*** jemcevoy has joined #openstack-kolla | 00:00 | |
*** Pavo has quit IRC | 00:00 | |
*** Pavo has joined #openstack-kolla | 00:02 | |
*** sdake has joined #openstack-kolla | 00:04 | |
jemcevoy | inc0 are you there? | 00:04 |
---|---|---|
*** kristian__ has joined #openstack-kolla | 00:08 | |
*** kristian__ has quit IRC | 00:09 | |
*** kristia__ has joined #openstack-kolla | 00:09 | |
*** goldyfruit has quit IRC | 00:12 | |
inc0 | jemcevoy, now I am | 00:13 |
*** kristia__ has quit IRC | 00:13 | |
jemcevoy | inc0, Are you attending the kube road show next Tuesday at 2pm to 5pm | 00:18 |
inc0 | yeah | 00:19 |
inc0 | see you there I assume?:) | 00:19 |
jemcevoy | It is a ten minute walk from my place if you wanna park here... | 00:19 |
inc0 | good to know:) | 00:20 |
inc0 | we'll probably carpool with jascott1 | 00:21 |
*** hfu has joined #openstack-kolla | 00:21 | |
inc0 | jascott1, can we use your car? we won't fit into my... | 00:21 |
jascott1 | no convertible ride? :) | 00:21 |
jascott1 | ya np, i might even clean it | 00:21 |
jemcevoy | I would need your license plate number to get a parking permit | 00:21 |
inc0 | no, unless you and qwang decide who runs in trunk | 00:21 |
sbezverk | kfox1111: ping | 00:21 |
inc0 | and this is not a overly large trunk | 00:22 |
kfox1111 | sbezverk: hey. whats up? | 00:22 |
jascott1 | jemcevoy i sent you PM thanks | 00:22 |
sbezverk | kfox1111: I am seeing one error and so far I failed to undersatnd where it is coming from: check this log: http://logs.openstack.org/50/428150/14/check/gate-kolla-kubernetes-deploy-centos-binary-2-helm-entrypoint-nv/60565a6/console.html#_2017-02-03_23_33_17_591530 | 00:23 |
*** emispur has quit IRC | 00:23 | |
qwang | inc0: I can't fit into a trunk ;( | 00:23 |
inc0 | yeah, I figured...neither can jascott1 | 00:24 |
sbezverk | kfox1111: it says that rabbitmq service already exist, but it is not at least not in kolla namespace. | 00:24 |
jascott1 | inc0 there is another trunk option :) | 00:24 |
jascott1 | ive already driven it :) | 00:25 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding compute-kit chart https://review.openstack.org/428150 | 00:25 |
kfox1111 | sbezverk: http://logs.openstack.org/50/428150/14/check/gate-kolla-kubernetes-deploy-centos-binary-2-helm-entrypoint-nv/60565a6/logs/svc/kolla-rabbitmq.txt | 00:25 |
kfox1111 | seems to exist | 00:25 |
inc0 | jascott1, we could hook this one https://www.thule.com/en-us/us/strollers/multisport-trailers/thule-chariot-cougar-1-_-10100534 | 00:25 |
inc0 | that would make everyone riding in convertiblew | 00:25 |
qwang | inc0: jascott1 can we take max | 00:26 |
jascott1 | <12 hours later> | 00:26 |
inc0 | nah, I'm just kidding | 00:26 |
sbezverk | kfox1111: but if you check the log above get svc -n kolla return no object found | 00:27 |
inc0 | we'll carpool | 00:27 |
inc0 | just...not with my car;) | 00:27 |
jascott1 | gotta get ready for musical festivities.. everyone have a great weekend | 00:27 |
inc0 | have fun man | 00:27 |
jascott1 | all are welcome! | 00:28 |
inc0 | I have my own musical festivities planned for today;) | 00:29 |
kfox1111 | looking.... | 00:32 |
kfox1111 | sbezverk: thats kind of weird, yeah.... | 00:33 |
*** dave-mcc_ has joined #openstack-kolla | 00:37 | |
*** dave-mccowan has quit IRC | 00:37 | |
kfox1111 | sbezverk: kind of at a loss. | 00:38 |
kfox1111 | helm bug maybe? | 00:38 |
*** lamt has quit IRC | 00:39 | |
kfox1111 | jascott1: could you look at sbezverk's issue? | 00:40 |
*** lamt has joined #openstack-kolla | 00:41 | |
kfox1111 | it looks like no services exist, then maybe get created as part of the release but then its failing to release. | 00:41 |
*** saneax is now known as saneax-_-|AFK | 00:42 | |
*** sbezverk has quit IRC | 00:42 | |
*** dave-mcc_ has quit IRC | 00:43 | |
*** salv-orlando has quit IRC | 00:43 | |
*** sbezverk has joined #openstack-kolla | 00:44 | |
*** lamt has quit IRC | 00:45 | |
sbezverk | kfox1111: more complex things are getting, more issues we see :( | 00:47 |
*** dave-mccowan has joined #openstack-kolla | 00:51 | |
kfox1111 | sbezverk: not unexpected from pretty new system. :/ | 00:53 |
kfox1111 | the whole single package thing was just an idea to see how far we could push helm. not sure it was going to work. | 00:54 |
kfox1111 | would be pretty neat if it worked, but not sure how useful it would be to normal users anyway. | 00:54 |
*** mattmceuen has joined #openstack-kolla | 00:55 | |
kfox1111 | but interesting to give to them anyway so they can fix it. | 00:55 |
kfox1111 | they will see things like that eventually from other apps then just openstack too. | 00:55 |
*** zhurong has joined #openstack-kolla | 00:55 | |
kfox1111 | sbezverk: | 00:57 |
kfox1111 | oh, just a thought... | 00:57 |
kfox1111 | this may be a different issue though. | 00:57 |
bmace_ | sdake: kfox1111: fyi, got almost all the openstack services running (at least the pods show as running) in the new dev setup container. had to step through some of the wait areas manually but looking great for early next week for a review. looks to be just a new docker file and a couple of new scripts in tools. | 00:57 |
kfox1111 | sbezverk: so, in the code, your creating a full blown openstack, then deleting it , then adding a new one. there still is a chance the workflow itself is messing with things. | 00:57 |
kfox1111 | sbezverk: relooking at the code, it looks like delete is async. if thats true, then the wait_for_pod_termination kicks in, but there isn't a wait_for_svc_termination? | 00:58 |
sbezverk | kfox1111: actually I found a broblem :( | 00:58 |
kfox1111 | bmace_: cool. :) | 00:58 |
sbezverk | I run helm with dry-run and it looks so wierd | 00:58 |
bmace_ | sbezverk: you hitting that in helm? because i hit that a couple of times, or similar, when i did a delete without --purge.. doesn't show up in list but can't create a new one. | 00:59 |
*** huzhengchuan has joined #openstack-kolla | 00:59 | |
sbezverk | bmace_: no it is not that | 00:59 |
sbezverk | kfox1111: check the latest entry point job logm by some reason multiple copies of rabbitmq svc gets rendered | 01:00 |
kfox1111 | k. sec | 01:00 |
*** dave-mccowan has quit IRC | 01:03 | |
*** rwallner has joined #openstack-kolla | 01:03 | |
kfox1111 | sbezverk: I see two rabbit svc's, but only due to there being a rabbit-svc and rabbit-mgmt svc. | 01:07 |
kfox1111 | that svc chart's a little weird. | 01:07 |
*** rwallner has quit IRC | 01:07 | |
bmace_ | kfox1111: is there something that horizon waits on before it gets an external kubernetes ip? my horizon pod seems to be running ok. don't see any errors in it, but i don't get any external-ip for it in the kubectl get services output.. all the other api pods look to have appropriate external ips | 01:08 |
kfox1111 | might have to configure it to export it. | 01:08 |
kfox1111 | yeah, looks like port_external is defaulting to false on it. | 01:09 |
sbezverk | kfox1111: check this out http://paste.openstack.org/show/597657/ | 01:10 |
sbezverk | I think it is element name issue :-) | 01:10 |
sbezverk | by some reason mariadb uses rabbitmq name | 01:10 |
kfox1111 | oh.... | 01:10 |
*** unicell has quit IRC | 01:11 | |
bmace_ | kfox1111: kk, was following the flow in ceph_workflow_service.sh and figured horizon would be exposed since i know you were recently adding horizon health tests into the gate. | 01:11 |
kfox1111 | this is a case where the values in the global tree might be overlapping. | 01:11 |
kfox1111 | we never see it in service deployments, but when all combined at the same time, they conflict. | 01:11 |
