*** yingjun has joined #openstack-kolla | 00:01 | |
srwilkers | gerrit acting weird for anyone else? trying to leave comments on a review and its just hanging | 00:03 |
---|---|---|
*** harlowja has quit IRC | 00:06 | |
*** ayoung has quit IRC | 00:08 | |
kfox1111 | srwilkers: working for me right now. | 00:08 |
kfox1111 | maybe transient issue? | 00:09 |
srwilkers | kfox1111, yeah. cleared up. jascott1 may have multiple comments in his heat engine review though :p | 00:12 |
*** Pavo has quit IRC | 00:12 | |
jascott1 | hold off plz I am converting to common | 00:12 |
jascott1 | common_statefulset that is | 00:13 |
*** Pavo has joined #openstack-kolla | 00:13 | |
srwilkers | :) | 00:15 |
portdirect | srwilkers: FYI, we use mage_full for consistency | 00:18 |
*** portdirect is now known as portdirect_away | 00:18 | |
*** portdirect_away has quit IRC | 00:19 | |
*** gomarivera has joined #openstack-kolla | 00:25 | |
*** Jeffrey4l_ is now known as Jeffrey4l | 00:28 | |
*** l4yerffeJ__ is now known as l4yerffeJ | 00:28 | |
*** lrensing has joined #openstack-kolla | 00:30 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for cinder keystone endpoints https://review.openstack.org/410923 | 00:30 |
*** yingjun has quit IRC | 00:31 | |
*** yingjun has joined #openstack-kolla | 00:32 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm chart for cinder-scheduler https://review.openstack.org/410525 | 00:33 |
*** harlowja has joined #openstack-kolla | 00:34 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance api and glance registry pods https://review.openstack.org/410481 | 00:42 |
*** ayoung has joined #openstack-kolla | 00:54 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Change bootstrap delete invocations to helm delete instead https://review.openstack.org/411548 | 00:56 |
*** saneax is now known as saneax-_-|AFK | 00:58 | |
*** jianyi has joined #openstack-kolla | 00:59 | |
*** sdake_ has quit IRC | 01:00 | |
*** unicell has quit IRC | 01:04 | |
*** tovin07_ has joined #openstack-kolla | 01:09 | |
*** tovin07_ has quit IRC | 01:11 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron dhcp/metadata agents. https://review.openstack.org/411493 | 01:13 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for cinder keystone endpoints https://review.openstack.org/410923 | 01:13 |
*** zhubingbing_ has joined #openstack-kolla | 01:13 | |
*** sdake has joined #openstack-kolla | 01:14 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: [wip] collect logs https://review.openstack.org/411403 | 01:15 |
*** tovin07_ has joined #openstack-kolla | 01:16 | |
*** huzhengchuan has joined #openstack-kolla | 01:16 | |
*** zhangqiankun has joined #openstack-kolla | 01:16 | |
*** duonghq has joined #openstack-kolla | 01:16 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance api and glance registry pods https://review.openstack.org/410481 | 01:16 |
duonghq | morning guys | 01:16 |
srwilkers | morning duonghq :) | 01:17 |
duonghq | sup srwilkers | 01:17 |
srwilkers | not much duonghq, how are you? | 01:17 |
duonghq | I'm finding my way to the Internet | 01:18 |
*** HyperJohnGraham has joined #openstack-kolla | 01:19 | |
*** unicell has joined #openstack-kolla | 01:20 | |
srwilkers | duonghq, did you lose it? ;) | 01:22 |
duonghq | srwilkers, somebody break the highway | 01:23 |
srwilkers | duonghq, i hate it when that happens | 01:24 |
Pavo | wow epel repo mirrors in kolla-build are horribad | 01:24 |
Pavo | every single mirror is 404 | 01:25 |
duonghq | srwilkers, agree with you, | 01:25 |
*** harlowja has quit IRC | 01:25 | |
duonghq | Pavo: it's fault of bestmirror plugins | 01:26 |
Pavo | how to fix? | 01:26 |
duonghq | Pavo, I have to use syncrepo of CentOS to make local repo | 01:26 |
duonghq | and add it to kolla-build repo | 01:26 |
Pavo | how do you do that | 01:26 |
duonghq | or you can select manually and override default repo | 01:27 |
duonghq | I use template-override | 01:27 |
duonghq | with header override and copy repo file to the base image | 01:27 |
Pavo | I tried adding the epel.repo in the docker/base/ directory and adding rpm_setup_config = epel.repo to /etc/kolla/kolla-build.conf and it adds it but the GPG key doesn't match | 01:28 |
Pavo | not sure what you mean duonghq | 01:28 |
duonghq | Pavo, disable the gpg check in your custom repo file | 01:29 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron dhcp/metadata agents. https://review.openstack.org/411493 | 01:29 |
openstackgerrit | Justin Scott proposed openstack/kolla-kubernetes: WIP Helmize Heat-engine https://review.openstack.org/410048 | 01:29 |
Pavo | ok will try that duonghq | 01:29 |
*** jtriley has joined #openstack-kolla | 01:30 | |
duonghq | Pavo, I override the epel.repo which is installed when epel package is installed with template override | 01:30 |
Pavo | mind sharing your template override? | 01:30 |
Pavo | and where does it go? | 01:30 |
duonghq | Pavo, I have meeting right now, brb in ~30mins. | 01:31 |
duonghq | sorry guy | 01:31 |
Pavo | ok np | 01:31 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm-ize Cinder api pod https://review.openstack.org/411556 | 01:31 |
*** zhangqiankun has quit IRC | 01:33 | |
*** sdake has quit IRC | 01:35 | |
*** sdake has joined #openstack-kolla | 01:38 | |
jascott1 | sdake ping | 01:43 |
*** jianyi has quit IRC | 01:43 | |
Pavo | if a image build says "binary not yet available for centos" do I build it from source? | 01:45 |
*** jianyi has joined #openstack-kolla | 01:47 | |
openstackgerrit | Justin Scott proposed openstack/kolla-kubernetes: WIP Helmize Heat-engine https://review.openstack.org/410048 | 01:48 |
*** JRobinson__ is now known as JRobinson__afk | 01:48 | |
jascott1 | srwilkers that was embarrassing :) | 01:49 |
*** v1k0d3n has quit IRC | 01:50 | |
jascott1 | srwilkers kfox1111 thought it should be a stateful | 01:51 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm chart for cinder-scheduler https://review.openstack.org/410525 | 01:51 |
*** zhubingbing_ has quit IRC | 01:52 | |
srwilkers | srwilkers, it happens man. i cant tell you how many times ive left something from a copy/paste in | 01:52 |
srwilkers | oops jascott1 ^ | 01:52 |
*** zhubingbing_ has joined #openstack-kolla | 01:52 | |
jascott1 | ya had an older imageBuilt line that was foobard | 01:52 |
jascott1 | srwilkers are you in NC? | 01:53 |
srwilkers | jascott1, nope. im in STL unfortunately | 01:53 |
jascott1 | been there a couple times :) | 01:54 |
jascott1 | to see concerts mostly | 01:54 |
srwilkers | jascott1, oh cool. yeah, its a pretty boring place. hoping to move away at some point, but we'll see | 01:54 |
jascott1 | im originally from SE but really dig the PNW | 01:55 |
jascott1 | it sells itself so I wont go on a rant ;) | 01:55 |
*** sayantan_ has joined #openstack-kolla | 01:56 | |
srwilkers | haha. ive had my eye on either the PNW or NC | 01:56 |
Pavo | when is the new fixes for murano gonna be added to pip package? | 01:56 |
srwilkers | NC has the benefit of a similar cost of living to saint louis | 01:56 |
*** tonanhngo has quit IRC | 01:57 | |
jascott1 | I lived in Western NC and its great for the SE but just try the west coast before you decide | 01:58 |
*** sdake has quit IRC | 02:00 | |
*** tonanhngo has joined #openstack-kolla | 02:01 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP-PoC Glance service orchestration with entrypoint https://review.openstack.org/410736 | 02:01 |
*** mdnadeem has joined #openstack-kolla | 02:03 | |
Jeffrey4l | morning guys | 02:04 |
*** unicell has quit IRC | 02:05 | |
*** tonanhngo has quit IRC | 02:05 | |
sbezverk | hi Jeffrey4l | 02:06 |
jascott1 | Good morning Jeffrey4l | 02:06 |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 02:07 |
*** mdnadeem has quit IRC | 02:07 | |
Jeffrey4l | hey sbezverk | 02:08 |
Jeffrey4l | sup jascott1 | 02:08 |
srwilkers | hey Jeffrey4l :) | 02:08 |
kfox1111 | hi | 02:08 |
sbezverk | it is cool, kolla IRC channel is up for 24 hours :-) | 02:08 |
Jeffrey4l | yep. | 02:09 |
srwilkers | sbezverk, thats what i like about it :) | 02:09 |
srwilkers | always people to chat with when im feeling the insomnia | 02:09 |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 02:09 |
duonghq | Pavo, around? | 02:11 |
sbezverk | srwilkers: here is one thing that will give you insomnia for sure. I run helm install --dry-run --debug for one chart and from time to time rendering generates different results. Try to explain me that :-) | 02:11 |
*** tovin07__ has joined #openstack-kolla | 02:11 | |
Pavo | yeah I'm here duonghq looks like removing the GPG check is working | 02:12 |
sbezverk | srwilkers: I just do Arrow Up Enter, wihtout changing anything | 02:12 |
*** Pavo has quit IRC | 02:12 | |
*** Pavo has joined #openstack-kolla | 02:12 | |
duonghq | Pavo, nice | 02:13 |
Pavo | duonghq if a image build says binary not yet available for centos can the image be built from source and still used from centos base? | 02:13 |
*** Pavo has quit IRC | 02:13 | |
*** Pavo has joined #openstack-kolla | 02:14 | |
kfox1111 | sbezverk: is it just an orderign thing? | 02:14 |
kfox1111 | sbezverk: cause yaml by default doesn't garantee ordering. | 02:14 |
duonghq | Pavo, you can build, packaging it and put in your local registry | 02:14 |
Pavo | from source right? | 02:14 |
Pavo | because it keeps saying there is no bifrost image made for centos yet, and was wanting to try it out | 02:15 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm Cinder Backup https://review.openstack.org/411570 | 02:15 |
duonghq | Pavo, sorry I lost the context | 02:15 |
duonghq | can you describe? | 02:15 |
srwilkers | sbezverk, man, that would drive me crazy trying to figure that out | 02:15 |
Pavo | I am doing a kolla-build and some of the images its trying to build say binary not yet available for centos because I am building from binary and not source, my question is can I rebuild those images from source only? | 02:16 |
sbezverk | kfox1111: do not think it is ordering because value changes only if other variable is empty | 02:17 |
*** tonanhngo has joined #openstack-kolla | 02:17 | |
sbezverk | is you check PS I just pushed | 02:17 |
sbezverk | value rendered for "name" changes between "glance" and "glance-test" | 02:18 |
kfox1111 | sbezverk: whats the url? | 02:18 |
openstackgerrit | sayantani proposed openstack/kolla: Clean up kolla-ansible related files from Kolla https://review.openstack.org/406429 | 02:18 |
sbezverk | kfox1111: https://review.openstack.org/410736 | 02:19 |
kfox1111 | sbezverk: thx | 02:20 |
*** sayantan_ has quit IRC | 02:20 | |
*** sayantan_ has joined #openstack-kolla | 02:20 | |
*** tonanhngo has quit IRC | 02:21 | |
kfox1111 | looking... | 02:22 |
openstackgerrit | Merged openstack/kolla: Move openvswitch db file into docker volume https://review.openstack.org/410733 | 02:22 |
*** sdake has joined #openstack-kolla | 02:22 | |
srwilkers | hey sdake | 02:23 |
kfox1111 | sbezverk: ok. so, your seeing name change when just rerunning the same command, such as: helm install --debug --dry-run /tmp/glance-3.0.0-1.tgz | 02:26 |
kfox1111 | ? | 02:26 |
*** gomarivera has quit IRC | 02:26 | |
kfox1111 | I'm seeing the ordering change between pvc and pv and visa versa. | 02:26 |
sbezverk | kfox1111: no I do not use tar I just run helm install --dry-run --debug helm/service/glance | 02:30 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Move openvswitch db file into docker volume https://review.openstack.org/411575 | 02:30 |
kfox1111 | sbezverk: shouldn't matter beetween chart dir and tarball? | 02:30 |
*** Jeffrey4l has quit IRC | 02:31 | |
*** Jeffrey4l has joined #openstack-kolla | 02:31 | |
kfox1111 | sbezverk: did you make any changes to the common pv file? | 02:31 |
kfox1111 | I see you passing pvName. | 02:31 |
*** Jeffrey4l has quit IRC | 02:33 | |
*** Jeffrey4l has joined #openstack-kolla | 02:33 | |
sbezverk | kfox1111: I use common_pv common_pvc from glance/template, I moved them there so I could easily edit them.. | 02:33 |
kfox1111 | yeah. just wondering if there were any other edits then just the changed files you checked in. | 02:33 |
sbezverk | kfox1111: see this paste http://paste.openstack.org/show/592569/ | 02:34 |
*** sayantan_ has quit IRC | 02:35 | |
*** sayantan_ has joined #openstack-kolla | 02:35 | |
kfox1111 | sbezverk: k. that looks really bad. :/ | 02:36 |
kfox1111 | did you make any chages to common_pv? | 02:36 |
kfox1111 | I'm not sure how you could get that with just useing the checked in _common_pv.yaml | 02:36 |
kfox1111 | the other thing to check, is what helm version? | 02:36 |
kfox1111 | with the version of common_pv in trunk, I get metadata: name: glance always. | 02:37 |
openstackgerrit | sayantani proposed openstack/kolla-ansible: Fix kolla-ansible gate https://review.openstack.org/411587 | 02:37 |
openstackgerrit | sayantani proposed openstack/kolla-ansible: Fix kolla-ansible gate https://review.openstack.org/411587 | 02:39 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for cinder keystone endpoints https://review.openstack.org/410923 | 02:40 |
*** yingjun has quit IRC | 02:40 | |
sbezverk | kfox1111: I run 2.1.0 | 02:41 |
sbezverk | I will play more with this tomorrow and engage helm people, that should make them to scratch their heads ;-) | 02:41 |
openstackgerrit | sayantani proposed openstack/kolla: Clean up kolla-ansible related files from Kolla https://review.openstack.org/406429 | 02:41 |
sbezverk | I am off for now, have a great evening gents, ttyl | 02:42 |
kfox1111 | I've seen that kind of thing before, | 02:42 |
kfox1111 | when I defined a var in one tmeplate, then used it in another. | 02:42 |
kfox1111 | dependong on template ordering, saw different things. | 02:42 |
*** sdake has quit IRC | 02:42 | |
kfox1111 | but never seen it when the var was defined in one and a macro called from another. | 02:42 |
kfox1111 | sbezverk: if you could post your common_pv before you go, that would be great. | 02:43 |
kfox1111 | sbezverk: have a good night. | 02:43 |
*** sdake has joined #openstack-kolla | 02:43 | |
kfox1111 | if I had to guess, I'd say you were initing a var outside of the macro? | 02:45 |
sayantan_ | Hi guys | 02:47 |
sayantan_ | srwilkers made the changes you suggested to the quickstart guide | 02:47 |
srwilkers | sayantan_, awesome. ill take a look | 02:47 |
sayantan_ | Amazing | 02:47 |
srwilkers | sbezverk, good night :) | 02:47 |
sayantan_ | Thanks srwilkers for reviewing it over and over again | 02:48 |
sayantan_ | Really appreciate your effort. It takes a lot of patience when it comes to docs :) | 02:48 |
srwilkers | sayantan_, happy to help. :) not something to thank me for | 02:48 |
sayantan_ | haha. Your contribution to it has made it better :) | 02:49 |
sayantan_ | I know you are about to say that's what you are here for | 02:49 |
sayantan_ | :P | 02:49 |
*** sayantan_ has quit IRC | 02:49 | |
*** eaguilar has joined #openstack-kolla | 02:49 | |
*** sayantan_ has joined #openstack-kolla | 02:50 | |
*** zhangqiankun has joined #openstack-kolla | 02:50 | |
srwilkers | sayantan_, looks good to me | 02:50 |
sayantan_ | great :) | 02:50 |
sayantan_ | Thanks | 02:51 |
*** yingjun has joined #openstack-kolla | 02:51 | |
*** zhangqiankun has quit IRC | 02:53 | |
*** zhangqiankun has joined #openstack-kolla | 02:53 | |
*** tovin07_ is now known as tovin07 | 02:58 | |
kfox1111 | arg.... | 02:59 |
kfox1111 | helms debugging rally leaves much to be desired sometimes.... | 02:59 |
*** sayantan_ has quit IRC | 03:00 | |
*** JRobinson__afk is now known as JRobinson__ | 03:00 | |
*** sayantan_ has joined #openstack-kolla | 03:00 | |
openstackgerrit | jianyi proposed openstack/kolla: Add zun container https://review.openstack.org/410616 | 03:00 |
kfox1111 | ah. found it. | 03:00 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize neutron dhcp/metadata agents. https://review.openstack.org/411493 | 03:01 |
openstackgerrit | jianyi proposed openstack/kolla: Add zun container https://review.openstack.org/410616 | 03:03 |
*** sdake has quit IRC | 03:03 | |
*** fragatina has quit IRC | 03:07 | |
*** fragatina has joined #openstack-kolla | 03:08 | |
*** fragatina has quit IRC | 03:12 | |
*** fragatina has joined #openstack-kolla | 03:12 | |
jianyi | Please everybody review this https://review.openstack.org/#/c/410616/ Good afternoon, everyone. Thank you all haha | 03:13 |
jianyi | :) | 03:14 |
*** fragatin_ has joined #openstack-kolla | 03:15 | |
*** fragatina has quit IRC | 03:15 | |
*** sdake has joined #openstack-kolla | 03:16 | |
*** fragatin_ has quit IRC | 03:18 | |
*** fragatina has joined #openstack-kolla | 03:19 | |
*** tyrola has quit IRC | 03:20 | |
*** sayantan_ has quit IRC | 03:22 | |
*** sayantan_ has joined #openstack-kolla | 03:23 | |
*** fragatina has quit IRC | 03:24 | |
*** tyrola has joined #openstack-kolla | 03:29 | |
*** dupengfei has joined #openstack-kolla | 03:30 | |
dupengfei | FOR Help~, helm can't work, When doing "helm ls", the tiller crashes, and helm report : "Error: transport is closing" | 03:31 |
