| sbezverk | basically ironic appears does not use any RPC calls | 00:00 |
|---|---|---|
| sbezverk | so statefulsets are not required | 00:00 |
| kfox1111 | none? hmmm.. | 00:00 |
| *** ccrouch has quit IRC | 00:00 | |
| sbezverk | yep | 00:00 |
| kfox1111 | whats the ironic api / conductor thing in the digram then? | 00:00 |
| sbezverk | I confirmed it today | 00:00 |
| sbezverk | ironic api will be deployment | 00:01 |
| kfox1111 | https://docs.openstack.org/developer/ironic/_images/deployment_architecture_2.png | 00:01 |
| sbezverk | conductor with sidecar tftp is daemonset | 00:01 |
| kfox1111 | yeah, ironic api shoudl work as a deployment. | 00:01 |
| sbezverk | there is another piece I have never heard is inspector | 00:02 |
| sbezverk | no diea what it does though | 00:02 |
| kfox1111 | there was a tool that netbooted nodes, then scraped info out of them. | 00:03 |
| kfox1111 | maybe thats its name now? | 00:03 |
| kfox1111 | is there a no-op ironic power driver? | 00:03 |
| kfox1111 | or a shell script one? | 00:04 |
| *** saneax-_-|AFK has joined #openstack-kolla | 00:04 | |
| kfox1111 | just looking at the diagram, I'd think conductor/dnsmasq pod would be a daemonset. | 00:04 |
| kfox1111 | Not sure dnsmasq used in this way would be able to ever float around nodes. | 00:05 |
| kfox1111 | and it also may assume once a conductor writes boot files to disk, they stay there, | 00:05 |
| kfox1111 | so emptyDir may have unintended sideeffects. | 00:05 |
| sbezverk | kfox1111: do not know.. about script | 00:06 |
| *** hyakuhei has joined #openstack-kolla | 00:06 | |
| kfox1111 | would be interesting to find out. we could spawn a vm in the gate, and pxeboot it. | 00:06 |
| sbezverk | but dns must be only one because dhcp server inside would be conflicting with other dhcp servers | 00:06 |
| kfox1111 | to test out functionality. | 00:06 |
| kfox1111 | pxe needs a dns server though. | 00:07 |
| sbezverk | kfox1111: I will ask tomorrow how ironic people do it at the gate for testing | 00:07 |
| kfox1111 | it may be though that they support external dhcp. | 00:07 |
| kfox1111 | in which case, I'd split the use cases apart to different objects. | 00:07 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 00:07 |
| kfox1111 | k | 00:07 |
| *** lamt has quit IRC | 00:08 | |
| *** jascott1_ has joined #openstack-kolla | 00:13 | |
| sbezverk | kfox1111: right that is why dns/dhcp goes into its own pod | 00:14 |
| *** fragatin_ has quit IRC | 00:15 | |
| *** fragatina has joined #openstack-kolla | 00:15 | |
| *** eaguilar has quit IRC | 00:21 | |
| *** hfu has joined #openstack-kolla | 00:28 | |
| kfox1111 | I would be surprised though if ironic conductor didn't write files out directly where dnsmasq reads them though. | 00:29 |
| kfox1111 | so they may need to be coupled together rather tightly. | 00:29 |
| kfox1111 | (pod level) | 00:29 |
| *** bpattewar__ has joined #openstack-kolla | 00:32 | |
| sbezverk | kfox1111: hmm maybe, will have to double check that.. finally I got configmaps generated.. :) | 00:33 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 00:41 |
| kfox1111 | If I remember correctly, they do some kind of concensus ring for ha membership. | 00:41 |
| kfox1111 | or they use to. | 00:42 |
| kfox1111 | so I think they assume nodes are rather long lived. | 00:42 |
| kfox1111 | so daemonset net=host may be the best bet. | 00:42 |
| kfox1111 | bmace: looks good. wf'ed. | 00:42 |
| kfox1111 | sbezverk: so, I've played with this a bit before: https://github.com/openstack/pyghmi | 00:43 |
| sbezverk | kfox1111: yeah, all of ironic needs to be net-host | 00:43 |
| kfox1111 | probably want to see if we can get that into the gate. | 00:43 |
| sbezverk | they all listen on a separate phys interface | 00:44 |
| sbezverk | connected to baremetal farm | 00:44 |
| bmace | kfox1111: kk, thanks.. i tossed in a re-check to get all green, not sure if that will mess with the merge or not, but it was all green before and i don't think my whitespace nuke from the releasenotes broke it :) | 00:44 |
| kfox1111 | bmace: yeah. the changeset doesn't touch anything that will effect that gate. so I'm not worried about it. | 00:44 |
| kfox1111 | sbezverk: we should be able to launch a tiny vm in libvirt and plumb it up with this: https://github.com/openstack/pyghmi/blob/master/bin/virshbmc | 00:46 |
| kfox1111 | then hand it over to ironic for testing. | 00:46 |
| openstackgerrit | yuhui proposed openstack/kolla master: [Trivial]Add/Correct some annotation words https://review.openstack.org/429442 | 00:46 |
| *** tovin07_ has quit IRC | 00:46 | |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Added support for containerized dev bringup https://review.openstack.org/429295 | 00:47 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 00:48 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Initial Mon https://review.openstack.org/410641 | 00:48 |
| kfox1111 | bmace: congrats. and thanks. :) | 00:48 |
| bmace | thanks to you guys also. nice to get it in but i know there is still a long way to go to get everything awesome and consistent. | 00:49 |
| *** jtriley has joined #openstack-kolla | 00:49 | |
| kfox1111 | +1 | 00:49 |
| sbezverk | bmace: Super job man, it tooks time but stability worth it :) | 00:57 |
| bmace | i certainly hope it at least helps folks get up to speed faster | 00:57 |
| bmace | and thanks :) | 00:58 |
| sbezverk | kfox1111: I checked ansible's ironic and dnsmask does not share any mounts with other containers | 01:00 |
| kfox1111 | very strange. | 01:00 |
| kfox1111 | do they have a dedictated tftp server then? | 01:01 |
| sbezverk | kfox1111: please double check :) I might missed something | 01:01 |
| sbezverk | kfox1111: yes, pxe is tftp server and I plun to run it as a side car container to conductor | 01:02 |
| kfox1111 | yeah. they do have a dedicated container it looks like: | 01:02 |
| kfox1111 | templates/ironic-pxe.json.j2 | 01:02 |
| sbezverk | kfox1111: so api and dnsmasq as deployment, and conductor with pxe/tftp as a daemon set | 01:03 |
| *** hfu has quit IRC | 01:03 | |
| kfox1111 | sounds doable, though still skeptical of dnsmasq. | 01:03 |
| kfox1111 | is it just dns or dns+dhcp? | 01:03 |
| *** hfu has joined #openstack-kolla | 01:04 | |
| kfox1111 | hmm.. looks like in kolla-ansible it just does a static pool of ip's. | 01:05 |
| sbezverk | it is just dhcp | 01:05 |
| sbezverk | kfox1111: original design was to use separate from neutron dhcp | 01:05 |
| kfox1111 | so, either a daemonset would work, or a deployment with net=host would work. | 01:05 |
| kfox1111 | and a hostport pinned so the deployment didn't coschedule. | 01:06 |
| sbezverk | kfox1111: cool, do you think you might be able to use it in your env? | 01:06 |
| kfox1111 | I'd like to. :) | 01:07 |
| kfox1111 | I'd probably recommend the daemonset, as its a bit simpler, and k8s is going to get rolling daemonset upgrades soonish. so the other benifits of deployments will go away. | 01:08 |
| sbezverk | kfox1111: ok I have no preference, daemonset then for dnsmasq as well | 01:09 |
| kfox1111 | looks like its so generic it could be used in general as a dns server for my house, or whatever. :) | 01:11 |
| *** cuongnv has joined #openstack-kolla | 01:12 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Initial Mon https://review.openstack.org/410641 | 01:14 |
| sbezverk | kfox1111: do you remember where we change mariadb's address to use mariadb dns name instead? | 01:16 |
| kfox1111 | not off hand. | 01:16 |
| sdake | bmace ++ :) | 01:16 |
| *** zhubingbing_ has joined #openstack-kolla | 01:16 | |
| sdake | sup zhenguo | 01:17 |
| sdake | zhubingbing_ i mean to say :) | 01:17 |
| zhubingbing_ | hi sdake | 01:17 |
| zhubingbing_ | http://paste.openstack.org/show/598919/ | 01:17 |
| kfox1111 | zhubingbing_: the closest docs we have is the just merged devenv docs. | 01:17 |
| sbezverk | kfox1111: found it, it is in the gate config | 01:18 |
| kfox1111 | zhubingbing_: nothing so simple as ` kolla-ansible deploy ` yet though. | 01:18 |
| zhubingbing_ | create-configmaps.sh i don’t find it | 01:18 |
| zhubingbing_ | i know | 01:18 |
| sdake | @zhubingbing_ hmm | 01:18 |
| sdake | @zhubingbing_ moment let me hunt for it | 01:18 |
| zhubingbing_ | thank you very much | 01:19 |
| kfox1111 | sbezverk: does dhcp really come from that container though? | 01:19 |
| zhubingbing_ | and kolla-service-start.sh | 01:19 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 01:19 |
| kfox1111 | the docs you pointed me at has: Example 1: PXE Boot and iSCSI Deploy Process | 01:19 |
| kfox1111 | says neutron does the dhcp stuff, and it enables/disables it dynamically. | 01:19 |