kfox1111 | bmace_: yeah, we may want to change that default. | 01:11 |
kfox1111 | but for now, looks like its internal only. | 01:12 |
kfox1111 | by default. | 01:12 |
*** eaguilar has quit IRC | 01:12 | |
sbezverk | kfox1111: I have computed values will have to find conflicting entries | 01:12 |
openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla-ansible master: Add release note regarding nova upgrade https://review.openstack.org/429138 | 01:12 |
kfox1111 | sbezverk: ah, yea,h, its element name related... | 01:12 |
kfox1111 | putting element_name all the way at the top makes sence for service packages, but not a compute kit. | 01:13 |
kfox1111 | sbezverk: I think, in the values.yaml for the compute kit, we just neeed to set the element_name: mariadb down a bit further | 01:13 |
kfox1111 | global.kolla.mariadb.all: element_name: mariadb | 01:13 |
kfox1111 | for rabbit, mariadb, and memcached. | 01:14 |
sbezverk | kfox1111: Yep, I think so | 01:15 |
sbezverk | I will edit values.yaml and add these values | 01:16 |
kfox1111 | cool. | 01:17 |
*** huzhengchuan has quit IRC | 01:19 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding compute-kit chart https://review.openstack.org/428150 | 01:20 |
sbezverk | kfox1111: let's see if it helps. I bet sdake will love this chart ;) whole opensatck with one shot | 01:21 |
sdake | sbezverk it is actually inc0 that will love that chart | 01:21 |
sdake | ;-) | 01:21 |
kfox1111 | sbezverk: +1 | 01:22 |
kfox1111 | sbezverk: plus one more idea. | 01:22 |
kfox1111 | I'd like to see genconfig ran on the dev env, imported into the k8s, | 01:22 |
kfox1111 | then pull the configmaps back out and dumped into helm/tests/devenv-configmaps/templates/* | 01:23 |
sbezverk | kfox111: I hope you do not mean to run ansible genconfig | 01:23 |
*** rwallner has joined #openstack-kolla | 01:23 | |
kfox1111 | we could then for the dev env, just load up the devenv-configmaps package, and the compute-kit package, and should be most of the way there. :) | 01:23 |
kfox1111 | we can then discuss how to make those cnfnigmaps more templated/split up at the ptg? | 01:24 |
kfox1111 | it would only work for the dev env, but would be a good way to kickstart the dev env? | 01:24 |
sbezverk | kfox1111: sure, it is long discussion for sure | 01:24 |
kfox1111 | bmace_: what do you think of that idea? | 01:25 |
inc0 | helm install kolla/openstack?:) | 01:26 |
* inc0 approves | 01:26 | |
kfox1111 | inc0: yeah. | 01:27 |
kfox1111 | all of compute kit at a shot. :) | 01:27 |
*** rwallner has quit IRC | 01:28 | |
sbezverk | inc0: helm install helm/compute-kits/compute-kit-0.4.0 --namespace kolla | 01:28 |
inc0 | all the product manager types would love that - and that drives adoption | 01:28 |
kfox1111 | inc0: not sure how useful it is for prod. | 01:28 |
inc0 | kfox1111, kolla-ansible deploy is pretty damn useful;) | 01:28 |
kfox1111 | as upgrading will be really oneshot/not much opertunity for upgrade orchestration. | 01:28 |
kfox1111 | but for testing, pretty sweet. | 01:28 |
sbezverk | kfox1111: I think for aio type of installation or lab, it will be super | 01:28 |
inc0 | but yeah, in simple scenerio, not that much | 01:29 |
kfox1111 | +1 | 01:29 |
*** adrian_otto has quit IRC | 01:29 | |
kfox1111 | inc0: agreed kolla-ansible deploy is pretty sweet. but not sure helm yet can do that in one shot with its orchestration primitives. | 01:29 |
inc0 | if we can get to same level of simplicity as kolla-ansible | 01:29 |
inc0 | we win | 01:29 |
kfox1111 | totally agree. but thats asking a lot of helm. | 01:30 |
inc0 | maybe not today | 01:30 |
inc0 | but if we know what is missing | 01:30 |
kfox1111 | +1 | 01:30 |
inc0 | we can always fill the gaps | 01:30 |
inc0 | it's opensource | 01:30 |
kfox1111 | so, playing with prometheus a bit. very interesting. | 01:31 |
inc0 | here's the deal, we need something that can be demoed to not-so-technical project managers too;) | 01:31 |
kfox1111 | totally agree. | 01:31 |
*** huzhengchuan has joined #openstack-kolla | 01:31 | |
inc0 | is it? one of things I wanted to add to kolla-ansible next release is monitoring done right | 01:31 |
kfox1111 | they seem to have dynamic monitoring down in a way that makes sence. | 01:32 |
kfox1111 | I really liked ansible for its support for plugable dynamic inventory. | 01:32 |
inc0 | I had some discussion about prometheus today with snap author | 01:32 |
kfox1111 | nagios' and friends have been a huge pain due to its static inventory assumptions. | 01:32 |
kfox1111 | but they have a dynamic inventory plugin for k8s. :) | 01:33 |
kfox1111 | so it can discover services/pods. | 01:33 |
inc0 | but it won't really do host-level monitoring | 01:33 |
kfox1111 | it can do that too. | 01:33 |
inc0 | we'll see anyway | 01:33 |
kfox1111 | a daemonset with the host level monitoring svc, | 01:33 |
inc0 | I want to have kickass diagnostic in kolla really | 01:34 |
kfox1111 | and maybe the dynamic service can pull from kube node list. | 01:34 |
kfox1111 | +1 | 01:34 |
inc0 | we did good job with log mgmt | 01:34 |
kfox1111 | kubernetes makes having monitoring much more important. | 01:34 |
inc0 | add monitoring | 01:34 |
Pavo | have you guys seen ansible -containers? | 01:34 |
kfox1111 | for logs and for checks. | 01:35 |
inc0 | Pavo, what do you mean? | 01:35 |
kfox1111 | Pavo: yeah. :/ | 01:35 |
Pavo | the project ansible containers | 01:35 |
Pavo | it seems pretty sweet | 01:35 |
kfox1111 | inc: ansible wants to do container orchestration themselves. :/ | 01:36 |
inc0 | ahh...yet another docker swarm? | 01:36 |
Pavo | kind of | 01:36 |
inc0 | eh. | 01:36 |
kfox1111 | yeah. :/ | 01:37 |
inc0 | if you ask me, that's footshot | 01:37 |
inc0 | they'd do better if they'd provide kick ass modules for k8s | 01:37 |
Pavo | it kind of does the same thing as kolla-build but instead of building using Dokcerfile or bash scripts inside the container it uses ansible playbooks to configure the container | 01:37 |
inc0 | and try to provide alternative to helm this way | 01:37 |
inc0 | ahh so it's container management and building? | 01:38 |
inc0 | duh | 01:38 |
inc0 | anyway, I gtg | 01:38 |
inc0 | talk to you guys next week | 01:38 |
kfox1111 | it does that too. | 01:38 |
Pavo | seeya | 01:38 |
sdake | kfox1111 your client is busted yet again | 01:38 |
inc0 | I'll read about it a bit:) thansk Pavo for pointing that | 01:38 |
kfox1111 | its both a build thing, but an orchestration too. | 01:38 |
inc0 | duh | 01:38 |
kfox1111 | sdake: your vpn's busted again. :) | 01:38 |
Pavo | yeah | 01:38 |
*** inc0 has quit IRC | 01:39 | |
sdake | kfox1111 its continously busted | 01:39 |
sbezverk | sdake: vpn for last couple of days really sucked | 01:40 |
sdake | sbezverk ya - i dunno what the prolem is - anyconnect is a pile of steam | 01:40 |
sdake | and the cvo is too slow | 01:41 |
sdake | can't win - i have to pick the less damned solution which at the moemnt is anyconnect | 01:41 |
sbezverk | sdake: yeah they rate limit traffic :( | 01:41 |
sdake | sbezverk i got the rate limiting turned off | 01:41 |
sdake | the hardware however is slow | 01:42 |
bmace_ | kfox1111: i think we chatted about the static configs the other day. i don't have a big problem with them if you think they are super useful.. but for dev i think it is also useful to generate them dynamically from the active repos, so right now you can tweak things, spin back up the dev container and see them reflected in your environment. | 01:42 |
sdake | bmace_ i think static configs would be a good place to start prior to ptg | 01:42 |
