| *** dims has joined #openstack-kolla | 00:00 | |
| *** jistr has quit IRC | 00:01 | |
| *** jistr has joined #openstack-kolla | 00:03 | |
| *** sayantani01 has quit IRC | 00:09 | |
| *** salv-orl_ has quit IRC | 00:15 | |
| kfox1111 | sdake: it looks like k8s's support for backing pv's with ceph rbd relies on kernel support which is flaky. :/ | 00:15 |
|---|---|---|
| sdake | i see - so it works most of the time but not all of the time? | 00:15 |
| kfox1111 | there were two cases SamYaple said coudl cause problems. | 00:16 |
| kfox1111 | if the volumes try to mount before ceph starts. | 00:16 |
| kfox1111 | or if ceph enters a bad state for a while. | 00:17 |
| kfox1111 | the former might be able to be worked around. | 00:17 |
| kfox1111 | the latter, is a big problem. | 00:17 |
| kfox1111 | I wouldn't want to my production db's dependent on it. | 00:17 |
| kfox1111 | ceph support in glance/nova/cinder is uneffected. | 00:17 |
| kfox1111 | it only really effects rabbit/mariadb. | 00:17 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Add a workaround for docker-engine 1.13.0 https://review.openstack.org/423568 | 00:24 |
| sdake | is nfs an option or iscsi? | 00:25 |
| sdake | what do we have in use now, ceph? | 00:26 |
| portdirect | nfs would not be great | 00:26 |
| portdirect | iscsi would be a really good option for this | 00:27 |
| kfox1111 | nfs is usually a scalability/ha problem. | 00:27 |
| kfox1111 | same with iscsi. | 00:27 |
| portdirect | smae with local... | 00:27 |
| kfox1111 | local is, unless you use clustering. | 00:27 |
| portdirect | yeah but we would be wouldnt we? | 00:27 |
| sdake | ok - so i think nfs is pretty terrible myself - wondering what it would take to enable all of the options | 00:28 |
| sdake | so its just configurable and the operator could choose what they want | 00:28 |
| kfox1111 | portdirect: I think we probably should. | 00:28 |
| kfox1111 | cephfs is another option. | 00:28 |
| kfox1111 | but performance will be a bit worse then nfs. but doesn't have the ha problem. | 00:28 |
| sdake | nfs can be setup in HA mode, but ya, nfs is a pile of steam | 00:28 |
| portdirect | im not sure cephfs is a good for mariadb | 00:29 |
| sdake | so in kolla-ansible - which is my only experience with ceph | 00:29 |
| sdake | ceph was only used for glance/nova/cinder via the RBD driver | 00:29 |
| sdake | in kolla-kubernetes, ther eare two ceph's running 9from my understanding, which is limited) | 00:29 |
| kfox1111 | yeah. so it would be closer to kolla-ansible then. local storage with clustering. | 00:29 |
| sdake | there is one that provides storage for the kubernetes PVS | 00:29 |
| sdake | and one that provides storage internally | 00:29 |
| kfox1111 | two different ways its used. one ceph. | 00:30 |
| portdirect | sdake: same ceph for both atm | 00:30 |
| sdake | can someone validate the above from me, or am I misunderstanding how persistent storage is handle? | 00:30 |
| kfox1111 | right. one for pvs for k8s, | 00:30 |
| kfox1111 | and the rest for nova/glance/cinder. | 00:30 |
| sdake | portdirect and kfox1111 you guys just siad opposite things :) | 00:30 |
| kfox1111 | the problem is only with the pvc's. | 00:30 |
| portdirect | kfox1111: could you clarify | 00:31 |
| kfox1111 | k8s is using kernel based rbd for k8s pv's. | 00:31 |
| kfox1111 | nova/glance/cinder is using librbd. | 00:31 |
| kfox1111 | the instability is with the kernel based rbd support. | 00:31 |
| portdirect | kfox1111: but to help sdake, could you calarifyt that they are both coming from the same ceph cluster atm | 00:32 |
| portdirect | and you are talking now about how that cluster is consumed in two different ways | 00:32 |
| sdake | sounds like same ceph cluster, two different access mechanisms | 00:32 |
| portdirect | yes | 00:33 |
| kfox1111 | portdirect: right. | 00:33 |
| kfox1111 | once ceph was handling both use cases. | 00:33 |
| sdake | cool thanks for clearing that up | 00:33 |
| sdake | and that is what we have at present - and it works most of the time :) | 00:33 |
| sbezverk | portdirect: I tried iscsi and hit issue | 00:33 |
| sdake | and maybe we need to transition to something else, so it works all of the time | 00:33 |
| portdirect | 60% of the time it works every time | 00:33 |
| sbezverk | I opened an issue with kube people, so far no answer | 00:33 |
| portdirect | hmm :( | 00:34 |
| portdirect | ive been using iscsi with a driver w wrote very well | 00:34 |
| portdirect | but not k8s own driver | 00:34 |
| *** zhurong has joined #openstack-kolla | 00:34 | |
| kfox1111 | portdirect: flexdriver? | 00:35 |
| sdake | can we enable a general approach to storage backing where anything canbe used? | 00:35 |
| portdirect | HDS now reommend it with some of their arrays :D | 00:35 |
| sbezverk | https://github.com/kubernetes/kubernetes/issues/39159 | 00:35 |
| portdirect | kfox1111: yes | 00:35 |
| *** goldyfruit has quit IRC | 00:35 | |
| sdake | or is that not a thing | 00:35 |
| sdake | reading the FM it says there are multiple storage backends | 00:35 |
| kfox1111 | sdake: its pretty muc htat way now. its just the underlying options are less then ideal at the moment. :/ | 00:35 |
| portdirect | sdake: thats what 1.5 PVC lets us do | 00:35 |
| sbezverk | portdirect: I like iscsi very much, it just does not work :( | 00:35 |
| kfox1111 | isci's hard to make ha though. | 00:36 |
| kfox1111 | so defeats the purpose for this use case I think. | 00:36 |
| sdake | ok - so ignoring the actual making it ha | 00:36 |
| portdirect | sbezverk: https://youtu.be/h1et3FR6zis?t=28m40s | 00:36 |
| sdake | and assuming that is someone elses problem (TM) | 00:36 |
| kfox1111 | unless you actually have ha iscsi externally from another vendor. | 00:36 |
| sdake | it sound slike with k8s 1.5 we can have a general approach to storage | 00:37 |
| kfox1111 | yeah. but you need a capable storage backend. | 00:37 |
| kfox1111 | ceph rbd seems to be a bit lacking at the moment the way k8s supports it. :/ | 00:37 |
| sdake | right so we could say something like "kolla-kubernetes coudl work with XYZ" | 00:37 |
| sdake | however in the gate we test with "Y" | 00:38 |
| sbezverk | kfox1111: not every solution must be software solution, you can build h/w redundant storage cluster and offer iscsi connectivity | 00:38 |
| kfox1111 | well, this is a batteries included issue I think. | 00:38 |
| sdake | kfox1111 right - we test with "Y" | 00:38 |
| sbezverk | kfox1111: it is just people are looking for "sheap" things | 00:38 |
| kfox1111 | sbezverk: right. thats what I meant. netap for example sells an appliance that would work. | 00:38 |
| sbezverk | "cheap" | 00:38 |
| sdake | sbezverk the word you are looking for is cheap :) | 00:38 |
| sdake | sbezverk sorry i thought you were making a play on words ;) | 00:38 |
| kfox1111 | netapp is the opposite of cheep though. | 00:38 |
| kfox1111 | next best solution I can think of is a cluster with local storage. | 00:40 |
| sbezverk | kfox1111: I think, and maybe I am totally wrong, when you build mission critical env, cheep, should not be in first 5 reason for technology selections ;) | 00:40 |
| portdirect | sbezverk: the acutal driver is here https://github.com/portdirect/harbor/blob/f82d7ded096e27cda8df8a5e03860a789066e41e/docker/kubernetes/kubernetes-kubelet/assets/usr/libexec/kubernetes/kubelet-plugins/volume/exec/cinder/cinder | 00:40 |
| kfox1111 | sbezverk: there's levels of sla and cheep. :) | 00:40 |
| sbezverk | portdirect: I do not think cinder driver and iscsi driver are the same thing?!? | 00:41 |
| portdirect | kfox1111: betteies included are nice, but this is satrting to get a bit out of scope is is not | 00:41 |
| portdirect | sbezverk: that does both :) | 00:41 |
| portdirect | uses cinder to create and manage iscsi | 00:42 |
| sbezverk | portdirect: hm interesting.. | 00:42 |
| kfox1111 | portdirect: which being out of scope? | 00:42 |
| portdirect | i think that we should try and use the k8s sortage options, so that people deplying it can bring whatever the want to the party | 00:43 |
| sdake | portdirect ++ | 00:43 |
| kfox1111 | how would you back mariadb with cinder, which needs a mariadb? | 00:43 |
| kfox1111 | portdirect: I don't disagree with that. | 00:43 |
| kfox1111 | put local storage is an option too. | 00:43 |
| portdirect | and speding too much time on building out ha storage moves us towards being a distro | 00:43 |
| portdirect | rather than a set of building blocks | 00:43 |
| kfox1111 | nothing stopping you from deploying a mariadb on pv's if you have a suitable one. | 00:43 |
| *** caowei has joined #openstack-kolla | 00:43 | |
| kfox1111 | not arguing we should remove support for them. | 00:43 |
| sdake | portdirect - so what i'm after is how to get kolla-kubernetes stood up in general :) | 00:44 |
| sdake | portdirect so i can document it | 00:44 |
| portdirect | what I'm saying though it im not sure it should be a blocker | 00:44 |
| sdake | portdirect so others can deveop for it | 00:44 |
| portdirect | sdake: exactly | 00:44 |
| kfox1111 | I am pretty sure I'll need a more ha db then just a single node can provide. | 00:44 |
| sdake | portdirect it doesn't have to be pretty or fantastic, it just has to sort of work :) | 00:44 |
| kfox1111 | so I'll probably contribute the local storage cluster code. | 00:44 |
| portdirect | kfox1111: we all will, but kolla-k8s doesnt need | 00:44 |
| portdirect | sdake: we are in total agreement | 00:45 |
| kfox1111 | if all of us do, why wouldn't others? | 00:45 |
| kfox1111 | sounds like it belongs under the kolla-kubernetes umbrella then. | 00:45 |
| portdirect | beacsue we need it - but ha kubernetes storage is not reeally kolla-k8s's problem in my opinion | 00:45 |
| kfox1111 | kolla-ansible is already doing it. | 00:45 |
| sdake | already doing which? :) | 00:45 |
| kfox1111 | I'm not arguing we implement a sotrage system from scratch. | 00:45 |
| kfox1111 | just supporting galera with local storage. | 00:46 |
| kfox1111 | kolla-ansible already does. | 00:46 |
| sdake | kolla-ansible uses bindmounts for galera | 00:46 |
| kfox1111 | maybe we are talking about differen things? | 00:46 |
| kfox1111 | sdake: exactly. | 00:46 |
| sdake | is that what ou mean by local storage? | 00:46 |
| kfox1111 | which are prety much the same as k8s host path's. | 00:46 |
| kfox1111 | yeah. | 00:46 |
| portdirect | so lets do that - but via pvc annotations | 00:47 |
| portdirect | and not include anything else in the scope of kolla-k8s | 00:47 |
| portdirect | ie do all sotage section via the annotation | 00:47 |
| kfox1111 | was just going to do it as an if conditional? | 00:47 |
| portdirect | this lets operators use whatever they want | 00:47 |
| sdake | portdirect would that allow ops to choose whatever storage backend they want? | 00:48 |
| portdirect | yes | 00:48 |
| sdake | wfm | 00:48 |
| kfox1111 | {{if local_storage }} host_path {{else}} pv: name: ... {{end}} | 00:48 |
| kfox1111 | ? | 00:48 |
| portdirect | no | 00:48 |
| portdirect | thats the old wat | 00:48 |
| portdirect | way | 00:48 |
| kfox1111 | there isn't a new way that works with hostpaths I think? | 00:49 |
| sbezverk | portdirect: if you have config example of working iscsi storage option for kube, pleae let me know.. | 00:49 |
| portdirect | what we should o is delacre a storage class | 00:49 |
| sdake | wow my mothers computer has a bunch of crap on it | 00:49 |
| kfox1111 | how do yho udeclare a storage class for local storage? | 00:49 |
| portdirect | and then use the annotation to select it | 00:49 |
| portdirect | 2 sec (im sure you can) | 00:49 |
| portdirect | minikube does i think | 00:49 |
| kfox1111 | oh. that would be cool if that works. | 00:50 |
| *** caowei has quit IRC | 00:50 | |
| *** caowei has joined #openstack-kolla | 00:52 | |