dupengfei | kubernetes: 1.4.6 helm: 2.0.2 tiller: 2.0.1 | 03:32 |
kfox1111 | crashes or isn't available yet? | 03:32 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance keystone endpoints https://review.openstack.org/411607 | 03:32 |
kfox1111 | does: | 03:32 |
dupengfei | At first ,tiller is Running, and when i do helm ls , it crashes | 03:33 |
kfox1111 | kubectl get pods --namespace kube-system | 03:33 |
kfox1111 | show it ready? | 03:33 |
dupengfei | tiller-deploy-3023369516-8oup6 1/1 Running 3 8m | 03:33 |
srwilkers | ive heard others having this same issue today | 03:33 |
srwilkers | havent heard if theyve got resolution yet though | 03:33 |
kfox1111 | and when you run it again, it incremetns to 4? | 03:33 |
*** huzhengchuan has quit IRC | 03:33 | |
dupengfei | Yes | 03:33 |
dupengfei | What 's your version of kubernetes and helm ? | 03:34 |
srwilkers | dupengfei, id suggest heading to the kubernetes helm slack channel | 03:34 |
srwilkers | they might be able to help as well | 03:34 |
Pavo | is there away to add murano horizon plugin manually in the horizon container? | 03:35 |
kfox1111 | might want to kubectl delete tiller-deploy and re helm init so the versions match. | 03:35 |
dupengfei | srwilkers: id like to address it quickly | 03:35 |
*** sdake has quit IRC | 03:35 | |
srwilkers | dupengfei, theyd be the ones who could address it quickly for you as its their product | 03:36 |
dupengfei | kfox1111: Have you even used helm? Can you tell me your kubernetes version and helm version? | 03:36 |
*** jianyi has left #openstack-kolla | 03:36 | |
*** sdake has joined #openstack-kolla | 03:36 | |
dupengfei | srwilkers: I think, when i use the same version as yours, it may works | 03:36 |
dupengfei | kfox1111: Have you ever used helm? Can you tell me your kubernetes version and helm version? | 03:37 |
*** ccesario has quit IRC | 03:37 | |
kfox1111 | I've used k8s since 1.2 and helm since 2.0.0beta2. | 03:38 |
kfox1111 | running k8s 1.5.1 and helm 2.1.0 now. | 03:38 |
kfox1111 | helm 2.0.2 is intended for k8s 1.4.x | 03:38 |
dupengfei | Is "k8s 1.5.1 and helm 2.1.0" suitable for our kolla-kubernetes ? | 03:39 |
dupengfei | kfox1111: Is "k8s 1.5.1 and helm 2.1.0" suitable for our kolla-kubernetes ? | 03:39 |
dupengfei | kfox1111: If so, i may change my framework to that. | 03:40 |
kfox1111 | trun know needs those versions. | 03:40 |
kfox1111 | trunk | 03:40 |
dupengfei | trunk? you mean master? | 03:40 |
kfox1111 | yeah. master. | 03:40 |
dupengfei | OK, i will change to 1.5.1 and 2.1.0 | 03:41 |
*** srwilkers is now known as srwilkers_zzz | 03:43 | |
kfox1111 | the gate's really backed up right now. :/ | 03:47 |
*** srwilkers_zzz has quit IRC | 03:48 | |
*** eaguilar has quit IRC | 03:48 | |
*** diogogmt has quit IRC | 03:54 | |
*** portdirect_away has joined #openstack-kolla | 03:57 | |
*** portdirect_away is now known as portdirect | 03:57 | |
kfox1111 | wb | 03:58 |
portdirect | hey o/ | 03:58 |
portdirect | so helm plugins are kinda awesome | 03:59 |
portdirect | https://github.com/portdirect/marina | 03:59 |
kfox1111 | yeah? | 03:59 |
kfox1111 | looking... | 04:00 |
kfox1111 | does it plug into the server or the client? | 04:01 |
portdirect | should have a plugin that shells out that ceph into a bunch of micro packages soon, ie helm micro foo | 04:02 |
portdirect | client | 04:02 |
portdirect | https://github.com/portdirect/marina/blob/master/helm-plugins/ceph/ceph | 04:02 |
duonghq | portdirect, how about this ps: https://review.openstack.org/#/c/410696/ | 04:02 |
kfox1111 | looks bassically like it just subexec's the plugin? | 04:02 |
portdirect | yup | 04:02 |
*** jtriley has quit IRC | 04:02 | |
duonghq | portdirect, are you working on this, or may I take over this? | 04:02 |
portdirect | duonghq: you are welcome to take it over | 04:03 |
duonghq | portdirect, got it | 04:03 |
portdirect | :) | 04:03 |
kfox1111 | does it provide anything else then "helm foo" rather then the user just calling "foo"? | 04:03 |
portdirect | kfox1111, a bunch of env vars, but not much else atm | 04:03 |
portdirect | so very basic | 04:04 |
*** sayantan_ has quit IRC | 04:04 | |
kfox1111 | ah. k. | 04:04 |
*** ccesario has joined #openstack-kolla | 04:04 | |
portdirect | but by wrapping a docker container you can essentially abstract out whatever you want, so for spliting up a package I'm gonna get it to run a container that walks thorugh a chart and then generates new ones on the fly. | 04:05 |
portdirect | does that seem sane? | 04:05 |
portdirect | this way the issue of variable propogation becomes much easier i think? | 04:06 |
kfox1111 | seems sane to me. | 04:06 |
kfox1111 | would proably try to do the same kind of thing tiller did. | 04:06 |
kfox1111 | load the continer into k8s and then rpc to it. | 04:06 |
kfox1111 | thats basically the operator aproach. | 04:06 |
portdirect | hmm. that could work | 04:07 |
kfox1111 | that would be pretty slick for mariadb/rabbitmq in particular. | 04:07 |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 04:07 |
*** jtriley has joined #openstack-kolla | 04:07 | |
kfox1111 | would be able to manage the cluster memebership, do backups, etc via the helm cli. | 04:07 |
portdirect | yeah think so | 04:08 |
kfox1111 | ceph too, as its pretty stateful. | 04:08 |
portdirect | nice - right need to get ovn/freeipa over then can look into getting kolla running :) | 04:11 |
kfox1111 | cool. let me know how it goes. :) | 04:12 |
*** Pavo has quit IRC | 04:12 | |
portdirect | wiil do :) | 04:12 |
*** sayantan_ has joined #openstack-kolla | 04:12 | |
*** yingjun has quit IRC | 04:13 | |
*** lrensing has quit IRC | 04:14 | |
sdake | sup peeps | 04:15 |
sdake | daughter had chorus concert tonight | 04:15 |
sdake | she rocked it | 04:15 |
*** v1k0d3n has joined #openstack-kolla | 04:15 | |
sdake | freshman in highschool | 04:15 |
*** Pavo has joined #openstack-kolla | 04:17 | |
jascott1 | cool man | 04:19 |
duonghq | cool | 04:19 |
*** sayantan_ has quit IRC | 04:20 | |
jascott1 | thats great experience for many reasons | 04:20 |
*** sayantan_ has joined #openstack-kolla | 04:20 | |
portdirect | hey sdake! thats awesome man :) | 04:22 |
kfox1111 | srwilker-: https://review.openstack.org/#/c/411607/1 looks to be a var not renamed. | 04:24 |
kfox1111 | arg... gate failure after an hour of waiting. :/ | 04:25 |
kfox1111 | the gate's really really busy. :/ | 04:26 |
*** dave-mccowan has quit IRC | 04:26 | |
*** yingjun has joined #openstack-kolla | 04:38 | |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 04:39 |
*** diogogmt has joined #openstack-kolla | 04:40 | |
*** bjolo_ has joined #openstack-kolla | 04:42 | |
openstackgerrit | jimmygc proposed openstack/kolla: Readability improvement to base image https://review.openstack.org/411154 | 04:44 |
sdake | ya it rocked | 04:45 |
sdake | kfox1111 you can check zuul via zuul.openstack.com | 04:46 |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 04:46 |
sdake | com/org | 04:46 |
duonghq | do we have any special technical reason for choosing fluentd over logstash? | 04:46 |
sdake | duonghq yes | 04:46 |
sdake | duonghq the idea of our kolla-ansible arch is as follows | 04:46 |
sdake | heka->ES->kibana | 04:47 |
sdake | we are moving that to fluentd->ES->kibana | 04:47 |
sdake | logstash doesn't scale | 04:47 |
sdake | and serves no valueable purpose in kolla's design | 04:47 |
sdake | its just overhead | 04:47 |
sdake | logstash-forwarder is deprecated | 04:47 |
sdake | the choices are a few - stick with heka (with is EOL), use logstash-forwarder (which is deprecated), use fluentd, or use ... ther eis something else | 04:48 |
tovin07 | logstash-forwarder -->> beats | 04:48 |
sdake | right the something else is beats | 04:48 |
tovin07 | <in new elastic stack (5.0)> | 04:48 |
sdake | we have decided on fluentd for the moment | 04:48 |
sdake | the point is we don't need logstash, and logstash-forwarder is eol | 04:49 |
sdake | all we really need is somethign to forward logs to es | 04:49 |
*** ccesario has quit IRC | 04:49 | |
sdake | if thats fleuntd or beats, idc | 04:49 |
sdake | but i think the community cares it to be fluentd :) | 04:49 |
duonghq | thank you for your explanation | 04:50 |
sdake | sure happy to help | 04:50 |
sdake | there was some discussion about it on the ml in teh past | 04:50 |
tovin07 | thanks :D | 04:50 |
sdake | where the choices wer emade | 04:50 |
sdake | and in various meeting logs | 04:50 |
kfox1111 | back. | 04:50 |
sdake | if it doesn't work for some technical reason, we will need to reconsider | 04:51 |
*** jtriley has quit IRC | 04:51 | |
sdake | Jeffrey4l ping | 04:51 |
duonghq | roger | 04:51 |
Jeffrey4l | sdake, pong | 04:51 |
sdake | Jeffrey4l had you planned to tag 0.4.0 of kolla-kubernetes today? | 04:52 |
kfox1111 | filebeat could do it. | 04:52 |
Jeffrey4l | no idea. depend on yours. kolla-k8s is not following the milestone cycle. sdake | 04:52 |
kfox1111 | I have yet to find a log forwarder that is sane. :/ | 04:53 |
sdake | Jeffrey4l - kfox1111 is 0.4.0 in a taggable state? | 04:53 |
kfox1111 | the log forwarder really needs a log roller built into it. | 04:53 |
Jeffrey4l | so will kolla-k8s release its next tag? | 04:53 |
kfox1111 | otherwise, the log roller will not know whats sent or not. | 04:53 |
sdake | kfox1111 as in master | 04:53 |
kfox1111 | sdake: Jeffrey4l if it can wait about a week, we'll be in much better shape. | 04:53 |
duonghq | kfox1111, you mean log rotate? | 04:53 |
sdake | kfox1111 the release team shuts down over the holidays | 04:54 |
kfox1111 | we're almost done with the microservice conversion. | 04:54 |
kfox1111 | duonghq: yeah. | 04:54 |
sdake | kfox1111 i'd say tag what we got if its pretty solid | 04:54 |
sdake | and from what I can tell its pretty solid | 04:54 |
sdake | minus the recent changes to the deps | 04:54 |
kfox1111 | sdake: depends on what you want... it is able to pass the gate checks, which are pretty extensive, but, | 04:54 |
kfox1111 | it requires a somewhat mix of old and new deployment stuff to actually do that now. | 04:55 |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 04:55 |
kfox1111 | so, pretty in progress. | 04:55 |
sdake | kfox1111 thats ok - people likely won't use the milestones directly since they aer pre-release software (indicated by the 0 = x stream) | 04:55 |
kfox1111 | Id kind of recommend waiting till after the holiday then rather then cut something so difficult to use as to not be worth trying? | 04:55 |
sdake | we can do tha ttoo | 04:56 |
sdake | up to you | 04:56 |
sdake | needs to be a wednesday | 04:56 |
sdake | pick a day any day | 04:56 |
kfox1111 | k. I'd vote wait then, and cut as soon as they come back. | 04:56 |
kfox1111 | we should have something much more solid then. | 04:56 |
sdake | and i'll adjust the tracker | 04:56 |
kfox1111 | jan 4? | 04:57 |
sdake | sounds good | 04:57 |
kfox1111 | k. | 04:57 |
sdake | Jeffrey4l wfu? | 04:57 |
*** qiliang27 has quit IRC | 04:57 | |
Jeffrey4l | sdake, yep. | 04:57 |
kfox1111 | cool. | 04:57 |
Jeffrey4l | please ping me when kolla-k8s is ready for next tag, kfox1111 | 04:58 |
sdake | https://launchpad.net/kolla-kubernetes/+milestone/0.4.0 | 04:59 |
sdake | Jeffrey4l - use tracker to judge that i think | 04:59 |
Jeffrey4l | OK. | 04:59 |
sdake | but ya - keep comms up with Jeffrey4l - he is our release liason | 04:59 |
sdake | please hav ea look at the milestone | 05:00 |
sdake | inc0 should be back on the 20th - although I think intc has a shutdown as well | 05:01 |
sdake | which means 23rd-4th | 05:01 |
sdake | not really sure | 05:01 |
sdake | he may work thru shutdown or they may not have a shutdown | 05:01 |
sdake | idc, i'm not working thru the shutdown - have shit to do around the house :) | 05:01 |
*** mdnadeem has joined #openstack-kolla | 05:02 | |
kfox1111 | sdake: +1. | 05:02 |
kfox1111 | Jeffrey4l: will do | 05:02 |
sdake | like apparently hang christmas lights | 05:02 |
sdake | 1 week before christmas | 05:02 |
sdake | at my daughter's demand | 05:02 |
openstackgerrit | Merged openstack/kolla-kubernetes: Helm-ize Cinder api pod https://review.openstack.org/411556 | 05:03 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Helmize Heat-engine https://review.openstack.org/410048 | 05:04 |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 05:12 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: [wip] collect logs https://review.openstack.org/411403 | 05:14 |
*** unicell has joined #openstack-kolla | 05:15 | |
openstackgerrit | Merged openstack/kolla-kubernetes: Change bootstrap delete invocations to helm delete instead https://review.openstack.org/411548 | 05:15 |
*** qiliang27 has joined #openstack-kolla | 05:15 | |
*** sayanta__ has joined #openstack-kolla | 05:22 | |
*** sdake_ has joined #openstack-kolla | 05:24 | |
*** sdake has quit IRC | 05:25 | |
*** sayantan_ has quit IRC | 05:26 | |
*** fragatina has joined #openstack-kolla | 05:26 | |
kfox1111 | this is definitely the time of cycle a cache would come in handy. :/ | 05:28 |
*** jtriley has joined #openstack-kolla | 05:32 | |
kfox1111 | gotta head out. nite all. | 05:34 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add freezer ansible role https://review.openstack.org/402224 | 05:35 |
*** v1k0d3n has quit IRC | 05:35 | |
*** fragatina has quit IRC | 05:36 | |
*** fragatina has joined #openstack-kolla | 05:36 | |
*** gomarivera has joined #openstack-kolla | 05:37 | |
*** diogogmt has quit IRC | 05:39 | |
*** saneax-_-|AFK is now known as saneax | 05:44 | |
*** jtriley has quit IRC | 05:50 | |
*** rbergeron has quit IRC | 05:51 | |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 05:52 |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 05:56 |
openstackgerrit | jimmygc proposed openstack/kolla: Readability improvement to base image https://review.openstack.org/411154 | 06:00 |
*** imcsk8 has quit IRC | 06:08 | |
*** imcsk8 has joined #openstack-kolla | 06:08 | |
*** Pavo has quit IRC | 06:12 | |
*** gomarivera has quit IRC | 06:15 | |
*** jtriley has joined #openstack-kolla | 06:16 | |
*** Pavo has joined #openstack-kolla | 06:17 | |
openstackgerrit | Justin Scott proposed openstack/kolla-kubernetes: Helmize Heat-engine https://review.openstack.org/410048 | 06:18 |
*** jianyi has joined #openstack-kolla | 06:20 | |
*** jtriley has quit IRC | 06:21 | |
*** HyperJohnGraham has quit IRC | 06:23 | |
*** rbergeron has joined #openstack-kolla | 06:24 | |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 06:29 |
bjolo | morning | 06:30 |
*** gomarivera has joined #openstack-kolla | 06:31 | |
*** sdake_ has quit IRC | 06:32 | |
*** jtriley has joined #openstack-kolla | 06:33 | |
*** sdake has joined #openstack-kolla | 06:34 | |
zhangshuai | Who ever used swift in kolla? | 06:40 |
*** msimonin has joined #openstack-kolla | 06:42 | |
openstackgerrit | jimmygc proposed openstack/kolla: Update swift-guide.rst https://review.openstack.org/411652 | 06:44 |
*** mliima has quit IRC | 06:45 | |
*** jtriley has quit IRC | 06:46 | |
*** portdirect_ has joined #openstack-kolla | 06:48 | |
*** tovin07__ has quit IRC | 06:48 | |
*** ccesario has joined #openstack-kolla | 06:50 | |
*** jtriley has joined #openstack-kolla | 07:00 | |
*** gomarivera has quit IRC | 07:00 | |
*** sdake has quit IRC | 07:03 | |
*** jtriley has quit IRC | 07:06 | |
*** kollabot has quit IRC | 07:20 | |
openstackgerrit | sayantani proposed openstack/kolla-ansible: [WIP]Fix kolla-ansible gate https://review.openstack.org/411587 | 07:23 |
*** jtriley has joined #openstack-kolla | 07:31 | |
openstackgerrit | jimmygc proposed openstack/kolla: Correct dependency in quickstart.rst https://review.openstack.org/411678 | 07:38 |
*** jtriley has quit IRC | 07:39 | |
*** JRobinson__ has quit IRC | 07:45 | |
*** portdirects_lapt has joined #openstack-kolla | 07:50 | |
openstackgerrit | sayantani proposed openstack/kolla: Clean up kolla-ansible related files from Kolla https://review.openstack.org/406429 | 07:55 |
*** pcaruana has joined #openstack-kolla | 07:56 | |
*** bjolo_ has quit IRC | 08:01 | |
*** yatin has quit IRC | 08:02 | |
*** sacharya has quit IRC | 08:06 | |
bjolo | zhangshuai, tried once just for testing. | 08:07 |
bjolo | did not quite get it to work | 08:07 |
*** sacharya has joined #openstack-kolla | 08:08 | |
bjolo | swift authentication with keystone was broken somehow | 08:08 |
bjolo | only worked for admin user | 08:08 |
bjolo | we run external ceph with radosgw in production | 08:08 |