sdake | bmace_ any chance you can make that happen | 01:42 |
sdake | I'll maek the docs happen in return ;) | 01:43 |
bmace_ | kfox1111: right now it makes an assumption that you have kolla-kubernetes, kolla-ansible and halcyon-vagrant-kubernetes checked out into ~devel/ and so it is pretty easy to test changes in any of those repos against a full bring up. | 01:43 |
sdake | to get us off of ansible entirely or kolla-kubernetes | 01:43 |
bmace_ | sdake: i just need to work on some of the wait logic and add some variables into the ceph_service_workflow.sh script and should be good to put my wip review up on monday or tuesday. part of it is a genconfig. | 01:44 |
bmace_ | i do have a small tweak to genconfig in my environment because the head of kolla-ansible has a broken genconfig if you do it in an environment other than where you are running thanks to the fluentd changes. | 01:45 |
rhallisey | busy Firday night | 01:45 |
sdake | bmace_ guesss we need toget that fixed | 01:45 |
bmace_ | so i had to i think comment out one of the lines related to fluentd to get genconfig to run clean | 01:45 |
sdake | rhallisey agreed - i'm gettting ready to ptfo :) | 01:45 |
jemcevoy | g'night | 01:45 |
*** jemcevoy has quit IRC | 01:45 | |
*** sbezverk has quit IRC | 01:45 | |
bmace_ | sdake: yup. i planned to file a bug on it once i had my wip review up. | 01:46 |
*** zhurong has quit IRC | 01:48 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding compute-kit chart https://review.openstack.org/428150 | 01:48 |
*** sbezverk has joined #openstack-kolla | 01:49 | |
*** rhallisey has quit IRC | 01:50 | |
*** rhallisey has joined #openstack-kolla | 01:51 | |
openstackgerrit | qiankun.zhang proposed openstack/kolla-ansible master: Enable tacker_horizon when enable_tacker is yes https://review.openstack.org/407822 | 01:54 |
*** huzhengchuan has quit IRC | 01:57 | |
*** zhurong has joined #openstack-kolla | 01:58 | |
*** kristian__ has joined #openstack-kolla | 01:58 | |
*** Pavo has quit IRC | 02:01 | |
*** Pavo has joined #openstack-kolla | 02:02 | |
*** kristian__ has quit IRC | 02:03 | |
*** 32NAAEFGJ has joined #openstack-kolla | 02:03 | |
*** sdake has quit IRC | 02:07 | |
kfox1111 | bmace_: I submitted a fix for the fluentd thing in trunk yesterday? it merged? | 02:10 |
kfox1111 | should be better now. | 02:10 |
*** huzhengchuan has joined #openstack-kolla | 02:11 | |
*** sdake has joined #openstack-kolla | 02:12 | |
*** rhallisey has quit IRC | 02:14 | |
sdake | kfox1111 com-puter died of power failure | 02:17 |
kfox1111 | bummer | 02:19 |
bmace_ | kfox1111: ah, that is great to hear. i haven't updated my kolla-ansible bits since then. | 02:20 |
kfox1111 | yeah. trunk was pinned back to avoid bugs like that. | 02:20 |
kfox1111 | but I removed the pin to test what the final release of kolla-ansible would do. | 02:20 |
kfox1111 | hit 2 tiny bugs. | 02:20 |
kfox1111 | so fixed them. :) | 02:20 |
*** crushil has joined #openstack-kolla | 02:29 | |
*** rhallisey has joined #openstack-kolla | 02:30 | |
*** rhallisey has quit IRC | 02:32 | |
*** huzhengchuan has quit IRC | 02:41 | |
*** tyrola has quit IRC | 02:42 | |
*** v1k0d3n has quit IRC | 02:44 | |
*** tyrola has joined #openstack-kolla | 02:44 | |
*** zhurong has quit IRC | 02:45 | |
*** v1k0d3n has joined #openstack-kolla | 02:46 | |
*** v1k0d3n has quit IRC | 02:46 | |
*** v1k0d3n has joined #openstack-kolla | 02:47 | |
*** zhurong has joined #openstack-kolla | 02:48 | |
*** furlongm has joined #openstack-kolla | 02:49 | |
*** huzhengchuan has joined #openstack-kolla | 02:54 | |
*** Pavo has quit IRC | 03:09 | |
sdake | is qwang about? | 03:11 |
*** Pavo has joined #openstack-kolla | 03:12 | |
*** sdake_ has joined #openstack-kolla | 03:23 | |
*** Pavo has quit IRC | 03:25 | |
*** sdake has quit IRC | 03:27 | |
*** lamt has joined #openstack-kolla | 03:33 | |
*** kristian__ has joined #openstack-kolla | 03:44 | |
*** sdake has joined #openstack-kolla | 03:44 | |
*** sdake_ has quit IRC | 03:45 | |
*** sudipto_ has joined #openstack-kolla | 03:46 | |
*** sudipto has joined #openstack-kolla | 03:46 | |
*** kristian__ has quit IRC | 03:48 | |
*** jrich523 has quit IRC | 03:52 | |
*** MasterOfBugs has quit IRC | 03:53 | |
*** MasterOfBugs has joined #openstack-kolla | 03:53 | |
*** pramodrj07 has joined #openstack-kolla | 03:54 | |
*** MasterOfBugs has quit IRC | 03:54 | |
*** furlongm has quit IRC | 03:56 | |
*** furlongm has joined #openstack-kolla | 03:57 | |
*** furlongm has quit IRC | 03:58 | |
*** zhurong has quit IRC | 03:59 | |
*** furlongm has joined #openstack-kolla | 03:59 | |
*** pramodrj07 has quit IRC | 03:59 | |
*** jpstnh has joined #openstack-kolla | 04:00 | |
openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible master: Update the CONTRIBUTING.rst file https://review.openstack.org/429176 | 04:05 |
*** MasterOfBugs has joined #openstack-kolla | 04:12 | |
qwang | sdake: what's up? | 04:13 |
*** lamt has quit IRC | 04:19 | |
*** sdake_ has joined #openstack-kolla | 04:20 | |
*** sdake has quit IRC | 04:23 | |
*** rwallner has joined #openstack-kolla | 04:24 | |
*** rwallner has quit IRC | 04:29 | |
*** MasterOfBugs has quit IRC | 04:29 | |
*** pramodrj07 has joined #openstack-kolla | 04:29 | |
*** pramodrj07 has quit IRC | 04:30 | |
*** crushil has quit IRC | 04:33 | |
*** sdake_ has quit IRC | 04:34 | |
*** rwallner has joined #openstack-kolla | 04:43 | |
*** rwallner has quit IRC | 04:48 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: [wip] test scenario2 in gate https://review.openstack.org/412279 | 05:01 |
*** jpstnh has quit IRC | 05:08 | |
*** harlowja has quit IRC | 05:08 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Fix murano issue in horizon container https://review.openstack.org/429183 | 05:11 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: [wip] test scenario2 in gate https://review.openstack.org/412279 | 05:13 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding compute-kit chart https://review.openstack.org/428150 | 05:15 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Add Loading nfs_shares file without .j2 extension in cinder config https://review.openstack.org/411837 | 05:16 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Fluentd should not chown all files in /var/log/kolla https://review.openstack.org/427253 | 05:16 |
*** kristian__ has joined #openstack-kolla | 05:34 | |
*** kristian__ has quit IRC | 05:34 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfigure action for trove https://review.openstack.org/423990 | 05:43 |
*** rwallner has joined #openstack-kolla | 05:44 | |
*** sgordon has quit IRC | 05:49 | |
*** rwallner has quit IRC | 05:49 | |
*** sgordon has joined #openstack-kolla | 05:54 | |
*** kristian__ has joined #openstack-kolla | 06:00 | |
*** kristian__ has quit IRC | 06:00 | |
*** kristian__ has joined #openstack-kolla | 06:01 | |
*** kristian__ has quit IRC | 06:05 | |
openstackgerrit | Merged openstack/kolla-ansible master: Create a new container when restart a container when necessary https://review.openstack.org/427036 | 06:09 |
openstackgerrit | Merged openstack/kolla-ansible master: Apply new register method https://review.openstack.org/428717 | 06:16 |