| sbezverk | kfox1111: that was what I was told | 01:21 |
| sbezverk | kfox1111: they support two model | 01:21 |
| kfox1111 | k | 01:21 |
| sbezverk | dedicated dhcp and neutron's dhcp | 01:21 |
| kfox1111 | ah. | 01:22 |
| kfox1111 | so they probably do the netboot anyway, and then just return a bootloader that does local boot again. | 01:22 |
| sbezverk | kfox1111: they prefer dedicated because it allows to serve ip to servers with unknown mac | 01:22 |
| kfox1111 | ah. that makes sense. | 01:22 |
| *** MasterOfBugs has quit IRC | 01:26 | |
| *** caowei has joined #openstack-kolla | 01:27 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla master: Updated from global requirements https://review.openstack.org/431961 | 01:27 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-ansible master: Updated from global requirements https://review.openstack.org/431962 | 01:27 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-kubernetes master: Updated from global requirements https://review.openstack.org/434009 | 01:28 |
| kfox1111 | heh. we now have bots writing code? | 01:30 |
| kfox1111 | the end is near... :/ | 01:30 |
| kfox1111 | :) | 01:30 |
| sbezverk | Oh no no no, no requiereemtn update | 01:30 |
| kfox1111 | ws. | 01:30 |
| sdake | zhubingbing_ found it: https://drive.google.com/file/d/0B8q6xDPETSkHYTJ0SjNGWU5GYjQ/view?usp=sharing | 01:31 |
| sdake | zhubingbing_ let me know if you run into further problems | 01:32 |
| sdake | @sbezverk surprised at how quickly you cranked out an ironic implementation | 01:33 |
| kfox1111 | sdake: work in progress. ;) | 01:34 |
| sdake | kfox1111 yup i see | 01:34 |
| sdake | kfox1111 although it looks as though it has most of the structure | 01:34 |
| kfox1111 | yup. :) | 01:35 |
| *** duonghq has joined #openstack-kolla | 01:36 | |
| *** japestinho__ has quit IRC | 01:38 | |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla master: Using https instead of http https://review.openstack.org/430122 | 01:39 |
| duonghq | morning guys | 01:39 |
| bpattewar_ | duonghq good morning | 01:39 |
| *** tyrola has quit IRC | 01:40 | |
| sdake | hey duonghq | 01:41 |
| *** tovin07 has joined #openstack-kolla | 01:41 | |
| sdake | hey bpattewar_ | 01:41 |
| *** tyrola has joined #openstack-kolla | 01:41 | |
| sdake | yo, I'll pay for anyone that is into tattoos to take Michal our fearless PTL up on his offer on the mailing list ;-) | 01:42 |
| duonghq | sdake, you mention the mascot? | 01:43 |
| kfox1111 | so close on this ceph package.. grr | 01:43 |
| kfox1111 | doh. | 01:45 |
| kfox1111 | cause its hardcoded and not using the var I so painstakingly built. | 01:45 |
| kfox1111 | :/ | 01:45 |
| kfox1111 | time to go home I think. :) | 01:45 |
| duonghq | cya kfox1111 | 01:45 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Initial Mon https://review.openstack.org/410641 | 01:46 |
| sdake | duonghq yup | 01:47 |
| duonghq | zhubingbing_, can you review this: https://review.openstack.org/#/c/405589/ | 01:47 |
| sdake | kfox1111 I'm pretty sure I need a 6 pack after this week concludes ;) | 01:48 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 01:49 |
| sdake | @sbezverk I was thinking about the problem with genconfig and tagging, and it doesn't really matter if we want to add more helm services post tag, as we can just use master in the gate until genconfig is gone. We are already using master so it seems pretty straightforward. | 01:49 |
| *** crushil has quit IRC | 01:57 | |
| bmace | kfox1111: fyi, the upstream global requirements has 2 spaces not one, which is why mine did.. https://github.com/openstack/requirements/blob/master/global-requirements.txt | 02:03 |
| bmace | kfox1111: it is inconsistent, but the kubernetes one, and a lot of others, have 2.. i don't know which is more common | 02:03 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for mongodb https://review.openstack.org/434022 | 02:10 |
| *** jascott1_ has quit IRC | 02:11 | |
| *** jascott1_ has joined #openstack-kolla | 02:12 | |
| sdake | bmace kfox1111 it doesn't matter what you put in there, our global requirements sync automatically per a bot | 02:15 |
| *** jascott1_ has quit IRC | 02:16 | |
| sdake | bmace kfox1111 so if you have something inconsistent and you modify it to be more consistent in kolla's code base, it will naturally undo whatever consistency you had | 02:16 |
| sdake | bmace kfox1111 the place to fix that is in the global requirements repository | 02:16 |
| *** unicell has quit IRC | 02:17 | |
| bmace | sdake: yeah, looks like it is the bot that posted the change? just looks odd with it posting a change that is diff from what is on github.. and why would someone make like a one character nit change like that anyway. | 02:19 |
| sdake | bmace which requirement line | 02:20 |
| sdake | waht was the requirement entry? | 02:20 |
| bmace | https://review.openstack.org/#/c/434009/1/requirements.txt | 02:20 |
| bmace | nuked a whitespace from the kubernetes entry.. | 02:20 |
| bmace | seems odd when the main global requirements has mixed one two and even more spaces between package names and comments all over the place. | 02:21 |
| sdake | https://github.com/openstack/requirements/commit/c77986d9d2fdb1d98b1d2404e0b391162bef8e3b | 02:22 |
| duonghq | bmace, have you finished the kolla-kubernetes dev's doc? | 02:23 |
| sdake | bmace follow me to openstack-requirements plz | 02:23 |
| bmace | sdake: huh? https://github.com/openstack/requirements/blob/master/global-requirements.txt ? | 02:24 |
| bmace | duonghq: i did update the devenv docs as part of the commit | 02:24 |
| sdake | join #openstack-requirements bmace | 02:24 |
| bmace | ah | 02:24 |
| *** jtriley has quit IRC | 02:26 | |
| *** jpstnh has joined #openstack-kolla | 02:31 | |
| *** japestinho_ has joined #openstack-kolla | 02:31 | |
| *** japestinho__ has joined #openstack-kolla | 02:31 | |
| *** japestinho__ has quit IRC | 02:33 | |
| *** jpstnh has quit IRC | 02:33 | |
| sdake | zhubingbing_ those instructions I sent you are for a way to setup kubernetes without the recent changes bmace has submitted | 02:43 |
| sdake | zhubingbing_ that is what I was using before we had bmace 's work merged (I was using ironic standalone to enable the systems) | 02:44 |
| sdake | are we finally ready for a tag ;) | 02:44 |
| *** japestinho_ has quit IRC | 02:44 | |
| sdake | JITT | 02:44 |
| sdake | Jeffrey4l_ - I think kolla-kubernetes is ready to rock with 0.5.0 | 02:45 |
| *** bpattewar__ has quit IRC | 02:45 | |
| Jeffrey4l_ | nice. | 02:45 |
| sdake | Jeffrey4l_ and i think we are also ready for rc1 tag given that you wanted to tag monday :) | 02:45 |
| *** bpattewar_ has quit IRC | 02:45 | |
| Jeffrey4l_ | ok. i will create the tag and branch for all kolla three deliverables. | 02:46 |
| zhubingbing_ | sdake router | 02:46 |
| zhubingbing_ | roger | 02:46 |
| sdake | Jeffrey4l_ note kolla-kubernetes does not need a branch yet | 02:46 |
| sdake | Jeffrey4l_ we won't branch until pike I think | 02:46 |
| Jeffrey4l_ | yes. | 02:46 |
| sdake | yes we need a branch? | 02:46 |
| Jeffrey4l_ | yes. i understand. | 02:47 |
| sdake | roger | 02:47 |
| openstackgerrit | Merged openstack/kolla-ansible master: Fix X-Forward-Proto Header Deletion https://review.openstack.org/405589 | 02:49 |
| *** crushil has joined #openstack-kolla | 02:50 | |
| *** jpstnh has joined #openstack-kolla | 02:50 | |
| *** japestinho_ has joined #openstack-kolla | 02:50 | |
| sdake | duonghq are you ready to go with bmace 's dev env | 02:50 |
| sdake | portdirect - pingola are you alive | 02:51 |
| *** cuongnv has left #openstack-kolla | 02:53 | |
| *** cuongnv has joined #openstack-kolla | 02:53 | |
| *** ssurana has quit IRC | 02:54 | |
| duonghq | sdake, you mean the setup_dev_env script? | 02:54 |
| sdake | yup | 02:54 |
| duonghq | I looked for doc in rst format, I'm looking the script now | 02:54 |
| sdake | i guess we should call it something else then devenv | 02:54 |
| sdake | because halycon is the devenv | 02:54 |
| sdake | bmace have a clever name ? | 02:55 |
| *** yuanying has quit IRC | 02:55 | |
| duonghq | sdake, so, I think we can simplify the halcyon ansible? | 02:56 |
| *** crushil has quit IRC | 02:56 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: add version check 2 https://review.openstack.org/432300 | 02:57 |
| sdake | duonghq how do you mean | 02:57 |
| *** crushil has joined #openstack-kolla | 02:58 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Use ocata tag in config.py https://review.openstack.org/432297 | 03:00 |
| *** cuongnv_ has joined #openstack-kolla | 03:02 | |
| *** peterjenkins__ has joined #openstack-kolla | 03:04 | |