*** fragatina has quit IRC | 08:09 | |
*** zhugaoxiao has quit IRC | 08:09 | |
*** zhugaoxiao has joined #openstack-kolla | 08:10 | |
zhangshuai | bjolo:yes | 08:12 |
*** Pavo has quit IRC | 08:12 | |
*** sacharya has quit IRC | 08:12 | |
zhangshuai | I debug karbor in kolla, karbor project use swift as backend default. | 08:13 |
*** yatin has joined #openstack-kolla | 08:14 | |
zhangshuai | bjolo, I don't know ceph, I don't know how to configure it in kolla. could you give some example or advice? | 08:15 |
*** shardy has joined #openstack-kolla | 08:16 | |
Jeffrey4l | zhangshuai, kolla support ceph radosgw now. try it check this https://review.openstack.org/406971 | 08:17 |
*** Pavo has joined #openstack-kolla | 08:17 | |
Jeffrey4l | zhangshuai, enable_ceph_rgw_keystone should work. | 08:17 |
*** sayanta__ has quit IRC | 08:18 | |
*** sayantan_ has joined #openstack-kolla | 08:19 | |
zhangshuai | ok, I'll try it. | 08:20 |
*** sayantan_ has quit IRC | 08:23 | |
*** matrohon has joined #openstack-kolla | 08:23 | |
*** rsjethani has joined #openstack-kolla | 08:27 | |
*** jtriley has joined #openstack-kolla | 08:28 | |
rsjethani | Hi all, I am trying kolla for the first time. I am following quick start guide: http://docs.openstack.org/developer/kolla/newton/quickstart.html | 08:28 |
rsjethani | My problem is that I can see ceph and cinder images via docker images but I cannot see any containers for them? | 08:29 |
rsjethani | I haven't modified cinder config anywhere....let me know what more info you guys require? | 08:30 |
rsjethani | instance creation in failing probably because of this | 08:31 |
*** jianyi has quit IRC | 08:33 | |
*** jianyi has joined #openstack-kolla | 08:33 | |
*** msimonin has quit IRC | 08:35 | |
*** egonzalez90 has joined #openstack-kolla | 08:36 | |
egonzalez90 | morning | 08:36 |
rsjethani | morning egonzalez90 | 08:37 |
*** jtriley has quit IRC | 08:38 | |
rsjethani | I need help | 08:39 |
rsjethani | can't see cinder container running | 08:40 |
*** portdirects_lapt has quit IRC | 08:41 | |
*** hogepodge has quit IRC | 08:42 | |
egonzalez90 | rsjethani: docker ps -a | grep cinder | 08:44 |
egonzalez90 | is restarting/excited? | 08:44 |
*** hogepodge has joined #openstack-kolla | 08:44 | |
dupengfei | Hello, I use helm V2.1.0 and kubernetes 1.5.1 still failure, Anyone whose helm works well ? | 08:51 |
jascott1 | hi dupengfei, what problem are you having? | 08:54 |
*** sacharya has joined #openstack-kolla | 09:09 | |
*** Serlex has joined #openstack-kolla | 09:12 | |
*** sacharya has quit IRC | 09:13 | |
*** rmart04 has joined #openstack-kolla | 09:13 | |
rsjethani | egonzalex90: cinder is not showing up at all. no container named cinder or similar | 09:19 |
rsjethani | egonzalez90: cinder is not showing up at all. no container named cinder or similar | 09:20 |
egonzalez90 | umm did you enable_cinder True in globals? | 09:20 |
rsjethani | my build coomand was like: sudo kolla-build nova glance cinder keystone heat | 09:21 |
egonzalez90 | by default is disabled IIRC | 09:21 |
egonzalez90 | kolla-build dont start containers, just create the docker images | 09:22 |
egonzalez90 | kolla-ansible deploy is the step to deploy OpenStack with ansible | 09:22 |
rsjethani | cinder is a core service why is it disabled by defaullt? | 09:22 |
rsjethani | what should it do now to enable cinder | 09:23 |
egonzalez90 | did you used kolla-ansible to deploy? have any container running ATM? docker ps will tell you | 09:24 |
rsjethani | yes | 09:24 |
rsjethani | yes many containers are running i can use openstack client to explore things | 09:25 |
rsjethani | i also checked on horizon....things look good so far | 09:25 |
egonzalez90 | add enable_cinder: yes to globals.yml | 09:25 |
rsjethani | backend as lvm? | 09:26 |
rsjethani | if yes then i think i also need to enable cinder_volume_group. right? | 09:27 |
*** jtriley has joined #openstack-kolla | 09:27 | |
*** athomas has joined #openstack-kolla | 09:28 | |
rsjethani | egonzalez: how to reconfigure my stack? | 09:31 |
*** qiliang27 has quit IRC | 09:35 | |
*** gfidente has joined #openstack-kolla | 09:35 | |
*** gfidente has joined #openstack-kolla | 09:35 | |
*** yingjun has quit IRC | 09:36 | |
openstackgerrit | jianyi proposed openstack/kolla: Add zun container https://review.openstack.org/410616 | 09:37 |
*** jtriley has quit IRC | 09:37 | |
Serlex | so cisco is dropping openstack in march it seems -_- | 09:41 |
*** gfidente has quit IRC | 09:47 | |
*** jianyi has quit IRC | 09:48 | |
Jeffrey4l | Serlex, why do u say so? | 09:51 |
*** gfidente has joined #openstack-kolla | 09:53 | |
jascott1 | talking about this? http://www.geekwire.com/2016/another-one-bites-dust-cisco-discontinues-1b-cloud-initiative-aws-azure-others-expand/ | 09:58 |
jascott1 | night all | 09:58 |
*** duonghq has quit IRC | 10:01 | |
*** tovin07 has quit IRC | 10:03 | |
Jeffrey4l | good night. | 10:05 |
*** ccesario has quit IRC | 10:06 | |
*** vbel has quit IRC | 10:07 | |
*** Pavo has quit IRC | 10:12 | |
*** Pavo has joined #openstack-kolla | 10:17 | |
*** openstackgerrit has quit IRC | 10:18 | |
*** l4yerffeJ has quit IRC | 10:22 | |
*** openstackgerrit has joined #openstack-kolla | 10:24 | |
openstackgerrit | Vladislav Belogrudov proposed openstack/kolla-ansible: Normal users cannot use Swift https://review.openstack.org/411742 | 10:24 |
rsjethani | Serlex: yes they are | 10:25 |
rsjethani | I was part of a team which contributed to Cisco's intercloud... | 10:25 |
Serlex | that is a shame :( | 10:26 |
rsjethani | The whole business unit will be shut down | 10:26 |
Serlex | are you moving to another team? | 10:26 |
rsjethani | Yup | 10:26 |
*** jtriley has joined #openstack-kolla | 10:27 | |
rsjethani | HPE also transfering their openstack and cloudfoundry business to SUSE | 10:27 |
*** zhubingbing_ has quit IRC | 10:29 | |
rsjethani | How can i properly shutdown and start kolla deployment? Currently trying out all in one node deployment | 10:31 |
*** jtriley has quit IRC | 10:31 | |
*** jtriley has joined #openstack-kolla | 10:32 | |
*** v1k0d3n has joined #openstack-kolla | 10:36 | |
*** jtriley has quit IRC | 10:37 | |
SamYaple | rsjethani: you can stop all the containers to "shutdown" the node. to start it back up youll want to run the playbooks again to start things in the right order | 10:41 |
*** v1k0d3n has quit IRC | 10:42 | |
rsjethani | SamYaple: Do I have to manually check dependencies? | 10:42 |
rsjethani | isn't there a one click way to do it? | 10:42 |
SamYaple | run the playbooks again, thats "one click" yea | 10:45 |
rsjethani | sorry for newbie Qs, are you talking about running 'kolla-ansible deploy'? | 10:47 |
*** papacz has quit IRC | 10:47 | |
rsjethani | SamYaple:sorry for newbie questions, are you talking about running 'kolla-ansible deploy'? | 10:50 |
openstackgerrit | Javier Castillo AlcÃbar proposed openstack/kolla-ansible: Fix Murano/RabbitMQ problem with haproxy https://review.openstack.org/410825 | 10:50 |
SamYaple | rsjethani: yes | 10:52 |
openstackgerrit | Javier Castillo AlcÃbar proposed openstack/kolla-ansible: Fix Murano/RabbitMQ problem with haproxy https://review.openstack.org/410825 | 10:52 |
rsjethani | SamYaple: thanks let me try that. | 10:56 |
*** magicboiz has quit IRC | 10:56 | |
*** sacharya has joined #openstack-kolla | 11:06 | |
*** sacharya has quit IRC | 11:11 | |
*** rsjethani has quit IRC | 11:11 | |
*** mliima has joined #openstack-kolla | 11:12 | |
*** rsjethani has joined #openstack-kolla | 11:14 | |
*** zhubingbing has joined #openstack-kolla | 11:15 | |
mliima | morning guys | 11:19 |
openstackgerrit | Vladislav Belogrudov proposed openstack/kolla-ansible: Fix Murano agent RabbitMQ messaging https://review.openstack.org/411760 | 11:19 |
mliima | Jeffrey4l, did you try to reproduce showmount -e issue? | 11:19 |
openstackgerrit | Merged openstack/kolla-kubernetes: Helmize neutron dhcp/metadata agents. https://review.openstack.org/411493 | 11:22 |
*** dgonzalez has quit IRC | 11:22 | |
*** dgonzalez has joined #openstack-kolla | 11:24 | |
openstackgerrit | Merged openstack/kolla-kubernetes: Helmize Heat-engine https://review.openstack.org/410048 | 11:25 |
*** jtriley has joined #openstack-kolla | 11:26 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP-PoC Glance service orchestration with entrypoint https://review.openstack.org/410736 | 11:27 |
*** papacz has joined #openstack-kolla | 11:29 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP-PoC Glance service orchestration with entrypoint https://review.openstack.org/410736 | 11:29 |
*** jtriley has quit IRC | 11:31 | |
*** jtriley has joined #openstack-kolla | 11:32 | |
*** jtriley has quit IRC | 11:37 | |
*** eaguilar has joined #openstack-kolla | 11:42 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP-PoC Glance service orchestration with entrypoint https://review.openstack.org/410736 | 12:00 |
*** rsjethani has quit IRC | 12:00 | |
des_consolado | hey I'm having a problem with this https://gist.github.com/thomasodot/ed1cbdf0f22001b2aba80f6ac421b998 | 12:10 |
*** Pavo has quit IRC | 12:12 | |
des_consolado | this interface definitely exists on the compute01 host | 12:13 |
des_consolado | I'm using the combined interface approach, so in my globals.yml I have only the internal vip address and the network interface set. | 12:14 |
des_consolado | The network interface is enp0s8 and it will be that for all 3 hosts I'm trying to deal with | 12:14 |
*** Pavo has joined #openstack-kolla | 12:17 | |
egonzalez90 | using ansible -m setup to those nodes, appears ansible_enp0s8 as interface? | 12:23 |
*** jtriley has joined #openstack-kolla | 12:26 | |
*** eaguilar has quit IRC | 12:28 | |
*** eaguilar has joined #openstack-kolla | 12:30 | |
openstackgerrit | zhubingbing proposed openstack/kolla: Add fluent-plugin-grok-parser in fluentd https://review.openstack.org/410089 | 12:32 |
zhubingbing | sup egonzalez90> | 12:32 |
*** dave-mccowan has joined #openstack-kolla | 12:33 | |
*** jtriley has quit IRC | 12:35 | |
Jeffrey4l | mliima, sorry. still no. busy for the b2 release. | 12:39 |
mliima | no problem Jeffrey4l | 12:45 |
*** l4yerffeJ has joined #openstack-kolla | 12:45 | |
mliima | Jeffrey4l, https://review.openstack.org/#/c/411434/ | 12:45 |
mliima | it was fixed in kolla-ansible | 12:46 |
Jeffrey4l | mliima, but i tried run a normal container to try `showmount` it works. | 12:46 |
mliima | cinder-volume Jeffrey4l | 12:46 |
*** sayantan_ has joined #openstack-kolla | 12:46 | |
mliima | showmout works fine in nova-compute | 12:46 |
Jeffrey4l | mliima, can we use the same change-id for this? | 12:46 |
mliima | same environment | 12:46 |
Jeffrey4l | yep | 12:46 |
mliima | ok | 12:47 |
Jeffrey4l | during backport kolla-ansible to kolla, it should be better use cherry-pick -x | 12:47 |
mliima | hmm | 12:47 |
Jeffrey4l | you can clone these two repo into one folder. | 12:47 |
Jeffrey4l | because kolla-ansible is cut from kolla. | 12:48 |
Jeffrey4l | it is easy to manage during backport. | 12:48 |
mliima | well I did not know that --' | 12:48 |
mliima | hehe | 12:48 |
mliima | sorry | 12:48 |
Jeffrey4l | no need sorry. | 12:48 |
Jeffrey4l | i threw a mail in ML. pbourke give use solution. | 12:49 |
Jeffrey4l | http://lists.openstack.org/pipermail/openstack-dev/2016-November/107598.html | 12:50 |
mliima | regarding showmount error, on my environment when i run "showmount -e" in nova_container it works fine, but running in cinder_volume container i get a error :/ | 12:50 |
mliima | it doesn't make sense to me hahaha | 12:50 |
Jeffrey4l | sorry, this link http://lists.openstack.org/pipermail/openstack-dev/2016-November/107536.html | 12:51 |
Jeffrey4l | hrm ok. | 12:51 |
*** sayantan_ has quit IRC | 12:51 | |
des_consolado | egonzalez90: yes, sorry for slow reply: https://gist.github.com/thomasodot/1977642c76d30e3815aa3c9286b0602d | 12:51 |
des_consolado | does it need to be promiscuous? | 12:51 |
mliima | I'll to kick and set fire in all conteiners hhahaa | 12:52 |
des_consolado | seems doubtful | 12:52 |
*** gomarivera has joined #openstack-kolla | 12:53 | |
Jeffrey4l | mliima, could u explain what it means by "kick and set fire". are u means a real container? if so, i get it lol. | 12:54 |
*** msimonin has joined #openstack-kolla | 12:54 | |
mliima | no no Jeffrey4l, unfortunately lol | 12:55 |
mliima | I can get a old box and write cinder-volume | 12:56 |
Jeffrey4l | old box works? | 12:56 |
mliima | I can kick and set fire | 12:57 |
*** gomarivera has quit IRC | 12:57 | |
Jeffrey4l | sometimes it is hard for a non native speaker to understand ;/ | 12:58 |
mliima | oh, well "old box" means a shoe box | 12:59 |
mliima | let's forget that Jeffrey4l :) | 13:00 |
Jeffrey4l | OK. | 13:00 |
*** sacharya has joined #openstack-kolla | 13:07 | |
openstackgerrit | Mauricio Lima proposed openstack/kolla: Fix ceilometer not sending logs to heka https://review.openstack.org/411794 | 13:08 |
mliima | Works Jeffrey4l ^^ | 13:09 |
*** mdnadeem has quit IRC | 13:10 | |
*** mewald has joined #openstack-kolla | 13:10 | |
*** v1k0d3n has joined #openstack-kolla | 13:10 | |
*** mewald has left #openstack-kolla | 13:10 | |
*** msimonin has quit IRC | 13:11 | |
*** sacharya has quit IRC | 13:11 | |
*** mewald has joined #openstack-kolla | 13:12 | |
Jeffrey4l | nice | 13:16 |
Jeffrey4l | mliima, i can run `showmount -e` in cinder-volume container. | 13:16 |
mliima | hmm | 13:17 |
mliima | ok | 13:17 |
Jeffrey4l | but failed to create a new volume. checking. | 13:17 |
Jeffrey4l | mliima, got this | 13:17 |
Jeffrey4l | <des_consolado> seems doubtful | 13:17 |
Jeffrey4l | <Jeffrey4l> mliima, could u explain what it means by "kick and set fire". are u means a real container? if so, i get it lol. | 13:17 |
Jeffrey4l | <mliima> no no Jeffrey4l, unfortunately lol | 13:17 |
Jeffrey4l | <mliima> I can get a old box and write cinder-volume | 13:17 |
Jeffrey4l | <Jeffrey4l> old box works? | 13:17 |
Jeffrey4l | <mliima> I can kick and set fire | 13:17 |
Jeffrey4l | <Jeffrey4l> sometimes it is hard for a non native speaker to understand ;/ | 13:17 |
Jeffrey4l | <mliima> oh, well "old box" means a shoe box | 13:18 |
Jeffrey4l | <mliima> let's forget that Jeffrey4l :) | 13:18 |
Jeffrey4l | sorry. copy issue..... | 13:18 |
mliima | np hehe | 13:18 |
Jeffrey4l | mliima, get this error u"truncate: cannot open '/var/lib/cinder/mnt/12e0f59430c828ba29a629eac36fa990/volume-648229a1-3ef7-4241-bd52-a7c07c89f618' for writing: Permission denied\n" | 13:18 |
Jeffrey4l | seem only some folder owner issue. | 13:18 |
mliima | are you using master branch? | 13:19 |
Jeffrey4l | mliima, yep. | 13:19 |
mliima | ok | 13:19 |
mliima | ok Jeffrey4l | 13:19 |
Jeffrey4l | cinder-volume is built yesterday. | 13:19 |
mliima | if showmout -e works for you | 13:19 |
Jeffrey4l | after fix the owner issue. i created a new cinder volume successfully. | 13:19 |
Jeffrey4l | it works. | 13:19 |
mliima | the error must be in my network configuration | 13:20 |
*** l4yerffeJ has quit IRC | 13:20 | |
Jeffrey4l | is your nova-compute and cinder-volume on the same node? | 13:20 |
Jeffrey4l | btw i am using a AIO env. | 13:20 |
mliima | but i still not can understand why showmount works in nova_compute and not work in cinder_volume | 13:20 |
mliima | yes | 13:20 |
mliima | AIO | 13:20 |
Jeffrey4l | hrm. | 13:20 |
*** l4yerffeJ has joined #openstack-kolla | 13:21 | |
Jeffrey4l | try strace? | 13:21 |
mliima | well, thank you Jeffrey | 13:21 |
Jeffrey4l | np | 13:21 |
mliima | i'll update my images | 13:21 |
Jeffrey4l | why u force on nfs? | 13:21 |
mliima | and try again using master | 13:21 |
Jeffrey4l | focus* | 13:21 |
Jeffrey4l | it has bad performance, imo. | 13:21 |
mliima | I was just testing :) | 13:22 |
*** l4yerffeJ_ has joined #openstack-kolla | 13:22 | |
Jeffrey4l | cool. | 13:22 |
*** l4yerffeJ has quit IRC | 13:23 | |
mliima | I'll get a vacation on monday | 13:23 |
mliima | I need :) | 13:23 |
Jeffrey4l | how long it will be? | 13:24 |
*** msimonin has joined #openstack-kolla | 13:24 | |
mliima | 14 days | 13:24 |
Jeffrey4l | wow, nice | 13:24 |
Jeffrey4l | any plan? | 13:24 |
mliima | travel | 13:24 |
Jeffrey4l | i guess i will see more fewer guys from next week. | 13:25 |
mliima | travel and rest | 13:25 |
Jeffrey4l | cool | 13:25 |
mliima | see family :) | 13:25 |
mliima | btw Jeffrey4l, as core I need to tell to someone about my vacation? | 13:26 |
*** mkoderer has joined #openstack-kolla | 13:27 | |
*** fragatina has joined #openstack-kolla | 13:29 | |
dmsimard | mandre: added you on https://review.openstack.org/#/c/411520/ | 13:38 |
*** mnasiadka has joined #openstack-kolla | 13:39 | |
openstackgerrit | Özgür Caner proposed openstack/kolla: HA router fails using neutron-vpn-agent https://review.openstack.org/411808 | 13:39 |