*** harlowja has joined #openstack-kolla | 06:17 | |
*** salv-orlando has joined #openstack-kolla | 06:18 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: [wip] test scenario2 in gate https://review.openstack.org/412279 | 06:22 |
*** salv-orlando has quit IRC | 06:22 | |
*** sudipto has quit IRC | 06:33 | |
*** sudipto_ has quit IRC | 06:33 | |
*** salv-orlando has joined #openstack-kolla | 06:42 | |
*** kristian__ has joined #openstack-kolla | 06:43 | |
*** crushil has joined #openstack-kolla | 06:43 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfigure action for congress https://review.openstack.org/422690 | 06:44 |
*** haplo37_ has quit IRC | 06:45 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Remove unnecessary until in register.yml file https://review.openstack.org/428719 | 06:45 |
*** kristian__ has quit IRC | 06:47 | |
*** sudipto has joined #openstack-kolla | 06:50 | |
*** sudipto_ has joined #openstack-kolla | 06:50 | |
*** harlowja has quit IRC | 06:50 | |
*** kristian__ has joined #openstack-kolla | 06:51 | |
*** kristian__ has quit IRC | 06:56 | |
*** haplo37_ has joined #openstack-kolla | 06:56 | |
*** v1k0d3n has quit IRC | 06:56 | |
*** sgordon has quit IRC | 07:03 | |
*** sgordon has joined #openstack-kolla | 07:08 | |
*** sudipto has quit IRC | 07:11 | |
*** sudipto_ has quit IRC | 07:11 | |
*** sgordon has quit IRC | 07:20 | |
*** bmace_ has quit IRC | 07:21 | |
*** bmace_ has joined #openstack-kolla | 07:21 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Fix trove dashboard issue in horizon container https://review.openstack.org/429183 | 07:27 |
*** sgordon has joined #openstack-kolla | 07:28 | |
32NAAEFGJ | ping sdake | 07:33 |
*** 32NAAEFGJ has quit IRC | 07:33 | |
*** zhubingbing has joined #openstack-kolla | 07:34 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Fix parameter issue in cinder handlers https://review.openstack.org/427268 | 07:39 |
*** sudipto_ has joined #openstack-kolla | 07:39 | |
*** sudipto has joined #openstack-kolla | 07:39 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: [wip] test scenario2 in gate https://review.openstack.org/412279 | 07:43 |
*** salv-orl_ has joined #openstack-kolla | 07:49 | |
*** harlowja has joined #openstack-kolla | 07:50 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Fix parameter issue in cinder handlers https://review.openstack.org/427268 | 07:50 |
openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfigure action for congress https://review.openstack.org/422690 | 07:50 |
*** salv-orlando has quit IRC | 07:51 | |
*** sudipto has quit IRC | 07:53 | |
*** sudipto_ has quit IRC | 07:53 | |
*** crushil has quit IRC | 07:53 | |
*** harlowja has quit IRC | 07:57 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Fix parameter issue in cinder handlers https://review.openstack.org/427268 | 07:58 |
*** sdake has joined #openstack-kolla | 08:02 | |
sdake | qwang evening still around? | 08:03 |
*** sgordon has quit IRC | 08:09 | |
*** sgordon has joined #openstack-kolla | 08:16 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for kibana https://review.openstack.org/423158 | 08:27 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Install osprofiler in magnum base container https://review.openstack.org/429205 | 08:27 |
*** kristian__ has joined #openstack-kolla | 08:36 | |
*** kristian__ has quit IRC | 08:44 | |
*** kristian__ has joined #openstack-kolla | 08:48 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Fix trove dashboard issue in horizon container https://review.openstack.org/429183 | 09:01 |
*** rwallner has joined #openstack-kolla | 09:04 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: [wip] test scenario2 in gate https://review.openstack.org/412279 | 09:06 |
*** rwallner has quit IRC | 09:08 | |
*** kristian__ has quit IRC | 09:14 | |
*** Jeffrey4l has quit IRC | 09:20 | |
*** Jeffrey4l has joined #openstack-kolla | 09:26 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Fix the policy file path in horizon container https://review.openstack.org/429242 | 09:39 |
*** mgoddard_ has joined #openstack-kolla | 09:39 | |
*** mannamne has joined #openstack-kolla | 09:40 | |
*** hrw has quit IRC | 09:42 | |
*** hrw has joined #openstack-kolla | 09:43 | |
*** Jeffrey4l has quit IRC | 09:44 | |
*** Jeffrey4l has joined #openstack-kolla | 09:45 | |
*** mannamne has quit IRC | 09:46 | |
*** sdake_ has joined #openstack-kolla | 09:48 | |
*** sdake has quit IRC | 09:50 | |
*** skuda has joined #openstack-kolla | 09:54 | |
*** hfu has quit IRC | 10:01 | |
*** mannamne has joined #openstack-kolla | 10:10 | |
*** mgoddard_ has quit IRC | 10:31 | |
*** zhubingbing has quit IRC | 10:38 | |
*** zhurong has joined #openstack-kolla | 10:38 | |
*** zhubingbing has joined #openstack-kolla | 10:39 | |
*** huzhengchuan has quit IRC | 10:45 | |
*** Mech422 has joined #openstack-kolla | 10:48 | |
Mech422 | Hey guys - anyone had problems with keystone/ldap on newton? keystone is only trying to hit the ldap server like 1 out of every 4-8 queries - the rest just return no data | 10:49 |
Mech422 | hmm - looks like it might be an haproxy issue | 10:58 |
*** huzhengchuan has joined #openstack-kolla | 11:00 | |
*** zhubingbing has quit IRC | 11:18 | |
*** mdnadeem has joined #openstack-kolla | 11:35 | |
*** sdake_ has quit IRC | 11:43 | |
*** hfu has joined #openstack-kolla | 11:53 | |
*** hfu has quit IRC | 11:56 | |
*** pbourke has quit IRC | 12:03 | |
*** pbourke has joined #openstack-kolla | 12:04 | |
*** rhallisey has joined #openstack-kolla | 12:04 | |
*** salv-orl_ has quit IRC | 12:07 | |
*** zhurong has quit IRC | 12:15 | |
*** haplo37 has quit IRC | 12:16 | |
*** haplo37 has joined #openstack-kolla | 12:25 | |
*** jpstnh has joined #openstack-kolla | 12:39 | |
*** jascott1 has quit IRC | 12:41 | |
*** huzhengchuan has quit IRC | 12:54 | |
*** Jeffrey4l has quit IRC | 12:56 | |
*** Jeffrey4l has joined #openstack-kolla | 12:56 | |
*** sdake has joined #openstack-kolla | 13:06 | |
*** schwicht has quit IRC | 13:16 | |
*** hfu has joined #openstack-kolla | 13:23 | |
*** hfu has quit IRC | 13:31 | |
sdake | sbezverk you around | 13:31 |
*** hfu has joined #openstack-kolla | 13:32 | |
*** Pavo has joined #openstack-kolla | 13:36 | |
sdake | kfox1111 are you around | 13:36 |
*** sdake has quit IRC | 13:42 | |
*** sdake has joined #openstack-kolla | 13:42 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding compute-kit chart https://review.openstack.org/428150 | 13:50 |
*** hfu has quit IRC | 13:58 | |
*** hfu has joined #openstack-kolla | 13:59 | |
*** salv-orlando has joined #openstack-kolla | 14:00 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP Adding compute-kit chart https://review.openstack.org/428150 | 14:03 |
sbezverk | sdake: hi, what is up? | 14:03 |
sdake | sbezverk morning - if you ahve a moment to speak via irc about the mappings that need to be done via haproxy or some external load balancer i'd like to get back into that discussion | 14:04 |
sdake | the next thing i want to tackle is multi-node tagged controllers and computes | 14:05 |
sbezverk | sdake: sorry not good timing, I will leave in a short bit.. best I can suggest maybe tonight | 14:05 |
sdake | ok - sounds good -i'll be arond im sure | 14:06 |
sdake | define tonight in et :) | 14:06 |
sdake | plz | 14:06 |
sdake | sbezverk really waht i am after is how to get from 1 compute to many comoputes | 14:06 |
sdake | in my home lab | 14:06 |
sdake | I have ISCSI luns available to me, so that is an option (via my synology nas) | 14:07 |
sbezverk | sdake: 6-8pm est but not for sure.. | 14:07 |
sdake | ok - sounds good | 14:07 |
sdake | can you think about how i would use that synology nas insted of setting up an external ceph cluster | 14:07 |
sdake | one way to think of synology nas is it is a linux box making iscsi luns available | 14:08 |
sdake | since thatis what it does | 14:08 |