| *** robcresswell_ has joined #openstack-kolla | 03:04 | |
| duonghq | sdake, I saw that halcyon-kubernetes has some ansible bit | 03:05 |
| *** serverascode_ has joined #openstack-kolla | 03:05 | |
| *** PyroMani_ has joined #openstack-kolla | 03:05 | |
| *** tonyb_ has joined #openstack-kolla | 03:06 | |
| *** jamespag` has joined #openstack-kolla | 03:06 | |
| *** hrww has joined #openstack-kolla | 03:06 | |
| *** slagle_ has joined #openstack-kolla | 03:06 | |
| duonghq | sdake, I think we need re-organize the tools dir | 03:06 |
| *** jamespag` has quit IRC | 03:06 | |
| *** jamespag` has joined #openstack-kolla | 03:06 | |
| *** g3ek- has joined #openstack-kolla | 03:06 | |
| *** cinerama` has joined #openstack-kolla | 03:07 | |
| *** serverascode has quit IRC | 03:07 | |
| *** hrw has quit IRC | 03:07 | |
| *** retr0h has quit IRC | 03:07 | |
| *** jamespage has quit IRC | 03:07 | |
| *** Mech422 has quit IRC | 03:07 | |
| *** slagle has quit IRC | 03:07 | |
| *** crushil has quit IRC | 03:07 | |
| *** cuongnv has quit IRC | 03:07 | |
| *** tonyb has quit IRC | 03:07 | |
| *** alanmeadows has quit IRC | 03:07 | |
| *** g3ek has quit IRC | 03:07 | |
| *** negronjl has quit IRC | 03:07 | |
| *** robcresswell has quit IRC | 03:07 | |
| *** cinerama has quit IRC | 03:07 | |
| *** peterjenkins_ has quit IRC | 03:07 | |
| *** magicboiz has quit IRC | 03:07 | |
| *** PyroMani has quit IRC | 03:07 | |
| *** Mech422 has joined #openstack-kolla | 03:07 | |
| *** g3ek- is now known as g3ek | 03:07 | |
| *** hrww is now known as hrw | 03:07 | |
| *** magicboiz has joined #openstack-kolla | 03:08 | |
| *** crushil has joined #openstack-kolla | 03:08 | |
| *** retr0h has joined #openstack-kolla | 03:08 | |
| *** negronjl has joined #openstack-kolla | 03:09 | |
| *** robcresswell_ is now known as robcresswell | 03:09 | |
| sdake | duonghq which tools dirk you mean in kolla-ansible? | 03:09 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla stable/mitaka: Bump OpenStack service version https://review.openstack.org/434043 | 03:10 |
| duonghq | sdake, I mean the tools dir in kolla-kubernetes | 03:10 |
| sdake | duonghq ya that directory is a mess | 03:10 |
| sdake | duonghq although I think we have higher priority things to solve | 03:11 |
| duonghq | and the ansible bit in halcyon-kubernetes | 03:11 |
| duonghq | sdake, sure | 03:11 |
| sdake | re halcyon-kubernetes | 03:11 |
| sdake | I think there we can just leave it as is, and if it turns out to be unstable take a look at other options | 03:11 |
| sdake | kfox1111 today had mentioned the problem of gating, and if we can't gate the dev env, it is problematic | 03:12 |
| duonghq | I see that setup_dev_env use halcyon's env as base env? | 03:12 |
| sdake | duonghq it will use any kubernetes environment that has crednetials associated with it | 03:12 |
| sdake | duonghq it being setup_dev_env | 03:12 |
| sdake | bmace is that correct or incorrect? | 03:13 |
| duonghq | iirc bmace or you told that the kube2 should have 1GB RAM on it but I cannot find any doc about the facet in our code base | 03:13 |
| *** jtriley has joined #openstack-kolla | 03:13 | |
| *** jascott1_ has joined #openstack-kolla | 03:13 | |
| sdake | duonghq here is the basic concept | 03:13 |
| sdake | sup jascott1 | 03:13 |
| sdake | duonghq I think halycon is off to a good start for us | 03:14 |
| sdake | duonghq although I get the impression it is no longer maintained and may not ever be gated | 03:14 |
| sdake | duonghq I have heard that AT&T went to minikube exclusively | 03:14 |
| *** tyrola has quit IRC | 03:14 | |
| sdake | although we need a multinode environment to porperlly test things | 03:14 |
| duonghq | oh, quite sad | 03:14 |
| sdake | minikube is full of problems for us, so thats not an option either | 03:15 |
| *** crushil has quit IRC | 03:15 | |
| sdake | for the moment lets keep things in place as is | 03:15 |
| duonghq | we have not had "plan B"? as official recommended base dev env? | 03:15 |
| sdake | and revaluate at a later time | 03:15 |
| duonghq | sure | 03:15 |
| sdake | duonghq plan B there are a couple options | 03:15 |
| sdake | we actually ahve a plan B and a plan C | 03:15 |
| *** tyrola has joined #openstack-kolla | 03:16 | |
| duonghq | guess that I can stick w/ halcyon atm? | 03:16 |
| *** jascott1_ has quit IRC | 03:18 | |
| sdake | the plan B and plan C are: https://github.com/kubernetes-incubator/kargo and https://github.com/kubernetes/kops | 03:18 |
| sdake | I have concerns about kubeadm in general as a result of this issue tracker: https://github.com/kubernetes/kubernetes/issues/31711 | 03:18 |
| *** mannamne has quit IRC | 03:18 | |
| duonghq | sdake, nice | 03:18 |
| *** serverascode_ is now known as serverascode | 03:18 | |
| sdake | halcyon uses kubeadm | 03:19 |
| sdake | anything in the kubernetes-incubator or kubernetes repo should be our PLAN A and PLAN B | 03:19 |
| sdake | something outside of the kubernetes namespace should be plan C | 03:20 |
| *** saneax-_-|AFK is now known as saneax | 03:20 | |
| sdake | duonghq I htink we don't have to make this change immediately | 03:20 |
| sdake | duonghq will you be at the PTGs | 03:20 |
| *** zhuzeyu has joined #openstack-kolla | 03:20 | |
| duonghq | sdake, I have no luck this time | 03:20 |
| *** tonyb_ is now known as tonyb | 03:21 | |
| *** japestinho_ has quit IRC | 03:21 | |
| *** jpstnh has quit IRC | 03:21 | |
| sdake | duonghq there is always remote participation | 03:21 |
| duonghq | sure, if the connection is good enough | 03:21 |
| duonghq | I'll join remotely | 03:21 |
| sdake | yup inc0 is getting that setup | 03:21 |
| sdake | I offered webex to inc0 although he didn't seem interested - so ymmv :) | 03:22 |
| duonghq | ya, I heard that | 03:22 |
| *** saneax is now known as saneax-_-|AFK | 03:22 | |
| *** jtriley has quit IRC | 03:22 | |
| sdake | duonghq if your looking to make a big impact on the project now that we have a basic development environment in place that is functional | 03:23 |
| sdake | duonghq we have alot of work that needs be done | 03:23 |
| duonghq | okay | 03:24 |
| duonghq | sdake, I'm trying to get the dev env up, it'll take awhile | 03:26 |
| sdake | duonghq as an aside, our release liason is tagging 0.5.0 of kolla-kubernetes soon: https://launchpad.net/kolla-kubernetes/+milestone/0.5.0 | 03:32 |
| Jeffrey4l_ | sdake, release kolla-kubernetes 0.5 patch is pushed https://review.openstack.org/434053 | 03:32 |
| sdake | Jeffrey4l_ nice :) | 03:33 |
| Jeffrey4l_ | sdake, for kolla and kolla-ansible project, we need release rc2 before final release, right? | 03:33 |
| sdake | if there are new patches that are backported yes | 03:34 |
| sdake | otherwise no | 03:34 |
| Jeffrey4l_ | ok. | 03:34 |
| *** rafael_sisweb has quit IRC | 03:36 | |
| *** zhuzeyu has quit IRC | 03:37 | |
| *** zhuzeyu has joined #openstack-kolla | 03:39 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 03:39 |
| *** tovin07_ has joined #openstack-kolla | 03:43 | |
| *** Jeffrey4l_ has quit IRC | 03:43 | |
| *** Jeffrey4l_ has joined #openstack-kolla | 03:43 | |
| sbezverk | sdake: you are in abandoning mood on St. Valentines days ;) | 03:44 |
| sdake | sbezverk cleaning up queue | 03:44 |
| sdake | sbezverk we dont need all this extra crap :) | 03:44 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for mongodb https://review.openstack.org/434022 | 03:45 |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Updated from global requirements https://review.openstack.org/434009 | 03:45 |
| sbezverk | sdake: will check back later. hopefully ironic api will be up and running ;) | 03:45 |
| sdake | sbezverk that would be hotness :) | 03:46 |
| duonghq | bmace, which projects should the DEV_BASE contain? | 03:46 |
| *** ssurana has joined #openstack-kolla | 03:49 | |
| *** yingjun has joined #openstack-kolla | 03:49 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Rename kolla namespace to kolla_ansible https://review.openstack.org/434058 | 03:52 |
| *** prameswar has joined #openstack-kolla | 03:54 | |
| *** cuongnv_ is now known as cuongnv | 03:57 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Rename kolla namespace to kolla_ansible https://review.openstack.org/434058 | 03:57 |
| *** prameswar has quit IRC | 03:59 | |
| *** unicell has joined #openstack-kolla | 04:03 | |
| *** jtriley has joined #openstack-kolla | 04:03 | |
| *** jtriley has quit IRC | 04:07 | |
| *** yuanying has joined #openstack-kolla | 04:08 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Rename kolla namespace to kolla_ansible https://review.openstack.org/434058 | 04:09 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: [TEST][MASTER][KOLLA-ANSIBLE] https://review.openstack.org/422786 | 04:10 |