| openstackgerrit | Li Yingjun proposed openstack/kolla-kubernetes: Fix different repo issue for documentation https://review.openstack.org/405054 | 00:52 |
| sdake | portdirect were you able to get that pin sorted out | 00:53 |
| portdirect | yeah - going to move to upstream docker | 00:53 |
| portdirect | will be in on monday | 00:53 |
| sdake | portdirect i looked at the running vagrant, its already running docker-engine dude | 00:54 |
| sdake | portdirect that is what the problem is - docker-engine 1.13 is busted | 00:54 |
| sdake | docker-engine = upstream docker | 00:55 |
| sdake | docker = red hat docker | 00:55 |
| sdake | red hat is not on docker-engine 1.13 :) | 00:55 |
| sdake | (centos that is) | 00:55 |
| portdirect | damn, thought i'd pulled that out https://github.com/att-comdev/halcyon-kubernetes/blob/master/kube-deploy/roles/deploy-kube/tasks/centos.yml#L39 | 00:55 |
| portdirect | will have it fisec on monday regardless | 00:55 |
| sdake | portdirect cool - so i think the key is to pin to an older version of docker | 00:56 |
| sdake | like 0.12.6 | 00:56 |
| sdake | given the complexity of windows, I have no idea how people maintain their own windows computers | 00:57 |
| sdake | I can barely maintain my parents, and my version of maintenace involves a full restore from a pristine installed copy from a NAS :) | 00:57 |
| *** hfu has joined #openstack-kolla | 01:05 | |
| kfox1111 | portdirect: been able to find an example of a local storage class? | 01:08 |
| kfox1111 | I'm looking through the code base but not finding much. | 01:08 |
| portdirect | Same here, it seems there is some provisioning stuff in there, but no examples | 01:09 |
| kfox1111 | I fired up a new minikube, but kubectl get storageclass has none defined. | 01:10 |
| kfox1111 | hhhmm... though. looks like my minikube binary is a bit old... let me download a newer one. | 01:11 |
| portdirect | Same here | 01:11 |
| portdirect | Its the iso that matters not minikube itself really | 01:11 |
| sdake | iso as in cdrom? | 01:12 |
| kfox1111 | yeah. but easier to just download the new minikube and minikube delete; minikube start | 01:12 |
| portdirect | dims would know | 01:12 |
| portdirect | sdake minikube launches a custom iso with a modified version of kyperkube (localkube) for testing, used to be based on boot2docker, but now is a buildroot based distro that google/coreos made | 01:16 |
| sdake | i see | 01:16 |
| sdake | is it CI/CDed as kubernetes is released? | 01:17 |
| sdake | perhaps it has an old version of k8s | 01:17 |
| portdirect | nah - it runs any version you want | 01:17 |
| portdirect | for example | 01:17 |
| portdirect | https://www.irccloud.com/pastebin/HB2vBrOD/ | 01:18 |
| sdake | portdirect cool so minikbube is an integration wrapper much like the current dev env we have today, however, the devenv today is based upon kubeadm | 01:19 |
| portdirect | yup | 01:19 |
| portdirect | we needed kubadm based for multinode | 01:19 |
| kfox1111 | sdake: it was made to be the most simple way to deploy a test k8s. | 01:19 |
| kfox1111 | downlad the binary and run it. | 01:19 |
| kfox1111 | it lacks multinode support though, as most folks don't realy need it. | 01:19 |
| portdirect | but if we go local stoage and drop the ceph dep, then we could use it | 01:19 |
| kfox1111 | it works with ceph. | 01:20 |
| sdake | portdirect i think using kubeadm is fine | 01:20 |
| sdake | since kubeadm is multinode | 01:20 |
| sdake | although I don't really understand what kubeadm is either :) | 01:20 |
| sdake | we do want our dev env to virtualize multliple nodes | 01:20 |
| sdake | the gate may not needi t but devs do | 01:21 |
| * portdirect wonders if maybe kolla should use ansible | 01:21 | |
| sdake | portdirect which deliverable of kolla :) | 01:21 |
| portdirect | lol | 01:21 |
| portdirect | all of them | 01:22 |
| portdirect | :) | 01:22 |
| kfox1111 | the gate is multinode. ;) | 01:22 |
| sdake | so in the earlly days of k8s - 0.9.7 or so - we tried making kolla use k8s and failed | 01:22 |
| sdake | then we tried shell scripts and that worked sort of | 01:22 |
| sdake | then we tried compose and that worked better | 01:22 |
| sdake | then we went to ansible and that was a winner | 01:22 |
| *** masuberu has quit IRC | 01:22 | |
| sdake | k8s lacked host=pid and net=pid support - which were major blockers | 01:23 |
| portdirect | lol - you should have stuck with it :) (no i know im just being rude) | 01:23 |
| sdake | it was probably missing a whoel slew of other crap too that would have dragge don and on | 01:23 |
| portdirect | i fork your work back then | 01:23 |
| kfox1111 | ok. redeployed minikube. its 1.5.1 now. | 01:23 |
| kfox1111 | no storage clsases defined though. | 01:23 |
| portdirect | and got round that by just bind mounting to docker socker | 01:23 |
| sdake | so k8s lacks a storage class at all? | 01:24 |
| sdake | or we have to load the min | 01:24 |
| sdake | perhaps its some kind of plugin | 01:24 |
| kfox1111 | sdake: no. just looking to see if minikube provided any storageclasses out of the box for local storage. | 01:24 |
| sdake | just throwing darts ata wall here :) | 01:24 |
| kfox1111 | it does not seem to. | 01:24 |
| *** masuberu has joined #openstack-kolla | 01:24 | |
| kfox1111 | if there is a way to do pvc provisioning from local storage via storage classes, I can't figure out how to. | 01:25 |
| sdake | ok cna someone define the difference between kubeadm and kube-spray while we are on this topic? | 01:25 |
| portdirect | sdake | 01:25 |
| portdirect | not easily | 01:25 |
| portdirect | thats a bit like the difference between packstack and fuel | 01:25 |
| sdake | kubeadm is packstack in the anaology? | 01:26 |
| portdirect | no they both work :D | 01:26 |
| sdake | i actually quite liked packstack, I think if rht would have stuck with packstack, they woudl have won :) | 01:26 |
| portdirect | just different ways of deploying a cluster | 01:26 |
| * portdirect was reffering to fuel | 01:26 | |
| kfox1111 | sdake: kubeadm is under the k8s umbrella. | 01:26 |
| kfox1111 | like helm. | 01:27 |
| kfox1111 | its newer, so less advanced. | 01:27 |
| sdake | kfox1111 i know - so is kubespray | 01:27 |
| sdake | kubespray in the incubator | 01:27 |
| kfox1111 | kubespray is? | 01:27 |
| sdake | this is literall ythe first free 8 hour block of time i've had since late nov to take a look at anything | 01:27 |
| sdake | and I just want to get going :) | 01:28 |
| kfox1111 | oh. seems like it is. thats news to me. :) | 01:28 |
| sdake | the most meaningful contribution i can make to kolla-kubernetes at htis point is documenting how all this stuff works and making it easy for devs to onramp | 01:28 |
| sdake | if devs canm't onramp, ops wont be able to onramp either | 01:29 |
| kfox1111 | +1 | 01:29 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Split Openvswitch into own role https://review.openstack.org/416793 | 01:29 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Add OpenDaylight role https://review.openstack.org/416367 | 01:29 |
| sdake | not that devs are smarter then ops - just they are typically more motivated and stick with it longer | 01:29 |
| sdake | we need to have a 30 minute "here is how you go from aero to winning" | 01:29 |
| kfox1111 | we paid a high cost for helm. :/ | 01:29 |
| kfox1111 | sdake: and to dig through code to figure out how. | 01:29 |
| sdake | kfox1111 right - although that was the right move in retrospec | 01:29 |
| sdake | kfox1111 i have dug throug hteh gate code | 01:30 |
| portdirect | kfox1111: yeah but it helps to be +ve | 01:30 |
| sdake | where I am jammed up now is setting up ceph | 01:30 |
| portdirect | other projects (ones that im onvoved in) can go from nothing to fully deployed on bare metal in under 8 mins now | 01:30 |
| sdake | hence the questions aroudn storage | 01:30 |
| portdirect | 13 lines of bash | 01:30 |
| sdake | portdirect is this aic-helm or something else? | 01:30 |
| portdirect | boht opensatck-helm, and harbor | 01:31 |
| kfox1111 | I used this to deply my ceph: http://docs.ceph.com/docs/jewel/rados/deployment/ | 01:31 |
| portdirect | harbor take 45 mins | 01:31 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Add OpenDaylight role https://review.openstack.org/416367 | 01:31 |
| kfox1111 | portdirect: kolla-kubernetes will get there. like I said, we paid a HIGH cost for helm. | 01:31 |
| kfox1111 | :/ | 01:32 |
| sdake | why ahe one snowflake in your datacenter when you can have a blizzard :) | 01:32 |
| kfox1111 | I had totally working docs at one point. then it all fell apart due to 3 or 4 code refactors. :/ | 01:32 |
| sdake | kfox1111 refactors are a natural part of development | 01:32 |
| sdake | there are 3 things that make a project great | 01:32 |
| sdake | four things actualy | 01:32 |
| sdake | community, code, docs, and gating | 01:32 |
| kfox1111 | yeah. It was an investment in community to do helm. | 01:33 |
| sdake | any one of those doesn't materialize, the chair falls over and the operator or dev sitting on the chair cracks his skull open | 01:33 |
| kfox1111 | but at the same time, it prevents more devs from starting up, as the docs bitrotted. | 01:33 |
| portdirect | none of the openstack on k8s projects has all four | 01:33 |
| sdake | fortunately docs are fixable | 01:33 |
| sdake | community is really hard to fix | 01:33 |
| kfox1111 | so short term, it hurt the community. | 01:33 |
| kfox1111 | yeah. | 01:33 |
| kfox1111 | ture. | 01:33 |
| kfox1111 | true | 01:33 |
| sdake | so lets fix the docs | 01:34 |
| sdake | i'm doing that | 01:34 |
| sdake | where i need help is storage :) | 01:34 |
| sdake | i think i can put together a 10-20 line shell script that does the job | 01:34 |
| kfox1111 | yup. | 01:34 |
| sdake | it may be longer - who knows | 01:34 |
| sdake | i have managed to stand up kolla-kubernetes on friday in 3 hours wihtout storage and it seemed to all work | 01:35 |
| sdake | minus the containers failing to start and whatnot :) | 01:35 |
| portdirect | lol - but getting the templates to render is agood 1st step | 01:36 |
| sdake | the dev env doc is perfect now, just need the quickstart to be so | 01:36 |
| sdake | portdirect i have been taking alot of notes and have history recorded and formulating a shell script to do the job | 01:36 |
| portdirect | sweet :) | 01:36 |
| sdake | i think our top mission to solve - right now - is how to get me unblocked so I can get eveyrone else unblocked :) | 01:37 |
| portdirect | If you push a DNM, ps it would be great to go over it | 01:37 |
| kfox1111 | sdake: so, where are you stuck? | 01:37 |
| sdake | i relaize not eveyrone is blocked | 01:37 |
| sdake | but eveyrone using the gate to test is super suboptiaml | 01:37 |
| sdake | kfox1111 stuck on how to setup storage | 01:37 |
| portdirect | i'sd like to be able to run koll-k8s as well | 01:37 |
| kfox1111 | no, thats too generic. | 01:37 |
| kfox1111 | what step of "deploy ceph" are you stuck on? | 01:37 |
| kfox1111 | or is it, step 1, 'pick a deployment tool'? | 01:37 |
| sdake | i haven't deployed ceph - ran out of time on friday | 01:37 |
| sdake | i got through all the gating and got everything written down | 01:37 |
| kfox1111 | ah. ok. | 01:38 |
| kfox1111 | so, stuck on time. need time to try deploying it? | 01:38 |
| sdake | it sounds now like we don't want to deploy ceph | 01:38 |
| sdake | but want to use a local storage thingamagic :) | 01:38 |
| kfox1111 | we need to write that support. | 01:38 |
| kfox1111 | and I still highly recomment backing nova/glance/cinder with ceph. | 01:39 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Split Openvswitch into own role https://review.openstack.org/416793 | 01:39 |
| portdirect | kfox, whould you help make this work https://github.com/openstack/kolla-kubernetes/blob/master/tests/bin/build_test_ceph.sh ouside of the gate (for devs) | 01:39 |
