways. | 19:10 |
kfox1111 | like it leaves a job container around to let you look at the logs. | 19:10 |
sdake | kfox1111 right - hence the responsibility question | 19:10 |
sdake | kubernetes intentionally likely leaves that around exactly for the use case you described | 19:10 |
sdake | hwoever if someone does a helm delete --purge they likely want their cluster gced | 19:11 |
kfox1111 | helm delete purge should clean it up I think? | 19:11 |
*** signed8bit has joined #openstack-kolla | 19:12 | |
*** matrohon has joined #openstack-kolla | 19:13 | |
*** signed8b_ has quit IRC | 19:15 | |
*** signed8b_ has joined #openstack-kolla | 19:15 | |
*** signed8bit has quit IRC | 19:18 | |
sdake | kfox1111 not sure it does - although not sure it doesn't | 19:18 |
sdake | kfox1111 i've tried so many things locally I've lost track :) | 19:18 |
kfox1111 | the finished job container stays around as long as the job does. | 19:18 |
kfox1111 | purging the release should delete the job. | 19:18 |
kfox1111 | which will delete the container. | 19:19 |
*** shashank_t_ has joined #openstack-kolla | 19:19 | |
sdake | kubernetes on deltion of a job deletes the container from docker ps -a? | 19:19 |
kfox1111 | I believe so. | 19:19 |
sdake | cool | 19:19 |
sdake | i'll see if I can validate that - although my list of validations is growing exponentially :) | 19:19 |
kfox1111 | I think there is a blueprint for it caching logs, so it can delete the container when actually done, | 19:20 |
kfox1111 | rather hten keep it around so that kubectl logs finished-job still works. | 19:20 |
*** mgoddard has quit IRC | 19:21 | |
sbezverk | kfox1111: I do not like the idea of deleting jobs, because with them you lose all logs for that task, it makes harder to troubleshoot.. | 19:21 |
sdake | sbezverk jobs only delete on a helm purge | 19:21 |
sdake | which is ok - since you are purging the install | 19:22 |
openstackgerrit | Merged openstack/kolla-ansible master: [deploy-guide] Moving the quick start to the deploy folder https://review.openstack.org/448509 | 19:22 |
sdake | sbezverk is it ok to make the ironic gate non-experimental? | 19:23 |
kfox1111 | sbezverk: if those logs are important, they should be saved somewhere less transient then k8s. | 19:23 |
kfox1111 | sdake: the centos ironic gate is ready I think. | 19:23 |
sbezverk | sdake: no, ubuntu is not working properly | 19:23 |
sdake | sbezverk ok - let me know and ill submit a change for it - or alternately you can | 19:24 |
sbezverk | and I do not see any close solutions | 19:24 |
sdake | sbezverk i know ubuntu is troublesome for ironic atm | 19:24 |
sdake | sbezverk do you hae any idea what is wrong - perhaps I can hep? | 19:24 |
sbezverk | I do have idea of what is wrong but I need Ubuntu/libvert expert to help. I ahev already filed the bug for ubuntu and they are looking into it | 19:26 |
sdake | sbezverk i'm not a libvrt expert but have a good understanding of libvirt in a container envrionment (expert level) | 19:29 |
sdake | sbezverk my general guess is its some sort of permissions issue | 19:29 |
sdake | since we downgraded livirt to 1.3.1 in ubuntu however centos runs 2.5 just fine | 19:30 |
sbezverk | sdake" that much even I can say ;) | 19:30 |
sdake | libvirt in ubuntu behaves differently - reads different files | 19:30 |
sdake | sbezverk if its a permission error point me at the log | 19:30 |
sdake | sbezverk i can help | 19:30 |
sdake | sbezverk i've sepnt 3 years of my flife jerking aroudn getting libvirt to run in containers with nova :) | 19:31 |
sdake | with all the various version changes that happen along the way | 19:31 |