| *** ssurana has quit IRC | 04:11 | |
| *** jrobinson is now known as jrobinson-afk | 04:17 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 04:17 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Fix doc format https://review.openstack.org/434062 | 04:17 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 04:26 |
| *** mdnadeem has joined #openstack-kolla | 04:41 | |
| *** jascott1_ has joined #openstack-kolla | 04:43 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 04:47 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for mongodb https://review.openstack.org/434022 | 04:47 |
| *** jascott1_ has quit IRC | 04:48 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: add version check 2 https://review.openstack.org/432300 | 04:51 |
| Jeffrey4l_ | sdake, please review https://review.openstack.org/434062 | 04:51 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for mongodb https://review.openstack.org/434022 | 04:52 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Rename kolla namespace to kolla_ansible https://review.openstack.org/434058 | 04:52 |
| *** jrobinson-afk is now known as jrobinson | 04:53 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Fix doc format https://review.openstack.org/434062 | 04:59 |
| *** alanmeadows has joined #openstack-kolla | 05:01 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Change Barbican default secret store to pkcs11 https://review.openstack.org/430137 | 05:02 |
| *** imcsk8_ has quit IRC | 05:08 | |
| bmace | duonghq: halcyon-vagrant-kubernetes, kolla-kubernetes and kolla-ansible | 05:11 |
| *** sudipto has joined #openstack-kolla | 05:13 | |
| *** sudipto_ has joined #openstack-kolla | 05:13 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for mariadb https://review.openstack.org/433480 | 05:16 |
| *** imcsk8 has joined #openstack-kolla | 05:16 | |
| *** tyrola has quit IRC | 05:20 | |
| *** tyrola has joined #openstack-kolla | 05:21 | |
| *** jpstnh has joined #openstack-kolla | 05:30 | |
| *** v1k0d3n has quit IRC | 05:31 | |
| *** fragatina has quit IRC | 05:37 | |
| *** skramaja has joined #openstack-kolla | 05:38 | |
| *** unicell has joined #openstack-kolla | 05:46 | |
| *** ssurana has joined #openstack-kolla | 05:49 | |
| *** krtaylor has joined #openstack-kolla | 05:55 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Rename kolla namespace to kolla_ansible https://review.openstack.org/434058 | 05:58 |
| *** l4yerffeJ has joined #openstack-kolla | 06:00 | |
| sp_ | portdirect: facing some issue in kollak8s setup | 06:04 |
| sp_ | portdirect: vagrant up getting hang after below installation progress | 06:05 |
| sp_ | TASK [kube-join : join workers to the initialized kubernetes cluster] ********** | 06:05 |
| sp_ | skipping: [kube1] | 06:05 |
| *** fragatina has joined #openstack-kolla | 06:09 | |
| *** salv-orlando has joined #openstack-kolla | 06:10 | |
| *** pcaruana has joined #openstack-kolla | 06:20 | |
| *** cuongnv has quit IRC | 06:23 | |
| *** krtaylor has quit IRC | 06:24 | |
| *** cuongnv has joined #openstack-kolla | 06:24 | |
| *** satyar has joined #openstack-kolla | 06:29 | |
| *** Jeffrey4l__ has joined #openstack-kolla | 06:30 | |
| *** l4yerffeJ__ has joined #openstack-kolla | 06:30 | |
| *** l4yerffeJ_ has quit IRC | 06:34 | |
| *** Jeffrey4l_ has quit IRC | 06:34 | |
| *** l4yerffeJ has quit IRC | 06:34 | |
| *** l4yerffeJ has joined #openstack-kolla | 06:34 | |
| duonghq | sp_, it's normal, | 06:36 |
| *** ssurana has quit IRC | 06:36 | |
| sp_ | duonghq: thx for response. what could be the problem though? | 06:37 |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes master: Helm charts for heat-create-db and heat-manage-db https://review.openstack.org/414496 | 06:37 |
| duonghq | it's waiting for something | 06:38 |
| duonghq | sp_, I used to kill the ansible and run it manually | 06:38 |
| *** jpstnh has quit IRC | 06:39 | |
| sp_ | duonghq: you mean $$$ansible-playbook halcyon-kubernetes/kube-deploy/kube-deploy.yml | 06:39 |
| sp_ | ? | 06:39 |
| sp_ | right ? | 06:39 |
| *** saneax-_-|AFK is now known as saneax | 06:39 | |
| duonghq | yup | 06:39 |
| duonghq | like that | 06:39 |
| *** tonanhngo has quit IRC | 06:40 | |
| *** krtaylor has joined #openstack-kolla | 06:40 | |
| sp_ | duonghq: i tried same beforehand but faced the same earlier | 06:40 |
| sp_ | duonghq: seems like it is a network issue | 06:40 |
| sp_ | to me | 06:40 |
| sp_ | I mean need proper ip address assignment to vms | 06:41 |
| *** jrobinson has quit IRC | 06:41 | |
| duonghq | sp_, I don't think so | 06:43 |
| sp_ | duonghq: so what can be the probable issue with this ? | 06:43 |
| *** yingjun has quit IRC | 06:46 | |
| duonghq | you can pass --verbose to ansible-playbook | 06:46 |
| *** jeffrey4l_ircc has joined #openstack-kolla | 06:47 | |
| *** jascott1_ has joined #openstack-kolla | 06:55 | |
| *** yingjun has joined #openstack-kolla | 06:55 | |
| *** jascott1_ has quit IRC | 07:00 | |
| *** salv-orl_ has joined #openstack-kolla | 07:01 | |
| *** salv-orlando has quit IRC | 07:03 | |
| *** jascott1_ has joined #openstack-kolla | 07:20 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Optimize reconfiguration for karbor https://review.openstack.org/424160 | 07:26 |
| *** jascott1_ has quit IRC | 07:31 | |
| *** jascott1_ has joined #openstack-kolla | 07:31 | |
| *** jascott1_ has quit IRC | 07:35 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Optimize reconfiguration for collectd https://review.openstack.org/424133 | 07:37 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Add the missing option for gloabols.yml https://review.openstack.org/434116 | 07:48 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Add the missing option for gloabols.yml https://review.openstack.org/434116 | 07:49 |
| *** bmace has quit IRC | 07:51 | |
| *** saneax is now known as saneax-_-|AFK | 07:58 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Fix the read mariadb log permission issue for fluentd https://review.openstack.org/434120 | 08:02 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Fix the read mariadb log permission issue for fluentd https://review.openstack.org/434120 | 08:03 |
| *** salv-orl_ has quit IRC | 08:10 | |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible master: Modify the parameter of recreate_or_restart_container action https://review.openstack.org/434124 | 08:10 |
| *** salv-orlando has joined #openstack-kolla | 08:10 | |
| *** bmace has joined #openstack-kolla | 08:10 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Disable elasticsearch output when enable_elasticsearch is no in fluent https://review.openstack.org/434128 | 08:11 |
| *** athomas has joined #openstack-kolla | 08:14 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Re-enable grafana gpgcheck https://review.openstack.org/434129 | 08:15 |
| Jeffrey4l__ | sdake, ping | 08:16 |
| sdake | sup Jeffrey4l__ | 08:16 |
| Jeffrey4l__ | sdake, please review this https://review.openstack.org/434058 | 08:16 |
| sdake | Jeffrey4l__ not sure what to make of that | 08:17 |
| Jeffrey4l__ | some issue during split repo. | 08:17 |
| Jeffrey4l__ | we need use different namespace for kolla and kolla-ansible | 08:17 |
| sdake | oh right | 08:17 |
| sdake | sorry really tired | 08:17 |
| sdake | moment let m elook | 08:17 |
| Jeffrey4l__ | otherwise, this two packages will be installed into site-packages/kolla folder | 08:18 |
| Jeffrey4l__ | thanks ;) | 08:18 |
| sdake | Jeffrey4l__ did you do a pip install . on this patch? | 08:19 |
| Jeffrey4l__ | yep. | 08:19 |
| sdake | cool | 08:19 |
| sdake | everything was ok? | 08:19 |
| Jeffrey4l__ | i think so. | 08:21 |
| Jeffrey4l__ | tox -exx will install the patch too. | 08:21 |
| *** fragatina has quit IRC | 08:21 | |
| *** unicell has quit IRC | 08:22 | |
| *** shardy has joined #openstack-kolla | 08:24 | |
| *** ipsecguy has joined #openstack-kolla | 08:28 | |
| Jeffrey4l__ | sdake, webex support linux now? i just now tried, still failed. any information? | 08:29 |
| Jeffrey4l__ | link? | 08:29 |
| sdake | you have to use the web link | 08:29 |
| sdake | err the web option | 08:29 |
| Jeffrey4l__ | what is "web option"? | 08:31 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Rename kolla namespace to kolla_ansible https://review.openstack.org/434058 | 08:35 |
| *** matrohon has joined #openstack-kolla | 08:37 | |
| *** l4yerffeJ__ has quit IRC | 08:38 | |
| *** saneax-_-|AFK is now known as saneax | 08:43 | |
| *** matrohon has quit IRC | 08:43 | |
| *** egonzalez has joined #openstack-kolla | 08:46 | |