sdake | i'm keen to get more then 1 copmute node and 1 master | 14:09 |
sdake | i'd like 1 controller 2 computes + pvs backed via iscsi NAS | 14:09 |
sbezverk | sdake: let me know when you get it working, would be interested to see | 14:10 |
sdake | well I need your help :) | 14:11 |
sdake | thats what i wanted to chat about | 14:11 |
sbezverk | I think it is not doable, but who knows | 14:11 |
sdake | i see | 14:11 |
sdake | how about nfs? | 14:11 |
sdake | (synology also has nfs support) | 14:11 |
sdake | the storage drivers for kubernetes have a variety of storage types | 14:12 |
*** ferrarixx599 has joined #openstack-kolla | 14:12 | |
sdake | what i have a deire to avoid is setting up a ceph cluster on 3 nodes and then setting up a kubernetes cluster no 3 nodes | 14:12 |
sdake | since I only have 3 nodes of hardware in my home lab | 14:12 |
ferrarixx599 | docker IRC channel asking for password? How to register? | 14:13 |
ferrarixx599 | my kolla installation having problems with docker | 14:13 |
sdake | ferrarixx599 which version of docker-engine are you using | 14:13 |
sdake | ferrarixx599 i think 13.0 is broken | 14:14 |
ferrarixx599 | latest. 1.13.0 | 14:14 |
sdake | i think you can get 1.12.6 working | 14:14 |
ferrarixx599 | you mean 1.13 doesn't work with kolla ? | 14:14 |
sdake | i mean 1.13.0 is broken | 14:14 |
sdake | as in DOA | 14:14 |
sdake | however it could be that 1.13.0 works for many people and just not kolla | 14:15 |
ferrarixx599 | the specs given version (in kolla mitaka release) >= 1.10.0 | 14:15 |
sdake | ferrarixx599 yup, we need a 1= 1.13.0 in there | 14:15 |
sdake | I believe its fixed in 1.13.1-rc1+ | 14:15 |
ferrarixx599 | oh okay. | 14:15 |
ferrarixx599 | experimental right? | 14:15 |
sdake | ferrarixx599 moment i'll submit a patch to fix the docs | 14:15 |
ferrarixx599 | cool! | 14:15 |
sdake | ferrarixx599 tbh, I'd just use 1.10.latest z and be done with it | 14:15 |
sdake | kolla doesn't use any newer features of docker | 14:16 |
sdake | 1.10.latest.z is gtg | 14:16 |
ferrarixx599 | thanks. how to docker irc btw? | 14:18 |
Pavo | you have to have a registered nick | 14:21 |
Pavo | ferrarixx599 type /nickserv register email password | 14:22 |
Pavo | then after you verfiy your email type /nickserv identify username password | 14:23 |
ferrarixx599 | thank you! :) | 14:23 |
Pavo | np | 14:23 |
Mech422 | sdake: if your just doing 'lab' stuff - why not just co-habit your ceph cluster with your nodes ? | 14:26 |
*** ferrarixx599 has left #openstack-kolla | 14:26 | |
Mech422 | sdake: especially if its SSD backed storage, its plenty fast enough for testing | 14:26 |
*** ferrarixx599 has joined #openstack-kolla | 14:27 | |
*** ferrarixx599 has left #openstack-kolla | 14:27 | |
openstackgerrit | Steven Dake proposed openstack/kolla-ansible master: Update kolla-ansible docker-version https://review.openstack.org/429277 | 14:27 |
sdake | Mech422 i coudl do that, i really dont want to run ceph alongside it though as its a pain in the ass to setup | 14:27 |
Mech422 | sdake: really? ceph is one of the less painful things for me to setup | 14:29 |
Mech422 | sdake: ceph-deploy has gotten pretty decent | 14:29 |
sdake | Mech422 i dont want to learn how to adminster ceph outside kubernetes hwen i can click 1 button and have an iscsi lun | 14:29 |
Mech422 | sdake: ahh | 14:29 |
sdake | i've got enough on my plate without spending weeks learning the ins and outs of ceph | 14:29 |
*** hfu has quit IRC | 14:30 | |
sdake | i think others will feel the same ;-) | 14:30 |
Mech422 | sdake: heh - well for lab/testing stuff, there's not really much to learn - basically 'Nuke' and 'Install' :-) | 14:30 |
sdake | idea is to get a dev env setup quickly without alot of hassle for devs to get work done | 14:30 |
Mech422 | yeah - I go the other way...I like to make my lab stuff as close to prod as I can (reasonably) achieve | 14:31 |
sdake | i think prod could use nfs and iscsi | 14:31 |
sdake | or iscsi atleast | 14:31 |
Mech422 | it would also be cool to try the emc one... | 14:32 |
Mech422 | forget what its called now - supposed to be really fast | 14:32 |
*** huzhengchuan has joined #openstack-kolla | 14:36 | |
*** zhurong has joined #openstack-kolla | 14:37 | |
sdake | iscsi is iscsi | 14:39 |
sdake | i we have an iscsi workflow | 14:39 |
sdake | what I am unclear aabout is whether that iscsi workflow works multinode | 14:40 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute-kit chart https://review.openstack.org/428150 | 14:48 |
Mech422 | sdake: I know nothing about iscsi - I thought if you mounted it in the host os it just looked/acted like any other scsi drive as far as apps. were concerned ? | 14:49 |
sdake | Mech422 no idea how it integrates with kubernetes | 14:50 |
sdake | Mech422 that is the questoin for sbezverk | 14:50 |
Mech422 | sdake: yeah, thats what I meant - I thought the idea was it didn't need special handling - it 'just worked' like LVM or MDRaid ? | 14:51 |
Mech422 | anyway, be interesting to hear how it works out for ya - might be fun to play with at some point :-) | 14:53 |
*** mannamne has quit IRC | 14:57 | |
*** huzhengchuan has quit IRC | 15:04 | |
*** Pavo has quit IRC | 15:05 | |
*** mdnadeem has quit IRC | 15:12 | |
*** salv-orlando has quit IRC | 15:14 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for kibana https://review.openstack.org/423158 | 15:16 |
*** crushil has joined #openstack-kolla | 15:26 | |
*** lamt has joined #openstack-kolla | 15:31 | |
Jeffrey4l | pbourke, around? abour https://review.openstack.org/428124 , the centos binary failed due to the registry in tarballs.o.o site is not updated. ubuntu + binary failed due to ubuntu cloud archive is using old code. | 15:32 |
*** lamt has quit IRC | 15:51 | |
*** v1k0d3n has joined #openstack-kolla | 16:00 | |
*** zhurong has quit IRC | 16:04 | |
SamYaple | Jeffrey4l: ive said since the beginning, youre going to have a bad time using tarballs.o.o like that. it doesnt update as quickly as kolla gates would need all over the world | 16:12 |
*** lamt has joined #openstack-kolla | 16:18 | |
*** jpstnh has quit IRC | 16:35 | |
*** zhubingbing has joined #openstack-kolla | 16:42 | |
qwang | sdake: sup | 16:49 |
*** lamt has quit IRC | 16:50 | |
*** lamt has joined #openstack-kolla | 16:53 | |
*** crushil has quit IRC | 16:56 | |
*** zhubingbing has quit IRC | 17:01 | |
*** sdake has quit IRC | 17:04 | |
*** crushil has joined #openstack-kolla | 17:07 | |
*** apuimedo has quit IRC | 17:08 | |
*** zhubingbing has joined #openstack-kolla | 17:14 | |
*** apuimedo has joined #openstack-kolla | 17:15 | |
*** lamt has quit IRC | 17:24 | |
*** lamt has joined #openstack-kolla | 17:26 | |
*** salv-orlando has joined #openstack-kolla | 17:31 | |
*** salv-orlando has quit IRC | 17:36 | |
*** lamt has quit IRC | 17:37 | |
sbezverk | kolla-kubernets: check out the following log, single helm chart deployed all these ;) http://logs.openstack.org/50/428150/21/check/gate-kolla-kubernetes-deploy-centos-binary-2-helm-entrypoint-nv/600687a/console.html#_2017-02-04_15_14_05_715025 | 17:37 |
sbezverk | please welcome compute-kit :) | 17:38 |
*** crushil has quit IRC | 18:00 | |
kfox1111 | morning. | 18:18 |
kfox1111 | sbezverk: nice. :) | 18:19 |
kfox1111 | congrats. :) | 18:20 |
*** sdake has joined #openstack-kolla | 18:23 | |
sdake | hey qwang | 18:24 |
sdake | can you link the submission your doing for the panel? | 18:24 |
kfox1111 | sdake: see sbezverk's compute kit package. :) | 18:26 |
sdake | kfox1111 will check it out | 18:27 |