dmsimard | mandre: there's also https://review.openstack.org/#/c/407165/ I would like to eventually follow up on and apply Jeffrey4l's suggestions/refactor properly | 13:39 |
dmsimard | In a future patchset, that is | 13:40 |
SamYaple | fyi for those interested. I have laid out a basic repo following how im trying to build 'kolla-salt' | 13:42 |
SamYaple | https://github.com/SamYaple/kolla-salt | 13:42 |
*** Bico_Fino has joined #openstack-kolla | 13:42 | |
SamYaple | once i confirm all tasks to deploy are definetly doable with saltstack im going to begin the work to get it official | 13:42 |
egonzalez90 | mliima: I think don't need to notify unless you're going to be out for long time | 13:44 |
mliima | just 14 days :) | 13:44 |
SamYaple | for reference, here is the entire memcache "role" in salt. https://github.com/SamYaple/kolla-salt/blob/master/kolla/memcached/init.sls | 13:45 |
egonzalez90 | I don't think so | 13:45 |
SamYaple | complete with container restart when config file changes | 13:45 |
egonzalez90 | mliima: enjoy it ;) | 13:45 |
openstackgerrit | Yann Provost proposed openstack/kolla-ansible: Add storage interface address https://review.openstack.org/411432 | 13:48 |
*** rhallisey has joined #openstack-kolla | 13:49 | |
dmsimard | SamYaple: oh boy, salt ? :p | 13:49 |
dmsimard | SamYaple: is that serious or an experiment ? | 13:49 |
SamYaple | dmsimard: lets put it this way, the ansible implementation in kolla started in my project "yaodu". that project started with salt | 13:50 |
SamYaple | dmsimard: salt jsut wasnt quite there yet. but now.... | 13:50 |
dmsimard | I ask because https://github.com/openstack/openstack-salt mostly died stillborn | 13:50 |
SamYaple | dmsimard: yea. but thats not using imaging | 13:51 |
*** mnasiadka has quit IRC | 13:51 | |
SamYaple | that project is closer to openstack-ansible than kolla-salt would be to it | 13:51 |
SamYaple | i think at aleast | 13:51 |
berendt | kolla-salt, pretty cool :) | 13:53 |
SamYaple | if you havent seen it berendt and dmsimard, i did start an etherpad laying out my specific goals for kolla-salt https://etherpad.openstack.org/p/kolla-salt | 13:54 |
SamYaple | im open to suggestions of course | 13:54 |
* dmsimard has admittedly never tried salt | 13:59 | |
*** bachp has quit IRC | 14:02 | |
Jeffrey4l | mliima, roger. | 14:05 |
Jeffrey4l | no idea. iirc, i only see one core send mail about his vocation in ml. | 14:06 |
*** sayantan_ has joined #openstack-kolla | 14:07 | |
*** Pavo has quit IRC | 14:12 | |
Jeffrey4l | mliima, ping | 14:13 |
Jeffrey4l | would u mind to use nfs_shares as customer filename rather than nfs_shares.j2? because we do not use *.j2 patter in /etc/kolla/config folder. | 14:14 |
*** dave-mccowan has quit IRC | 14:14 | |
SamYaple | dmsimard: well all you need to know is that it can be written in entirely pure python | 14:15 |
openstackgerrit | Merged openstack/kolla-ansible: Add PyYAML to requirements.txt, dependency of genpwd https://review.openstack.org/411520 | 14:16 |
mliima | no Jeffrey4l | 14:16 |
Jeffrey4l | cool i will push a patch to make this change. | 14:17 |
*** Pavo has joined #openstack-kolla | 14:17 | |
Jeffrey4l | another patch, could u give a glance? https://review.openstack.org/411403 it almost done. | 14:18 |
Jeffrey4l | i will file a bug, and this change will backport to all stable branch. | 14:18 |
*** bachp has joined #openstack-kolla | 14:19 | |
*** fguillot has joined #openstack-kolla | 14:23 | |
*** HyperJohnGraham has joined #openstack-kolla | 14:27 | |
*** rsjethani has joined #openstack-kolla | 14:29 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Collect logs at the end of gate https://review.openstack.org/411403 | 14:29 |
*** Bico_Fino has quit IRC | 14:31 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Remove the .j2 suffix in nfs_shares file https://review.openstack.org/411837 | 14:32 |
*** msimonin has quit IRC | 14:35 | |
kfox1111 | morning. | 14:35 |
*** goldyfruit has joined #openstack-kolla | 14:37 | |
*** Bico_Fino has joined #openstack-kolla | 14:38 | |
SamYaple | morning kfox1111 | 14:39 |
openstackgerrit | Merged openstack/kolla-ansible: Some tasks fail if NIC names differ https://review.openstack.org/411193 | 14:39 |
* berendt squashed some bug reports | 14:40 | |
awiddersheim | Jeffrey4l: removing.j2 will affect | 14:42 |
awiddersheim | how we run things | 14:42 |
Jeffrey4l | awiddersheim, it do not matter. it will be still render as jinja template. just the name is changed. | 14:43 |
awiddersheim | I understand that but | 14:43 |
*** TxGirlGeek has joined #openstack-kolla | 14:43 | |
Jeffrey4l | because we do not use *.j2 pattern in that folder. | 14:43 |
awiddersheim | the way we configure everything is we put everything in a inventory<site name> directory | 14:43 |
awiddersheim | which would include the custom configs | 14:43 |
awiddersheim | and then we run ansible-playbook -i inventory/<site name> | 14:44 |
awiddersheim | and ansible is pretty good about finding all of the inventory files, group_vars etc | 14:44 |
awiddersheim | but can wrongly interpret some files in that directory as being invnetory files | 14:44 |
awiddersheim | which oyu can fix by telling ansible to ignore certain extensions | 14:44 |
*** Jeffrey4l has left #openstack-kolla | 14:45 | |
*** Jeffrey4l has joined #openstack-kolla | 14:45 | |
SamYaple | if its a jinja2 file, it should keep the .j2 extention by convention | 14:45 |
Jeffrey4l | SamYaple, we are talking the file in /etc/kolla/config folder, which is created by end-user. | 14:45 |
mliima | well, after copy the .j2 or any extension will be ignored | 14:45 |
Jeffrey4l | we do not use that patter in this folder. and it is meaningless for end-user to guest what j2 is. | 14:46 |
SamYaple | Jeffrey4l: oh. then i dont think any of those have .j2 | 14:46 |
SamYaple | yea | 14:46 |
SamYaple | i agree with Jeffrey4l | 14:46 |
awiddersheim | Jeffrey4l: removing the extension | 14:46 |
SamYaple | thats been a thing since almost the beginning of ansible | 14:46 |
awiddersheim | will break our prod sites :( | 14:46 |
SamYaple | or kolla-ansible rather | 14:46 |
*** srwilkers has joined #openstack-kolla | 14:46 | |
Jeffrey4l | yep. | 14:46 |
srwilkers | good morning | 14:46 |
mliima | awiddersheim, you can add it in your env | 14:47 |
mliima | :) | 14:47 |
awiddersheim | then kolla won't copy it | 14:47 |
mliima | just a .j2 | 14:47 |
Jeffrey4l | awiddersheim, no nfs_share is added in O cycle. it is OK to make this kina change. | 14:47 |
*** jtriley has joined #openstack-kolla | 14:47 | |
Jeffrey4l | awiddersheim, i just changed the source filename, just the dest. | 14:47 |
awiddersheim | can we add a .conf extension to it? | 14:47 |
awiddersheim | I understand what you changed | 14:47 |
awiddersheim | it still breaks my use case | 14:48 |
Jeffrey4l | awiddersheim, i do not think so. in the official doc, it use nfs_shares name, so let us make it. | 14:48 |
*** l4yerffeJ_ has quit IRC | 14:48 | |
SamYaple | awiddersheim: im confused, weve had this convention in place since the beginning | 14:48 |
awiddersheim | let me explain again | 14:48 |
awiddersheim | what I'm doing | 14:48 |
SamYaple | is it just that nfs_share is the only thing you are using? | 14:48 |
*** l4yerffeJ_ has joined #openstack-kolla | 14:48 | |
awiddersheim | so we lay out our inventory and configs in a directory structure like inventory/<site name> | 14:49 |
awiddersheim | and then in you would find all the config overrides | 14:49 |
awiddersheim | in inventory/<site name>/configs | 14:49 |
*** Bico_Fino has quit IRC | 14:50 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance keystone endpoints https://review.openstack.org/411607 | 14:50 |
Jeffrey4l | you are deploy multi openstack by using one inventory folder? | 14:50 |
awiddersheim | no | 14:50 |
awiddersheim | we deploy multi openstack | 14:50 |
awiddersheim | but each site | 14:50 |
awiddersheim | gets it's own folder | 14:50 |
*** l4yerffeJ_ has quit IRC | 14:50 | |
awiddersheim | with it's own overrdie configs | 14:50 |
awiddersheim | and then we give ansible the entire site directory | 14:51 |
awiddersheim | as the inventory | 14:51 |
Jeffrey4l | hrm seems i do not get the point. why inventory/<site name>/configs rather than using /etc/kolla/config | 14:51 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm Cinder Backup https://review.openstack.org/411570 | 14:51 |
awiddersheim | beacuse we have multiple sites | 14:51 |
awiddersheim | each with their own set of configs | 14:51 |
SamYaple | awiddersheim: i understand your point here | 14:51 |
Jeffrey4l | so you change the node_custom_config variable for each site? | 14:52 |
awiddersheim | yes | 14:52 |
awiddersheim | exactly | 14:52 |
SamYaple | awiddersheim: i will say though, this has been in place since 1.0.0 and we cant jsut change it | 14:52 |
awiddersheim | I'm fine with not having .j2 but I need | 14:52 |
awiddersheim | atleast some extension on the source | 14:52 |
*** dave-mccowan has joined #openstack-kolla | 14:52 | |
Jeffrey4l | nice solution. | 14:52 |
awiddersheim | so that I can tell ansible to ignore it | 14:52 |
awiddersheim | with the inventory_ignore_extensions | 14:52 |
awiddersheim | so ansible doesn't try to read my source nfs_shares file | 14:53 |
awiddersheim | as an ansible invenotry file by mistake | 14:53 |
awiddersheim | right now we set it as such | 14:53 |
awiddersheim | inventory_ignore_extensions = ~, .orig, .bak, .ini, .cfg, .retry, .pyc, .pyo, .yml, .md, .conf, .cnf, .j2 | 14:53 |
Jeffrey4l | if you want to ignore it, why not just remove it ? | 14:53 |
Jeffrey4l | oh got it. | 14:53 |
awiddersheim | well I do need to copy nfs_shares over for my environments | 14:53 |
awiddersheim | I added this file originally btw | 14:54 |
awiddersheim | and added the .j2 there specifically for this | 14:54 |
SamYaple | awiddersheim: then youll need to augment all the of the config copy stuff to look at both no extention and a .j2 or .conf extention | 14:54 |
SamYaple | we can import as many files as we want | 14:54 |
awiddersheim | SamYaple: not sure what you mean | 14:54 |
SamYaple | the specific section that Jeffrey4l modified in that patch | 14:55 |
*** gomarivera has joined #openstack-kolla | 14:55 | |
SamYaple | it can have both the non-extentioned files, and files with extentions | 14:55 |
Jeffrey4l | awiddersheim, there will be other file without extension. | 14:55 |
SamYaple | as well as all of the other similiarly copied ones | 14:55 |
Jeffrey4l | does the inventory_ignore_extensions support folder? | 14:56 |
Jeffrey4l | it is wired that only nfs_shared need a j2 filer but other custom file not. | 14:56 |
awiddersheim | Jeffrey4l: I have not run into any fiels without extensions yet that I need | 14:56 |
awiddersheim | Jeffrey4l: most of the other custom files are .conf | 14:56 |
awiddersheim | Jeffrey4l: No I dont' think it supports folders unfortunatel | 14:57 |
Jeffrey4l | awiddersheim, it is. but we do not add extra extension to custom file ;( | 14:57 |
SamYaple | Jeffrey4l: we can. there is nothing stopping us | 14:57 |
Jeffrey4l | awiddersheim, check horizon config.yml file. | 14:58 |
Jeffrey4l | it use bare local_settings filename too, without any extension. | 14:58 |
SamYaple | Jeffrey4l: thats unrelated | 14:58 |
*** Bico_Fino has joined #openstack-kolla | 14:58 | |
awiddersheim | Jeffrey4l: I beleive you that there are other files without extensions | 14:59 |
SamYaple | and thats is convention. horizon looks for tah tfile without an extention | 14:59 |
Jeffrey4l | if we add it, it will be confused to end-user. he will know nothing why this file need .j2 and local_settings file not. | 14:59 |
SamYaple | i think its dumb. but its django | 14:59 |
*** mewald has quit IRC | 14:59 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm chart for cinder-scheduler https://review.openstack.org/410525 | 14:59 |
SamYaple | Jeffrey4l: i didnt say we _remove_ the ability to do it without an extention | 14:59 |
*** gomarivera has quit IRC | 14:59 | |
SamYaple | we can look for both, with and without extention | 14:59 |
*** coolsvap has quit IRC | 14:59 | |
awiddersheim | +1 | 14:59 |
Jeffrey4l | hrm ;/ | 15:00 |
SamYaple | we cant remove it anyway. without deprecation | 15:00 |
awiddersheim | maybe I need to sell you guys on how we use kolla and setup our inventory | 15:00 |
SamYaple | a .conf or .ini extention would seem appropriate | 15:00 |
mliima | I still agree with Jeffrey4l | 15:00 |
SamYaple | rather than a .j2 | 15:00 |
mliima | :/ | 15:00 |
awiddersheim | it works quite well for multiple sites | 15:00 |
awiddersheim | distinct sites | 15:00 |
openstackgerrit | Merged openstack/kolla: Add run_gate.sh script for CI https://review.openstack.org/410593 | 15:00 |
SamYaple | awiddersheim: nah ive had the same thought about the extention thing. you can blame me for writing it that way | 15:00 |
SamYaple | awiddersheim: i wasnt using my 20/20 hindsight vision | 15:01 |
awiddersheim | it happens | 15:01 |
awiddersheim | I'm fine with cahnging the extension | 15:01 |
*** goldyfruit has quit IRC | 15:01 | |
awiddersheim | or looking for both with our without | 15:01 |
awiddersheim | the .j2 does not matter to me | 15:02 |
awiddersheim | it was just an extension | 15:02 |
awiddersheim | that I could use | 15:02 |
*** ocaner has joined #openstack-kolla | 15:02 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance api and glance registry pods https://review.openstack.org/410481 | 15:02 |
Jeffrey4l | hrm, i have to say this is added since Newton. so we can not just remove it . ;( | 15:02 |
Jeffrey4l | but we need have agreement, what the filename in custom folder will be like? should we break the rule/policy? | 15:03 |
SamYaple | awiddersheim: Jeffrey4l we can do a glob-style lookup of tiles | 15:03 |
SamYaple | nfs_shares* will match | 15:03 |
SamYaple | s/tiles/files/ | 15:03 |
awiddersheim | that is a simple solution | 15:03 |
SamYaple | Jeffrey4l: that would also solve your "this was added in newton" problem | 15:04 |
Jeffrey4l | what will we do for next such file? SamYaple | 15:04 |
SamYaple | what do you mean "next file" ? | 15:04 |
SamYaple | it will find all files that match taht glob | 15:05 |
Jeffrey4l | I prefer to use find [ nfs_shares.j2, nfs_shares ]. | 15:05 |
Jeffrey4l | more explicit. | 15:05 |
SamYaple | well you were acting like you didnt want that :P | 15:05 |
Jeffrey4l | ignore this case. | 15:05 |
Jeffrey4l | ignore the back compatible issue. | 15:06 |
kfox1111 | sbezverk: awake? | 15:06 |
Jeffrey4l | will we use the nfs_shares* pattern forever? | 15:06 |
SamYaple | Jeffrey4l: yea why not | 15:06 |
*** portdirect has quit IRC | 15:06 | |
SamYaple | would allow people to do nfs_shares.01 nfs_shares.02 | 15:07 |
SamYaple | people like stuff like that | 15:07 |
*** portdirect has joined #openstack-kolla | 15:07 | |
SamYaple | its litereally no additional work for us | 15:07 |
Jeffrey4l | or follow the rule: do not add extra extension for custom file. | 15:07 |
SamYaple | which is already broken | 15:07 |
SamYaple | and was never a rule | 15:07 |
Jeffrey4l | it is not hard. but we need agreement. | 15:07 |
Jeffrey4l | we can fix it now. | 15:08 |
*** sacharya has joined #openstack-kolla | 15:08 | |
SamYaple | like i said Jeffrey4l. i implemented that before _any_ of our templates had .j2 | 15:08 |
SamYaple | it was really early on | 15:08 |
SamYaple | please dont take my mistake as gospel | 15:08 |
Jeffrey4l | if not, we need document every service, what's the custom filename will looks like. | 15:08 |
SamYaple | if we glob nfs_shares*, we dont need to document anything beyond whats already been documented | 15:09 |
Jeffrey4l | SamYaple, so will we need add * to all the files in custom config folder? | 15:09 |
SamYaple | yea | 15:09 |
SamYaple | i havent heard a reason not too | 15:09 |
Bico_Fino | Jeffrey4l: Hi, about the bug 1650345, the patch can be applied to Newton branch or just Ocata? | 15:10 |
openstack | bug 1650345 in kolla "Ironic Inspector DBAPIError" [High,Confirmed] https://launchpad.net/bugs/1650345 - Assigned to Jeffrey Zhang (jeffrey4l) | 15:10 |
Jeffrey4l | Bico_Fino, it is kinda big feature. so shouldn't be backport bo N | 15:12 |
Jeffrey4l | to N* | 15:12 |
*** sacharya has quit IRC | 15:13 | |
Bico_Fino | I see. | 15:13 |
Bico_Fino | We already fixed some parts in Newton. | 15:13 |
Jeffrey4l | Bico_Fino, cool. do u have a fix on master? | 15:14 |
Bico_Fino | Jeffrey4l: not yet, when we got it functional I will wrap everything and make a patch. | 15:15 |
Jeffrey4l | Bico_Fino, i have a patch. and it almost done. did u check it? | 15:16 |
Bico_Fino | Jeffrey4l: Just checking first with you. | 15:16 |
Bico_Fino | Jeffrey4l: Going check it and try apply it to 3.0.1 | 15:16 |