| hrw | morning | 08:46 |
| *** yuanying has quit IRC | 08:47 | |
| *** yuanying has joined #openstack-kolla | 08:47 | |
| *** yuanying has quit IRC | 08:47 | |
| *** jeckxie has joined #openstack-kolla | 08:48 | |
| *** zioproto has joined #openstack-kolla | 08:49 | |
| *** jeckxie has left #openstack-kolla | 08:49 | |
| egonzalez | morning hrw | 08:49 |
| *** matrohon has joined #openstack-kolla | 08:52 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Add the missing option for gloabols.yml https://review.openstack.org/434116 | 08:53 |
| *** jascott1_ has joined #openstack-kolla | 08:54 | |
| *** Serlex has joined #openstack-kolla | 08:54 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Disable elasticsearch output when enable_elasticsearch is no in fluent https://review.openstack.org/434128 | 08:58 |
| openstackgerrit | Merged openstack/kolla-ansible master: Add default roles used by Barbican https://review.openstack.org/422643 | 08:58 |
| *** jascott1_ has quit IRC | 08:59 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Disable elasticsearch output when enable_elasticsearch is no in fluentd https://review.openstack.org/434128 | 08:59 |
| *** jeckxie has joined #openstack-kolla | 09:04 | |
| *** jeckxie has quit IRC | 09:06 | |
| *** jeffrey4l_ircc has quit IRC | 09:07 | |
| *** mgoddard has joined #openstack-kolla | 09:11 | |
| *** gfidente has joined #openstack-kolla | 09:13 | |
| *** gfidente has quit IRC | 09:13 | |
| *** gfidente has joined #openstack-kolla | 09:13 | |
| *** yingjun has quit IRC | 09:17 | |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla-ansible master: Fix typo in all.yml https://review.openstack.org/434159 | 09:17 |
| *** laki has joined #openstack-kolla | 09:18 | |
| laki | Hello | 09:18 |
| laki | Is there any default volume-group name for lvm backend enabled by kolla ? | 09:18 |
| egonzalez | laki: cinder-volumes https://github.com/openstack/kolla-ansible/blob/master/doc/cinder-guide.rst | 09:22 |
| *** salv-orlando has quit IRC | 09:23 | |
| laki | thx | 09:23 |
| *** zhuzeyu has quit IRC | 09:24 | |
| *** yuanying has joined #openstack-kolla | 09:26 | |
| *** ipsecguy has quit IRC | 09:34 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Disable elasticsearch output when enable_elasticsearch is no in fluentd https://review.openstack.org/434128 | 09:35 |
| *** hrw has quit IRC | 09:38 | |
| openstackgerrit | Wei Cao proposed openstack/kolla-ansible master: Fix freezer auth issue https://review.openstack.org/434174 | 09:39 |
| *** hrw has joined #openstack-kolla | 09:42 | |
| *** hfu has quit IRC | 09:53 | |
| des_consolado | wow this is disgusting the error I'm getting from kolla-ansible deployment this morning :( http://paste.openstack.org/show/598975/ | 09:53 |
| des_consolado | something to do with the Setting sysctl values play in neutron | 09:53 |
| *** salv-orlando has joined #openstack-kolla | 09:54 | |
| *** tovin07_ has quit IRC | 09:56 | |
| egonzalez | des_consolado: host_in_groups var was touched yesterday in defaults/main.yml | 09:59 |
| hrw | let's kill jessie(-backports) from kolla... long live the stretch! | 10:00 |
| *** laki has quit IRC | 10:01 | |
| *** hyakuhei has quit IRC | 10:01 | |
| *** hyakuhei has joined #openstack-kolla | 10:01 | |
| *** hyakuhei has quit IRC | 10:01 | |
| *** hyakuhei has joined #openstack-kolla | 10:01 | |
| *** laki has joined #openstack-kolla | 10:03 | |
| *** cuongnv has quit IRC | 10:04 | |
| satyar | hi egonzalez | 10:15 |
| egonzalez | sup satyar | 10:15 |
| satyar | can i have the cinder volume (backend iscsi) at a different node and the nova compute on a different node | 10:16 |
| satyar | will the boot from volume will work? | 10:16 |
| egonzalez | satyar: should work, iscsi container are started in both storage and compute nodes | 10:19 |
| *** sambetts|afk is now known as sambetts | 10:19 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-ansible master: Updated from global requirements https://review.openstack.org/431962 | 10:19 |
| satyar | yes | 10:20 |
| openstackgerrit | Wei Cao proposed openstack/kolla-ansible master: Fix freezer auth issue https://review.openstack.org/434174 | 10:22 |
| *** duonghq has quit IRC | 10:22 | |
| des_consolado | gah, now errors building 3.0.3 images, 'Unhandled error when running BuildTask', looks like it's coming from this line: dest_archive = os.path.join(image.path, source['name'] + '-archive') | 10:24 |
| des_consolado | KeyError: 'name' | 10:24 |
| satyar | egonzalez is the multipathd required at both storage and compuet? | 10:24 |
| satyar | compute | 10:24 |
| egonzalez | satyar: as far i can see in inventory only in compute | 10:25 |
| *** satyar has quit IRC | 10:27 | |
| *** hfu has joined #openstack-kolla | 10:30 | |
| *** salv-orlando has quit IRC | 10:31 | |
| openstackgerrit | OpenStack Release Bot proposed openstack/kolla master: Update reno for stable/ocata https://review.openstack.org/434207 | 10:33 |
| des_consolado | oh ocata is out now? | 10:34 |
| *** mannamne has joined #openstack-kolla | 10:35 | |
| egonzalez | des_consolado: not branched yet | 10:36 |
| egonzalez | des_consolado: kolla is branched kolla-* not yet | 10:39 |
| des_consolado | egonzalez: kolla- only has master anyway doesn't it | 10:40 |
| des_consolado | well it has some tags too... | 10:41 |
| egonzalez | des_consolado: yes for now, but will be branched with ocata soon | 10:41 |
| des_consolado | oh that will be nice | 10:42 |
| *** zhubingbing_ has quit IRC | 10:45 | |
| hrw | guys, how to get into container used during building an image? "./tools/build.py -b debian -r1 --keep openstack-base" fails for me and I want to get into image used and check why | 10:54 |
| hrw | INFO:kolla.image.build.base: ---> Running in 10c2e6ddea85 | 10:55 |
| hrw | docker run -i -t 10c2e6ddea85 /bin/bash | 10:55 |
| hrw | not work ;( | 10:55 |
| hrw | ok, other id did | 10:57 |
| *** athomas has quit IRC | 11:00 | |
| *** sudipto_ has quit IRC | 11:01 | |
| *** sudipto has quit IRC | 11:01 | |
| *** athomas has joined #openstack-kolla | 11:08 | |
| *** hfu has quit IRC | 11:09 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Run simple_cell_setup after the new container is up during upgrade https://review.openstack.org/431508 | 11:09 |
| hrw | Benedikt Trefzer: whoever you are... I hate you | 11:12 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Run simple_cell_setup after the new container is up during upgrade https://review.openstack.org/431508 | 11:12 |
| *** hfu has joined #openstack-kolla | 11:12 | |
| *** salv-orlando has joined #openstack-kolla | 11:13 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Run simple_cell_setup after the new container is up during upgrade https://review.openstack.org/431508 | 11:13 |
| openstackgerrit | Wei Cao proposed openstack/kolla-ansible master: Change freezer service type https://review.openstack.org/434217 | 11:14 |
| egonzalez | hrw: LOL | 11:15 |
| hrw | egonzalez: I was wondering why debian/stretch was not working for me for hours | 11:15 |
| hrw | egonzalez: then looked at apt preferences injected by kolla. they blocked any installs | 11:15 |
| *** caowei has quit IRC | 11:16 | |
| Jeffrey4l__ | egonzalez, cloud u review https://review.openstack.org/434058 which is critical. | 11:20 |
| *** Jeffrey4l__ is now known as Jeffrey4l | 11:22 | |
| *** jeffrey4l_ircc has joined #openstack-kolla | 11:22 | |
| *** crushil has joined #openstack-kolla | 11:25 | |
| *** magicboiz has quit IRC | 11:26 | |
| *** magicboiz has joined #openstack-kolla | 11:28 | |
| *** hfu has quit IRC | 11:31 | |
| *** v1k0d3n has joined #openstack-kolla | 11:32 | |
| *** laki has quit IRC | 11:35 | |
| *** qwe has joined #openstack-kolla | 11:36 | |
| *** qwe has quit IRC | 11:40 | |
| hrw | whoever invented --logs-dir was genius | 11:44 |
| *** qwe has joined #openstack-kolla | 11:46 | |
| *** salv-orlando has quit IRC | 11:48 | |
| *** qwe has quit IRC | 11:49 | |
| *** mdorman has quit IRC | 11:53 | |
| *** mdorman has joined #openstack-kolla | 11:55 | |
| des_consolado | egonzalez: do you know what to do about that error that you said host_in_groups was changed? I'm back there now lol | 12:02 |
| des_consolado | looks like it was Jeffrey4l with commit a1901d4264aa63b195ff1ccbcded5e63bf31c1ce | 12:02 |
| egonzalez | i have not checked into it yet | 12:03 |
| des_consolado | well it's definitely broken deployment for me | 12:05 |
| *** crushil has quit IRC | 12:06 | |
| *** Kdecherf has quit IRC | 12:09 | |
| Jeffrey4l | des_consolado, what error message did u got? | 12:11 |
| *** salv-orlando has joined #openstack-kolla | 12:11 | |