sdake | kfox1111 quick q re iscsi as a storage backend for kubernetes | 18:27 |
kfox1111 | very cool. if we can get a devenv configmaps package too, it should be pretty easy. :) | 18:27 |
sdake | kfox1111 is that a thing? | 18:27 |
kfox1111 | yeah. sbezverk implemnted an iscsi backend. | 18:27 |
sdake | he told me earlier this morning that wont work for multinode?? | 18:28 |
kfox1111 | oh. wait. for k8s itself? | 18:28 |
sdake | yup | 18:28 |
kfox1111 | non starter. | 18:28 |
sdake | mor edetail plz | 18:28 |
qwang | sdake: I haven't submitted it yet. I need Justin to fill his speaker's bio | 18:28 |
sdake | qwang cool can you link the etherpad you have | 18:29 |
sdake | qwang i wanted to look it over | 18:29 |
kfox1111 | sdake: you'd need basically cinder to provision the iscsi volumes for you. | 18:29 |
qwang | https://etherpad.openstack.org/p/openstack-on-helm | 18:29 |
sdake | kfox1111 i have a NAS with iscsi support in my home lab | 18:29 |
kfox1111 | but cinder doesn't support manual mounts as far as I know. | 18:29 |
qwang | sdake: thank you for taking a look | 18:29 |
kfox1111 | sdake: thats a different issue, but still an issue. | 18:29 |
kfox1111 | so, there is no standard api for iscsi volume provisioning, except for cinder. | 18:30 |
kfox1111 | but we couldn't use cinder for this case. :/ | 18:30 |
kfox1111 | you could manually provision iscsi volumes though, and import them into k8s. | 18:30 |
kfox1111 | kolla-kubernetes would be able to use them. | 18:30 |
sdake | qwang i made some adjustments | 18:33 |
sdake | qwang you can keep them or ditch em :) | 18:34 |
*** salv-orlando has joined #openstack-kolla | 18:34 | |
sdake | kfox1111 ok - so in my synology nas, I can create ISCSI volumes | 18:34 |
qwang | sdake: thank you | 18:34 |
sdake | kfox1111 would one iscsi volume be sufficient for all of kubernetes? | 18:35 |
sdake | or do I need a slew of them | 18:35 |
sdake | kfox1111 i want to avoid setting up ceph on my dev machines - yet i wnt multinode capability | 18:35 |
kfox1111 | sdake: if you are using ceph too, you would need two pv's. if not, at least 3. | 18:35 |
sdake | i don't plan to use ceph | 18:36 |
kfox1111 | so, 3 | 18:36 |
sdake | 3 iscsi volumes then? | 18:36 |
kfox1111 | yeah. so make 3 iscsi volumes via synology's api. then create 3 pv's in kubernetes pointing at the volumes. | 18:37 |
sdake | synology's api afaik is a GUI :) | 18:37 |
sdake | what are the 3 pvs used for? | 18:37 |
kfox1111 | yeah. most iscsi appliences are. | 18:37 |
kfox1111 | so... pvs... | 18:38 |
*** skuda has quit IRC | 18:38 | |
kfox1111 | iscsi woudl only support static pv's. how that works is the following: | 18:38 |
sdake | kfox1111 - this is what i have - although an older model: https://www.amazon.com/Synology-Network-Attached-Storage-DS216play/dp/B015JQAWW0/ref=sr_1_2?s=pc&ie=UTF8&qid=1486233489&sr=1-2&keywords=synology+nas | 18:38 |
kfox1111 | a k8s admin creates a pv by creating a document with a description of how to attach to a volume along with its size. | 18:38 |
sdake | document/manifest? | 18:39 |
*** salv-orlando has quit IRC | 18:39 | |
kfox1111 | whenever a user wants a pv, the user creates a pvc which states how much space they need. | 18:39 |
kfox1111 | k8s then looks at all the defined/unused pv's, picks one, and binds it to the pvc. | 18:39 |
kfox1111 | the pvc can be mounted into any pod for that user. | 18:40 |
kfox1111 | so a pvc is kind of like a cinder volume in a tenant. | 18:40 |
kfox1111 | https://kubernetes.io/docs/user-guide/persistent-volumes/#types-of-persistent-volumes | 18:41 |
kfox1111 | look at the example under: "Persistent Volumes" | 18:41 |
kfox1111 | and you'd replace nfs with the iscsi information for a specific volume. | 18:41 |
kfox1111 | and make 3 of them. | 18:41 |
sdake | i saw that, nfs might be easier | 18:42 |
kfox1111 | yes/no. | 18:43 |
sdake | what do the 3 pvs represent in openstack? | 18:43 |
kfox1111 | nfs will probably cause problems with mariadb? | 18:43 |
kfox1111 | and I don't know if it will lock right. | 18:43 |
kfox1111 | might be worth a try just to see though. | 18:43 |
sdake | agreed | 18:44 |
kfox1111 | there are 3 stateful services. mariadb, rabbitmq, and glance-api. | 18:44 |
sdake | wot about nova and cinder? | 18:44 |
kfox1111 | without ceph, | 18:44 |
kfox1111 | nova is always local storage, | 18:44 |
kfox1111 | and cinder is local storage too. | 18:45 |
sdake | woudlnt it make more sense for those to be on nfs/iscsi or whatever? | 18:45 |
kfox1111 | icsi can't be shared. | 18:45 |
sdake | nfs can be | 18:45 |
kfox1111 | nfs doesn't scale very well without expensive hardware. | 18:46 |
kfox1111 | but for very small systems, might make sense. | 18:46 |
sdake | understood | 18:46 |
sdake | nfs is also not HA | 18:46 |
kfox1111 | should be possible with afew tweaks to the templates. | 18:46 |
sdake | want to make the external storage as flexible as possible | 18:46 |
kfox1111 | never done an nfs deployment so not sure what it would take. | 18:46 |
sdake | my synology nas can also export nfs volumes ;-) | 18:46 |
kfox1111 | yeah. nfs is pretty darned easy. :) | 18:47 |
sdake | synology nas is just linux with a nice gui and a bunch of apps and whatnot built in | 18:47 |
kfox1111 | though if someone sold ceph appliences, it wuold be just as easy for that use case... :) | 18:47 |
kfox1111 | ah. | 18:47 |
sdake | interesting point | 18:47 |
kfox1111 | yeah. one of my friends runs something similar. | 18:47 |
sdake | *startup idea* :) | 18:48 |
kfox1111 | yup. | 18:48 |
sdake | ok well i'll see if nfs works for those 3 services | 18:49 |
sdake | and go from there | 18:49 |
sdake | if nfs doesn't work for those 3, it surely wont be suitable for nova and cinder | 18:49 |
kfox1111 | well, actualy the opposite I think. | 18:49 |
kfox1111 | I think it will work well for nova/cinder, as openstack itself supports those use cases. | 18:49 |
kfox1111 | has for a while. we may just need a bit of template tweaks to enable nfs mount on nova-compute container. | 18:50 |
kfox1111 | backing a mariadb with nfs though, I'm not so sure. | 18:50 |
sdake | so what your suggesting is using iscsi for those 3 non-shared, and nfs for the 2 shared? | 18:50 |
kfox1111 | glance might work. | 18:50 |
kfox1111 | sdake: yeah, that might be easier. | 18:50 |
sdake | i've got this cloud.yaml that sbezverk gave me, although i have no idea how it was generated | 18:51 |
*** lamt has joined #openstack-kolla | 18:52 | |
sdake | what is the downside of using nfs for the 5 pvcs? | 18:53 |
kfox1111 | tools/build_example_yaml.py | 18:53 |
kfox1111 | gotta ensure they dont overlap, | 18:53 |
kfox1111 | and locking works. | 18:53 |
sdake | overlap - how do you mean | 18:53 |
sdake | nfs is supposed to b eposix compliant | 18:53 |
sdake | whether it is or not, remains to be seen :) | 18:53 |
kfox1111 | heh. nfs is not posix compliant. | 18:53 |
sdake | posix coplaince = fcntl() works | 18:54 |
kfox1111 | its darned near impossible for a network fs to be so. :) | 18:54 |
kfox1111 | fcntl sort of works simetimes perhaps. ;) | 18:54 |
kfox1111 | luster got posix support much closer then nfs, but still has a few gotchas. :/ | 18:54 |
kfox1111 | posix is just way to local disk heavy in design. | 18:54 |
kfox1111 | anyway. I digress. | 18:54 |
kfox1111 | give each volume a different subdir in nfs. | 18:55 |
kfox1111 | at least. | 18:55 |
sdake | yes of course | 18:55 |
kfox1111 | don't want to accidently have one pod touching another pods data. | 18:55 |
kfox1111 | thats all I meant by nonoverlapping. | 18:55 |
sdake | i see | 18:55 |