Jeffrey4l | Bico_Fino, cool | 15:16 |
*** goldyfruit has joined #openstack-kolla | 15:17 | |
Jeffrey4l | SamYaple, I like explicit not glob. it is common that when operator wants make some change to test, he rename his custom nova.conf to nova.conf.bak for backup. than make some change in nova.conf. in your the glob version, it will cause some issue in this case. | 15:19 |
SamYaple | Jeffrey4l: not if we say it is a .d type directory, which are all very common | 15:21 |
*** jgriffith is now known as jgriffith_AutoAw | 15:22 | |
Jeffrey4l | i know .d type directory. .d type will include *.conf file, not a* file. so it still works when i rename it to .bak file. | 15:22 |
*** jgriffith_AutoAw is now known as jgriffith | 15:22 | |
Jeffrey4l | and most openstack project haven't the .d concept. | 15:23 |
Jeffrey4l | for the nfs_shares, i am OK to add supporting nfs_shares and nfs_shares.j2 file. but i do not want to add such pattern to all configration file. | 15:24 |
*** dave-mccowan has quit IRC | 15:27 | |
*** Bico_Fino has quit IRC | 15:30 | |
*** rmart04 has quit IRC | 15:32 | |
SamYaple | Jeffrey4l: actually i think most of them do have .d support now | 15:33 |
SamYaple | anyway, like i said im not realyl affected by this | 15:33 |
SamYaple | i was just giving suggestions | 15:33 |
Jeffrey4l | roger. i will implement support nfs_shares and nfs_shares.j2 file for back compatible and unify | 15:34 |
*** srwilkers has quit IRC | 15:39 | |
*** dave-mccowan has joined #openstack-kolla | 15:41 | |
*** l4yerffeJ_ has joined #openstack-kolla | 15:47 | |
*** tonanhngo has joined #openstack-kolla | 15:48 | |
*** tonanhngo has quit IRC | 15:48 | |
*** tonanhngo has joined #openstack-kolla | 15:49 | |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 15:54 |
*** srwilkers has joined #openstack-kolla | 16:01 | |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 16:01 |
*** egonzalez90 has quit IRC | 16:02 | |
*** sdake has joined #openstack-kolla | 16:03 | |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 16:03 |
srwilkers | morning everyone o/ | 16:04 |
srwilkers | happy friday | 16:04 |
sayantan_ | Morning guys :) | 16:06 |
sayantan_ | Morning srwilkers | 16:06 |
*** eaguilar has quit IRC | 16:06 | |
*** saneax is now known as saneax-_-|AFK | 16:07 | |
sbezverk | good morning | 16:08 |
sayantan_ | Have been working on the quickstart guide documentation. It would be great to have some eyes on it :) https://review.openstack.org/#/c/404993/ | 16:09 |
sayantan_ | Let's get this rolling :) We need better docs! | 16:10 |
SamYaple | sayantan_: you wont find anyone here disagreeing with oyu | 16:12 |
*** Pavo has quit IRC | 16:12 | |
sayantan_ | SamYaple I agree :) | 16:13 |
SamYaple | sayantan_: youll want to check with the openstack docs team. you shouldnt be copying the docs like that | 16:14 |
openstackgerrit | sayantani proposed openstack/kolla: massive re-work on Quick Start Guide https://review.openstack.org/404993 | 16:14 |
SamYaple | there are already ways to generate multiple docs based on different distros (its how upstream docs work) | 16:14 |
SamYaple | i think the team is openstack-manauls | 16:14 |
sayantan_ | oh | 16:15 |
sayantan_ | I will seek help from these guys today then | 16:15 |
SamYaple | sayantan_: yea like all of this http://docs.openstack.org/project-install-guide/newton/ | 16:15 |
SamYaple | thats based on the same templates with the distro specific bits mixed in | 16:15 |
sayantan_ | SamYaple That looks neat. Thanks for the input. Will work on it | 16:16 |
*** Pavo has joined #openstack-kolla | 16:17 | |
*** Bico_Fino has joined #openstack-kolla | 16:23 | |
*** bmace has quit IRC | 16:24 | |
*** bmace has joined #openstack-kolla | 16:24 | |
*** pcaruana has quit IRC | 16:25 | |
*** Bico_Fino has quit IRC | 16:28 | |
SamYaple | portdirect: you has showed some interest in the salt thing. here is the entire memcached 'role' https://github.com/SamYaple/kolla-salt/blob/master/kolla/memcached/init.sls | 16:29 |
srwilkers | morning sayantan_ :) | 16:31 |
srwilkers | hey sbezverk :) happy friday | 16:31 |
sdake | morning peeps | 16:31 |
srwilkers | morning sdake :) | 16:31 |
sdake | ya no doubt - happy friday :) | 16:31 |
sdake | I'm starting my friday early | 16:31 |
sdake | just wanted to drop in and say hi :) | 16:31 |
sdake | got a million things to do around the house | 16:31 |
sayantan_ | Hi sdake | 16:31 |
sdake | sup sayantan_ | 16:32 |
*** adrian_otto has joined #openstack-kolla | 16:32 | |
sayantan_ | working with the openstack-manuals team on dividing the guide according to distros | 16:32 |
srwilkers | portdirect, hey man o/ | 16:32 |
*** HyperJohnGraham has quit IRC | 16:34 | |
*** Bico_Fino has joined #openstack-kolla | 16:35 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance keystone endpoints https://review.openstack.org/411607 | 16:36 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm chart for cinder-scheduler https://review.openstack.org/410525 | 16:39 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance api and glance registry pods https://review.openstack.org/410481 | 16:41 |
*** sayantan_ has quit IRC | 16:44 | |
mliima | somebody using ubuntu to deploy kolla AIO? | 16:44 |
SamYaple | mliima: i have, why do you ask? | 16:45 |
*** rhallisey has quit IRC | 16:47 | |
mliima | well sam, when i run showmount -e in cinder_volume i get this error "clnt_create: RPC: Unknown host", but in same environment, same deploy, if i run "showmount -e" in nova_compute it works | 16:47 |
mliima | it doesn't make any sense to me | 16:48 |
SamYaple | its proabbly the different mounts mounted into the container | 16:49 |
mliima | bid mounts? | 16:49 |
SamYaple | yea | 16:49 |
mliima | hm... | 16:49 |
mliima | i'll check | 16:49 |
SamYaple | if i had to guess of the top of my head, id guess /dev | 16:50 |
SamYaple | just because it is mounts | 16:51 |
*** matrohon has quit IRC | 16:51 | |
*** msimonin has joined #openstack-kolla | 16:51 | |
mliima | https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/nova/tasks/start_compute.yml#L35-L36 | 16:52 |
mliima | https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/cinder/tasks/start.yml#L31-L32 | 16:52 |
mliima | :/ | 16:53 |
SamYaple | cinder-volume ipc mode is host | 16:54 |
SamYaple | for some reason.... | 16:54 |
SamYaple | probably iscsid related | 16:55 |
mliima | do you think that it can be related? | 16:55 |
mliima | ipc_mode with my issue? | 16:55 |
SamYaple | i think thats the only differnce :) | 16:56 |
SamYaple | so maybe | 16:56 |
*** gomarivera has joined #openstack-kolla | 16:56 | |
mliima | i know if I install openvswitch in cinder_volume, showmout works fine, but i don't want to do taht | 16:56 |
mliima | ok, i'll remove it | 16:57 |
SamYaple | oh. maybe its a dependancy missing then | 16:57 |
SamYaple | strace it | 16:57 |
SamYaple | im out of ideas. i would strace it now if i were you. figure out what it is | 16:58 |
SamYaple | RPC: unknown host seems to be related to hostname, so maybe /etc/hosts is somethign you should check too | 16:58 |
mliima | because in nova_compute it works fine | 16:59 |
mliima | AIO deployment | 16:59 |
mliima | same host | 16:59 |
mliima | Jeffrey4l, tried to reproduce it, and works fine, but i don't know if him used ubuntu or centos | 16:59 |
mliima | anyway, thank you SamYaple | 16:59 |
*** Serlex has quit IRC | 17:00 | |
*** sayantan_ has joined #openstack-kolla | 17:00 | |
*** Edigleison has joined #openstack-kolla | 17:00 | |
mliima | i'll use my "workaround" at least until find a better fix | 17:01 |
mliima | :) | 17:01 |
*** gomarivera has quit IRC | 17:01 | |
*** gomarivera has joined #openstack-kolla | 17:07 | |
*** adrian_otto has quit IRC | 17:08 | |
*** sayantan_ has quit IRC | 17:08 | |
*** sacharya has joined #openstack-kolla | 17:09 | |
*** sayantan_ has joined #openstack-kolla | 17:09 | |
openstackgerrit | Merged openstack/kolla: Move openvswitch db file into docker volume https://review.openstack.org/411575 | 17:09 |
*** Edigleison has quit IRC | 17:12 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance keystone endpoints https://review.openstack.org/411607 | 17:13 |
*** jgriffith is now known as jgriffith_AutoAw | 17:13 | |
*** jgriffith_AutoAw is now known as jgriffith | 17:14 | |
*** eaguilar has joined #openstack-kolla | 17:15 | |
dmsimard | sean-k-mooney: o/ | 17:19 |
dmsimard | nevermind, answered my own question | 17:20 |
*** athomas has quit IRC | 17:20 | |
klindgren | having a problem running kolla-ansible out of venv. The bash script keeps populating the venv directory as the basedir, for playbooks, When I attempt to overide the playbook play it appears to be ignored. Specifically because actions re populate the the PLAYBOOK variable using ${BASEDIR} - so I am confused by what the point of being able to specify the playbook file. Shouldn't this be changed to be able to specify the playbo | 17:21 |
klindgren | ok basedir instead? | 17:21 |
klindgren | IE it defaults at: https://github.com/openstack/kolla-ansible/blob/master/tools/kolla-ansible#L70 | 17:21 |
klindgren | gets updated at: https://github.com/openstack/kolla-ansible/blob/master/tools/kolla-ansible#L87 | 17:22 |
klindgren | completly ignored at: https://github.com/openstack/kolla-ansible/blob/master/tools/kolla-ansible#L177 | 17:22 |
SamYaple | klindgren: all of that looks correct to me | 17:28 |
dmsimard | klindgren: I'm curious about running kolla-ansible out of a venv too. I have a different use case but I feel it's clunky nonetheless -- in fact I have no interest in using the kolla-ansible binary because I'm interested in including the roles and tasks natively since I use kolla-ansible as part of an Ansible playbook | 17:28 |
SamYaple | klindgren: im confused as to your exact issue | 17:28 |
*** sdake has quit IRC | 17:28 | |
klindgren | SamYaple, # kolla-ansible -i glance-only -p /root/kolla-ansible/ansible/kolla-host.yml -t glance bootstrap-servers | 17:29 |
klindgren | Bootstraping servers : ansible-playbook -i glance-only -e @/etc/kolla/globals.yml -e @/etc/kolla/passwords.yml -e CONFIG_DIR=/etc/kolla --tags glance -e action=bootstrap-servers /root/kolla-ansible/venv/ansible/kolla-host.yml | 17:29 |
klindgren | ERROR! the playbook: /root/kolla-ansible/venv/ansible/kolla-host.yml could not be found | 17:29 |
klindgren | I specify that the playbook is /root/kolla-ansible/ansible/kolla-host.yml | 17:29 |
klindgren | that is isgnored | 17:29 |
klindgren | and instead BASEDIR/ansible/kolla-host.yml is used | 17:30 |
klindgren | because of: https://github.com/openstack/kolla-ansible/blob/master/tools/kolla-ansible#L177 | 17:30 |
klindgren | so whats the point of specifying -p or --playbook | 17:30 |
klindgren | if its going to be ignored | 17:30 |
SamYaple | klindgren: i think there is some confusion here as to what you are specifying there | 17:31 |
SamYaple | that option is for your "site.yml" type playbook | 17:31 |
SamYaple | im not saying this isnt a bug, but that option isnt meant to be used with "bootstrap-servers" or the other options | 17:32 |
SamYaple | really its only for deploy. | 17:32 |
SamYaple | im sure no one has a problem with you chanign that | 17:33 |
*** eaguilar has quit IRC | 17:33 | |
SamYaple | personally, i hate kolla-ansible script. it was never supposed to be in bash | 17:33 |
dmsimard | SamYaple: right, it's a glorified wrapper over ansible-playbook | 17:33 |
*** unicell has quit IRC | 17:33 | |
*** eaguilar has joined #openstack-kolla | 17:33 | |
SamYaple | dmsimard: we were supposed to have a glorified wrapper over asible-playbook.... in python | 17:34 |
SamYaple | that also did password generate and otehr types of checking | 17:34 |
SamYaple | and all kinds of stuff. | 17:34 |
dmsimard | I'm working on a playbook that sets up kolla-ansible and then includes bootstrap-server natively | 17:34 |
SamYaple | i actually fought to not have the kolla-ansible bash script even merged | 17:34 |
dmsimard | without the kolla-ansible script at all | 17:34 |
*** lamt has joined #openstack-kolla | 17:34 | |
SamYaple | i said if we merge the bash version, we will never write a different one lol | 17:35 |
klindgren | so I can add a new option for playbook basedir? | 17:35 |
*** ayoung has quit IRC | 17:35 | |
SamYaple | klindgren: you can refactor that thing until your heart is content. but it looks like what you really want is a '--basedir' option | 17:35 |
SamYaple | or something like that | 17:36 |
SamYaple | if you _really_ want to know why its like that, its because most people don't `pip install kolla-ansible`, but rather run it from a git clone'd directory (and thats how the script was originally written) | 17:37 |
*** adrian_otto has joined #openstack-kolla | 17:37 | |
klindgren | I didn't either - I jsut ran python setup.py install | 17:37 |
klindgren | from within a venv | 17:38 |
SamYaple | thats teh same thing | 17:38 |
klindgren | right but I think if I did that from a venv it would have the ansible-kolla stuff in the venv | 17:38 |
SamYaple | if you ran tools/kolla-ansible youd find it using that local directory | 17:38 |
klindgren | and not outside the folder | 17:38 |
klindgren | ok - let me play with it, and stop doing something on my end and see if I still want to do up an additional field. | 17:40 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm Chart for Cinder Volume https://review.openstack.org/411927 | 17:40 |
openstackgerrit | David Moreau Simard proposed openstack/kolla-ansible: WIP: Refactor the baremetal role https://review.openstack.org/411930 | 17:43 |
sean-k-mooney | dmsimard: hi | 17:44 |
*** vhosakot has joined #openstack-kolla | 17:44 | |
dmsimard | sean-k-mooney: maybe you want to review ^, it's a WIP need to do some additional testing since that role is not tested in gate jobs | 17:44 |
*** ocaner has quit IRC | 17:44 | |
sean-k-mooney | sure. that role was not complet thogh | 17:45 |
*** portdirect_ has quit IRC | 17:45 | |
sean-k-mooney | it was ment to be extended to setup the build host and also do several other operations on the host in pike | 17:45 |
dmsimard | I'm not really adding anything to it | 17:46 |
dmsimard | just cleaning it up | 17:46 |
dmsimard | it should be no-op | 17:47 |
dmsimard | ok I lied, the only real change is installing docker-py from packages instead of from pip | 17:47 |
sean-k-mooney | i will conflict with the change i was planning on added of ovs-dpdk this cycle but ill start reviewing what you have and see if i can build on it | 17:47 |
dmsimard | what's ovs-dpdk have to do with boostrapping a build/deploy node ? | 17:48 |
sean-k-mooney | i need to allocate hugepages and do driver configuration on the host | 17:48 |
dmsimard | I feel like that would be a better fit in wherever ovs-dpdk is installed/setup imo | 17:49 |
SamYaple | sean-k-mooney: really, hugepages is a great tuning that should be done anyway | 17:49 |
SamYaple | sean-k-mooney: we run VMs. they love hugepages | 17:49 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP-PoC Glance service orchestration with entrypoint https://review.openstack.org/410736 | 17:49 |
sean-k-mooney | SamYaple: yes they doo | 17:49 |
sean-k-mooney | dmsimard: the baremetal role was intended to do all host prep before kolla deploys openstack | 17:50 |
sean-k-mooney | dmsimard: or at least the bootstrap-servers was | 17:50 |
sean-k-mooney | it could be a sperate role | 17:50 |
SamYaple | sean-k-mooney: im running dpdk + neutron with openvswitch firewall at my house now! (havent combined the two yet) | 17:51 |
dmsimard | sean-k-mooney: yeah imo this ought to be a role that installs the dependencies for kolla build/deploy to work | 17:51 |
sean-k-mooney | dmsimard: https://review.openstack.org/#/c/408876/ is how i was going to adress the hugepage requirement | 17:51 |
SamYaple | sean-k-mooney: this is outside of kolla | 17:51 |
sean-k-mooney | SamYaple: its not outside the scope of the kolla host playbook | 17:52 |
sean-k-mooney | SamYaple: its outside the scope of the kolla deploy playbook though | 17:52 |
SamYaple | sean-k-mooney: no no i mean i was running dpdk outside of kolla (non-kolla openstack dpeloy) | 17:53 |
dmsimard | sean-k-mooney: basically where I'm headed right now is to 1) cleanup the bootstrapping 2) leverage that to setup to nodepool gate node 3) cleanup the gate testing setup | 17:53 |
sean-k-mooney | SamYaple: oh well you can do that with devstack we provide a plugin for that, trippleo and fuel can deploy it too | 17:53 |
SamYaple | yep | 17:53 |
sean-k-mooney | SamYaple: i am hoping to land support for ovs-dpdk this cycle though so maybe you can use it in a month or so | 17:54 |
SamYaple | awesome | 17:54 |
sean-k-mooney | SamYaple: https://review.openstack.org/#/q/topic:bp/ovs-dpdk | 17:55 |