| *** rwallner has joined #openstack-kolla | 12:12 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Disable elasticsearch output when enable_elasticsearch is no in fluentd https://review.openstack.org/434128 | 12:12 |
| *** satyar has joined #openstack-kolla | 12:17 | |
| des_consolado | Jeffrey4l: http://paste.openstack.org/show/599007/ | 12:19 |
| Jeffrey4l | des_consolado, ansible version? | 12:19 |
| des_consolado | 2.2.1.0 | 12:19 |
| Jeffrey4l | does this work before yestoday? | 12:20 |
| des_consolado | I think so yes, I've just done a git revert, testing deployment now, I'm expecting it to go through successfully, I'll let you know when it does | 12:20 |
| Jeffrey4l | cool thanks. | 12:21 |
| berendt | making xenserver usable with kolla is not that easy :( | 12:22 |
| berendt | neutron is a bitch | 12:22 |
| Jeffrey4l | never tried xen.. | 12:23 |
| berendt | Jeffrey4l: never do it ;) | 12:23 |
| Jeffrey4l | i will lolo | 12:24 |
| des_consolado | gah do you guys use AUFS and get lots of "no space left on device" errors when simply trying to start containers? | 12:24 |
| *** Kdecherf has joined #openstack-kolla | 12:24 | |
| Jeffrey4l | des_consolado, run out of inode? | 12:24 |
| des_consolado | this affects me a lot and I keep having to restart hosts and delete /var/lib/docker and it's just like damn, can we really use this shit in production | 12:24 |
| berendt | des_consolado: you have enough free space on your /var/lib/docker drive? | 12:24 |
| Jeffrey4l | imo, btrfs is better. | 12:25 |
| berendt | overlayfs is also working fine | 12:25 |
| des_consolado | these hosts are massive... here is my df -h http://paste.openstack.org/show/599008/ | 12:25 |
| Jeffrey4l | try df -i | 12:25 |
| des_consolado | iuse is all at 1% | 12:26 |
| Jeffrey4l | hrm then no idea ;( | 12:26 |
| *** lucker has quit IRC | 12:29 | |
| des_consolado | oh was a bunch of old mounted shit | 12:33 |
| des_consolado | restarting the docker daemon fixed it | 12:33 |
| des_consolado | maybe the destroy role's cleanup_host.yml should restart the docker daemon? hmm.. | 12:34 |
| egonzalez | hrw and others interested in debian http://lists.openstack.org/pipermail/openstack-dev/2017-February/112220.html | 12:52 |
| des_consolado | Jeffrey4l: https://bugs.launchpad.net/kolla-ansible/+bug/1664962 confirmed this disappears when reverting that commit | 12:55 |
| openstack | Launchpad bug 1664962 in kolla-ansible "Offending line in 'Setting sysctl values' play in Neutron role" [Undecided,New] | 12:55 |
| berendt | egonzalez: hmm.. does this affect our delayed debian deprecation? | 13:03 |
| egonzalez | berendt: if finally remove Ocata packages yes | 13:03 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 13:03 |
| egonzalez | also affect ubuntu, as some of packages are made by debian | 13:03 |
| berendt | let's wait and hope that anybody hires zigo | 13:04 |
| *** slagle_ is now known as slagle | 13:06 | |
| hrw | egonzalez: :( | 13:08 |
| hrw | berendt: I just did work on getting debian in kolla to be great again | 13:09 |
| Jeffrey4l | globals.yml file egonzalez ? | 13:11 |
| Jeffrey4l | sorry pined wrong person. des_consolado ^^ | 13:11 |
| egonzalez | des_consolado: I deployed an hour ago and had no issue with that task | 13:12 |
| egonzalez | also upgraded several times today | 13:12 |
| Jeffrey4l | egonzalez, are u using the latest code? that patch is pushed yesterday. | 13:12 |
| egonzalez | yeah, I rebased master | 13:12 |
| des_consolado | Jeffrey4l: http://termbin.com/3fq4 my globals.yml I don't think there's anything too unusual | 13:19 |
| Jeffrey4l | des_consolado, inventory file? | 13:20 |
| des_consolado | http://termbin.com/j8wq pretty standard too, I did just change this though, before I had network00 also being a control node (as well as a network node) | 13:20 |
| Jeffrey4l | wired. it works in gate and my env and egonzalez's test | 13:22 |
| hrw | hm few days of playing with kolla and laptop has 1513 docker images. have to finish fixing and finally make some use | 13:22 |
| Jeffrey4l | i tried by using latest ansible 2.2.1.0 which works, too. | 13:23 |
| des_consolado | I'm running kolla-ansible with kolla-ansible/tools/kolla-ansible | 13:23 |
| egonzalez | hrw: maybe have a bunch of dangling images | 13:23 |
| *** sudipto_ has joined #openstack-kolla | 13:23 | |
| *** sudipto has joined #openstack-kolla | 13:23 | |
| des_consolado | rather than the pip-installed binary. Can't remember why, I have it in my head that it's better this way lol | 13:23 |
| berendt | egonzalez: yes.. | 13:23 |
| hrw | egonzalez: of course. 1091 kolla ones | 13:23 |
| Jeffrey4l | fyi, i useful script https://github.com/ZZROTDesign/docker-clean to clean useless docker images | 13:24 |
| berendt | Jeffrey4l: thanks | 13:25 |
| Jeffrey4l | des_consolado, does ./tools/kolla-ansible genconfig -vvv works? | 13:25 |
| hrw | Jeffrey4l: thx | 13:25 |
| des_consolado | Jeffrey4l: with or without the commit? | 13:25 |
| Jeffrey4l | des_consolado, with ( use the latest code ) | 13:25 |
| Jeffrey4l | des_consolado, if it do not works, please copy the logs out with more context line. | 13:27 |
| Jeffrey4l | the lines you added in bug description containers some "$" char and "\n" char. | 13:28 |
| *** jheroux has joined #openstack-kolla | 13:28 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: Fixing compute kit name https://review.openstack.org/434268 | 13:29 |
| hrw | 1228 docker images from kolla. centos (binary, source), debian/jessie-backports (binary, source), debian/stretch (binary), ubuntu (binary, source). | 13:31 |
| *** salv-orlando has quit IRC | 13:34 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 13:38 |
| *** srwilkers has joined #openstack-kolla | 13:43 | |
| des_consolado | Jeffrey4l: these are the errors from the genconfig -vvv, this is exactly how ansible outputs this by default and copied perfectly, it's pretty horrible to read though because it's all on one line, is this okay, or what do you usually do with this kind of output before sharing it? http://paste.openstack.org/show/599029/ | 13:44 |
| *** lamt has joined #openstack-kolla | 13:45 | |
| srwilkers | morning p/ | 13:45 |
| *** mgiles has joined #openstack-kolla | 13:45 | |
| jistr | inc0, portdirect: hi, do you guys know some workaround for the lack of ability to request "shared" hostPath volumes in k8s? i haven't been successful in making network namespaces propagate between containers in k8s deployments, was wondering if/how you were able to achieve that | 13:45 |
| *** rwallner has quit IRC | 13:45 | |
| jistr | https://github.com/openstack/kolla-kubernetes/blob/d1e8d95a915309aa0d5a33cc6f9a7cdf4aeda8f7/services/neutron/neutron-l3-agent-daemonset.yml.j2#L112 | 13:45 |
| Jeffrey4l | des_consolado, this is fine. let me check it. | 13:45 |
| *** rhallisey has joined #openstack-kolla | 13:46 | |
| *** fguillot has joined #openstack-kolla | 13:46 | |
| *** rwallner has joined #openstack-kolla | 13:48 | |
| sbezverk | jistr: the issue is still there but it does not prevent neutron from working | 13:48 |
| Jeffrey4l | des_consolado, past out "pip freeze" ? | 13:49 |
| sbezverk | the only thing you cannot do is to access them from the host | 13:49 |
| *** lamt has quit IRC | 13:49 | |
| Jeffrey4l | des_consolado, then try to remove these two lines and try https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/neutron/tasks/config.yml#L13,L14 | 13:51 |
| jistr | sbezverk: hmm in my case it seemed to give me trouble... the netns seemed to be created by one of the agents, i think it was L3 agent creating the netns for router... and then OVS agent wasn't able to see the netns and work with it, which made Neutron port creation not work in general | 13:52 |
| *** rhallisey has quit IRC | 13:52 | |
| *** zhubingbing has joined #openstack-kolla | 13:56 | |
| portdirect | jistr: this is a current pain point with k8s, and actually what initially drove me to use OVN for my work | 13:59 |
| portdirect | there's a pr in with k8s upstream, that has been around for ages to help resolve this, I'll see if I can dig it up | 14:00 |
| *** yingjun has joined #openstack-kolla | 14:01 | |
| portdirect | for some other things I've done I've sidestepped the issue by bind mounting the docker socket into a pod and then using that to launch the container I've required (eg neutron l3 agent) but again this is a horrible hack :/ | 14:01 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 14:02 |
| *** eaguilar has joined #openstack-kolla | 14:02 | |
| *** crushil has joined #openstack-kolla | 14:03 | |
| sbezverk | jistr: could you be more specific about youe problem? we have VMs running at the gate on kolla-kube with full connectivity.. | 14:03 |