kfox1111 | so the root of one pv cant see data from another pv. | 18:55 |
sdake | nfs is listed as readwritemany | 19:01 |
sdake | iscsi is only listed as readwriteonce | 19:01 |
sdake | in the docs | 19:01 |
sdake | so nfs it is | 19:01 |
sdake | how do I get register the pv manifest with nfs setup in kubernetes? | 19:02 |
sdake | get// | 19:02 |
*** lamt has quit IRC | 19:02 | |
sdake | assuming i can magically sort out how to create 3 nfs volumes for the PVCs required by kolla-kkubernetes, how do I tell kubernetes about them? | 19:03 |
*** sdake has quit IRC | 19:09 | |
*** sdake has joined #openstack-kolla | 19:10 | |
kfox1111 | readwritemany is not really desirable for this case tough. | 19:10 |
kfox1111 | cause if you get two mariadb's both trying to use the same pvc, your going to have a bad day. | 19:10 |
kfox1111 | readwriteonce should prevent 2 pods from ever trying to run in that config. | 19:11 |
*** Pavo has joined #openstack-kolla | 19:11 | |
sdake | how does mariadb know to map a pv with readwriteonce to a pvc ? | 19:16 |
sdake | i have magically created 3 nfs volumes :) | 19:16 |
sdake | ratheer to the correct pvc | 19:17 |
sdake | hmm i think you know what I'm after :) | 19:17 |
sdake | how doees the mariadb pvc know to map to a pv with readwriteonce | 19:17 |
kfox1111 | you make some pvc's and then theyh try and bind randomly to free pv's. | 19:18 |
sdake | so then how to make nova work with nfs? | 19:18 |
sdake | since it requires readwritemany? | 19:18 |
kfox1111 | for this use case, I don't think you want any pv's with readwritemany. | 19:18 |
kfox1111 | hmm... | 19:19 |
kfox1111 | that is a good question... | 19:19 |
sdake | yes - i'm good at asking those ;) | 19:19 |
kfox1111 | maybe a pvc would still work for that use case. | 19:19 |
kfox1111 | and in that case, you would want one with readwritemany... | 19:19 |
kfox1111 | we need to look at the pvc spec... | 19:19 |
sdake | where is that located | 19:19 |
kfox1111 | ah. | 19:20 |
kfox1111 | you can specify the requerd readwritemany in the pvc too. | 19:20 |
kfox1111 | so it will exclude pv's that don't match. | 19:20 |
sdake | nice | 19:20 |
sdake | i'd expect that would be the caes | 19:20 |
kfox1111 | https://kubernetes.io/docs/user-guide/persistent-volumes/#persistentvolumeclaims | 19:20 |
sdake | case | 19:20 |
sdake | well i have tested thenfs mount | 19:24 |
sdake | it works from linux to my snology mount | 19:24 |
kfox1111 | k. | 19:24 |
kfox1111 | k8s just does kernel mounts, so should mostly work. | 19:24 |
sdake | now how to get the spec into kubernetes | 19:24 |
kfox1111 | though k8s lacks the ability to specify many nfs mount options. :/ | 19:24 |
kfox1111 | need to create some pv yaml's. | 19:25 |
kfox1111 | then kubectl create -f foo.yaml | 19:25 |
sdake | cool | 19:25 |
sdake | is there a pv yaml example hiding somewhere for the lazy :) | 19:25 |
kfox1111 | on that website | 19:25 |
sdake | thanks i'll hunt i down | 19:25 |
*** salv-orlando has joined #openstack-kolla | 19:35 | |
*** salv-orlando has quit IRC | 19:39 | |
kfox1111 | sdake: know when the final openstack icons will be released? | 19:43 |
sdake | if you mean the mascots, ptg is the target for stickers and whatnot | 19:43 |
kfox1111 | yeah. ok. cool. | 19:43 |
kfox1111 | helm has a spot for icons in the charts. | 19:43 |
sdake | although alot of teams are not happy with their mascotts i hear | 19:43 |
sdake | kfox1111 hotness :) | 19:44 |
sdake | so it may be after the ptg | 19:44 |
sdake | our mascot kicks ass | 19:44 |
kfox1111 | part of the problem I think is they have been showing the projects the icons in isolation. | 19:44 |
kfox1111 | they may be better looking when shown all as a set. | 19:44 |
sdake | "THESE ARE NOT YOUR FRIENDS THEY ARE YOUR COMPETITION" | 19:51 |
sdake | enders game such a fantastic movie | 19:52 |
sdake | what is fascinating about that statement above is later they are all friends :) | 19:52 |
*** Pavo has quit IRC | 19:55 | |
sdake | portdirect_away when you are about struggling to get the devenv setup - runnign into the same error i had before I believe | 20:07 |
*** jascott1 has joined #openstack-kolla | 20:16 | |
*** jascott1 has quit IRC | 20:20 | |
kfox1111 | sdake: thought you had it working? | 20:29 |
kfox1111 | whats broke? | 20:29 |
*** schwicht_at_work has joined #openstack-kolla | 20:33 | |
*** salv-orlando has joined #openstack-kolla | 20:36 | |
*** salv-orlando has quit IRC | 20:40 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: dump configmaps in the gate for testing. https://review.openstack.org/429292 | 20:55 |
*** david-lyle_ has joined #openstack-kolla | 21:02 | |
*** david-lyle has quit IRC | 21:05 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: dump configmaps in the gate for testing. https://review.openstack.org/429292 | 21:07 |
*** Pavo has joined #openstack-kolla | 21:08 | |
*** Pavo has quit IRC | 21:14 | |
*** mannamne has joined #openstack-kolla | 21:17 | |
*** dave-mccowan has joined #openstack-kolla | 21:17 | |
kfox1111 | bmace_: around? | 21:27 |
*** skuda has joined #openstack-kolla | 21:35 | |
*** Jeffrey4l has quit IRC | 21:35 | |
*** Jeffrey4l has joined #openstack-kolla | 21:35 | |
*** mannamne has quit IRC | 21:36 | |
*** dave-mccowan has quit IRC | 21:45 | |
*** mannamne has joined #openstack-kolla | 21:46 | |
*** dave-mccowan has joined #openstack-kolla | 21:48 | |
*** schwicht_at_work has quit IRC | 21:51 | |
sdake | kfox1111 I had a real hardware model working - not virt | 21:53 |
sdake | kfox1111 here is the describe herror: | 21:53 |
kfox1111 | ah ok. | 21:54 |
sdake | hmm it looks like after i went to sleep for a while itcame up | 21:54 |
kfox1111 | heh. self healed. :) | 21:55 |
sdake | [sdake@minime-03 halcyon-kubernetes]$ kubectl get pods | 21:55 |
sdake | NAME READY STATUS RESTARTS AGE | 21:55 |
sdake | 3b2da751-1d55-44e0-9317-f3294b2f31ad 1/1 Running 0 1h | 21:55 |
kfox1111 | do a --all-namespaces | 21:57 |
sdake | all running | 21:57 |
sdake | (with weave) | 21:57 |
kfox1111 | nice. | 21:57 |
kfox1111 | and kolla running too, or just the base env? | 21:58 |
sdake | just bae end | 21:58 |
sdake | i just woke u pdude :) | 21:58 |
kfox1111 | k | 21:58 |
sdake | i am going to try that next | 21:58 |
kfox1111 | cool. | 21:58 |
sdake | once i cna figure out how to get all the files I need for my docs changes | 21:58 |
sdake | superbowl party tomorrow | 21:59 |
*** jascott1 has joined #openstack-kolla | 22:00 | |
*** dave-mccowan has quit IRC | 22:05 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute-kit chart https://review.openstack.org/428150 | 22:21 |
*** breitz has quit IRC | 22:27 | |
*** sdake_ has joined #openstack-kolla | 22:27 | |
*** breitz has joined #openstack-kolla | 22:28 | |
*** sdake has quit IRC | 22:29 | |
*** sdake_ has quit IRC | 22:31 | |
*** sdake_ has joined #openstack-kolla | 22:34 | |
sdake_ | kfox1111 my vpn is busted | 22:34 |
* sdake_ curses cisco | 22:35 | |
*** salv-orlando has joined #openstack-kolla | 22:38 | |
bmace_ | kfox1111: whats up? I'm semi-around. | 22:39 |
*** salv-orlando has quit IRC | 22:42 | |
*** skuda has quit IRC | 22:55 | |
*** klindgren_ has joined #openstack-kolla | 22:58 | |
*** klindgren has quit IRC | 23:00 | |
*** laki has joined #openstack-kolla | 23:01 | |
laki | Hello | 23:01 |
sdake_ | hey laki | 23:01 |
sdake_ | what can I do for ya | 23:01 |
sdake_ | "Who green-lit the launch?" | 23:02 |
laki | got a question ... during "default" build of openstack ( 3.0.2 ) cinder was missed | 23:02 |
sdake_ | Apocolpyse is a pretty good movie too | 23:02 |