sean-k-mooney | it actully works today but i have a few hacks i need to fix | 17:55 |
*** sayantan_ has quit IRC | 17:56 | |
*** rhallisey has joined #openstack-kolla | 17:57 | |
sean-k-mooney | the main one is running qemu instnace in the same group as ovs-dpdk so that they can acess the vhost-user sockets | 17:57 |
sean-k-mooney | dmsimard: yes that does sound good. but again the boot-strap server action in the future will do more then jsut install the dependcies for kolla-deploy | 17:59 |
sean-k-mooney | dmsimard: it will only do the addtional options if you request them though | 17:59 |
dmsimard | sean-k-mooney: then we should abstract the bootstrap into a bootstrap role and keep the baremetal role for some other things on top | 17:59 |
dmsimard | We need something to "just" bootstrap the node for kolla-build/kolla-deploy | 18:00 |
sean-k-mooney | dmsimard: well the intent is that if you doont aske ti to do addtional work by enabling it in the global.yml it wont | 18:00 |
*** lamt has quit IRC | 18:00 | |
dmsimard | so the dpdk stuff will default to false ? | 18:01 |
sean-k-mooney | yep | 18:01 |
sean-k-mooney | https://review.openstack.org/#/c/408876/1/ansible/roles/baremetal/tasks/grub.yml | 18:01 |
dmsimard | I'm okay with that | 18:01 |
sean-k-mooney | if you look at the kernel_args they only activate that code path if it defiend and it is not by default | 18:02 |
*** zhugaoxiao has quit IRC | 18:02 | |
sean-k-mooney | so none of the addtional fuctiontionality i want to introduce will execut unless you explictly set a value to true or defient it. it will all be off by default | 18:03 |
*** zhugaoxiao has joined #openstack-kolla | 18:03 | |
sean-k-mooney | my intent is to keep the workflow the same and just enable things like configuring the kernel commandline if the use request that | 18:04 |
*** pbourke has quit IRC | 18:05 | |
*** unicell has joined #openstack-kolla | 18:05 | |
*** pbourke has joined #openstack-kolla | 18:05 | |
*** wwriverrat_ has joined #openstack-kolla | 18:07 | |
portdirect | evening o/ | 18:09 |
SamYaple | o/ | 18:10 |
*** Pavo has quit IRC | 18:12 | |
portdirect | SamYaple: cheers for charing the salt stuff, I've never really played with it, but have always been impressed by the speed seems to get through things | 18:13 |
openstackgerrit | Merged openstack/kolla-kubernetes: Helm chart for cinder-scheduler https://review.openstack.org/410525 | 18:14 |
SamYaple | portdirect: yea right now ive fully implemented rabbitmq (and memcache, what littel there is). it takes less than a minute to run fully, and only 5 seconds to run when nothing changes | 18:14 |
*** JoseMello has joined #openstack-kolla | 18:15 | |
sean-k-mooney | as in a 5 second run with the equivalent of kolla-ansible deploy | 18:15 |
portdirect | SamYaple: from humble beginings come great things :) I'm going through the rabbit stuff now - looks good! | 18:16 |
sean-k-mooney | do you have reconfigure support? | 18:16 |
SamYaple | sean-k-mooney: right now, baked in, it restarts containers when config changes are detected (per host) | 18:16 |
*** Pavo has joined #openstack-kolla | 18:17 | |
sean-k-mooney | SamYaple: cool the reconfigre action in kolla-ansible is still rather slow | 18:18 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance keystone endpoints https://review.openstack.org/411607 | 18:19 |
SamYaple | sean-k-mooney: reconfigure takes ~5 seconds. most of that is just waiting for the container to stop | 18:19 |
SamYaple | i suppose it could take longer depending on what the service must do to shut down | 18:19 |
SamYaple | portdirect: i really like the macros and how the help to DRY-ify the code | 18:21 |
portdirect | yeah i was literally just looking at them | 18:21 |
sean-k-mooney | SamYaple: kolla-ansible currently spend most of its time trying to figure out what changed and what need to be updated | 18:21 |
SamYaple | sean-k-mooney: indeed. its partly because we dont use handlers, and partly because we _cant_ use handlers :P | 18:22 |
SamYaple | ssh-ing per task doesnt help | 18:22 |
SamYaple | portdirect: mind you this is all _very_ rough code. im not sure what is goign to stick. what is readable. what is _reasonable_ | 18:22 |
SamYaple | all that stuff | 18:22 |
sean-k-mooney | yes that is true. we should perhaps consider alling out how to use pipline more clearly in the docs | 18:23 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm Chart for Cinder Volume https://review.openstack.org/411927 | 18:23 |
sean-k-mooney | anyway got to run. im actully already on vacation but i need to rebase some patches before i go | 18:23 |
SamYaple | bye sean-k-mooney | 18:24 |
SamYaple | good catching up | 18:24 |
sean-k-mooney | yep ill take a look at some of the kolla-salt stuff at a later date too. it sound interesting | 18:25 |
sean-k-mooney | eventurally ill get a chance too looka at kolla-k8s | 18:25 |
sean-k-mooney | eventurally... | 18:25 |
SamYaple | :) | 18:26 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm Cinder Backup https://review.openstack.org/411570 | 18:26 |
SamYaple | sean-k-mooney: i can write "tasks" in pure python with salt | 18:26 |
SamYaple | very powerful | 18:26 |
*** adrian_otto has quit IRC | 18:30 | |
*** jgriffith is now known as jgriffith_AutoAw | 18:33 | |
*** sdake has joined #openstack-kolla | 18:33 | |
*** vhosakot has quit IRC | 18:34 | |
*** goldyfruit has quit IRC | 18:38 | |
Bico_Fino | Jeffrey4l: Tried the patch. http://paste.openstack.org/show/592671/ | 18:39 |
v1k0d3n | SamYaple: you probably know the kolla-build process better than most. seems like you did some work on that, is that right? | 18:40 |
*** goldyfruit has joined #openstack-kolla | 18:40 | |
SamYaple | v1k0d3n: i was around for alot of the work, yea | 18:42 |
v1k0d3n | reason i ask, is because i want to customize kolla-build.conf. if i remove [ITEM] will it still work correctly? or would the missing headings break the build process? | 18:42 |
*** vhosakot has joined #openstack-kolla | 18:43 | |
SamYaple | so htats more of a question for oslo.config, but im 95% sure you can only specify they items you want to _override_. | 18:43 |
SamYaple | the rest would only get defaults | 18:43 |
SamYaple | (thats what i do) | 18:43 |
v1k0d3n | and i am admittedly using a hacky container to do this automatically. | 18:43 |
v1k0d3n | and i only want to define overrides. | 18:44 |
v1k0d3n | so basically... | 18:44 |
SamYaple | yea. thats exactly what i do | 18:44 |
*** haplo37_ has quit IRC | 18:44 | |
v1k0d3n | what i want to do...is instead of doing this: https://github.com/att-comdev/dockerfiles/blob/master/kolla-builder/scripts/entrypoint.sh#L15-L19 | 18:44 |
portdirect | v1k0d3n: sams right re the olso.config - you only need to spec overrides from the default | 18:45 |
v1k0d3n | i want to just start off fresh, rather than setting things. i've even thought about allowing a developer to import their own file using an ENV override. (this may be easier). | 18:45 |
v1k0d3n | ok. so the default headings in the kolla-build.conf in the etc/ dir is just an example and the headings could just be removed...right? | 18:46 |
v1k0d3n | i think i'm understanding correctly, just being clear (because i'm pretty slow). lol | 18:46 |
portdirect | v1k0d3n: yeah thats what i do for all my stuff, i never use the config files, but generate them using crusidi and env cars every time c acontainer/pod starts | 18:47 |
v1k0d3n | sorry for asking oslo questions too btw. | 18:47 |
v1k0d3n | ah, so i don't even need to copy that file. that helps. | 18:47 |
SamYaple | v1k0d3n: no its fine. i was just letting you know we use oslo.config and thats how oslo.config works | 18:47 |
v1k0d3n | step removed. | 18:47 |
v1k0d3n | makes sense. | 18:47 |
*** fragatina has quit IRC | 18:47 | |
v1k0d3n | thanks SamYaple | 18:47 |
v1k0d3n | the oslo stuff is extremely detailed. | 18:48 |
v1k0d3n | docs are good. almost overkill for what i'm really looking for. | 18:48 |
*** mliima has quit IRC | 18:49 | |
SamYaple | portdirect: https://docs.saltstack.com/en/latest/ref/states/all/salt.states.zk_concurrency.html | 18:49 |
SamYaple | portdirect: thats going to allow fine control over service reboots | 18:49 |
v1k0d3n | and as far as kolla is concerned (this is probably also an oslo question), i can also just point to custom keystone base GIT repo and that should be just fine with kolla image builds? | 18:50 |
*** tonanhngo has quit IRC | 18:50 | |
*** gfidente is now known as gfidente|afk | 18:51 | |
portdirect | SamYaple: *that* is nice, i need to look into ZMQ security but i kinda wanna play | 18:52 |
*** JoseMello has quit IRC | 18:52 | |
*** haplo37_ has joined #openstack-kolla | 18:54 | |
*** sdake has quit IRC | 18:54 | |
*** Pavo has quit IRC | 18:54 | |
*** Pavo has joined #openstack-kolla | 18:55 | |
*** jgriffith_AutoAw is now known as jgriffith | 18:55 | |
*** JoseMello has joined #openstack-kolla | 18:56 | |
*** sdake has joined #openstack-kolla | 18:56 | |
*** gomarivera has quit IRC | 18:57 | |
v1k0d3n | SamYaple: so did you refactor kolla for salt use? | 18:58 |
v1k0d3n | if so, very slick. salt is impressive. | 18:58 |
*** mliima has joined #openstack-kolla | 18:58 | |
v1k0d3n | i've seen you mention a couple of times. | 18:58 |
SamYaple | v1k0d3n: no im working on an all new deploy using saltstack | 18:59 |
v1k0d3n | awesome! | 19:00 |
SamYaple | portdirect: its as secure as ssh, if not more due to not being ssh :P | 19:00 |
v1k0d3n | project in gh or garrit, or personal? | 19:00 |
SamYaple | v1k0d3n: if all goes to plan, it will be official project | 19:00 |
portdirect | SamYaple: well in that case :) when i get time expect me to be pestering you :P | 19:01 |
SamYaple | portdirect: it uses public key auth just like ssh, but you must expliciltly except keys and such | 19:01 |
SamYaple | accept* | 19:01 |
SamYaple | portdirect: clemson university is actually using it without ssh on the host at all | 19:01 |
SamYaple | all commands go through salt | 19:01 |
portdirect | nice! | 19:01 |
v1k0d3n | that's awesome! love seeing options pop up for opinionated deployments (CAPS). | 19:02 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance keystone endpoints https://review.openstack.org/411607 | 19:02 |
v1k0d3n | core projects you want everyone working in one, but for CAPS that just makes total sense. | 19:02 |
v1k0d3n | deployments by nature is opinionated. it has to be. | 19:02 |
v1k0d3n | btw you're close. in charlotte here. | 19:02 |
v1k0d3n | brother in law went to clemson | 19:02 |
SamYaple | im actually in greenville, sc | 19:03 |
portdirect | I've been playing with this helm plugins - I kinda like how getto they are, means you can do pretty much anything you want: https://github.com/portdirect/marina/ | 19:03 |
SamYaple | i just know a bit about clemson | 19:03 |
v1k0d3n | nice to have some other southerners in here ;) | 19:03 |
v1k0d3n | gotcha | 19:03 |
*** JoseMello has quit IRC | 19:04 | |
Pavo | SamYaple you only like an hour from me then | 19:09 |
SamYaple | Pavo: NO YOURE ONE HOUR FROM ME! | 19:10 |
Pavo | haha | 19:11 |
sbezverk | man helm templating is so buggy :-( I almost wist to come back to jinja2 | 19:11 |
jascott1 | I saw the band FrogWings in Greenville SC (allman bros spin off) about 20 yrs ago | 19:12 |
SamYaple | sbezverk: lol i dont know what helm is by based on the frequency of patchsets i want to stay away | 19:12 |
jascott1 | haha | 19:13 |
*** zhubingbing has quit IRC | 19:15 | |
*** sdake has quit IRC | 19:16 | |
*** JoseMello has joined #openstack-kolla | 19:16 | |
sbezverk | SamYaple: more scary is the number os issue we need to file.. | 19:17 |
*** sdake has joined #openstack-kolla | 19:17 | |
*** Bico_Fino has quit IRC | 19:17 | |
SamYaple | :( | 19:18 |
portdirect | sbezverk: I actually kinda think its almost a good thing :/ though rubbish in the short term we are both young projects - and they are listening to what we have to say | 19:21 |
portdirect | sbezverk: hopefully you dont quote me on that in six months time to make me feel bad! :D | 19:22 |
sbezverk | portdirect: yeah but I would really prefer that somebody else hit these obstacles is it too egoistic ;-) ? | 19:24 |
portdirect | i do actually feel sorry for you. everytime i find a problem, i find you have already encountered it or filed a bug report :) sbezverk, all your bugs belong to him. | 19:26 |
srwilkers | oh hey sbezverk, jascott1 | 19:28 |
sbezverk | portdirect: Oh I know what I need to do :-) I will take a couple of days off, so you can move forward and start hitting helm | 19:28 |
portdirect | no!! | 19:29 |
jascott1 | srwilkers hey hows things? | 19:29 |
*** sayantan_ has joined #openstack-kolla | 19:29 | |
srwilkers | pretty good man. how about you? | 19:30 |
*** vhosakot has quit IRC | 19:30 | |
*** mliima has quit IRC | 19:30 | |
sbezverk | srwilkers: hey man, when do you start contributing to helm, I already have a wish list for you ;-) | 19:33 |
jascott1 | srwilkers preparing to goto ATT store and my daughter's phone upgraded :/ other than that im good | 19:33 |
srwilkers | sbezverk, going to start tonight and over the weekend. ;) | 19:34 |
srwilkers | jascott1, oh man, im sorry. im an employee, and they still bug me about dropping my unlimited data every time i go in | 19:34 |
jascott1 | i have unlimited as well | 19:35 |
jascott1 | and my wife | 19:35 |
jascott1 | daughter doesnt tho | 19:35 |
jascott1 | frankly I dont use it and dont want to pay for it anymore | 19:35 |
jascott1 | maybe you know a way, but i end up using regular amount and paying like $230 for 3 lines/mo | 19:36 |
srwilkers | yeah, i dont blame you. i burn through it like crazy. a typical month for me is anywhere between 15-18 gigs or so | 19:36 |
srwilkers | oh wow | 19:36 |
*** lamt has joined #openstack-kolla | 19:36 | |
srwilkers | hey lamt | 19:36 |
jascott1 | i thnk i can get son on and have 4 smart phones for less than that! | 19:36 |
jascott1 | cant do tethering on unlimeted either and some other gotacha iirc | 19:37 |
*** sdake has quit IRC | 19:37 | |
jascott1 | or can do tethering but its like 40$/mo or something | 19:37 |
jascott1 | anyway enough about my phone bill sorry everyone | 19:37 |
srwilkers | let the rage flow through you | 19:38 |
jascott1 | lol | 19:38 |
lamt | srwilkers o/ | 19:38 |
kfox1111 | I'm on the super secret tmobile walmart plan. | 19:38 |
jascott1 | hmm might bring my sith body guards | 19:38 |
kfox1111 | like 140 minutes a month, unlimited text/data, $30 bucks a month. | 19:39 |
srwilkers | hey kfox1111 :) | 19:39 |
kfox1111 | hi. :) | 19:39 |
kfox1111 | or something like that. cant remember the exact number of minutes. | 19:39 |
kfox1111 | I hardly ever call anyone on the phone though so a sweet deal. | 19:39 |
srwilkers | yeah im the same way. i hate talking on the phone :( | 19:39 |
jascott1 | right | 19:39 |
jascott1 | most men do imo :) | 19:40 |
jascott1 | know i do... | 19:40 |
*** jmccarthy has quit IRC | 19:40 | |
kfox1111 | work uses lync so all of that's free over ip. | 19:40 |
kfox1111 | and most other places use hangouts, zoom, webex, etc. | 19:40 |
kfox1111 | also all free. | 19:40 |
*** Bico_Fino has joined #openstack-kolla | 19:41 | |
srwilkers | yeah. most of the people i talk to almost use slack exclusively, and you can do calls over slack now | 19:41 |
kfox1111 | oh. slack can now? | 19:41 |
*** rhallisey has quit IRC | 19:41 | |
*** jmccarthy has joined #openstack-kolla | 19:41 | |
kfox1111 | sweet. | 19:41 |
kfox1111 | I use slack a lot too. | 19:41 |
kfox1111 | never noticed that option. | 19:41 |
jascott1 | they have video too now i read but never used | 19:41 |
jascott1 | maybe thats premium idk. assumed it was all p2p browser based anyway | 19:42 |
*** sdake has joined #openstack-kolla | 19:42 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP-PoC Glance service orchestration with entrypoint https://review.openstack.org/410736 | 19:42 |
kfox1111 | ah. that would make sense. | 19:42 |
*** Bico_Fino has quit IRC | 19:46 | |
*** lamt has quit IRC | 19:48 | |
*** lamt has joined #openstack-kolla | 19:51 | |
*** srwilkers is now known as srwilkers_brb | 19:51 | |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 19:54 |
*** fragatina has joined #openstack-kolla | 19:55 | |
openstackgerrit | Merged openstack/kolla-kubernetes: Helm charts for glance keystone endpoints https://review.openstack.org/411607 | 19:56 |
kfox1111 | almost all microserviced. | 19:57 |
kfox1111 | mostly just endpoints at this point. | 19:57 |
kfox1111 | :) | 19:57 |
srwilkers_brb | yup :) just gotta tie up the glance bits today on my end | 19:57 |
srwilkers_brb | brb, foosball table calls | 19:57 |
kfox1111 | awesome. :) | 19:57 |
kfox1111 | srwilkers_brb: have fun. :) | 19:57 |