| sdake | portdirect cool i hadn't seen that file | 01:40 |
| sdake | portdirect seen/got to it yet :) | 01:40 |
| portdirect | thats how the gte runs ceph | 01:40 |
| kfox1111 | thats the majority of the logic. | 01:40 |
| portdirect | i mageed to make it run outside with a lot of sed | 01:40 |
| portdirect | once | 01:40 |
| kfox1111 | the rest is tests/bin/setup_gate_loopback.sh | 01:41 |
| sdake | portdirect ready to devconf :) | 01:41 |
| portdirect | not yet - gona get the pp done monday | 01:41 |
| portdirect | got a lot to fit in | 01:42 |
| sdake | i'm sure it will be great | 01:42 |
| kfox1111 | tools/setup_simple_ceph_users.sh, and tools/setup-ceph-secrets.sh | 01:42 |
| sdake | because thats how you roll :) | 01:42 |
| kfox1111 | I tried to seperate out the last two files so they could be used with an external ceph. | 01:42 |
| *** nihilifer has quit IRC | 01:42 | |
| kfox1111 | though never tried that yet. | 01:42 |
| portdirect | basicly gonna give 10 mins to each: why?, apline images & building in dockerhub, deploying k8s/rpm-ostree, freeipa/ipsion & keystone, demo | 01:43 |
| sdake | i've got a 2 hour meetup in tempe in february | 01:44 |
| sdake | need to write the abstract | 01:44 |
| * sdake groans | 01:44 | |
| kfox1111 | sdake. theres another option. | 01:44 |
| kfox1111 | if you only want to do a two node dev env, one controller one compute, | 01:44 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Split Openvswitch into own role https://review.openstack.org/416793 | 01:45 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Add OpenDaylight role https://review.openstack.org/416367 | 01:45 |
| kfox1111 | you could use the iscsi branch of the gate instead. | 01:45 |
| kfox1111 | then you don't need ceph at all. | 01:45 |
| sdake | kfox1111 i am after multinode for devs, as the dev env simulates 4 nodes | 01:45 |
| kfox1111 | what do the 4 nodes do? | 01:45 |
| *** nihilifer has joined #openstack-kolla | 01:45 | |
| sdake | one is controller 3 are assigned whatever roels you like - the current docs recommend compute nad controlelr on all 3 nodes | 01:46 |
| portdirect | kfox1111: currently nothing... 1 controller. 3 nodes | 01:46 |
| sdake | roles/labels | 01:46 |
| kfox1111 | one k8s controller, and 3 k8s compute nodes then? | 01:46 |
| portdirect | we set it up that way when we imagined that we could get it running outside of gat fater | 01:46 |
| sdake | kfox1111 right | 01:46 |
| kfox1111 | k. | 01:46 |
| kfox1111 | so the 3 nodes coudl be used, for 1 openstack controller and 2 computes | 01:47 |
| kfox1111 | or 3 controllers/computes combined. | 01:47 |
| portdirect | the idea was the 2nd | 01:47 |
| kfox1111 | or 1 controller, one nn, and one compute | 01:47 |
| sdake | the docs indicate the 2nd model | 01:47 |
| kfox1111 | k. | 01:47 |
| portdirect | as this would let us test ha of everything at lowest hardware cost | 01:47 |
| sdake | ya - should run in 16-32gb of ram | 01:47 |
| sdake | on one machine | 01:47 |
| kfox1111 | portdirect: I guess the issue is, the iscsi gate will work on one controller,l but not 3. | 01:48 |
| kfox1111 | but doesn't need ceph. | 01:48 |
| portdirect | thats fine | 01:48 |
| kfox1111 | so the 1 controller 2 compute would be quicker to document. | 01:48 |
| portdirect | we can use labels to do that easily | 01:49 |
| kfox1111 | depends I guess on just how soon you want to onboard new devs. | 01:49 |
| sdake | kfox1111 got 2 coming right now - just got a linked in email two fulltimer devs coming to kolla-kubernetes | 01:49 |
| kfox1111 | the test ceph in the gate will only run on one node too. | 01:49 |
| sdake | they are looking for how to get engaged | 01:49 |
| sdake | and I'm like "groan - i can't even get engaged:)" | 01:49 |
| portdirect | kfox1111: i dont want to get political - but others are joining other projects beasue of this | 01:50 |
| kfox1111 | ceph-deploy could do a 3 node full ha deployment pretty easily. | 01:50 |
| kfox1111 | portdirect: yeah, I get that. | 01:50 |
| kfox1111 | like I said, helm was not withouit cost. :/ | 01:50 |
| sdake | ok lets forget the past | 01:50 |
| sdake | and focus on teh future | 01:50 |
| kfox1111 | yup. | 01:50 |
| kfox1111 | so how quickly do yo uwant a dev env. | 01:50 |
| kfox1111 | ? | 01:50 |
| sdake | so the future is - we need docs, and right now :) | 01:50 |
| kfox1111 | wha tradeoffs are ok, vs what we shoudl delay for? | 01:51 |
| sdake | i am committed to writing the docs | 01:51 |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Split Openvswitch into own role https://review.openstack.org/416793 | 01:51 |
| sdake | i'd like it to match up with the devenv we have already documented | 01:51 |
| sdake | sinc ethat works well | 01:51 |
| sdake | so tradeoffs | 01:51 |
| kfox1111 | if you want something in the next day or two, I'd recommend doing 1 controoler, 2 compute, and basing it on the iscsi_workflow for now. | 01:51 |
| kfox1111 | you can skip ceph alltogether then. | 01:52 |
| kfox1111 | it sounds like you are really close to being able to do that. | 01:52 |
| kfox1111 | then, once thats done, and the pressure is off, we can work on smoothing out the ceph deployment stuff. | 01:53 |
| kfox1111 | (and maybe new folks can help with that) | 01:53 |
| sdake | ok - so thinking down those lines | 01:55 |
| sdake | where we recommend 1 controller 2 computes | 01:55 |
| sdake | is the PV hardcoded in the dev env to use ceph portdirect ^^? | 01:56 |
| kfox1111 | its just changing the node lables. | 01:56 |
| kfox1111 | sdake: no, its flexible. | 01:56 |
| kfox1111 | the iscsi gate uses localpath pv's. | 01:56 |
| sdake | kfox1111 i was spekaing about the halcyon devenv we are using | 01:59 |
| sdake | lacyon dev env may or may not make assumptions about the storage in use | 01:59 |
| sdake | i'm unclear if this is the case | 01:59 |
| sdake | portdirect could you confirm? | 01:59 |
| kfox1111 | oh. | 02:00 |
| *** sayantani01 has joined #openstack-kolla | 02:00 | |
| openstackgerrit | Marcus Williams proposed openstack/kolla-ansible: Add OpenDaylight role https://review.openstack.org/416367 | 02:02 |
| portdirect | no halcyon is 100% stock k8s | 02:02 |
| portdirect | with the expetion of adding things required for ceph support if slected | 02:03 |
| portdirect | but it does not preculde using any other storage | 02:03 |
| portdirect | or alter how you would do it | 02:03 |
| kfox1111 | nice. | 02:03 |
| sdake | portdirect does that mean we could potentially use the iscsi workflow then? | 02:04 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 02:04 |
| sdake | external to the dev env? | 02:04 |
| portdirect | yes | 02:04 |
| sdake | i recall something about halcyon not supporting PVCs | 02:04 |
| sdake | is that fixed? | 02:04 |
| portdirect | that was k8s not supporting pvcs at the time (1.4.6) so yes fixed now | 02:05 |
| sdake | cool | 02:05 |
| kfox1111 | portdirect: you talking about dynamic pvc's? | 02:05 |
| sdake | so i shoudl be able to get iscsi workflow working pretty easily and document that in a quickstart | 02:05 |
| portdirect | the isse there comes from the use of k8s in public clouds | 02:05 |
| kfox1111 | cause pvc's have been supported for a while. using them in 1.3 still on one system. | 02:05 |
| kfox1111 | ah. | 02:05 |
| portdirect | where is you set up a k8s cluster on aws/gce/azure etcd you have pvc support out of the box | 02:05 |
| portdirect | but not on bare metal | 02:06 |
| kfox1111 | the storageclass stuff? | 02:06 |
| portdirect | and halycon is bare metal (in vm's) | 02:06 |
| sdake | ok - so lets focus on baby steps with a goal of having a quickstart in reviewable state by monday or tuesday | 02:06 |
| portdirect | either sotage class orf via the cloud provider | 02:06 |
| sdake | I'll write the docs myself | 02:06 |
| sdake | I have monday and tuesday mostly available | 02:06 |
| kfox1111 | portdirect: ok. that makes sense. | 02:06 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 02:07 |
| sdake | would an iscsi backend with 1 controller 2 computes be suitable for dev kfox1111 and portdirect ? | 02:07 |
| portdirect | kfox1111: youve leant to ready my jabberings :) I could ahrdly make that out :) | 02:07 |
| portdirect | sdake if it starts its sutuable for dev :) | 02:07 |
| portdirect | they can help make it better | 02:08 |
| portdirect | if you can get it to a working keystone | 02:08 |
| kfox1111 | +1. | 02:08 |
| kfox1111 | it will have all the bits to test things out, fix bugs, and add new features. | 02:08 |
| kfox1111 | and worst case, they can contribute and check against ceph in the gate. | 02:09 |
| kfox1111 | so yea, it whould help a lot I think. | 02:09 |
| sdake | i like whould :) | 02:09 |
| sdake | i'll have to use that | 02:09 |
| sdake | ;) | 02:09 |
| kfox1111 | then we will l have more time to get ceph support into it. | 02:09 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 02:19 |
| *** Syffs has quit IRC | 02:32 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 02:32 |
| *** hfu has quit IRC | 02:44 | |
| *** jrich523 has quit IRC | 02:45 | |
| *** masuberu has quit IRC | 02:50 | |
| *** sp___ has joined #openstack-kolla | 02:53 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 02:55 |
| Jeffrey4l_ | sdake, re release note. yes. i will add the link. | 02:56 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix add kubernetes image release note format https://review.openstack.org/423739 | 03:26 |
| *** hfu has joined #openstack-kolla | 03:28 | |
| sdake | Jeffrey4l_ thanks :) | 03:29 |
| sdake | Jeffrey4l_ someone asked for it today | 03:29 |
| sdake | or maybe it was yesterday - for the newotn 3.0.2 release notes | 03:29 |
| sdake | and i noticed they arn't published | 03:29 |
| Jeffrey4l_ | sdake, could u review it? https://review.openstack.org/423737 | 03:29 |
| *** sudipto has joined #openstack-kolla | 03:32 | |
| *** sudipto_ has joined #openstack-kolla | 03:32 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add chrony ansible role https://review.openstack.org/399355 | 03:32 |
| openstackgerrit | jianyi proposed openstack/kolla-ansible: Add zun ansible role https://review.openstack.org/416179 | 03:34 |
| *** hfu has quit IRC | 03:34 | |
| *** zhurong has quit IRC | 03:43 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimize reconfigure action for nova https://review.openstack.org/407416 | 03:47 |
| *** salv-orlando has joined #openstack-kolla | 03:48 | |
| *** salv-orlando has quit IRC | 03:52 | |
| *** hfu has joined #openstack-kolla | 03:56 | |
| *** hfu has quit IRC | 04:06 | |
| sdake | Jeffrey4l_ i did review | 04:07 |
| sdake | it lgtm however i am unclear on the unreleased.html thing | 04:07 |
| sdake | Jeffrey4l_ the release core reviewers will know what is correct | 04:07 |
| Jeffrey4l_ | roger. i think, during developing, release-notes should be "unreleased.html" and should move to ocata.html at release. | 04:08 |
| sdake | Jeffrey4l_ cool - the release team can review - they will know for certain | 04:09 |
| sdake | Jeffrey4l_ i did add a note tho to the review | 04:09 |
| Jeffrey4l_ | i see. thanks. | 04:09 |
| *** sudipto has quit IRC | 04:09 | |
| *** sudipto_ has quit IRC | 04:09 | |
| sdake | Jeffrey4l_ can you review this plz https://review.openstack.org/#/c/423633/ | 04:12 |
| Jeffrey4l_ | sdake, done. | 04:13 |
| sdake | Jeffrey4l_ tx | 04:13 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Fix mysql database user of ceilometer mysql https://review.openstack.org/423744 | 04:16 |
| *** caowei has quit IRC | 04:24 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Fix mysql database user of ceilometer https://review.openstack.org/423744 | 04:32 |
| *** sp___ has quit IRC | 04:34 | |