sdake | thousands of hours spent looking at these problems | 19:31 |
sbezverk | sdake: here you go: http://logs.openstack.org/03/449803/25/experimental/gate-kolla-kubernetes-deploy-ubuntu-source-3-ironic-nv/1025efe/console.html#_2017-03-28_05_34_30_657355 | 19:31 |
vhosakot | sdake: what is the 4 finger magic on mac you're talking about? :) easy scrolling using four fingers? | 19:32 |
sdake | pid file denied | 19:32 |
sdake | vhosakot you can put windows in the top so they take up a full scren | 19:32 |
*** lamt has quit IRC | 19:32 | |
kfox1111 | " | 19:32 |
kfox1111 | [kubernetes-announce] k8s v1.6.0 is live! | 19:32 |
kfox1111 | " | 19:32 |
sdake | then use four fingers to scroll across yoru desktop | 19:32 |
vhosakot | sdake: ah ok cool | 19:32 |
sdake | kfox1111 BRING THE PAIN | 19:32 |
kfox1111 | YEAH. BRING IT! :) | 19:33 |
* kfox1111 jumps up and down excitedly | 19:33 | |
kfox1111 | ok... enough redbull for me.... | 19:33 |
* sdake crawls up in a ball on the floor | 19:33 | |
*** jmccarthy has joined #openstack-kolla | 19:34 | |
kfox1111 | hehe. | 19:35 |
sdake | sbezverk i've got roughly one hour before i have torture session at dentist prep | 19:35 |
*** jmccarthy has quit IRC | 19:35 | |
kfox1111 | 5000 nodes. nice. :) | 19:35 |
sdake | sbezverk if your up for a webex, I can guide you through what I know | 19:35 |
*** iceyao has joined #openstack-kolla | 19:35 | |
sdake | sbezverk as in how to debug further | 19:35 |
sdake | sbezverk this really sin't an issue with libvirt - more of a cnofig issue i believe | 19:36 |
sdake | why we are not using 2.5 I don't know | 19:36 |
sdake | but I suspect we need conditionals around libvirt for ubuntu vs centos | 19:36 |
sbezverk | sdake: we are using 2.5 everywhere now, same results | 19:37 |
sdake | sbezverk i thought libvirt was on 1.3.1 on ubuntu? | 19:37 |
sbezverk | nope it turned out the host was using 1.3.1 but containers were using 2.5 | 19:38 |
sbezverk | now I have the same version on both same result | 19:38 |
sdake | are you doing hthsi in your local en or in the gate? | 19:39 |
sbezverk | sdake: unfortunately I do not have time for webex | 19:39 |
sdake | sbezverk ok | 19:39 |
sbezverk | I will wait input from Ubuntu and it is at the gate | 19:39 |
*** lamt has joined #openstack-kolla | 19:39 | |
sdake | sbezverk well here are the things you want to capture | 19:39 |
sdake | in the gate | 19:39 |
sdake | yo uwant to capture all contents of /etc/libvirt | 19:39 |
sdake | you can to capture ls -lR of /etc/libvirt | 19:40 |
sdake | can/want | 19:40 |
sdake | you want to detmerine where that pid file is being written | 19:40 |
*** iceyao has quit IRC | 19:40 | |
sdake | you want to determine the permissions of the dirctory wheret he pid file is being written | 19:40 |
sdake | you can do all of this in the gate | 19:40 |
sdake | can you do that - and I can provide further analysis | 19:41 |
sdake | to determine where pid file is written yo uwant to run continer with strace | 19:41 |
sdake | ubuntu upstream wont really help here for this problem | 19:41 |
sdake | they don't have th domain expertise | 19:42 |
sdake | this is really a kolla expert problem not a libvirt expert problem | 19:42 |
sdake | sbezverk if its available recommend watcing rogue one | 19:42 |
sdake | star wars movie | 19:43 |
sdake | not sure if its in .ca yet | 19:43 |
sdake | sbezverk you may enjoy it | 19:43 |
sbezverk | I will but very little time | 19:43 |
sbezverk | available, got ot run.. | 19:43 |
sdake | sbezverk ttyl - thanks for the hep today on the dns stuff | 19:43 |