| jistr | portdirect: hehe ok, thanks for the info | 14:04 |
| jistr | sbezverk: i'll try to bring up the env again to get the exact errors | 14:04 |
| jistr | but in general | 14:04 |
| jistr | sbezverk: i saw qrouter / qdhcp network namespaces in L3 / dhcp agents respectively, but not in OVS agent. And OVS agent printed something in the sense that it tried to perform operations on the qr-.... etc. interfaces, but failed. It knew that those interfaces indeed existed, but i could only see them from the agent that created them (L3 / dhcp), but not from OVS agent. | 14:06 |
| portdirect | jistr: if its the /dev/netns issue - you may find that you actually don't need to use the hosts, I may be totally of base here, but i seem to remember that (at least some) of the neutron agents create them as required on restart. | 14:07 |
| *** rhallisey has joined #openstack-kolla | 14:08 | |
| jistr | portdirect: yea they were created just fine (as you say, worst case they appeared when i restarted the L3/dhcp agents), but OVS agent gave me errors. It tried to work with the interfaces but couldn't, i presume it was because it couldn't find them. | 14:10 |
| jistr | maybe this could differ by ML2 backend used... i actually went with GRE on OVS to begin with, b/c that's what i'm most familiar with to date, i'm no networking expert | 14:11 |
| sbezverk | jistr: it should not matter, in the gate we use ml2/vxlan | 14:12 |
| sbezverk | works well | 14:12 |
| portdirect | ^^ i tend to use vxlan, but i think the issues would most likley be the same | 14:12 |
| sbezverk | jistr: next time when you bring env up, can you try it? | 14:12 |
| sdake | hey portdirect - we released 0.5.0 :) | 14:13 |
| srwilkers | nice | 14:13 |
| portdirect | sdake: whoop :) | 14:13 |
| sdake | portdirect nice work during this sprint :) | 14:13 |
| jistr | hmm ok, strange... yea i'll try to poke around the deployment i have a bit more, thanks a bunch for the info sbezverk and portdirect | 14:13 |
| sdake | srwilkers - sorry i thought you already knew | 14:13 |
| sdake | portdirect it did include a rift on your devenv that you kevin and I worked on - bmace reworked it into a functional thing and it merged | 14:14 |
| sdake | gotta get the kids to school | 14:14 |
| sbezverk | sdake: see something unusual ;)? http://paste.openstack.org/show/599036/ | 14:14 |
| srwilkers | how…ironic ;) | 14:15 |
| sbezverk | :) | 14:15 |
| sbezverk | gents please ack this one https://review.openstack.org/434268 | 14:16 |
| sbezverk | at the last minute we renamed compute_kit | 14:16 |
| sbezverk | but missed one spot | 14:16 |
| sbezverk | this will return more greeny to the gate | 14:16 |
| srwilkers | green gates = good gates | 14:17 |
| srwilkers | makes me happy | 14:17 |
| *** salv-orlando has joined #openstack-kolla | 14:18 | |
| *** salv-orlando has quit IRC | 14:18 | |
| sbezverk | srwilkers: +1 | 14:18 |
| *** salv-orlando has joined #openstack-kolla | 14:18 | |
| *** lamt has joined #openstack-kolla | 14:20 | |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Fixing compute kit name https://review.openstack.org/434268 | 14:22 |
| *** srwilkers is now known as srwilkers_coffee | 14:22 | |
| openstackgerrit | Mark Giles proposed openstack/kolla-kubernetes master: Fix minor typos in compute kit test script https://review.openstack.org/434310 | 14:24 |
| *** proto_ has joined #openstack-kolla | 14:26 | |
| *** proto_ has quit IRC | 14:27 | |
| *** zioproto has quit IRC | 14:28 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 14:32 |
| *** yingjun has quit IRC | 14:32 | |
| *** lamt has quit IRC | 14:32 | |
| Jeffrey4l | egonzalez, re upgrade cell issue. what it means by "cell None is not created" | 14:34 |
| egonzalez | Jeffrey4l: if have a Ocata environment up do: nova exec nova_api nova-manage cell_v2 list_cells --verbose | 14:35 |
| egonzalez | is the cell where old instances are running | 14:35 |
| *** hfu has joined #openstack-kolla | 14:37 | |
| egonzalez | Jeffrey4l: don't know why with simple_cell_setup in newton is not created, but with Ocata it is :/ | 14:37 |
| sdake | sbezverk hotness :) | 14:37 |
| Jeffrey4l | roger. removed the re-run simple cell setup in handlers, the cell is not created. | 14:38 |
| sdake | sbezverk 0.5.0 has been tagged, however, we can rename anything prior to 1.0.0 :) | 14:38 |
| *** srwilkers_coffee has quit IRC | 14:39 | |
| sdake | sbezverk what was compute kit renamed to? I must have missed that | 14:39 |
| *** jtriley has joined #openstack-kolla | 14:41 | |
| des_consolado | Jeffrey4l: sorry was on lunch, heres the pip freeze, http://termbin.com/l64l and I'm just re-running the genconfig without those 2 lines now. | 14:41 |
| sbezverk | sdake: it used be compute-kit-0.5.0 | 14:41 |
| sbezverk | but then we decided to keep just compute-kit and do not build multiple versions | 14:42 |
| Jeffrey4l | des_consolado, you are using jinaj2 2.9.5? ansible should not works with that? | 14:42 |
| sdake | right since its a collection of services | 14:42 |
| sdake | makes complete sense to me | 14:42 |
| des_consolado | oh dammit lolll | 14:42 |
| kfox1111 | morning. | 14:42 |
| des_consolado | I ran into that before as well grrrrr | 14:42 |
| sdake | hey peeps - I a mgoing to be incommunicado today - as I have a pretty long day wrt this meetup | 14:42 |
| Jeffrey4l | ansible should raise some error when he work with 2.9.5, iirc. | 14:43 |
| sdake | if you want to attend virtually you can - its sort of like a meetup and webinar and live stream mashed up into one ;) | 14:43 |
| *** yingjun has joined #openstack-kolla | 14:43 | |
| kfox1111 | sdake: fully buzwords compliant it seems. :) | 14:44 |
| egonzalez | inc0: ping re https://bugs.launchpad.net/kolla-ansible/+bug/1664674 | 14:44 |
| openstack | Launchpad bug 1664674 in kolla-ansible "kolla-ansible tries to restart containers while genconfig" [Critical,New] | 14:44 |
| sdake | kfox1111 i am not doing a buzzword presentation | 14:44 |
| kfox1111 | oh. ok. :) | 14:44 |
| sdake | kfox1111 i am talking about real technology and real things not nonsense ;) | 14:44 |
| sdake | if someone asks me to define the buzzword cloud native, I will do so however :) | 14:45 |
| sbezverk | kfox1111: morning | 14:45 |
| kfox1111 | sbezverk: morning. | 14:45 |
| *** hfu has quit IRC | 14:45 | |
| sdake | https://www.meetup.com/PhoenixCloud/events/237219755/ | 14:47 |
| sdake | here is the meetup info for those interested - although the webex info is that in that event descriptoin - instead that is on the ml | 14:47 |
| des_consolado | so jinja2 2.8.x is what we want I guess? | 14:47 |
| *** lamt has joined #openstack-kolla | 14:49 | |
| Jeffrey4l | des_consolado, yep try 2.8.x first. and see whether the issue gone. | 14:51 |
| egonzalez | weird, I have Jinja2==2.9.5 and works :/ | 14:52 |
| *** srwilkers has joined #openstack-kolla | 14:52 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 14:55 |
| *** lamt has quit IRC | 14:56 | |
| *** lamt has joined #openstack-kolla | 14:56 | |
| *** bpattewar_ has joined #openstack-kolla | 14:58 | |
| des_consolado | sdake: that looks cool where are the webex details? | 14:58 |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Fix minor typos in compute kit test script https://review.openstack.org/434310 | 14:58 |
| des_consolado | maybe I have to register on this redhatexecutiveevents site | 14:59 |
| *** adrian_otto has joined #openstack-kolla | 15:01 | |
| *** ccrouch has joined #openstack-kolla | 15:02 | |
| *** mgiles has quit IRC | 15:02 | |
| *** haplo37_ has quit IRC | 15:03 | |
| *** mgiles has joined #openstack-kolla | 15:03 | |
| *** Pavo has joined #openstack-kolla | 15:05 | |
| *** haplo37_ has joined #openstack-kolla | 15:06 | |
| *** duonghq has joined #openstack-kolla | 15:10 | |
| duonghq | evening guys | 15:10 |
| duonghq | I guess that we still have meeting today? | 15:10 |
| hrw | in 49 minutes, right? | 15:11 |
| egonzalez | hrw: yep | 15:12 |
| duonghq | thanks | 15:13 |
| zhubingbing | r；） | 15:14 |
| zhubingbing | guys | 15:14 |
| * hrw -> checking/comparing lists of build images | 15:15 | |
| *** JoseMello has joined #openstack-kolla | 15:15 | |
| *** bpattewar has quit IRC | 15:16 | |
| *** jeffrey4l_ircc has quit IRC | 15:17 | |
| *** zioproto has joined #openstack-kolla | 15:20 | |
| *** lamt has quit IRC | 15:23 | |
| duonghq | hi zhubingbing | 15:24 |
| zhubingbing | hi duonghq | 15:24 |
| *** srwilkers has quit IRC | 15:24 | |
| *** adrian_otto has quit IRC | 15:27 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes master: Helm charts for heat-create-db and heat-manage-db https://review.openstack.org/414496 | 15:27 |