| openstackgerrit | Merged openstack/kolla-ansible: Update the upgrade.yml for trove https://review.openstack.org/423142 | 04:37 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for gnocchi https://review.openstack.org/423133 | 04:48 |
| *** salv-orlando has joined #openstack-kolla | 04:49 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for gnocchi https://review.openstack.org/423133 | 04:53 |
| *** salv-orlando has quit IRC | 04:53 | |
| openstackgerrit | Merged openstack/kolla: Install karbor dashboard into horizon image https://review.openstack.org/423575 | 04:54 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for searchlight https://review.openstack.org/423590 | 04:57 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 05:02 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 05:05 |
| *** saneax-_-|AFK is now known as saneax | 05:07 | |
| openstackgerrit | Merged openstack/kolla: Kubernetes: Rename Images to allow use with native deployment tools https://review.openstack.org/423633 | 05:10 |
| *** sacharya has joined #openstack-kolla | 05:11 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for aodh https://review.openstack.org/423670 | 05:11 |
| *** yangzb09 has joined #openstack-kolla | 05:13 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for cloudkitty https://review.openstack.org/422516 | 05:14 |
| *** mnaser has quit IRC | 05:15 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for senlin https://review.openstack.org/423336 | 05:16 |
| *** caowei has joined #openstack-kolla | 05:18 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for senlin https://review.openstack.org/423336 | 05:19 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for cloudkitty https://review.openstack.org/422516 | 05:19 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for aodh https://review.openstack.org/423670 | 05:20 |
| *** sdake has quit IRC | 05:21 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 05:21 |
| *** mnaser has joined #openstack-kolla | 05:24 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for gnocchi https://review.openstack.org/423133 | 05:25 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 05:27 |
| *** goldyfruit has joined #openstack-kolla | 05:37 | |
| *** mnaser has quit IRC | 05:49 | |
| *** salv-orlando has joined #openstack-kolla | 05:49 | |
| *** salv-orlando has quit IRC | 05:54 | |
| *** mnaser has joined #openstack-kolla | 05:57 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for ceilometer https://review.openstack.org/423606 | 05:58 |
| *** sdake has joined #openstack-kolla | 06:05 | |
| *** trinaths has joined #openstack-kolla | 06:09 | |
| *** SridharP has joined #openstack-kolla | 06:10 | |
| *** SridharP1 has joined #openstack-kolla | 06:19 | |
| *** SridharP has quit IRC | 06:21 | |
| *** saneax is now known as saneax-_-|AFK | 06:22 | |
| *** goldyfruit has quit IRC | 06:23 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for ceilometer https://review.openstack.org/423606 | 06:28 |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes: Horizon service package https://review.openstack.org/423016 | 06:31 |
| openstackgerrit | Yao Lu proposed openstack/kolla: Change docker restart policy https://review.openstack.org/423026 | 06:34 |
| *** zhurong has joined #openstack-kolla | 06:37 | |
| openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Add freezer ansible role https://review.openstack.org/402224 | 06:40 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for gnocchi https://review.openstack.org/423133 | 06:41 |
| *** zhurong has quit IRC | 06:42 | |
| *** zhurong has joined #openstack-kolla | 06:42 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 06:46 |
| *** salv-orlando has joined #openstack-kolla | 06:50 | |
| *** zhubingbing has joined #openstack-kolla | 06:53 | |
| *** salv-orlando has quit IRC | 06:55 | |
| *** hachi has joined #openstack-kolla | 07:01 | |
| openstackgerrit | yuhui proposed openstack/kolla: Added oslo_log replacing logging library,more file https://review.openstack.org/421199 | 07:02 |
| openstackgerrit | yuhui proposed openstack/kolla: Added oslo_log replacing logging library,more file https://review.openstack.org/421199 | 07:03 |
| *** yangzb09_ has joined #openstack-kolla | 07:13 | |
| *** yangzb09 has quit IRC | 07:16 | |
| *** saneax-_-|AFK is now known as saneax | 07:27 | |
| openstackgerrit | yuhui proposed openstack/kolla: Added oslo_log replacing logging library,more file https://review.openstack.org/421199 | 07:31 |
| *** sacharya has quit IRC | 07:33 | |
| openstackgerrit | yuhui proposed openstack/kolla: Added oslo_log replacing logging library,more file https://review.openstack.org/421199 | 07:33 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: [TEST] https://review.openstack.org/423766 | 07:33 |
| *** bmace_ has quit IRC | 07:47 | |
| *** bmace_ has joined #openstack-kolla | 07:47 | |
| *** sp___ has joined #openstack-kolla | 07:50 | |
| *** salv-orlando has joined #openstack-kolla | 07:51 | |
| *** salv-orlando has quit IRC | 07:56 | |
| *** SridharP1 has quit IRC | 07:58 | |
| *** SridharP has joined #openstack-kolla | 08:00 | |
| *** sdake has quit IRC | 08:02 | |
| *** harish23r has joined #openstack-kolla | 08:02 | |
| zhubingbing | hello guys | 08:02 |
| zhubingbing | https://review.openstack.org/#/c/423573/ | 08:02 |
| zhubingbing | who can help me look it ? | 08:02 |
| harish23r | hey guys! | 08:03 |
| harish23r | can anyone help me out with my all in one deployment error? while using "kolla-ansible deploy" command | 08:03 |
| harish23r | This is my error: | 08:04 |
| harish23r | TASK: [common | starting heka container] ************************************* | 08:04 |
| harish23r | failed: [localhost] => {"changed": true, "failed": true} | 08:04 |
| harish23r | msg: APIError(HTTPError(u'500 Server Error: Internal Server Error for url: http+docker://localunixsocket/v1.25/conteiners/create?name=heka',),) | 08:04 |
| harish23r | FATAL: all hosts have failed -- aborting | 08:04 |
| harish23r | localhost: 0k=28 changed=0 unreachable=0 failed=1 | 08:04 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for gnocchi https://review.openstack.org/423133 | 08:08 |
| Jeffrey4l_ | harish23r, did it works before? | 08:13 |
| harish23r | nope. | 08:13 |
| Jeffrey4l_ | try use "docker pull" to pull the image, does it work? | 08:13 |
| *** sdake has joined #openstack-kolla | 08:14 | |
| harish23r | yes. the registry operation works | 08:14 |
| Jeffrey4l_ | harish23r, try downgrade the docker-engine to 1.12.6, kolla do not support docker-engine 1.13 now. | 08:14 |
| harish23r | how did you identify my version? | 08:15 |
| harish23r | btw, i tried both versions. 1.12 and 1.13. same result. | 08:15 |
| Jeffrey4l_ | harish23r, you are using 1.25 api. which provides by docker 1.13. | 08:16 |
| Jeffrey4l_ | hrm. how about docker run, does it work? | 08:16 |
| harish23r | I see. But both gave the same error. while starting heka container. | 08:16 |
| harish23r | docker run works fine. | 08:16 |
| harish23r | i used docker run to start the registry. | 08:16 |
| openstackgerrit | Merged openstack/kolla-ansible: Add Karbor doc link in README.rst https://review.openstack.org/422995 | 08:17 |
| harish23r | What does the error actually mean? its a docker error right? | 08:17 |
| Jeffrey4l_ | could u get more specific error? check the docker logs. | 08:17 |
| harish23r | on a different machine now. | 08:17 |
| Jeffrey4l_ | i do not know. docker engine just return that. | 08:17 |
| harish23r | i ran the command: "kolla-ansible deploy" | 08:18 |
| harish23r | it shows many lines and finally stopped with that error. | 08:18 |
| openstackgerrit | Wei Cao proposed openstack/kolla: Add dragonflow container https://review.openstack.org/418193 | 08:18 |
| Jeffrey4l_ | check the docker.log / journalctl -u docker -f , to check what's going on. | 08:18 |
| harish23r | okay. | 08:19 |
| harish23r | i tried it earlier. but im very new to docker & kolla | 08:20 |
| harish23r | couldn't understand much | 08:20 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimize reconfigure action for nova https://review.openstack.org/407416 | 08:21 |
| Jeffrey4l_ | harish23r, try to get more error message . | 08:22 |
| Jeffrey4l_ | from docker engine's log | 08:22 |
| Jeffrey4l_ | show me you "docker info" harish23r | 08:22 |
| harish23r | just on another machine now. i'll get back soon! | 08:22 |
| harish23r | thanks Jeffrey41_ | 08:23 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 08:23 |
| Jeffrey4l_ | sdake, could u ack this one https://review.openstack.org/423739 | 08:23 |
| sdake | Jeffrey4l_ acked | 08:24 |
| Jeffrey4l_ | thanks. | 08:24 |
| sdake | Jeffrey4l_ did release notes not previously render properly? | 08:25 |
| Jeffrey4l_ | yep. at least the format is OK. | 08:25 |
| Jeffrey4l_ | but i have no idea why 3.0.0.0bx still on the current series releaset note sections. http://docs.openstack.org/releasenotes/kolla/ | 08:26 |
| Jeffrey4l_ | sdake, | 08:26 |
| *** williamcaban has quit IRC | 08:27 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for ceilometer https://review.openstack.org/423606 | 08:31 |
| openstackgerrit | yuhui proposed openstack/kolla: Added oslo_log replacing logging library,more file https://review.openstack.org/421199 | 08:33 |
| *** williamcaban has joined #openstack-kolla | 08:36 | |
| harish23r | btw, Jeffrey41_ , I'm using a local registry. I've build and pushed the required images to the registry. | 08:36 |
| harish23r | docker images, lists the images in the registry. | 08:36 |
| harish23r | do you want me to check if "docker pull" works with the local registry? | 08:37 |
| *** williamcaban has quit IRC | 08:40 | |
| openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Load karbor dashboard https://review.openstack.org/423576 | 08:41 |
| Jeffrey4l_ | harish23r, yep. need check whether docker works without kolla. | 08:45 |
| *** saneax is now known as saneax-_-|AFK | 08:46 | |
| sp___ | sdake: Jeffrey4l_ I know you both have crunch of time. but it would be great if you review this for BP completion of kolla coverage. https://review.openstack.org/#/c/422444/ | 08:49 |
| Jeffrey4l_ | sorry. i will review it now | 08:49 |
| sp___ | Jeffrey4l_: np, tx | 08:50 |
| *** sayantani01 has quit IRC | 08:52 | |
| *** salv-orlando has joined #openstack-kolla | 08:52 | |
| *** sdake has quit IRC | 08:53 | |
| *** salv-orlando has quit IRC | 08:57 | |
| *** harish23r has quit IRC | 09:00 | |
| Jeffrey4l_ | sp___, done. some nit. | 09:00 |
| sp___ | Jeffrey4l_: tx | 09:00 |
| Jeffrey4l_ | np. | 09:00 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for mistral https://review.openstack.org/423787 | 09:06 |
| *** lamt has quit IRC | 09:17 | |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible: Use required_if in kolla_docker.py https://review.openstack.org/423788 | 09:25 |
| openstackgerrit | Merged openstack/kolla: Fix add kubernetes image release note format https://review.openstack.org/423739 | 09:30 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimize reconfigure action for nova https://review.openstack.org/407416 | 09:32 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for mistral https://review.openstack.org/423787 | 09:40 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for mistral https://review.openstack.org/423787 | 09:48 |
| *** zhurong has quit IRC | 09:49 | |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible: Use required_if in kolla_docker.py https://review.openstack.org/423788 | 09:50 |
| *** salv-orlando has joined #openstack-kolla | 09:53 | |
| openstackgerrit | Sridhar Pothuganti proposed openstack/kolla-kubernetes: Correction of file path in the developer environment page https://review.openstack.org/423793 | 09:55 |
| *** salv-orlando has quit IRC | 09:58 | |
| openstackgerrit | yuhui proposed openstack/kolla: Added oslo_log replacing logging library,more file https://review.openstack.org/421199 | 09:58 |