sdake | sbezverk that i think closed out the questions I had for the most part | 19:44 |
SamYaple | sbezverk: just fyi, if im reading that correctly, you cant be running libvirt on the host AND in the container. thats explicity forbidden | 19:44 |
sbezverk | SamYaple: libvirtd runs only in the cotnainer | 19:45 |
sbezverk | host uses a socket created by the container | 19:45 |
sbezverk | to communicate | 19:45 |
SamYaple | sbezverk: ah. i see. yea the best way to do that is with `docker exec` | 19:46 |
SamYaple | personally, i setup an alias `alias virsh='docker exec -it nova-libvirt virsh'` | 19:46 |
SamYaple | working with two versions of libvirt was always a nightmare, so now it doesnt live on my host at all | 19:46 |
sdake | sbezverk pretty sure i nkolla-kubenetes we decided sockets didn't work well for comms and went to tcp model | 19:46 |
sbezverk | SamYaple: that will not work, because virtual bmc communicates with libvirt via socket/pid | 19:47 |
sbezverk | and virtual bmc run only on host | 19:47 |
SamYaple | sbezverk: ok cool. i wasn't sure of the full scenario of what you are asking before. glad you have all the details in line now | 19:47 |
sdake | sbezverk virtualbmc is writing the pidfile for libvirt? | 19:47 |
sbezverk | sdake: how virtual bmc can write pid file for libvirt? | 19:49 |
sdake | sbezverk not sure - igot impression virtual bmc communicates with libvirt | 19:49 |
sdake | the problem is libvirt can't write the pidfile | 19:49 |
sbezverk | sdake: it uses it and socket to communicate with it | 19:49 |
sbezverk | sdake: atm I am not at that stage yet | 19:50 |
sbezverk | right now I cannot even define VM via virsh | 19:50 |
sbezverk | which is perfectly fine with centos | 19:50 |
sdake | sbezverk ok - cool so no nova integration problems | 19:50 |
sdake | need ls -ld on directory wehre pidfile is being written | 19:51 |
sdake | and ls -lR on directory where pidfile is being written | 19:51 |
sdake | as well as all of the other stuff above i requested :) | 19:51 |
sdake | get me the information - i can diagnose from there | 19:51 |
kfox1111 | yay. init-containers graduated! :) | 19:52 |
* kfox1111 does a happy dance | 19:53 | |
kfox1111 | no more json encoded strings. :) | 19:53 |
sdake | cool | 19:54 |
sdake | kfox1111 does that rid of us all the json? | 19:54 |
kfox1111 | yeah. | 19:55 |
* sdake does happy jig :) | 19:55 | |
*** alop has quit IRC | 19:55 | |
kfox1111 | wil make the templates sooo much nicer to look at. :) | 19:55 |
kfox1111 | meeting next. bbl. | 19:55 |
sdake | gott ajet as well - ttyl | 19:55 |
sdake | BB-in8 hrs or so :) | 19:56 |
*** manheim has joined #openstack-kolla | 20:01 | |
*** manheim has quit IRC | 20:02 | |
*** adrian_otto has joined #openstack-kolla | 20:08 | |
*** srwilkers has quit IRC | 20:09 | |
*** eanylin has quit IRC | 20:17 | |
*** eanylin has joined #openstack-kolla | 20:17 | |
*** rwallner has quit IRC | 20:21 | |
*** rwallner_ has joined #openstack-kolla | 20:27 | |
*** rwallner_ has quit IRC | 20:27 | |
*** rwallner has joined #openstack-kolla | 20:31 | |
*** rwallner has quit IRC | 20:36 | |
*** satyar_ has joined #openstack-kolla | 20:49 | |
*** satyar has quit IRC | 20:52 | |
*** sambetts is now known as sambetts|afk | 20:55 | |
*** rhallisey has quit IRC | 20:55 | |
*** matrohon has quit IRC | 21:02 | |
*** haplo37 has quit IRC | 21:12 | |
*** vhosakot_ has joined #openstack-kolla | 21:20 | |
*** haplo37 has joined #openstack-kolla | 21:21 | |
*** vhosakot has quit IRC | 21:22 | |
*** nathharp has joined #openstack-kolla | 21:22 | |
*** jrobinson has joined #openstack-kolla | 21:24 | |
*** esharao has quit IRC | 21:28 | |
*** vhosakot_ has quit IRC | 21:28 | |
*** vhosakot has joined #openstack-kolla | 21:30 | |
*** satyar_ has quit IRC | 21:30 | |
*** nathharp has quit IRC | 21:35 | |
*** sayantan_ has quit IRC | 21:36 | |
*** gfidente|afk has quit IRC | 21:38 | |
*** fooliouno has quit IRC | 21:45 | |
*** new2kolla has quit IRC | 21:46 | |
*** ipsecguy has quit IRC | 21:47 | |
*** adrian_otto has quit IRC | 21:53 | |
*** unicell has quit IRC | 21:56 | |
*** unicell has joined #openstack-kolla | 22:01 | |
kfox1111 | yay. service token less containers made it to 1.6 too. | 22:07 |
kfox1111 | automountServiceAccountToken field. | 22:07 |
kfox1111 | we probably should set that liberally on our pods. | 22:07 |
*** iceyao has joined #openstack-kolla | 22:16 | |
*** shashank_t_ has quit IRC | 22:16 | |
*** shashank_t_ has joined #openstack-kolla | 22:17 | |
*** iceyao has quit IRC | 22:20 | |
*** shashank_t_ has quit IRC | 22:22 | |
*** manheim has joined #openstack-kolla | 22:22 | |
kfox1111 | sbezverk: https://github.com/kubernetes/kubernetes/pull/41597 | 22:37 |
kfox1111 | not sure how much I trust it. but its there. | 22:37 |
sbezverk | kfox1111: cool, I hope dns fix is also in.. | 22:38 |
*** eanylin has quit IRC | 22:38 | |
*** shashank_t_ has joined #openstack-kolla | 22:38 | |
*** eanylin has joined #openstack-kolla | 22:39 | |
*** schwicht has quit IRC | 22:39 | |
kfox1111 | sbezverk: havent seen a changelog note for it yet, and I'm almost through the list. | 22:39 |
kfox1111 | :/ | 22:39 |
sbezverk | hm, I was told that it would make it, as there was big demand.. | 22:41 |
kfox1111 | maybe they forgot the release note. | 22:42 |
*** shashank_t_ has quit IRC | 22:43 | |
*** signed8b_ is now known as signed8bit_Zzz | 22:44 | |
*** lamt has quit IRC | 22:46 | |
* kfox1111 hopes 1.6 shows up in the rpm repo soon... | 22:47 | |
*** vhosakot has quit IRC | 22:49 | |
*** david-lyle has quit IRC | 22:51 | |
*** manheim has quit IRC | 22:51 | |
*** shashank_t_ has joined #openstack-kolla | 22:58 | |
*** lamt has joined #openstack-kolla | 23:00 | |
kfox1111 | sbezverk: it woudl be interesting to adapt this for rbd support too: https://github.com/kubernetes-incubator/external-storage/blob/master/ceph/cephfs/cephfs-provisioner.go | 23:01 |
kfox1111 | wouldn't need a custom build apiserver then. | 23:02 |
*** schwicht has joined #openstack-kolla | 23:02 | |
*** caoyuan has joined #openstack-kolla | 23:03 | |
kolla-slack | <sbezverk> what do you mean custom build api server | 23:09 |
kolla-slack | <sbezverk> ? | 23:09 |
*** adrian_otto has joined #openstack-kolla | 23:16 | |
*** lamt has quit IRC | 23:17 | |
kfox1111 | sbezverk: if we wanted to support dynamic provisioning of rbd volumes, the only way today is to have a rbd client in the kube-apiserver contaienr I think. | 23:18 |
kfox1111 | which is why it got added to the kolla images. | 23:18 |
sbezverk | Ah, no it is not in apiserver, it is in controller-manager container | 23:19 |
kfox1111 | but with this new, external dynamic provisioner functionality, it could be split out. | 23:19 |
kfox1111 | ah. ok. | 23:19 |
kfox1111 | either way. it decouples ceph from kube, which is a great thing. :) | 23:19 |
*** schwicht has quit IRC | 23:26 | |
*** schwicht has joined #openstack-kolla | 23:30 | |
*** dave-mccowan has quit IRC | 23:31 | |
*** schwicht has quit IRC | 23:33 | |
*** rwellum has quit IRC | 23:41 | |
*** erlon has joined #openstack-kolla | 23:43 | |
spsurya | morning guys | 23:48 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!