openstackgerrit | David Moreau Simard proposed openstack/kolla-ansible: WIP: Refactor the baremetal role https://review.openstack.org/411930 | 19:57 |
openstackgerrit | sayantani proposed openstack/kolla: Divide Quick Start Guide into 2 docs https://review.openstack.org/404993 | 19:59 |
*** rhallisey has joined #openstack-kolla | 19:59 | |
*** sdake has quit IRC | 20:02 | |
*** JoseMello has quit IRC | 20:02 | |
kfox1111 | alanmeadows: you around? | 20:03 |
alanmeadows | kfox1111: hey | 20:03 |
*** sdake has joined #openstack-kolla | 20:04 | |
*** JoseMello has joined #openstack-kolla | 20:05 | |
sayantan_ | Hi srwilkers | 20:07 |
*** eaguilar has quit IRC | 20:09 | |
*** jgriffith is now known as jgriffith_AutoAw | 20:20 | |
*** jgriffith_AutoAw is now known as jgriffith | 20:23 | |
*** sdake has quit IRC | 20:24 | |
*** HyperJohnGraham has joined #openstack-kolla | 20:24 | |
*** srwilkers_brb is now known as srwilkers | 20:26 | |
*** sdake has joined #openstack-kolla | 20:26 | |
srwilkers | hey sayantan_ :) | 20:26 |
sayantan_ | Hey srwilkers. I do agree with you. That patch has become too large. Getting difficult for me to maintain as well. | 20:28 |
sayantan_ | I will take up dividing according to distros in a new patch set | 20:28 |
sayantan_ | Spoke with the documentation guys | 20:28 |
sayantan_ | they have an easier of doing it | 20:28 |
sayantan_ | As of now, have kept the operators and developers guides | 20:29 |
*** ayoung has joined #openstack-kolla | 20:29 | |
*** rhallisey has quit IRC | 20:33 | |
*** JoseMello has quit IRC | 20:39 | |
*** HyperJohnGraham has quit IRC | 20:43 | |
*** sdake has quit IRC | 20:46 | |
*** sdake has joined #openstack-kolla | 20:49 | |
*** tonanhngo has joined #openstack-kolla | 20:51 | |
*** tonanhngo has quit IRC | 20:52 | |
*** fragatina has quit IRC | 20:53 | |
v1k0d3n | portdirect: hey man you around? | 20:54 |
*** rsjethani has quit IRC | 20:54 | |
* portdirect turns off the lights and hides behind sofa | 20:55 | |
*** gomarivera has joined #openstack-kolla | 20:55 | |
*** Pavo has quit IRC | 20:55 | |
v1k0d3n | c'mon man. only throwing warm fuzzies. it's friday. | 20:55 |
*** mkoderer has quit IRC | 20:56 | |
*** gomarivera has quit IRC | 20:59 | |
*** Pavo has joined #openstack-kolla | 21:00 | |
srwilkers | oh hey v1k0d3n | 21:03 |
srwilkers | long time no chat | 21:03 |
v1k0d3n | no kidding! | 21:04 |
v1k0d3n | :) | 21:04 |
*** sdake has quit IRC | 21:09 | |
*** sdake has joined #openstack-kolla | 21:10 | |
*** sdake has quit IRC | 21:10 | |
*** sdake has joined #openstack-kolla | 21:10 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance api and glance registry pods https://review.openstack.org/410481 | 21:12 |
*** aric49 has joined #openstack-kolla | 21:13 | |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance api and glance registry pods https://review.openstack.org/410481 | 21:14 |
kfox1111 | srwilkers: there's a wip in the commit message of that. it shoudl be ready to merge I think when it passes tests. | 21:14 |
kfox1111 | can you pull that out? | 21:14 |
srwilkers | kfox1111, of course | 21:15 |
kfox1111 | thx. :) | 21:15 |
aric49 | Hi Folks, I am having an issue with a newly deployed kolla cluster on Ubuntu 16.04 baremetal servers. I am seeing that the rabbitmq container is restarting itself every few seconds on my controller nodes | 21:15 |
openstackgerrit | Steve Wilkerson proposed openstack/kolla-kubernetes: Helm charts for glance api and glance registry pods https://review.openstack.org/410481 | 21:15 |
*** lamt has quit IRC | 21:16 | |
aric49 | the docker ps -a status is showing: Restarting (0) About a minute ago | 21:16 |
aric49 | this is for kolla release: 3.0.1 | 21:16 |
srwilkers | heading home before this freezing rain gets in. i shall return | 21:18 |
*** jtriley has quit IRC | 21:19 | |
*** srwilkers has quit IRC | 21:19 | |
openstackgerrit | David Moreau Simard proposed openstack/kolla-ansible: Refactor the baremetal role https://review.openstack.org/411930 | 21:20 |
*** lamt has joined #openstack-kolla | 21:20 | |
*** sdake has quit IRC | 21:30 | |
*** gfidente|afk has quit IRC | 21:31 | |
*** sdake has joined #openstack-kolla | 21:32 | |
*** l4yerffeJ has joined #openstack-kolla | 21:32 | |
*** Jeffrey4l has quit IRC | 21:35 | |
*** l4yerffeJ_ has quit IRC | 21:35 | |
*** Jeffrey4l has joined #openstack-kolla | 21:36 | |
openstackgerrit | Merged openstack/kolla-kubernetes: Helm Cinder Backup https://review.openstack.org/411570 | 21:41 |
*** fguillot has quit IRC | 21:41 | |
*** adrian_otto has joined #openstack-kolla | 21:42 | |
*** msimonin has quit IRC | 21:43 | |
*** msimonin has joined #openstack-kolla | 21:47 | |
*** adrian_otto has quit IRC | 21:49 | |
sdake | @srwilker- i had unlimited data on verizon some time ago but they jerked around with teh plans so much i coudln't get my kids on the unlimited plan without giving up unlimited | 21:50 |
sdake | now my phone bill is 611 a month | 21:51 |
sdake | almost more then my mortgage payment (@740 a mo) | 21:51 |
jascott1 | sdake wow | 21:52 |
jascott1 | thought i was paying alot | 21:52 |
sdake | not helpful everyone keeps breaking their devices and needing replacements - even though we have insurance - which is a scam | 21:52 |
sdake | jascott1 ya - its gone up lately - need to have a look at the bill | 21:52 |
sdake | it was 400 wich was significant but tolerable | 21:52 |
sdake | we have a 30gb data plan | 21:52 |
sdake | my wife uses 22gb of it | 21:52 |
sdake | every month | 21:52 |
sdake | i use about 0.5 gb | 21:53 |
jascott1 | just go back from ATT, im paying less now but didnt have to drop much cash for my daughter's iphone 7 and they "threw in" ipad mini | 21:53 |
sdake | cool throw in is good | 21:53 |
sdake | at verizon I have to pay full cash, but gte free finacning over 24 months | 21:53 |
jascott1 | ha was just looking my wife used about 19GB i was at 2 somehow, idk I just mostly read HN and stuff | 21:53 |
*** vhosakot has joined #openstack-kolla | 21:54 | |
sdake | jascott1 i barely use my cell phone except google maps | 21:54 |
sdake | no the road, our data usage is very high however | 21:54 |
jascott1 | primary function: maps | 21:54 |
sdake | jascott1 when I travel, my phone is invaluable | 21:54 |
jascott1 | man when I go downtown my phone is the only chance I have of making it back ;) | 21:54 |
sdake | what sucks is i dropped my edge 7 and the screen is busted | 21:54 |
sdake | yet I suck it up because it just means more pain to fix | 21:55 |
jascott1 | ah bummer | 21:55 |
sdake | jascott1 ;-) | 21:55 |
sdake | jascott1 your in portland right? I can get aroudn portland pretty well without a map | 21:55 |
jascott1 | gotta iphone5 right here thats been spiderwebd, sitting here for a couple years | 21:55 |
sdake | jascott1 oregon is appealing to me because of portland | 21:55 |
jascott1 | ya i can get around the main downtown part but east side gets confusing quick | 21:55 |
sdake | mine is still functional although the glass doesn't "look pretty" | 21:56 |
jascott1 | "oh I will make a U turn." You are now going over the river | 21:56 |
sdake | jascott1 oh right - not sure where the east side is :) | 21:56 |
sdake | oh, I walk | 21:56 |
sdake | i'd possibly take uber now | 21:56 |
sdake | been a long time since i've been to portland actually | 21:56 |
v1k0d3n | SamYaple: you still around? have another container build question, since you know it really well. | 21:56 |
sdake | but i really like it and could live downtown carless | 21:56 |
jascott1 | we have a great train/public transit system ehre | 21:56 |
jascott1 | no turnstiles | 21:56 |
sdake | jascott1 right - arizona - that doesn't work because everything is so spread out | 21:57 |
sdake | jascott1 and its a 115 avg temperature in the summer months | 21:57 |
sdake | jascott1 have pubic transport, tkaes 2 hrs to get anywhere | 21:57 |
v1k0d3n | if i build keystone, there are a lot of other depends. and all i want is just keystone-base and no dependancies...is there a way to remove the linking without changing the kolla-build process? | 21:57 |
sdake | v1k0d3n no but there should be | 21:58 |
sdake | v1k0d3n i just am not quite sure how to make the implementation do that | 21:58 |
v1k0d3n | yeah :( in our case that's a bit deal. i would imaging for other big corps also. | 21:58 |
sdake | v1k0d3n the tricky part is not building but pulling the parent | 21:58 |
sdake | v1k0d3n its actually a big deal for alot of people | 21:58 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Cleanup ceph_workflow https://review.openstack.org/412004 | 21:58 |
sdake | godaddy has the same problem - perhaps harlowja has some ideas when he comes around | 21:58 |
sdake | v1k0d3n what needs to happen I htink is the ability to pull the parent image from somewhere rather then rebuild it | 21:59 |
v1k0d3n | for small/med i can see why they would want to have the other depends, but if you're a developer or big corp...you'd use ci anyway. | 21:59 |
sdake | v1k0d3n even medium businesses (100-500 employees) may use a ci pipeline | 22:00 |
sdake | small probably be satisified with teh current build model | 22:00 |
sdake | v1k0d3n i'm right here with you on buildling only leaf images | 22:02 |
SamYaple | v1k0d3n: you can build just keystone-base | 22:02 |
sdake | v1k0d3n if we could sort out how to do that, that would be huge | 22:02 |
sdake | SamYaple right so e.g. build nova-novncproxy | 22:02 |
*** msimonin has quit IRC | 22:02 | |
*** msimonin1 has joined #openstack-kolla | 22:02 | |
sdake | builds all the dependencies | 22:02 |
sdake | everyt ime | 22:02 |
sdake | in a ci pipeline, that sucks | 22:02 |
*** msimonin1 has quit IRC | 22:02 | |
sdake | it would be better if the deps were pulled | 22:02 |
SamYaple | nien | 22:02 |
SamYaple | thats long been decided | 22:02 |
SamYaple | otherwise all teh libraries may not be consistent | 22:03 |
*** unicell has quit IRC | 22:03 | |
SamYaple | until we no longer have a "base" image, it must build like this | 22:03 |
jascott1 | thats why we should use version pinning | 22:03 |
* jascott1 ducks | 22:03 | |
sdake | SamYaple idea is people provide their own repos | 22:03 |
sdake | SamYaple and own everything - then can build leaf images without that problem | 22:03 |
SamYaple | yea well then the build cache should cover that | 22:04 |
sdake | there is no build cache in a ci pipeline | 22:04 |
SamYaple | i mean its really easy to implement a --no-deps option | 22:04 |
SamYaple | its been rejected several times | 22:04 |
sdake | i dont recall ever seeing one implemented | 22:04 |
SamYaple | its two lines of code and youve been involved in the discussion before | 22:05 |
sdake | i think offering the capability with appropriate disclaimers makes alot of sense | 22:05 |
sdake | SamYaple possible i was | 22:05 |
sdake | SamYaple possible I was wrong if I supported not having a --no-deps implementation | 22:05 |
sdake | SamYaple tbh I can't remember what happend yesterday :) | 22:05 |
SamYaple | anyway, getting rid of the base layer as it is now and getting smaller images is the way to go | 22:05 |
SamYaple | that way all images are independant of eachother | 22:06 |
SamYaple | and this isn't a problem | 22:06 |
sdake | `how do we get rid of base? | 22:07 |
sdake | I'd be pro independent images :) | 22:07 |
sdake | the reason we have base now is because everything is so bloated | 22:07 |
sdake | and its not really our fault distros bloat their distros | 22:07 |
aric49 | hey guys, having an issue with the 3.0.1 RabbitMQ container. The container keeps restarting itself every few minutes --- has anyone encountered an issue like that? | 22:07 |
SamYaple | nah i got ubuntu keystone down to 40MB | 22:07 |
*** unicell has joined #openstack-kolla | 22:08 | |
sdake | did it involve non-native docker tooling? | 22:08 |
SamYaple | sdake: which docker themselves said is what we should be using | 22:08 |
jascott1 | aric49 anything in the logs? | 22:08 |
sdake | aric49 haven't seen that - which distro? | 22:08 |
aric49 | I am deploying CentOS binary | 22:09 |
sdake | SamYaple i missed that last part - docker said we should be doing which? | 22:09 |
sdake | aric49 centos binary works well for me | 22:09 |
aric49 | let me setup a gist with the log output | 22:09 |
sdake | SamYaple how do you get keystone down to 40gb? | 22:09 |
stevemar | mb! | 22:10 |
stevemar | :P | 22:10 |
sdake | right mb :) | 22:10 |
sdake | aric49 that would help | 22:10 |
aric49 | https://gist.github.com/aric49/54600e42ace4d68b9af56d626d14a938 | 22:10 |
*** jgriffith is now known as jgriffith_AutoAw | 22:11 | |
SamYaple | sdake: https://github.com/docker/docker/issues/332#issuecomment-232646326 | 22:11 |
sdake | v1k0d3n what sam proposes if it could be made to work would be perfect for your use case | 22:11 |
aric49 | it seems to just repeat what is in the gist every minute or so | 22:11 |
SamYaple | and that comment is mirrored elsewhere. the Dockerfile is not the _only_ way to build docker images | 22:11 |
aric49 | i can exec into the container, but it kicks me out after a few seconds | 22:11 |
SamYaple | aric49: did you run the prechecks? | 22:12 |
aric49 | yes | 22:12 |
*** adrian_otto has joined #openstack-kolla | 22:12 | |
aric49 | prechecks pass | 22:12 |
SamYaple | ok | 22:12 |
*** jgriffith_AutoAw is now known as jgriffith | 22:12 | |
aric49 | all openstack services are up, but i cant see any hypervisors or create any VMs.... when I do a docker ps -a, it shows the rabbitmq container in a "restarted" state | 22:13 |
SamYaple | yea i dont run with "restart_policy: always" for this very reason | 22:13 |
SamYaple | missing key details like exit codes | 22:13 |
aric49 | i have also tried rebuilding the image and redeploying | 22:14 |
klindgren | is HEKA a requirement to run kolla-ansible? | 22:15 |
klindgren | I dont see a way to disable it | 22:15 |
*** Pavo has quit IRC | 22:16 | |
SamYaple | klindgren: afaik kolla-ansible requires heka, at least as of mitaka | 22:17 |
SamYaple | i dont think that changed in newton | 22:17 |
klindgren | what does heka provide? | 22:17 |
SamYaple | logging stuff | 22:18 |
sdake | klindgren heka forwards logs to elasticsearch from each node | 22:19 |
sdake | kibana visualualizes the logs | 22:19 |
sdake | klindgren in other words, heka is for diagnostics | 22:19 |
klindgren | And if I am trying to slowly integrate kolla into an existing env and already have a solution for that? | 22:19 |
sdake | klindgren it could probably be made optional - but it is not at this point in time | 22:19 |
sdake | klindgren also we are replacing heka with fluentd in master | 22:20 |
sdake | klindgren (if it makes ocata) from my understanding | 22:20 |
SamYaple | klindgren: pretty much no for kolla-ansible. it would be a hard cut over. | 22:20 |
SamYaple | klindgren: i am working on kolla-salt, with one of its stated goals being able ot do that. but thats not now | 22:20 |
sdake | klindgren its a matter of code - however, its not a trivial thing | 22:20 |
klindgren | is their any other things that are required like heka? | 22:21 |
sdake | klindgren nothing pops to mind | 22:21 |
sdake | klindgren keep in mind i just got up for a 6 hour nap :) | 22:21 |
SamYaple | klindgren: the non-openstack services that get deployed by default are rabbitmq,mariadb, and memcached | 22:21 |
SamYaple | and heka of course | 22:22 |
SamYaple | i can't think of another one. might be missing one | 22:22 |
sdake | SamYaple i think that is correct | 22:22 |
sdake | heka is pretty much built in - the question klindgren - areou using elasticsearch at all? | 22:22 |
klindgren | so I already have rabbitmq, mariadb, memcached - is their a way for me to jsut supply what to use for config options to those and dont deploy them? | 22:22 |
sdake | klindgren you can disable everything but heka | 22:22 |
klindgren | we use logstash to pump logs into hadoop and kafka which then splits them also into elastic search | 22:23 |
sdake | klindgren however your on your own for managing rabbitmq/mariadb/memecached in this case | 22:23 |
klindgren | IE - this is a solved problem for us | 22:23 |
SamYaple | klindgren: yea, but youll need to _really_ understand the kolla config system for that to work | 22:23 |
sdake | klindgren you mean you use logstash-forwarded to pump logs into hadoop | 22:24 |
klindgren | we also need to manage the rabbitmq config/mariadb/memachaced system per nova cell | 22:24 |
sdake | klindgren logstash-forwarder is EOL | 22:24 |
klindgren | not exactly. logstash parses the logs, and technially it outputs it to some crappy nodejs thing that parses it and send the correct messages to some internal company api | 22:24 |
sdake | klindgren heka is also EOL - which is why we are moving to fluentd | 22:24 |
sdake | klindgren each node has to run a forwarder | 22:25 |