| *** Syffs has joined #openstack-kolla | 10:02 | |
| *** williamcaban has joined #openstack-kolla | 10:03 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for mistral https://review.openstack.org/423787 | 10:05 |
| *** williamcaban has quit IRC | 10:16 | |
| *** yangzb09_ has quit IRC | 10:22 | |
| trinaths | Hi | 10:25 |
| trinaths | While setting up the kolla-kibernetes development evironment, following the guide https://github.com/openstack/kolla-kubernetes/blob/master/doc/source/development-environment.rst#managing-and-interacting-with-the-environment , I got the following error, http://paste.openstack.org/show/595964/ . Please help me resolve the issue. I'm setting up the environement on Ubunut 14.04.5. Internet is connected to eth1 interface of the box. | 10:28 |
| trinaths | kube1: Configuring and enabling network interfaces... | 10:28 |
| trinaths | is giving the error. | 10:28 |
| trinaths | thanks in advance for the help | 10:29 |
| *** haplo37_ has quit IRC | 10:37 | |
| *** haplo37_ has joined #openstack-kolla | 10:37 | |
| *** zhubingbing has quit IRC | 10:47 | |
| *** zhubingbing has joined #openstack-kolla | 10:48 | |
| openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Code coverage improvement for kolla https://review.openstack.org/422444 | 10:51 |
| *** pq_ has quit IRC | 10:51 | |
| *** salv-orlando has joined #openstack-kolla | 10:54 | |
| *** salv-orlando has quit IRC | 10:58 | |
| openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Fix solum precheck https://review.openstack.org/423798 | 11:05 |
| *** caowei has quit IRC | 11:06 | |
| trinaths | the virtualbox version is 4.3 | 11:07 |
| *** williamcaban has joined #openstack-kolla | 11:13 | |
| *** williamcaban has quit IRC | 11:17 | |
| trinaths | Moving forward, found another issue, with kubectl. | 11:29 |
| trinaths | root@openstack-pike:/tmp/halcyon-vagrant-kubernetes# kubectl run -i -t $(uuidgen) --image=busybox --restart=Never | 11:29 |
| trinaths | error: Error loading config file "/root/.kube/config": yaml: mapping values are not allowed in this context | 11:29 |
| trinaths | did I miss something | 11:29 |
| trinaths | ? | 11:29 |
| *** salv-orlando has joined #openstack-kolla | 11:54 | |
| *** masber has quit IRC | 11:57 | |
| *** masber has joined #openstack-kolla | 11:58 | |
| *** yangzb09_ has joined #openstack-kolla | 12:18 | |
| trinaths | with CentOS it all went well. But MODULE_ARGS failed at Ansible. | 12:23 |
| trinaths | Here is the error http://paste.openstack.org/show/595971/ | 12:24 |
| trinaths | Its an ansbile error. I'm unable to solve it. | 12:24 |
| trinaths | any solution to it ? | 12:25 |
| zhubingbing | ok | 12:29 |
| zhubingbing | trinaths i look it | 12:30 |
| zhubingbing | you can upgrade ansible version | 12:30 |
| *** sp___ has quit IRC | 12:31 | |
| *** mgoddard has joined #openstack-kolla | 12:33 | |
| *** williamcaban has joined #openstack-kolla | 12:34 | |
| *** williamcaban has quit IRC | 12:39 | |
| *** sudipto_ has joined #openstack-kolla | 12:44 | |
| *** sudipto has joined #openstack-kolla | 12:44 | |
| *** salv-orlando has quit IRC | 12:55 | |
| *** tyrola has quit IRC | 12:57 | |
| *** tyrola has joined #openstack-kolla | 12:59 | |
| *** SridharP has quit IRC | 13:00 | |
| *** zhubingbing has quit IRC | 13:03 | |
| *** salv-orlando has joined #openstack-kolla | 13:04 | |
| openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Correct the groups for solum in haproxy.cfg.j2 https://review.openstack.org/420407 | 13:10 |
| *** SridharP has joined #openstack-kolla | 13:10 | |
| *** mgoddard has quit IRC | 13:10 | |
| trinaths | ansible version is 2.2.0 | 13:11 |
| trinaths | to which version I need to upgrade | 13:11 |
| *** SridharP has quit IRC | 13:11 | |
| trinaths | ? | 13:11 |
| *** esmiurium has quit IRC | 13:17 | |
| *** hieulq_ has joined #openstack-kolla | 13:18 | |
| *** hieulq_ has quit IRC | 13:21 | |
| *** esmiurium has joined #openstack-kolla | 13:24 | |
| *** zhubingbing has joined #openstack-kolla | 13:27 | |
| trinaths | zhubingbing: Hi | 13:29 |
| zhubingbing | hi | 13:29 |
| trinaths | zhubingbing: the ansible version is 2.2.0 | 13:29 |
| zhubingbing | docker version ？ | 13:30 |
| *** kristian__ has joined #openstack-kolla | 13:30 | |
| zhubingbing | you deploy k8s? | 13:30 |
| zhubingbing | sorry, I have not been deployed kolla-k8s | 13:31 |
| zhubingbing | ；（ | 13:31 |
| trinaths | zhubingbing: there is an issue with docker in my system. Please check this http://paste.openstack.org/show/595973/ | 13:31 |
| zhubingbing | ；） | 13:32 |
| zhubingbing | i know | 13:32 |
| trinaths | zhubingbing: so, how solve the docker issue | 13:32 |
| zhubingbing | apt-get remove docker-engine | 13:32 |
| trinaths | zhubingbing: there is no such package installed, http://paste.openstack.org/show/595974/ | 13:33 |
| *** kristian__ has quit IRC | 13:33 | |
| zhubingbing | apt-get remove docker | 13:34 |
| trinaths | zhubingbing: done | 13:34 |
| zhubingbing | apt-get -y install docker-engine | 13:34 |
| trinaths | zhubingbing: doing... | 13:36 |
| zhubingbing | ok | 13:36 |
| zhubingbing | wait finish | 13:37 |
| trinaths | zhubingbing: Docker version 1.13.0, build 49bf474 | 13:39 |
| zhubingbing | ok | 13:39 |
| zhubingbing | docker info | 13:39 |
| zhubingbing | you can continue deploy kolla-k8s | 13:40 |
| trinaths | zhubingbing: http://paste.openstack.org/show/595975/ | 13:40 |
| zhubingbing | there is an issue with docker ，i think it have been solove | 13:40 |
| trinaths | I;m following the documentation at: https://github.com/openstack/kolla-kubernetes/blob/master/doc/source/development-environment.rst#managing-and-interacting-with-the-environment | 13:40 |
| zhubingbing | ；） | 13:40 |
| zhubingbing | hmm | 13:41 |
| trinaths | zhubingbing: at $> vagrant up , I got these errors | 13:41 |
| zhubingbing | you can continue working for deploy | 13:41 |
| trinaths | zhubingbing: ok :) | 13:42 |
| zhubingbing | sorry, I have not been deployed kolla-k8s | 13:42 |
| zhubingbing | 8/5000 | 13:43 |
| zhubingbing | I will go on to deploy kolla-K8S ；） | 13:43 |
| trinaths | zhubingbing: okay. 8/5000 ? | 13:43 |
| zhubingbing | I sent the wrong message with 8/5000 | 13:44 |
| zhubingbing | ;) | 13:44 |
| trinaths | zhubingbing: :D | 13:44 |
| zhubingbing | ；） | 13:47 |
| *** sudipto has quit IRC | 13:49 | |
| *** sudipto_ has quit IRC | 13:49 | |
| *** mgoddard has joined #openstack-kolla | 13:51 | |
| *** sp___ has joined #openstack-kolla | 13:52 | |
| *** williamcaban has joined #openstack-kolla | 13:56 | |
| openstackgerrit | Wei Cao proposed openstack/kolla: Fix solum dashboard into horizon image. https://review.openstack.org/423837 | 13:58 |
| trinaths | zhubingbing: still the same error with ansible | 14:00 |
| trinaths | zhubingbing: fatal: [kube1]: FAILED! => {"changed": false, "failed": true, "module_stderr": "Shared connection to 127.0.0.1 closed.\r\n", "module_stdout": "Traceback (most recent call last):\r\n File \"/tmp/ansible_QYND1A/ansible_module_command.py\", line 212, in <module>\r\n main()\r\n File \"/tmp/ansible_QYND1A/ansible_module_command.py\", line 90, in main\r\n module = CommandModule(argument_spec=dict())\r\n File \"/tmp/ansible | 14:00 |
| *** williamcaban has quit IRC | 14:00 | |
| zhubingbing | ;) you can install docker-engine 1.20 | 14:01 |
| zhubingbing | you can try it | 14:01 |
| trinaths | oh okay | 14:01 |
| trinaths | zhubingbing: there is not version 1.20 for ubuntu | 14:05 |
| trinaths | zhubingbing: i'm using Ubuntu 14.04.5 | 14:05 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimize reconfigure action for nova https://review.openstack.org/407416 | 14:06 |
| sp___ | Jeffrey4l_: incorporated your comments please review it again, Now seems ok to me :( https://review.openstack.org/#/c/422444/9 | 14:07 |
| openstackgerrit | Wei Cao proposed openstack/kolla: Fix searchlight dashboard into horizon image. https://review.openstack.org/423845 | 14:09 |
| *** zhubingbing has quit IRC | 14:09 | |
| *** caowei has joined #openstack-kolla | 14:10 | |
| *** zhubingbing has joined #openstack-kolla | 14:10 | |
| Jeffrey4l_ | sp___, done. | 14:13 |
| sp___ | Jeffrey4l_: thanks | 14:14 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimize reconfigure action for nova https://review.openstack.org/407416 | 14:14 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Add chrony ansible role https://review.openstack.org/399355 | 14:15 |
| sbezverk | Jeffrey4l_: ping | 14:16 |
| Jeffrey4l_ | sbezverk, pong | 14:16 |
| sbezverk | Jeffrey4l_: do you see consistent gate failure on centos? ubuntu's jobs retrun success | 14:18 |
| Jeffrey4l_ | sbezverk, no. any link? | 14:18 |
| Jeffrey4l_ | or error message? | 14:18 |
| sbezverk | http://logs.openstack.org/16/423016/3/check/gate-kolla-kubernetes-deploy-centos-binary-2-iscsi-nv/eda3da3/console.html | 14:18 |
| sbezverk | Jeffrey4l_: and it is the same error on all cenots | 14:19 |
| sbezverk | like repo busted | 14:19 |
| *** caowei has quit IRC | 14:19 | |
| Jeffrey4l_ | yes. seems infra mirror is busted. | 14:19 |
| Jeffrey4l_ | http://mirror.iad.rax.openstack.org/epel | 14:19 |
| Jeffrey4l_ | may be a recheck works? | 14:20 |
| sbezverk | effrey4l_: yah, will try in a bit.. | 14:20 |
| *** zhubingbing has quit IRC | 14:20 | |
| Jeffrey4l_ | otherwise, should ping openstack-infra team to check what going on. | 14:20 |
| *** caowei has joined #openstack-kolla | 14:21 | |
| Jeffrey4l_ | i can download the file manually. seems it is OK now. | 14:21 |
| sbezverk | Jeffrey4l_: thanks for confirming.. I will run recheck to see | 14:21 |
| Jeffrey4l_ | np. | 14:22 |
| openstackgerrit | Wei Cao proposed openstack/kolla: Fix solum dashboard into horizon image. https://review.openstack.org/423837 | 14:25 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for octavia https://review.openstack.org/423846 | 14:29 |
| *** hieulq_ has joined #openstack-kolla | 14:29 | |
| *** SridharP has joined #openstack-kolla | 14:31 | |
| *** caowei has quit IRC | 14:32 | |
| *** williamcaban has joined #openstack-kolla | 14:33 | |
| *** williamcaban has quit IRC | 14:38 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for senlin https://review.openstack.org/423336 | 14:39 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for cloudkitty https://review.openstack.org/422516 | 14:41 |
| *** l4yerffeJ has quit IRC | 14:42 | |
| *** l4yerffeJ has joined #openstack-kolla | 14:42 | |
| *** mgoddard has quit IRC | 14:43 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for searchlight https://review.openstack.org/423590 | 14:44 |
| *** hieulq_ has quit IRC | 14:46 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for tacker https://review.openstack.org/422036 | 14:48 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 14:50 |
| *** breitz has quit IRC | 14:52 | |
| *** breitz has joined #openstack-kolla | 14:52 | |
| *** sdake has joined #openstack-kolla | 14:53 | |
| *** mgoddard has joined #openstack-kolla | 14:55 | |
| *** sdake_ has joined #openstack-kolla | 14:56 | |
| *** Mech422 has quit IRC | 14:57 | |
| *** sdake has quit IRC | 14:58 | |
| *** Jeffrey4l_ has quit IRC | 15:00 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add freezer ansible role https://review.openstack.org/402224 | 15:02 |
| *** n0isyn0i1e is now known as n0isyn0ise | 15:10 | |
| *** schwicht has joined #openstack-kolla | 15:16 | |
| openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Code coverage improvement for kolla https://review.openstack.org/422444 | 15:16 |
| *** SridharP has quit IRC | 15:22 | |
| *** salv-orlando has quit IRC | 15:24 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 15:29 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Optimize reconfiguration for tacker https://review.openstack.org/422036 | 15:31 |
| openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Code coverage improvement for kolla https://review.openstack.org/422444 | 15:34 |
| sdake_ | morning folks | 15:37 |
| kfox1111 | morning. | 15:38 |
| sdake_ | kfox1111 i was thinking about something last night - and wanted confirmation from you specifically | 15:38 |
| sdake_ | do you only test kolla-kubernetes in the gate, or do you also test outside of the gate? | 15:38 |
| kfox1111 | right now, my focus has been on the gate. sbezverk has been testing outside the gate. | 15:39 |
| sdake_ | kfox1111 cool - I see why you said that increasing the gate timers would slow down your productivity | 15:39 |
| kfox1111 | I wana make sure the base set of objects is sold and it doesnt' regress. | 15:39 |
| kfox1111 | yeah. | 15:39 |
| sdake_ | kfox1111 generally developing with a dependence on the gate is slow in general | 15:39 |
| kfox1111 | yeah. | 15:39 |
| kfox1111 | but, it has lead to a much improved gate too. | 15:40 |
| sdake_ | most devs use both tehcniques, test locally then gate | 15:40 |
| kfox1111 | so kind of a tradeoff. | 15:40 |
| kfox1111 | yeah. | 15:40 |
| sdake_ | agree, our gate is totally kick ass | 15:40 |
| sdake_ | by forcing yourself to only test against the gate, you have (mostly alone:) done a great job getting the gate into rockin shape | 15:40 |
| sdake_ | the kolla-kubernetes gate is superior to other gating in kolla | 15:40 |
| kfox1111 | if I hadn't had to debug a bunch of problems from within the gate, the gate logging wouldn't be so helpful now. :) | 15:40 |
| sdake_ | not that its a contest, just a reality | 15:40 |
| kfox1111 | yeah. | 15:41 |
| kfox1111 | an unintended sideaffect of depending on the gate for dev. | 15:41 |
| kfox1111 | interesting lesson learned. | 15:41 |
| sp___ | sdake_: morning | 15:43 |
| *** jmontleon has quit IRC | 15:47 | |
| *** jmontleon has joined #openstack-kolla | 15:49 | |
| sdake_ | sup sp___ | 15:49 |
| sdake_ | kfox1111 yup - although I think we want to get eveyrone into a position where they can develop locally or inside the gate | 15:50 |
| *** duonghq has joined #openstack-kolla | 15:50 | |
| duonghq | evening guys | 15:50 |
| sdake_ | sup duonghq | 15:50 |
| kfox1111 | sdake_: totally agree. | 15:50 |
| kfox1111 | just hasn't been a focus for anyone that has been contributing code. :/ | 15:51 |
| sdake_ | kfox1111 ok so hate to rehash the past; our conversation from last night, you believe iscsi is the fastest path to getting there? | 15:51 |
| sdake_ | its a focus for me :) | 15:51 |
| kfox1111 | yeah. if you want something asap, thats probably the fastest way there. | 15:51 |
| sdake_ | kfox1111 after our conversatoin trailed off last night i finished most of the job on tech support for my parents | 15:52 |
| kfox1111 | yeah. and I'm really grateful your working on it. :) | 15:52 |
| sdake_ | kfox1111 but then PTFOed as didnt' have energy to take a look at deploying kolla-kubernetes with the iscsi gate | 15:53 |
| sdake_ | I recall when doing the ceph work, there is a hardcode IP address in the gate jobs | 15:53 |
| sdake_ | (or bringing up ceph that is) | 15:53 |
| sdake_ | then I found this review piling around: | 15:53 |
| kfox1111 | no worries. only so much a single person can do. :) | 15:53 |
| kfox1111 | the gate hard codes some of its config in its config files, as they don't changne in the gate. | 15:54 |
| sdake_ | https://review.openstack.org/#/c/405720/ | 15:54 |
| sdake_ | kfox1111 tell me about it | 15:54 |
| kfox1111 | but the config files are changable. | 15:54 |
| sdake_ | the config files come from globals.yaml? | 15:54 |
| sdake_ | or somewhere else | 15:54 |
| duonghq | sbezverk, are you there? | 15:54 |
| sdake_ | unwrapping the union here | 15:54 |
| sdake_ | I think that review above definately doesn't work but is on the right track | 15:55 |
| sdake_ | I was going to use it as a starting point | 15:55 |
| kfox1111 | yeha, the configs mostly kolla/globals.yaml and a bit of kolla-kubernetes.yaml | 15:55 |
| kfox1111 | the gate's versions are mostly in tests/config/ | 15:55 |
| kfox1111 | they are just appended to the trun versions. | 15:55 |
| sdake_ | kfox1111 can you offer a review above that explains the steps that Pete went through - so we can get the gate operational fucntionality out of your brain into developer documentation? | 15:56 |
| kfox1111 | yeah. | 15:56 |
| sdake_ | cool - that I can work with :) | 15:56 |
| *** jmontleon has quit IRC | 15:56 | |
| kfox1111 | ah. he's working on d usting off the docker dev env idea. | 15:56 |
| kfox1111 | nice. | 15:56 |
| sdake_ | kfox1111 just looking at pete review last night before i trialed off to lala sleep time, I think pete had assumed the globals.yaml wasn't the config file | 15:57 |
| sdake_ | kfox1111 ya - I thnk thats acool idea, although I'd like to simplify it to outside of docker | 15:57 |
| sdake_ | kfox1111 and possibly tackle a dev env inside docker later | 15:57 |
| sdake_ | as in after the ptg | 15:57 |
| sdake_ | kfox1111 I just want to get something going as soon as feasible :) | 15:58 |
| kfox1111 | sdake_: when we figure out the right way to do configmap generation via helm, | 15:59 |
| kfox1111 | we can drop the need for kolla-genconfig, which will eleminate the need for a python based dev environment. | 15:59 |
| sdake_ | i dont know if we will do configmap generation via helm or via a separate tool | 15:59 |
| kfox1111 | yeah. still gotta figure that out. | 15:59 |
| sdake_ | has that been decided (the configmap generation) | 15:59 |
| sdake_ | i don't think helm can really do configmap generation | 15:59 |
| sdake_ | configmap generation is #2 in my list of things to sort out | 16:00 |
| kfox1111 | I think its one of those thigns we may want to support a batteries included option? | 16:00 |
| sdake_ | (for kolla, I have many lists;) | 16:00 |
| kfox1111 | external configmap generation shoudl be supported for sure. | 16:00 |
| kfox1111 | but maybe we support a betteries included set of helm configmap packages. | 16:00 |
| sdake_ | i guess i dont understand enough about how helm works to know if helm could do configmap generation | 16:00 |
| kfox1111 | I think I figured out a way to map helm yaml to openstack ini files generically. | 16:01 |
| sdake_ | what i've seen of other repos is people place configmap.tpl files in their charts | 16:01 |
| *** salv-orlando has joined #openstack-kolla | 16:01 | |
| kfox1111 | so theres the possibility to maybe do them unopinionatedly. | 16:01 |
| *** v1k0d3n has quit IRC | 16:01 | |
| sdake_ | and then operators are expected the fork the repo and hand-edit the files | 16:01 |
| sdake_ | i personally find this unappealing | 16:01 |
| kfox1111 | yeah me too. | 16:01 |
| sdake_ | what is the technique you had in mind for helm? | 16:02 |
| kfox1111 | its anti helm in my optionion. | 16:02 |
| kfox1111 | talking with the helm devs, they prefer the binary/source split, and users use binaries, not source. | 16:02 |
| sdake_ | its not just anti-helm, its anti-oeprator | 16:02 |
| *** jmontleon has joined #openstack-kolla | 16:02 | |
| kfox1111 | yeah. that forces operators to be dev/ops. | 16:02 |
| kfox1111 | disalows pure ops. | 16:02 |
| sdake_ | operators have enough on their plates already | 16:03 |
| kfox1111 | I'm dev/ops, but I have ops coworkers, so I totally undnerstand where they are coming from. | 16:03 |
| sdake_ | kolla's job should be to make their lives easier not harder ;) | 16:03 |
| kfox1111 | +1 | 16:03 |
| kfox1111 | we can figure out a way to have binary packages for configmap generation. | 16:03 |
| kfox1111 | that should make ops happy. | 16:03 |
| sdake_ | ok well le can chat about that later so i don't have to reload all the state in my brain (and you too) | 16:03 |
| kfox1111 | there shoudl be a way to do that I think. | 16:04 |
| sdake_ | after the dev env is working | 16:04 |
| kfox1111 | ok. | 16:04 |
| sdake_ | if you could provide a review of pete's devenv shell scripts in particular | 16:04 |
| sdake_ | (and also the docker code) | 16:04 |
| sdake_ | with an eye to "whats in the repo today" | 16:04 |
| sdake_ | vs "where we want it to be in the future" | 16:04 |
| sdake_ | that would help me the most understand how to provide a quickstart that works and is accurate | 16:05 |
| sdake_ | it may not be quick, but it will be a start :) | 16:05 |
| *** v1k0d3n has joined #openstack-kolla | 16:05 | |
| kfox1111 | yeah. I'm reviewing it now. | 16:05 |
| sdake_ | thanks dude :) | 16:05 |
| sdake_ | ok - i've got to go to my parents to finish tehe job on tedch support, however, whilethats finishing i willl ssh into my servers remotely from there and see if I can't get iscsi rolling from your review + code review of iscsi | 16:06 |
| sdake_ | of the iscsi gate | 16:06 |
| duonghq | srwilkers_, ping | 16:08 |
| *** Jeffrey4l_ has joined #openstack-kolla | 16:09 | |
| *** sdake_ has quit IRC | 16:15 | |
| *** schwicht has quit IRC | 16:22 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for aodh https://review.openstack.org/423670 | 16:32 |
| kfox1111 | portdirect: you awake? | 16:33 |
| duonghq | srwilkers_, ping | 16:33 |
| duonghq | anybody have idea why the kolla-kubernetes's gate is unstable? | 16:33 |
| kfox1111 | duonghq: what are you seeing? | 16:35 |
| duonghq | kfox1111, in my ps https://review.openstack.org/#/c/423016/3 | 16:35 |
| duonghq | after 3 recheck in a round, only 1 gate is red | 16:36 |
| duonghq | previous check failed due to some component like mariadb cannot start | 16:36 |
| duonghq | both init-element and svc | 16:36 |
| kfox1111 | yeah... flaky docker hub. :/ | 16:36 |
| kfox1111 | it would be really nice to get a container cache going. | 16:36 |
| duonghq | yup, the last check fail is http://logs.openstack.org/16/423016/3/check/gate-kolla-kubernetes-deploy-centos-binary-2-helm-entrypoint-nv/4e97eeb/console.html | 16:37 |
| duonghq | dial tcp: lookup registry-1.docker.io on 69.20.0.164:53: no such host | 16:37 |
| duonghq | 16:37 | |
| kfox1111 | as we're increasing the number of jobs, the frequency of hub issues is going up. | 16:37 |
| duonghq | ya | 16:37 |
| duonghq | do you think the docker hub is overload? | 16:37 |
| kfox1111 | not sure. they have been having some issues the last week or so with stability. | 16:38 |
| kfox1111 | not sure if its a bug, or a ddos attack, or just too muich regular traffic. | 16:38 |
| duonghq | okay | 16:39 |
| kfox1111 | yeah. I agree your ps looks like its passing all tests. just a flaky docker hub. | 16:41 |
| duonghq | so I'll recheck 1-2 times tomorrow | 16:42 |
| kfox1111 | duonghq: the ps is looking very good. one minir issue I think and its good for merge. | 16:43 |
| duonghq | kfox1111, can you tell me | 16:43 |
| kfox1111 | its in the review. | 16:44 |
| duonghq | kfox1111, just see, thanks | 16:44 |
| kfox1111 | np. :) | 16:45 |
| duonghq | but I think that it is just kind of soft dependency? | 16:45 |
| kfox1111 | maybe. the user can always turn it back off. just thinking in the general case, the user won't want their users to see horizon before authentication works. | 16:46 |
| kfox1111 | otherwise they might get support requests for "horizon is broken". | 16:46 |
| duonghq | so, does it need to put in requirements.yml? | 16:47 |
| kfox1111 | I don't htink horizon gives useful error messages if its up and keystone is down. | 16:47 |
| portdirect | o/ | 16:47 |
| kfox1111 | in the values.yaml file with the other deps. | 16:47 |
| kfox1111 | let me see if I can fidn the svc name for you. | 16:47 |
| kfox1111 | portdirect: looking through your ps for the dev env. looking very nice. :) | 16:47 |
| kfox1111 | wondering about forking the files though. | 16:47 |
| portdirect | which one? | 16:47 |
| kfox1111 | one of the reasons minikube stuff bitrotted, is because I couldn't figure out how to get it into the gate, to prevent regressions. | 16:48 |