laki | currently I'm progressing with building it | 23:02 |
laki | how should I later add it ? is there parameter kolla-ansible deploy cinder ? | 23:03 |
sdake_ | deploy is idempotent | 23:04 |
sdake_ | meaning you can just enable it in the config and be good to go | 23:04 |
sdake_ | bmace_ say i figured out what was wrong with my devenv - i just had to wait 20-30 minutes for vagrant up to get the boxes nito a working state | 23:05 |
*** david-lyle_ has quit IRC | 23:05 | |
*** g3ek has quit IRC | 23:06 | |
bmace_ | sdake_: the bringup can take a while, yeah. glad you got it running :) | 23:06 |
sdake_ | this is the first free time i've had to do anything helpful since november | 23:06 |
laki | I don't get it ... it's already build - I don't see it in the docker ps --all list | 23:06 |
sdake_ | so making the most of my saturday :) | 23:06 |
sdake_ | laki did you enable it in globals.yml? | 23:07 |
bmace_ | if you want i can post the WIP i have at this point, but there are a lot of manual wait spots where you would need to look at the script and wait appropriately until i have the python code written to do the auto-wait. | 23:07 |
*** david-lyle has joined #openstack-kolla | 23:07 | |
sdake_ | laki cinder doesn't default to on | 23:07 |
*** haplo37 has quit IRC | 23:07 | |
laki | but I don't want to build it from the beginning | 23:07 |
sdake_ | bmace_ works for me - i was going to put a review up for my version - just as long as we can separate it from the workflow in the gate i think we are good | 23:07 |
laki | I want to add just built cinder to already working | 23:07 |
laki | openstack | 23:07 |
sdake_ | laki right - deploy will do the job | 23:08 |
*** g3ek has joined #openstack-kolla | 23:08 | |
bmace_ | sdake_: currently i document in one of the scripts using the ceph_service_workflow.sh which will need a few tweaks to run in any env (moving some hard coded values to args and fixing the wait logic mostly) | 23:08 |
*** haplo37 has joined #openstack-kolla | 23:09 | |
laki | so I execute "docker-ansible deploy" and it won't mess the currently working openstack ? | 23:09 |
sdake_ | laki kolla-ansible deploy leaves existing stuff alone - hence its indepntent | 23:09 |
laki | clear | 23:09 |
bmace_ | sdake_: but other than that the ceph_service_workflow.sh works well. | 23:09 |
sdake_ | ok well whatever works - i am mostly after having people understand how the stuff works rather then spend 2 weeks decomposing the gat escripts which is a huge pita ;) | 23:09 |
bmace_ | sdake_: i'll toss up my review so you can take a look at it.. keep in mind it has a hardcoded proxy value in there right now for the oracle network.. easy enough to yank from the BUILD_ARGS but would be good to see what people think of the general approach | 23:10 |
sdake_ | bmace_ sounds good | 23:10 |
kfox1111 | sounds good to me too. | 23:10 |
openstackgerrit | BorneMace proposed openstack/kolla-kubernetes master: WIP: Added support for containerized dev bringup https://review.openstack.org/429295 | 23:11 |
sdake_ | the devenv that portdirect_away and kfox1111 made was an interestng idea - although I found it extremly compelx in practice to use | 23:11 |
*** david-lyle has quit IRC | 23:11 | |
bmace_ | sdake_: also, right now the container just does a wait forever loop when you start it up and you need to exec the script manually. this was mostly to make it easy for me to debug things. the setup_dev_env.sh script will obviously be done by default at container run time. | 23:12 |
kfox1111 | bmace_: yeah. have had to do that alot myself. | 23:12 |
laki | if this will work - I will be more then impressed | 23:13 |
kfox1111 | in the prehelm cli, I even added a -d container_name flag to automatically replace the containers startup with a wait looop. | 23:13 |
sdake_ | laki i'll bet you 1 AUD :) | 23:13 |
kfox1111 | should try and get that feature into helm. :) | 23:13 |
kfox1111 | would be generally useful to all helm deployments. | 23:13 |
bmace_ | overall i like the flow of it more than the old dev env stuff we had up though.. less new files and seems a bit simpler.. Dockerfile is in tools instead of kolla-k8s base dir and it works against local code trees instead of checking stuff out on the fly | 23:13 |
kfox1111 | nice. | 23:14 |
*** mannamne has quit IRC | 23:18 | |
laki | sdake_: dude | 23:21 |
laki | respect | 23:21 |
sdake_ | laki i am respectful | 23:22 |
laki | ;) thx | 23:23 |
openstackgerrit | BorneMace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for MariaDB https://review.openstack.org/429296 | 23:24 |
laki | hmm | 23:26 |
laki | I don't see it in horizon | 23:26 |
laki | but answer is other in gui | 23:26 |
openstackgerrit | BorneMace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for MariaDB for devenv https://review.openstack.org/429296 | 23:29 |
*** haplo37 has quit IRC | 23:29 | |
*** stevebaker has quit IRC | 23:31 | |
*** haplo37 has joined #openstack-kolla | 23:32 | |
*** Pavo has joined #openstack-kolla | 23:32 | |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for RabbitMQ for devenv https://review.openstack.org/429297 | 23:34 |
*** stevebaker has joined #openstack-kolla | 23:35 | |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for memcached for devenv https://review.openstack.org/429298 | 23:38 |
*** salv-orlando has joined #openstack-kolla | 23:38 | |
laki | thx - bye | 23:38 |
*** laki has quit IRC | 23:38 | |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for keepalived for devenv https://review.openstack.org/429299 | 23:42 |
*** salv-orlando has quit IRC | 23:43 | |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Keystone for devenv https://review.openstack.org/429300 | 23:45 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for resolv-conf for devenv https://review.openstack.org/429301 | 23:47 |
*** salv-orlando has joined #openstack-kolla | 23:48 | |
*** salv-orl_ has joined #openstack-kolla | 23:50 | |
*** salv-orlando has quit IRC | 23:53 | |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Openvswitch-db-server for devenv https://review.openstack.org/429302 | 23:53 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Openvswitch-vswitchd for devenv https://review.openstack.org/429303 | 23:53 |
*** salv-orl_ has quit IRC | 23:54 | |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Cinder-api for devenv https://review.openstack.org/429306 | 23:56 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Cinder-api-haproxy for devenv https://review.openstack.org/429307 | 23:56 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Cinder-backup for devenv https://review.openstack.org/429308 | 23:56 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Cinder-scheduler for devenv https://review.openstack.org/429309 | 23:56 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Cinder-volume for devenv https://review.openstack.org/429310 | 23:56 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Glance-api for devenv https://review.openstack.org/429311 | 23:58 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Glance-api-haproxy for devenv https://review.openstack.org/429312 | 23:58 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Glance-registry-haproxy for devenv https://review.openstack.org/429313 | 23:58 |
openstackgerrit | Borne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Glance-registry for devenv https://review.openstack.org/429314 | 23:58 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute-kit chart https://review.openstack.org/428150 | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!