sdake | that is what heka is | 22:25 |
sdake | a forwarder | 22:25 |
klindgren | which then puts the messages on kafka, which then splits it between hadoop and ES | 22:25 |
*** dave-mccowan has quit IRC | 22:25 | |
sdake | there are two parts to logstash | 22:25 |
jascott1 | oh hey sdake, can you add or give me one liner about what this is? https://blueprints.launchpad.net/kolla-kubernetes/+spec/helm-meta | 22:25 |
sdake | logstash (the daemon) and logstash-forwarder | 22:25 |
sdake | jascott1 helm install openstack | 22:26 |
jascott1 | ah | 22:26 |
jascott1 | and only that I take it? | 22:26 |
jascott1 | only deals with that package namespace 'openstack' | 22:26 |
klindgren | sdake, confused, we are just using the normal logstash-output plugin? | 22:27 |
klindgren | with running the logstash daemon per node, with centraly managed logstash config | 22:27 |
klindgren | to reduce the split/parsing from a centralized location. | 22:28 |
sdake | klindgren we are not using logstash-forwarder, we are using heka instead | 22:28 |
jascott1 | sdake mind if I add some description? iirc I checked and meta was not in that arch doc | 22:29 |
sdake | meta is part of layer 4 iirc | 22:30 |
sdake | meta/service/microservice are al part of layer 4 | 22:30 |
sdake | jascott1 do whatevery ou like with the blueprints :) | 22:31 |
*** adrian_otto has quit IRC | 22:31 | |
sdake | jascott1 i surely am not the only person that should be editing them or amintaining them :) | 22:31 |
sdake | thats a distributed community sort of thing | 22:31 |
jascott1 | aw yeah | 22:31 |
sdake | klindgren that logstash daemon per node is "logstash-forwarder" | 22:31 |
sdake | klindgren i would assume | 22:32 |
klindgren | its "logstash" | 22:32 |
sdake | klindgren if your running logstash per node, kolla's arch was definately not designed to handle that | 22:32 |
*** unicell has quit IRC | 22:32 | |
sdake | klindgren just woke up - so have to think on the fly here | 22:32 |
sdake | klindgren give me a bit to think about it | 22:32 |
jascott1 | cool yeah figured if I was confused others might be, false logic but still works out to be true often | 22:32 |
sdake | jascott1 hard eearned :) | 22:33 |
klindgren | np - again trying to figure out how to brownfield kolla/kolla-ansible here. so I totally expect some of these things | 22:33 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Entrypoint dependency init container https://review.openstack.org/412005 | 22:33 |
klindgren | because I really don't want to duplicate efforts for using ansible to deploy kolla containers to servers | 22:33 |
sdake | klindgren haer that | 22:35 |
sdake | klindgren not sure - might ping other folks not on a friday at the end of year :) | 22:36 |
*** cliles has quit IRC | 22:36 | |
sdake | klindgren i'll give athink to your problem - added to my gtd archive | 22:36 |
sdake | klindgren not sure i have any wise answers | 22:37 |
sdake | klindgren but maybe something will come to me after i return from pto (20th-4th) | 22:37 |
*** unicell has joined #openstack-kolla | 22:37 | |
*** jgriffith is now known as jgriffith_AutoAw | 22:38 | |
sdake | klindgren some of our community doesn't have a holiday shutdown enforced upon them and will be around during that period | 22:38 |
sdake | klindgren i'd keep trying on that Q | 22:38 |
sdake | klindgren another option is the mailing list | 22:38 |
sdake | klindgren might provide a link to this irc log in the ml post so people know what was being talked about | 22:39 |
sdake | SamYaple cpuguy says --squash was just merged | 22:42 |
openstackgerrit | Merged openstack/kolla: Install solum dashboard into horizon image https://review.openstack.org/408981 | 22:42 |
sdake | SamYaple without the need for external tooling | 22:42 |
sdake | SamYaple although annoyingly probably in docker 1.13 or something later | 22:43 |
klindgren | sdake, kk - will try just adding a conditional against centralized logging in the common includes for now | 22:43 |
sdake | klindgren the problem you have there is you need something to do the log forwarding on the nodes or captur ethe logs | 22:43 |
klindgren | I already have it | 22:44 |
SamYaple | sdake: you are missing the point. --squash wont get what we want. it will help a little (and its already taged in 1.12 i believe) | 22:44 |
SamYaple | external high level tooling is needed | 22:44 |
sdake | SamYaple just woke up dude - mind cutting me a break and explaining it to me :) | 22:44 |
sdake | cool | 22:44 |
sdake | thanks :) | 22:44 |
SamYaple | i mean, its been pointed out several times by several docker employees (including the linked comment) that the Dockerfile stuff won't cut it for advanced projects | 22:45 |
sdake | ya - I have read for putting images in the official docker registry - external tooling is out | 22:45 |
SamYaple | nope | 22:45 |
SamYaple | thats for _building_ in dockerhub | 22:45 |
SamYaple | you can push anythign you want | 22:45 |
SamYaple | we can't build in dockerhub anyway thanks to how we fetch tarballs and stuff | 22:45 |
sdake | yup aware of that problem | 22:46 |
SamYaple | technically build.py is external tooling | 22:46 |
sbezverk | portdirect: ping | 22:46 |
sdake | SamYaple totally agree | 22:46 |
portdirect | hey i was just about to jump into the whole build discussion :( whats up sbezverk? :) | 22:46 |
sdake | SamYaple thats what i'm getting at, our dockerfiles as is wont pass "inspection" | 22:46 |
sdake | for the official images | 22:47 |
SamYaple | anyway. the whole point is we can get our images tiny. but we cant use the, by dockers own admission, simplistic dockerfiles to do so. at least not entirely | 22:47 |
sdake | not sure we want to head down that route or not | 22:47 |
SamYaple | no reason to get rid of them, but we also dont need to be limited by them | 22:47 |
SamYaple | well, well never get them building in dockerhub. ive been down that path. its not practicle and will siginifcantly bloat sizes | 22:48 |
sbezverk | portdirect: even though I understand the reason for your suggestion I am concerned that it will create some problem | 22:48 |
sdake | seems like people not interested in even maintaining dockerhub at all | 22:48 |
sdake | ya - i agree re dockerhub builds | 22:48 |
sbezverk | portdirect: how developer would know that three lines do not come with macro? | 22:49 |
sdake | about the practical part - bloat sizes don't know but trust you say it is so it is | 22:49 |
SamYaple | we cant push our current images due to the size. we need to trim them down. to do that we must extend beyond Dockerfile since that is our limitation. this is how i can get a 40MB keystone image | 22:49 |
portdirect | ok - i'll through my comments into the ring. dockers --sqush (unless its changed) will make the situation potentially worse of us as it just take an image and flattens it to a single layer (essentially docker export - docker import), though I'm concerd that if we dump the dockerfile then we will make it much harder for ourselves and loose potential platform infexibility. | 22:50 |
SamYaple | portdirect: no this is a differntthing | 22:50 |
SamYaple | portdirect: this is official docker tooling | 22:50 |
SamYaple | portdirect: it squashes images in layers now | 22:50 |
SamYaple | errr in a dockerfile | 22:51 |
*** fragatin_ has joined #openstack-kolla | 22:51 | |
SamYaple | basically it builds a layer from the top of the dockerfile to the bottom | 22:51 |
portdirect | SamYaple: thats what i was refering to I've been following, and your updating my knowledge of that ps as i type so I'll jump back out :) | 22:51 |
*** dave-mccowan has joined #openstack-kolla | 22:51 | |
SamYaple | so rabbitmq would have 3 layers (more or less). ubuntu->base->rabbitmq | 22:51 |
sdake | portdirect more brains welcome :) | 22:51 |
SamYaple | not jsut a single layer | 22:51 |
portdirect | SamYaple: so it does what i had in that ps that i showed you nad i think you did the same? | 22:52 |
SamYaple | portdirect: nah that was different if I remember correctly. that that was not official squash | 22:52 |
portdirect | sbezverk: i would expect the end of that macro to go aws well - so a deveoper has to set up to annotations etc, but can just drop in that snippet of json without the yaml surrounding it | 22:53 |
SamYaple | portdirect: that was unoffical peoples attempts at it | 22:53 |
*** fragatin_ has quit IRC | 22:53 | |
SamYaple | portdirect: this would be part of the dockerclient build. like dc.build(squash=True) | 22:53 |
SamYaple | portdirect: but thats still not going to get us what we want unless we resign to install and uninstall build tools every layer (and that would _KILL_ IO and bandwidth) | 22:54 |
portdirect | SamYaple: sorry not explaining myelf well, I know that ws unofficia, buth docker now does the same thing? when that pull request first started over at docker it was just to a single layer per image, not per dockerfile | 22:54 |
portdirect | SamYaple: yeah - thats actually why i changed to alpine - as it was almost fast enough to make that practtical | 22:55 |
sbezverk | portdirect: how about three macros ? head,body and tail? | 22:55 |
SamYaple | portdirect: https://github.com/docker/docker/issues/332#issuecomment-257740418 | 22:55 |
portdirect | sbezverk: ok :) | 22:55 |
portdirect | though thats macro tastic... | 22:55 |
portdirect | sbezverk: my concern was cutting options, so 3 macros wfm. | 22:56 |
portdirect | SamYaple:nice | 22:56 |
*** dave-mccowan has quit IRC | 22:56 | |
*** fragatina has joined #openstack-kolla | 22:56 | |
SamYaple | portdirect: this will still leave us with about the same amount of bloat | 22:56 |
*** gomarivera has joined #openstack-kolla | 22:56 | |
SamYaple | portdirect: but now we can get rid of the && crap | 22:57 |
sdake | portdirect the issue is in a typical CI pipeline all that stuff has to be built - which takes 4-8 hours because there is no docker cache | 22:57 |
sdake | and its built over and over and over | 22:57 |
sdake | klindgren is that accurate? | 22:57 |
portdirect | sdake: i think it is | 22:57 |
sdake | so even --squash doesn't do the job | 22:57 |
SamYaple | right. and it wont until we dont have common layers | 22:58 |
sdake | SamYaple how do we get there? | 22:58 |
SamYaple | i started to detail a plan for that. | 22:58 |
portdirect | sdake: itw would prbably make it take even longer :( | 22:58 |
sdake | cool | 22:58 |
SamYaple | part of the issue is build _times_ | 22:58 |
sdake | i think alot fo people would be keen to see it | 22:58 |
SamYaple | layers solves build times | 22:58 |
SamYaple | so without layers our build times sucks | 22:59 |
portdirect | SamYaple: yeah a hich level overview would be great | 22:59 |
sdake | layers also solve pull times | 22:59 |
SamYaple | so what klindgren is complaining about wont go away | 22:59 |
SamYaple | not when keystone is 40MB it doesnt :P | 22:59 |
sdake | SamYaple right - which is why you suggested --no-deps and i suggested pulling the deps from the docker registry | 22:59 |
klindgren | sorry what am I complaining about? | 23:00 |
sdake | klindgren your not complaining :) | 23:00 |
klindgren | I dont remeber a complaint around docker layers and the numbers of them? | 23:00 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Entrypoint dependency init container https://review.openstack.org/412005 | 23:00 |
sdake | klindgren but you did mention the ci build times are a problem iirc - or perhaps that was harlowja | 23:00 |
SamYaple | klindgren: you got wrong-named | 23:00 |
SamYaple | v1k0d3n: is the one | 23:00 |
*** fragatina has quit IRC | 23:00 | |
portdirect | sbezverk: :) | 23:01 |
*** gomarivera has quit IRC | 23:01 | |
SamYaple | at a certain point people must realize taht to build keystone, you must install build-tools. then packages, then pip stuff, then keystone | 23:01 |
klindgren | Maybe that was harlow. I think I remember we were trying to investigate if we could be each project individually and re-use common layers. But it seems like that may not be possible. | 23:01 |
*** cliles has joined #openstack-kolla | 23:01 | |
SamYaple | that takes time | 23:01 |
SamYaple | klindgren: no it was v1k0d3n | 23:01 |
portdirect | sbezverk: though essential for sanity, lines 37-39 of your new ps made me laugh | 23:02 |
SamYaple | ok im out | 23:02 |
kfox1111 | portdirect: https://review.openstack.org/#/c/412005 look better to you? | 23:03 |
sbezverk | portdirect: consistency my friend ;-) | 23:03 |
sbezverk | from a to z | 23:03 |
portdirect | kfox1111: yeah thats the one i was saying made me laugh :) | 23:03 |
sbezverk | portdirect: think lego | 23:04 |
kfox1111 | portdirect: hehe. yeah. | 23:04 |
portdirect | sbezverk: good point, it always comes in a box bigger than the model :D | 23:04 |
*** unicell has quit IRC | 23:05 | |
portdirect | sbezverk: oh actually i do have a problem with it | 23:05 |
portdirect | sbezverk: not sure how to solve though :/ | 23:05 |
*** aric49 has quit IRC | 23:06 | |
portdirect | sbezverk: should there not be a comma at line 34, but only if there is another container? | 23:06 |
kfox1111 | portdirect: oh. yeah. | 23:06 |
kfox1111 | comma should be provided by the parent. | 23:06 |
portdirect | sbezverk: oh actually ignore me, thi init container could just go last? | 23:06 |
kfox1111 | so after the {{ include "..." . }}, | 23:07 |
sbezverk | yes | 23:07 |
sbezverk | that is the idea | 23:07 |
portdirect | oh panic averted :) | 23:07 |
sdake | klindgren it was v1k0d3n and harlowja :) | 23:07 |
sdake | klindgren point being we got alot of people that want that problem solved for the ci pipeline case :) | 23:08 |
portdirect | sdake: is harlowja, using dind as well? | 23:08 |
portdirect | *docker in docker | 23:08 |
sdake | portdirect not that I know of - you would have to ask him | 23:08 |
sdake | he isn't about atm - friday + holidays = people have christmas to deal with :) | 23:09 |
portdirect | k, cause thats a reall killer with the default options (vfs storage) | 23:09 |
kfox1111 | portdirect: look good to you? | 23:11 |
portdirect | what? sbezverk's patch | 23:11 |
kfox1111 | portdirect: yeah. | 23:12 |
portdirect | (just got a +1 from me) | 23:12 |
kfox1111 | k. | 23:12 |
*** TxGirlGeek has quit IRC | 23:12 | |
portdirect | :) | 23:12 |
*** saneax-_-|AFK is now known as saneax | 23:12 | |
*** adrian_otto has joined #openstack-kolla | 23:12 | |
v1k0d3n | hey | 23:13 |
v1k0d3n | whats all this now? | 23:13 |
v1k0d3n | reading a bit... | 23:13 |
portdirect | though thinking - can we get the best of both worlds? is it possible to combine the three into one in that file so if only an init was needed then you could have sbezverk's orginal version effectily? | 23:14 |
*** vhosakot has quit IRC | 23:16 | |
sbezverk | portdirect: we might need this container with other microservices where init container already exists | 23:16 |
klindgren | we aren't doing docker-in-docker | 23:16 |
SamYaple | klindgren: just run docker in a pod so you can run openshift to run tripleo to run openshift | 23:17 |
portdirect | sbezverk: yeah thats why i succested the change :) but was wondering if we should (or possible) add a macro that simple wraps the three it's now proken up into for simple use cases.. | 23:17 |
portdirect | SamYaple: yo dawg! | 23:18 |
klindgren | but only if we can run that in vms on openstack | 23:19 |
SamYaple | klindgren: easy. just use magnum | 23:20 |
openstackgerrit | Merged openstack/kolla-kubernetes: Entrypoint dependency init container https://review.openstack.org/412005 | 23:20 |
*** cliles has quit IRC | 23:21 | |
sdake | SamYaple with openstack with kubernetes on top :) | 23:21 |
sdake | DAWG | 23:22 |
sdake | we have t-1 month until ocata is cut IIRC, so a total refactor of the dockerfiles I think is not really ideal in this cycle :) | 23:23 |
SamYaple | no but static uid/gid need to land | 23:23 |
sdake | SamYaple right - i think Jeffrey4l is on to that part of things - but I am not up to speed on where its at, and tbh I don't have time to be up to speed on it | 23:24 |
sdake | SamYaple i have pto 20th-4th so wont be up to speed anytime soon on uid/gid mapping | 23:25 |
*** sayantan_ has quit IRC | 23:26 | |
*** cliles has joined #openstack-kolla | 23:27 | |
*** jgriffith_AutoAw is now known as jgriffith | 23:31 | |
*** cliles has quit IRC | 23:35 | |
*** cliles has joined #openstack-kolla | 23:37 | |
*** lamt has quit IRC | 23:37 | |
*** Pavo has joined #openstack-kolla | 23:42 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Some more dep container features https://review.openstack.org/412010 | 23:43 |
*** awiddersheim has quit IRC | 23:54 | |
*** gomarivera has joined #openstack-kolla | 23:57 | |
*** unicell has joined #openstack-kolla | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!