| kfox1111 | https://review.openstack.org/#/c/405720/ | 16:48 |
| *** hachi has quit IRC | 16:48 | |
| kfox1111 | in that ps, you fork a few files that we've still been changing relatively often. | 16:48 |
| kfox1111 | could we change it to add conditionals in the other files rather then forking? | 16:49 |
| sbezverk | good morning | 16:49 |
| kfox1111 | That way, its tested inside the gate for regressions, and less code to maintain. | 16:49 |
| duonghq | kfox1111, I think the service name is keystone-api? | 16:49 |
| duonghq | morning sbezverk | 16:51 |
| kfox1111 | duonghq: not sure. keystone is a little different. I'm looking through the code now. | 16:51 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Remove contrib directory https://review.openstack.org/422559 | 16:52 |
| *** duonghq__ has joined #openstack-kolla | 16:56 | |
| *** sdake has joined #openstack-kolla | 16:56 | |
| kfox1111 | finally found it. its talkign to keystone-public dns entry. | 16:59 |
| sbezverk | kfox1111: duonghq: I think if you want to depend on keystone this service name should be used: keystone-admin | 16:59 |
| kfox1111 | yeah, so keystone-public is the svc to depend on. | 16:59 |
| sbezverk | kfox1111: strange since horizon runs within a cluster why would it use public svc? | 17:00 |
| kfox1111 | sbezverk: I thought so too, but http://logs.openstack.org/96/333996/14/check/gate-kolla-kubernetes-deploy-centos-binary-2-ceph-multi-nv/1f4e1a1/logs/kolla/horizon/local_settings | 17:00 |
| kfox1111 | and search for OPENSTACK_KEYSTONE_URL. | 17:00 |
| sbezverk | in some cases public ip might not be even reachable if it would go through proxy | 17:00 |
| kfox1111 | I'd expect public ip's to always be reachable from the cluster? | 17:01 |
| kfox1111 | not sure what woudl break if that wasn't the case. | 17:01 |
| sbezverk | kfox1111: nope, I publish "public name" which is not resolved by internal cluster dns | 17:02 |
| *** yangzb09_ has quit IRC | 17:02 | |
| kfox1111 | ah. | 17:02 |
| kfox1111 | that makes sense. | 17:02 |
| kfox1111 | well, genconfig is doing something bad then. | 17:02 |
| sbezverk | this separate makes sense for extra sec protection | 17:02 |
| sbezverk | kfox1111: very possible | 17:02 |
| *** hachi has joined #openstack-kolla | 17:03 | |
| kfox1111 | sbezverk: can you please post a kolla-ansible ps to fix that? I think you know that process better then I do. | 17:04 |
| duonghq | so, keystone-admin is better solution? | 17:04 |
| kfox1111 | keystone-internal I think. | 17:04 |
| duonghq | nice | 17:05 |
| sbezverk | kfox1111: internal would work too | 17:06 |
| duonghq | but does it needs go to requirement yaml together, or just for k8s-entrypoint? | 17:06 |
| sbezverk | kfox1111: for kolla-ansible it might make sense | 17:06 |
| portdirect | kfox1111: sorry got caught up in something | 17:06 |
| sbezverk | becaseu they do not "kubernetes cluster with internal dns" | 17:06 |
| portdirect | kfox1111: your suggestion makes total sense - and where i would have tried to go if id had the time to finish it | 17:07 |
| portdirect | kfox1111: i think ideally a developer should be able to work locally - and pretty much only push to gerrit when they expect it to pass the gate | 17:07 |
| kfox1111 | portdirect: k. | 17:07 |
| kfox1111 | agreed. | 17:07 |
| portdirect | kfox1111: if we get there, then every possible workflow is pretty much accounted for | 17:08 |
| *** salv-orlando has quit IRC | 17:08 | |
| kfox1111 | so, to review that review, I'm going to have to do a local checkout and do a diff between the forked file and whats in the gate. | 17:08 |
| portdirect | i think sdake said he was interested in taking it over | 17:08 |
| kfox1111 | if not much has changed, I could add the conditionals if your really busy? | 17:08 |
| portdirect | that would be great | 17:08 |
| kfox1111 | then we'd co-author it? | 17:09 |
| portdirect | yeah man :) | 17:09 |
| sdake | portdirect take which over | 17:09 |
| kfox1111 | ok. cool. didn't want to step on your toes. | 17:09 |
| sdake | sorry ws doing tech support day 2 operations :) | 17:09 |
| sdake | let me read scrollback | 17:09 |
| portdirect | kfox1111: taking workload on is deffo not stepping on my toes :) | 17:09 |
| sdake | portdirect this is less then a military op i've got rollign at my parents :) | 17:09 |
| sdake | kfox1111 - ya we work as a team so its not like anyone owns any particular part of the code base | 17:09 |
| kfox1111 | portdirect: glad you feel that way. :) not everyone does. just checking. | 17:09 |
| portdirect | lol - things going south on yah | 17:10 |
| sdake | kfox1111 so - dont worry about toe bashing | 17:10 |
| portdirect | sdake: nuke it from orbit 'tis the only way to be sure | 17:10 |
| *** sudipto has joined #openstack-kolla | 17:10 | |
| *** sudipto_ has joined #openstack-kolla | 17:10 | |
| sdake | [10:08:24] <kfox1111>so, to review that review, I'm going to have to do a local checkout and do a diff between the forked file and whats in the gate. | 17:11 |
| kfox1111 | sdake: some devs would rather a ps be standaline, and just get reviews. just a procedural difference between devs. I get it. so just trying to play nice. | 17:11 |
| * portdirect if kfox touches my ps... mumble mumble... somthing something.. | 17:11 | |
| portdirect | :) | 17:11 |
| *** kbaegis has joined #openstack-kolla | 17:11 | |
| kfox1111 | yeah. that. :) | 17:11 |
| sdake | i was dc'ed driving to my parents house so missed the context of the convo | 17:11 |
| kfox1111 | sdake: the ps you pointed me at for review. | 17:11 |
| sdake | kfox1111 i didn't understand what you meant is why i pasted it | 17:11 |
| portdirect | sdake: kfox is gonna see how rotted it is | 17:12 |
| sdake | was going to ask a followup if you could expand on what you meant :) | 17:12 |
| kfox1111 | sdake: the ps forks some files that are used for the gate, and makes a few changes I think. | 17:12 |
| sdake | oh that particular patch | 17:12 |
| kfox1111 | its hard to see what changed, when the whole file changed. | 17:12 |
| kfox1111 | to really review it, I'll have to checkout the ps, then do a diff on the forked files. | 17:12 |
| sdake | kfox1111 cool if you want to co-author on it i'm sure portdirect wouldn't mind | 17:12 |
| sdake | the only person who really cares about co-authors is sambetts|afk | 17:12 |
| sdake | samyaple that is | 17:12 |
| sdake | kfox1111 is that something you could do today - or no time? | 17:13 |
| portdirect | kfox1111: from that you should be able waht we need to support halycon - but ideally it should be generic | 17:13 |
| portdirect | kfox1111: if you need anything just hit me up | 17:13 |
| kfox1111 | portdirect: k. I may hit you up to test it. | 17:15 |
| kfox1111 | I don't have a way to do so currently. | 17:15 |
| *** sudipto has quit IRC | 17:15 | |
| *** sudipto_ has quit IRC | 17:15 | |
| kfox1111 | so I'll just do the diff, add the if's and repost. | 17:15 |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes: Horizon service package https://review.openstack.org/423016 | 17:15 |
| duonghq | thank kfox1111 | 17:16 |
| portdirect | kfox1111: that would make my day :) | 17:16 |
| kfox1111 | np. :) | 17:17 |
| *** sayantani01 has joined #openstack-kolla | 17:17 | |
| sdake | kfox1111 cool getting something up that doesn't work is better then getting nothing up :) | 17:17 |
| kfox1111 | yup. | 17:19 |
| kfox1111 | portdirect: with this ps, are we going to switch it to iscsi for now? | 17:20 |
| portdirect | kfox1111: I'm easy on that front | 17:20 |
| kfox1111 | k. | 17:20 |
| portdirect | kfox1111: whatever is easiest | 17:20 |
| kfox1111 | k. | 17:21 |
| sbezverk | portdirect: have you made iscsi and storageclass working? | 17:21 |
| portdirect | never tried | 17:21 |
| portdirect | though I'd look at the openshift docs | 17:21 |
| kfox1111 | theres no infra in iscsi to provide iscsi for k8s though. | 17:22 |
| portdirect | as i think that the mostly use iscsi or gluster for on-prem | 17:22 |
| *** v1k0d3n has quit IRC | 17:22 | |
| sbezverk | kfox1111: what do you mean? | 17:22 |
| kfox1111 | portdirect: who mostly use? | 17:22 |
| portdirect | openshift | 17:22 |
| sbezverk | kfox1111: it supposed to work in theory, but I suspect there is a bug | 17:23 |
| kfox1111 | sbezverk: the gate has stuff for getting cinder to provision iscsi stuff. | 17:23 |
| kfox1111 | getting iscsi volumes provisioned without cinder is a lot more work. | 17:23 |
| *** v1k0d3n has joined #openstack-kolla | 17:23 | |
| kfox1111 | more work I think then getting ceph stuff finsihed off. | 17:23 |
| sbezverk | kfox1111: ok please keep me in the loop, I would really like to see it working.. | 17:24 |
| sp___ | sdake: it would be gr8 if you review https://review.openstack.org/#/c/422444/11 | 17:25 |
| kfox1111 | sbezverk: will do. I don't have any plans on implementing non cinder iscsi provisioning though. | 17:25 |
| *** goldyfruit has joined #openstack-kolla | 17:25 | |
| sbezverk | kfox1111: for cinder to provide this functionaity, does it need other components like keystone and mariadb or it is some sort of a special version? | 17:27 |
| portdirect | sbezverk: i've used cinder for this purpose (as in that think i showd you yesterday) it would need a kesytone and dbs for both | 17:28 |
| portdirect | sbezverk: its possible to run without any auth - but yuck | 17:28 |
| portdirect | (also never tried) | 17:28 |
| sbezverk | portdirect: but then it is really lame solution, you depend on opensatck to bring kube cluster to run opensatck.. brrr I would neven use this crap | 17:29 |
| portdirect | +2 | 17:29 |
| portdirect | yeah - I need to run the ks and cinder db stroage outside of itself | 17:29 |
| portdirect | i manged to get them to self host - but it was to scary | 17:30 |
| kfox1111 | and you would loose all ha support in the process. :/ | 17:30 |
| kfox1111 | as cinder doesn't support ha with iscsi. | 17:30 |
| portdirect | not nesisarity - but its too much work - if you want to do that just use tripple-o | 17:30 |
| sbezverk | kfox1111: idk in this case I would just make efforts to get just iscsi with kube work, no cinder and all extra crap.. | 17:31 |
| kfox1111 | I gave up long ago trying to make iscsi ha without just paying a vendor for an appliance. | 17:31 |
| kfox1111 | ceph handles that case usually much better. | 17:31 |
| portdirect | sbezverk: thats the approach that we should take | 17:32 |
| kfox1111 | I think the best soltuion is to just use galera clustering and skip the pv's altogether. | 17:32 |
| sbezverk | kfox1111: with ceph you will always have less performance as there is no "ceph" enabled NIC | 17:32 |
| portdirect | +1 | 17:32 |
| sbezverk | which is different that with iscsi optimized nics | 17:32 |
| *** mgoddard has quit IRC | 17:32 | |
| kfox1111 | sbezverk: true/false. | 17:32 |
| kfox1111 | stroage is usually the bottleneck these days, not network. | 17:33 |
| kfox1111 | and ceph can scale out much farther then the traditional iscsi appliance. | 17:33 |
| * portdirect rabbitholes | 17:33 | |
| kfox1111 | hehe. yeah. | 17:33 |
| *** portdirect is now known as portdirect_away | 17:33 | |
| sbezverk | kfox1111: never had to do any study on scalability vs specialized hardware fr for storage, but nothing beats specially designed forwarding acics in router's/switches world.. | 17:35 |
| kfox1111 | sbezverk: true. well, mostly. | 17:36 |
| kfox1111 | if you could have one switch with asics, | 17:36 |
| kfox1111 | or 100x without, eventually if you have enough parallel load, | 17:36 |
| kfox1111 | parallel will outperform single. | 17:37 |
| sbezverk | kfox1111: good luck with electricity bill ;) | 17:37 |
| kfox1111 | thats why all hpc systems no longer build snowflake systems, but compute clusters. :) | 17:37 |
| duonghq | night guys | 17:37 |
| *** duonghq has quit IRC | 17:37 | |
| kfox1111 | hpc use to use custom built procesors/mb's/networks. | 17:38 |
| kfox1111 | really fast, but bang for buck, | 17:38 |
| kfox1111 | the economy of scale of general purpose computing made it much much cheeper per flop to buy a bunch of cheep machines and tie them together. | 17:38 |
| sbezverk | kfox1111: right | 17:38 |
| kfox1111 | ceph I think is doing the same for storage. | 17:38 |
| kfox1111 | vendors make some really awesome iscsi appliances. | 17:39 |
| kfox1111 | but they are designed/priced like bigiron. | 17:39 |
| sbezverk | kfox1111: for computing, I would agree.. there are too too many variations of processing | 17:39 |
| kfox1111 | ceph is built like compute clusters. but just for storage. | 17:39 |
| *** sudipto_ has joined #openstack-kolla | 17:39 | |
| *** sudipto has joined #openstack-kolla | 17:39 | |
| kfox1111 | ceph doesn't solve all the worlds storage problems. but it seems to be in a good sweetspot for general storage stuff. | 17:40 |
| *** sudipto has quit IRC | 17:41 | |
| *** sudipto_ has quit IRC | 17:41 | |
| kfox1111 | it grows cheeper as it expands. | 17:41 |
| sbezverk | kfox1111: I agree growing ceph is easier, but since it is going through linux stack, the performace will always be limiting | 17:42 |
| kfox1111 | yeah, but that can be counteracted by suporting sharding of data across servers. | 17:42 |
| kfox1111 | 100 spiny disks will be faster then 10. | 17:43 |
| sbezverk | kfox1111: what is sharding? | 17:43 |
| kfox1111 | say you have a 1tb file. | 17:43 |
| kfox1111 | you can break it into 1g chunks, and put each on a different server set. | 17:43 |
| kfox1111 | then when you get the file back, you can stream all the chunks back in parallel. | 17:44 |
| kfox1111 | so you can scale your io past any disk/server can do. | 17:44 |
| kfox1111 | iscsi doesn't support that. | 17:44 |
| sbezverk | kfox1111: got it.. would it not create extra load when this 1tb file gets changed? curious which process controls sharding? | 17:45 |
| kfox1111 | only the chunks that get changed get the io. | 17:45 |
| kfox1111 | the sharding process is done by the client itself. | 17:46 |
| kfox1111 | and somewhat the mons. | 17:46 |
| sbezverk | kfox1111: right but figuring out where they are located, takes cpu cycles | 17:46 |
| kfox1111 | yeah. thats where the crushmap comes in. | 17:46 |
| kfox1111 | its a multi level hashing mechanism. | 17:46 |
| kfox1111 | very cleaver, and very fast. | 17:46 |
| kfox1111 | thats cephs real claim to fame. | 17:47 |
| sbezverk | kfox1111: I see, interesting.. | 17:47 |
| kfox1111 | so it basically lets you scale out your storage io and disk space the same way hpc scales out computation. | 17:47 |
| *** kbaegis has quit IRC | 17:48 | |
| *** sudipto_ has joined #openstack-kolla | 17:52 | |
| *** sudipto has joined #openstack-kolla | 17:52 | |
| *** kbaegis has joined #openstack-kolla | 17:53 | |
| sbezverk | kfox1111: horizon check failed duonghq PS | 17:53 |
| sbezverk | I do not see much logging generated by that curl | 17:54 |
| kfox1111 | sec... | 17:55 |
| sbezverk | kfox1111: but the url looks good to me.. | 17:56 |
| *** kbaegis has quit IRC | 17:57 | |
| kfox1111 | yeah. if I had to guess, I'd say kubernetes-entrypoint blocked it for a little bit? | 17:58 |
| kfox1111 | the horizon check doesn't currently check to see if its ready yet. | 17:59 |
| kfox1111 | so the added delay by the dependencies is not accounted for by the horizon test. | 17:59 |
| kfox1111 | probably need a check loop. | 18:00 |
| *** goldyfruit has quit IRC | 18:00 | |
| sbezverk | kfox1111: horizon check has been merged? | 18:04 |
| sbezverk | nvm my master was not up to date | 18:06 |
| *** sudipto has quit IRC | 18:06 | |
| *** sudipto_ has quit IRC | 18:06 | |
| kfox1111 | sbezverk: I just posted a bit of code to the review I think will fix it. | 18:09 |
| *** sacharya has joined #openstack-kolla | 18:09 | |
| sbezverk | kfox1111: looks good | 18:11 |
| sbezverk | maybe you should do it in a separate PS | 18:11 |
| sbezverk | then we can try to merge and re-check | 18:12 |
| kfox1111 | yeah. either way. | 18:12 |
| sbezverk | otherwise will have to wait at least 12 hours | 18:12 |
| kfox1111 | do you have time to do it? I'm trying to help portdirect/sdake with getting the dev env fixed up a bit. | 18:12 |
| sbezverk | kfox1111: actually no rush. let's wait | 18:13 |
| kfox1111 | ok. | 18:13 |
| sdake | yo folks - jsut got done doing the copy phase of the critical files for my mother's computer | 18:28 |
| sdake | going to eat - since i had no food yesterday at all | 18:28 |
| sdake | when i'm back, i should be ready to rock on testing out the dev env and getting it rolling | 18:29 |
| *** sp___ has quit IRC | 18:29 | |
| *** kbaegis has joined #openstack-kolla | 18:31 | |
| kfox1111 | sdake: I've got the tweaked ps just about ready to post. | 18:31 |
| *** sdake_ has joined #openstack-kolla | 18:31 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Add Development Setup Script for use with Kubernetes Helm https://review.openstack.org/405720 | 18:33 |
| *** sdake has quit IRC | 18:34 | |
| *** williamcaban has joined #openstack-kolla | 18:35 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Add Development Setup Script for use with Kubernetes Helm https://review.openstack.org/405720 | 18:36 |
| *** dave-mccowan has joined #openstack-kolla | 18:39 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 18:39 |
| *** williamcaban has quit IRC | 18:39 | |
| *** williamcaban has joined #openstack-kolla | 18:40 | |
| *** trinaths has quit IRC | 18:47 | |
| *** wirehead_ has quit IRC | 18:51 | |
| *** wirehead_ has joined #openstack-kolla | 18:52 | |
| *** sdake_ has quit IRC | 19:00 | |
| *** v1k0d3n has quit IRC | 19:01 | |
| *** sdake has joined #openstack-kolla | 19:02 | |
| *** wirehead_ has quit IRC | 19:03 | |
| *** wirehead_ has joined #openstack-kolla | 19:04 | |
| *** sdake has quit IRC | 19:10 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 19:13 |
| *** sdake has joined #openstack-kolla | 19:18 | |
| *** wirehead_ has quit IRC | 19:19 | |
| *** wirehead_ has joined #openstack-kolla | 19:19 | |
| kfox1111 | gota head out for a few. bbiab. | 19:20 |
| *** wirehead_ has quit IRC | 19:22 | |
| *** wirehead_ has joined #openstack-kolla | 19:23 | |
| *** v1k0d3n has joined #openstack-kolla | 19:24 | |
| *** wirehead_ has quit IRC | 19:27 | |
| *** wirehead_ has joined #openstack-kolla | 19:28 | |
| *** wirehead_ has quit IRC | 19:32 | |
| *** wirehead_ has joined #openstack-kolla | 19:33 | |
| *** goldyfruit has joined #openstack-kolla | 19:33 | |
| *** wirehead_ has quit IRC | 19:37 | |
| *** wirehead_ has joined #openstack-kolla | 19:38 | |
| *** lamt has joined #openstack-kolla | 19:39 | |
| *** kbaegis has quit IRC | 19:40 | |
| *** v1k0d3n has quit IRC | 19:42 | |
| *** v1k0d3n has joined #openstack-kolla | 19:43 | |
| *** v1k0d3n has quit IRC | 19:43 | |
| *** Syffs has quit IRC | 19:49 | |
| *** lamt has quit IRC | 19:53 | |
| *** duonghq__ has quit IRC | 19:54 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 19:59 |
| *** dave-mccowan has quit IRC | 20:07 | |
| *** wirehead_ has quit IRC | 20:11 | |
| *** apuimedo has quit IRC | 20:12 | |
| *** wirehead_ has joined #openstack-kolla | 20:12 | |
| *** Serlex has joined #openstack-kolla | 20:12 | |
| *** apuimedo has joined #openstack-kolla | 20:13 | |
| *** apuimedo has quit IRC | 20:20 | |
| *** apuimedo has joined #openstack-kolla | 20:20 | |
| *** apuimedo has quit IRC | 20:25 | |
| *** v1k0d3n has joined #openstack-kolla | 20:25 | |
| *** v1k0d3n has quit IRC | 20:25 | |
| *** v1k0d3n has joined #openstack-kolla | 20:26 | |
| *** apuimedo has joined #openstack-kolla | 20:27 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 20:29 |
| *** apuimedo has quit IRC | 20:31 | |
| *** apuimedo has joined #openstack-kolla | 20:33 | |
| *** kbaegis has joined #openstack-kolla | 20:34 | |
| kfox1111 | sdake: https://review.openstack.org/#/c/422950/ ? | 20:34 |
| *** sdake has quit IRC | 20:36 | |
| *** apuimedo has quit IRC | 20:38 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP - local registry in gates https://review.openstack.org/418550 | 20:38 |
| *** kbaegis has quit IRC | 20:38 | |
| *** goldyfruit has quit IRC | 20:39 | |
| *** apuimedo has joined #openstack-kolla | 20:40 | |
| *** kbaegis has joined #openstack-kolla | 20:43 | |
| *** apuimedo has quit IRC | 20:44 | |
| *** apuimedo has joined #openstack-kolla | 20:46 | |
| *** schwicht has joined #openstack-kolla | 20:47 | |
| *** lamt has joined #openstack-kolla | 20:48 | |
| *** apuimedo has quit IRC | 20:51 | |
| *** apuimedo has joined #openstack-kolla | 20:54 | |
| *** severion has joined #openstack-kolla | 20:55 | |
| *** severion has quit IRC | 20:56 | |
| *** apuimedo has quit IRC | 20:58 | |
| *** severion has joined #openstack-kolla | 20:59 | |
| *** v1k0d3n has quit IRC | 20:59 | |
| *** severion has quit IRC | 20:59 | |
| *** v1k0d3n has joined #openstack-kolla | 20:59 | |
| *** apuimedo has joined #openstack-kolla | 21:01 | |
| *** apuimedo has quit IRC | 21:08 | |
| *** sacharya has quit IRC | 21:08 | |
| *** apuimedo has joined #openstack-kolla | 21:09 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 21:09 |
| *** apuimedo has quit IRC | 21:13 | |
| *** apuimedo has joined #openstack-kolla | 21:16 | |
| *** apuimedo has quit IRC | 21:21 | |
| *** apuimedo has joined #openstack-kolla | 21:24 | |
| *** nihilifer has quit IRC | 21:27 | |
| *** nihilifer has joined #openstack-kolla | 21:28 | |
| *** apuimedo has quit IRC | 21:31 | |
| *** l4yerffeJ_ has joined #openstack-kolla | 21:34 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 21:34 |
| *** apuimedo has joined #openstack-kolla | 21:34 | |
| *** Jeffrey4l_ has quit IRC | 21:34 | |
| *** l4yerffeJ has quit IRC | 21:35 | |
| *** Jeffrey4l_ has joined #openstack-kolla | 21:35 | |
| *** sdake has joined #openstack-kolla | 21:46 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 22:05 |
| *** mgoddard has joined #openstack-kolla | 22:07 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP - local registry in gates https://review.openstack.org/418550 | 22:17 |
| *** kbaegis has quit IRC | 22:23 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: Support rolling upgrades of horizon https://review.openstack.org/333996 | 22:28 |
| *** mgoddard has quit IRC | 22:32 | |
| *** kbaegis has joined #openstack-kolla | 22:33 | |
| *** Serlex has quit IRC | 22:37 | |
| *** salv-orlando has joined #openstack-kolla | 22:38 | |
| *** sdake has quit IRC | 22:40 | |
| *** salv-orlando has quit IRC | 22:47 | |
| *** kbaegis has quit IRC | 22:53 | |
| *** kbaegis has joined #openstack-kolla | 22:54 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP - local registry in gates https://review.openstack.org/418550 | 22:56 |
| *** masber has quit IRC | 23:03 | |
| *** masber has joined #openstack-kolla | 23:03 | |
| *** saneax-_-|AFK is now known as saneax | 23:04 | |
| *** masber has quit IRC | 23:06 | |
| *** williamcaban has quit IRC | 23:06 | |
| *** masber has joined #openstack-kolla | 23:07 | |
| *** williamcaban has joined #openstack-kolla | 23:11 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Support rolling upgrades of horizon https://review.openstack.org/333996 | 23:13 |
| *** jrobinson has joined #openstack-kolla | 23:15 | |
| *** williamcaban has quit IRC | 23:16 | |
| *** sdake has joined #openstack-kolla | 23:22 | |
| openstackgerrit | Dan Prince proposed openstack/kolla: Allow mariadb log directory to be configurable https://review.openstack.org/423868 | 23:22 |
| *** kbaegis has quit IRC | 23:25 | |
| *** williamcaban has joined #openstack-kolla | 23:29 | |
| *** lamt has quit IRC | 23:31 | |
| *** williamcaban has quit IRC | 23:34 | |
| sdake | Jeffrey4l_ arey ou about | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!