| *** saneax is now known as saneax-_-|AFK | 15:28 | |
| *** adrian_otto has joined #openstack-kolla | 15:28 | |
| *** lamt has joined #openstack-kolla | 15:29 | |
| hrw | guys: I have a question about python-rtslib. Are you fine with using python-rtslib-fb instead? | 15:29 |
| *** srwilkers has joined #openstack-kolla | 15:30 | |
| des_consolado | this can't be right http://paste.openstack.org/show/599054/ - the amount of mounts on my compute node seems to increase exponentially | 15:30 |
| des_consolado | on the other hosts I just have a steady around 100 mounts | 15:30 |
| hrw | -fb exists in pip so can be installed from source where it is missing | 15:30 |
| *** srwilkers has left #openstack-kolla | 15:31 | |
| *** srwilkers has joined #openstack-kolla | 15:31 | |
| sdake | des_consolado webex details are here: http://lists.openstack.org/pipermail/openstack-dev/2017-February/112190.html | 15:33 |
| hrw | looks like it will be fine. Ubuntu uses rtslib one, centos has rtslib-fb so can debian | 15:36 |
| *** ccrouch has quit IRC | 15:38 | |
| *** mdnadeem has quit IRC | 15:40 | |
| Jeffrey4l | sdake, please review https://review.openstack.org/434058 | 15:43 |
| des_consolado | well it looks like it's the iscsid container that starts eating up mounts in some kind of loop like that... | 15:43 |
| openstackgerrit | Merged openstack/kolla-ansible master: Rename kolla namespace to kolla_ansible https://review.openstack.org/434058 | 15:50 |
| *** sp__ has joined #openstack-kolla | 15:50 | |
| *** rwallner has quit IRC | 15:53 | |
| *** ccrouch has joined #openstack-kolla | 15:54 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible master: Create nova cell0 before migrating nova database https://review.openstack.org/431508 | 15:56 |
| Jeffrey4l | egonzalez, i made some change for upgrade cell issue, could u check it again? ^^ | 15:56 |
| inc0 | good morning | 15:57 |
| srwilkers | morning inc0 | 15:57 |
| Jeffrey4l | moring inc0 | 15:57 |
| sp__ | morning inc0 | 15:57 |
| *** penfold has joined #openstack-kolla | 15:57 | |
| *** jascott1_ has joined #openstack-kolla | 15:58 | |
| penfold | #join #openstack-meeting-4 | 15:58 |
| egonzalez | Jeffrey4l: didn't get any error while upgrading without a simple_cell_setup and db sync before bootstrap? | 15:59 |
| *** rwallner has joined #openstack-kolla | 15:59 | |
| egonzalez | sometimes it failed and other not | 16:00 |
| zhubingbing | moring | 16:00 |
| Jeffrey4l | egonzalez, yep. the nova_cell0 and cell0 items in nova_api is required. Others are not neccessary before running database migrations. | 16:00 |
| Jeffrey4l | i made several test. which works. | 16:00 |
| *** vhosakot has joined #openstack-kolla | 16:00 | |
| egonzalez | Jeffrey4l: i'll test later | 16:01 |
| egonzalez | looks good | 16:01 |
| Jeffrey4l | in you patch, run simple_cell_setup does nothing. because cell0 is created. ( so u run simple_cell_setup tice ) | 16:01 |
| Jeffrey4l | twice* | 16:02 |
| egonzalez | having simple_cell_setup at the end will cause erros on vm provisioning while upgrading due no host are sync in any cell yet | 16:03 |
| egonzalez | thats why i moved simple_cell_setup after nova-api is up | 16:04 |
| des_consolado | oh I've been having problems with simple_cell_setup, when deploying I have to deploy twice or deploy and then run the nova-manage command manually and then deploy again or something | 16:06 |
| des_consolado | have any of you had problems with the iscsi container going crazy and eating resource? | 16:06 |
| egonzalez | des_consolado: should not be the case that, whats the issue with cells in deploy? | 16:07 |
| *** erhudy has joined #openstack-kolla | 16:14 | |
| des_consolado | trying to reproduce | 16:20 |
| Jeffrey4l | egonzalez, try my patch. i tested it, it works. | 16:20 |
| *** cinerama` is now known as cinerama | 16:23 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Use ocata tag in config.py https://review.openstack.org/432297 | 16:25 |
| des_consolado | really strange that the jinja2 downgrade fixed my issue with the latest commit on that neutron play but for you it runs fine with latest jinja2 though egonzalez | 16:26 |
| *** yingjun has quit IRC | 16:26 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla master: Use ocata tag in config.py https://review.openstack.org/434370 | 16:28 |
| *** crushil has quit IRC | 16:29 | |
| *** crushil has joined #openstack-kolla | 16:29 | |
| klindgren | gema, let me know if you have any ideas on healtchecking RPC containers. My Idea so far is basically 3 fold. 1.) For API's add the docker HEALTCHECK command 2.) For RPC services that support systemd-notify see if we can use that to alteast get the "I started ok" functionality, if the support morethan notify_once it would be possible to use that as a "ping" | 16:30 |
| gema | klindgren: yep, will think about it | 16:30 |
| *** sp__ has quit IRC | 16:30 | |
| gema | klindgren: can you give me an example of those? | 16:30 |
| gema | from our list | 16:30 |
| gema | so that I can look at the actual code | 16:30 |
| klindgren | 3.) Generic test to see if the container restarted with in ~10-20 seconds after being started. | 16:30 |
| klindgren | so oslo.service provides systemd_notify: https://github.com/openstack/oslo.service/blob/master/oslo_service/systemd.py#L59-L67 | 16:31 |
| gema | klindgren: makes me nervous to put such a time constraint to the overloaded gates | 16:31 |
| gema | klindgren: great, thanks | 16:31 |
| des_consolado | oh wtf... I just re-produced the error from before, the "Setting sysctl values" play one, using Jinja2 2.8.1 | 16:32 |
| vhosakot | bmace: sdake: my update.. haven't got a chance to test bmace's patch set yet... I'll keep you posted once I test it... | 16:32 |
| vhosakot | brb 15 mins | 16:33 |
| *** vhosakot has quit IRC | 16:33 | |
| gema | klindgren: sorry I am at the interop wg meeting as well, will look into this tomorrow :D | 16:34 |
| klindgren | gema, kk - also here is something that uses it: https://github.com/openstack/neutron/blob/e6ee6ad8ee0f392fa1700a86561c5934b94cac70/neutron/plugins/ml2/drivers/openvswitch/agent/ovs_neutron_agent.py#L322 | 16:37 |
| gema | klindgren: ack, thanks :D | 16:37 |
| *** adrian_otto has quit IRC | 16:39 | |
| *** zioproto has quit IRC | 16:39 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Disable elasticsearch output when enable_elasticsearch is no in fluentd https://review.openstack.org/434128 | 16:40 |
| openstackgerrit | Merged openstack/kolla-ansible master: Update the CONTRIBUTING.rst file https://review.openstack.org/429176 | 16:43 |
| kfox1111 | have they sent out atc's for the summit yet? | 16:44 |
| ccrouch | can someone point me at the original source for the mariadb configmap which is generated when deploying kolla-kubernetes? I thought it was somewhere in kolla-ansible but not sure? | 16:44 |
| *** mgoddard has quit IRC | 16:47 | |
| inc0 | des_consolado: you can set_sysctl: "no" in globals | 16:48 |
| *** hieulq_ has joined #openstack-kolla | 16:49 | |
| hrw | (venv-for-kolla)17:49 hrw@pinkiepie-centos:kolla$ for distro in debian ubuntu;do for type in binary source;do ./tools/build.py -b $distro -t $type -r1 --logs-dir logi/$distro-$type --base-image aarch64/$distro;done;done | 16:50 |
| des_consolado | inc0: presumably some functionality is lost with this though? | 16:51 |
| hrw | let's go! | 16:51 |
| *** zhubingbing has quit IRC | 16:51 | |
| inc0 | des_consolado: well we need to run few sysctl changes like make sure ipv4 forwarding is set | 16:53 |
| inc0 | for that you need to run tasks with become | 16:53 |
| inc0 | which we don't in this case, maybe worth adding a bug for it | 16:53 |
| hrw | krtaylor: https://github.com/hrw/kolla/tree/hrw/0215-test-all if you have some spare ppc64le cycles | 16:53 |
| *** srwilkers has quit IRC | 16:56 | |
| *** fragatina has joined #openstack-kolla | 16:56 | |
| * krtaylor looking | 16:56 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 16:57 |
| hrw | krtaylor: it should work. does not use any external repos nor arch tweaks | 16:58 |
| *** jascott1_ has quit IRC | 16:58 | |
| *** adrian_otto has joined #openstack-kolla | 16:59 | |
| *** jascott1_ has joined #openstack-kolla | 16:59 | |
| hrw | krtaylor: would be probably quicker than me requesting ppc64le machine at Red Hat | 16:59 |
| krtaylor | hehheh, hm... | 16:59 |
| krtaylor | you'd prob get one faster than me :) | 17:00 |
| hrw | krtaylor: btw - when z/arch support? | 17:01 |
| krtaylor | I am working on setting up an environment for testing atm (for ppc64le) | 17:01 |
| *** crushil has quit IRC | 17:01 | |
| *** rwellum has joined #openstack-kolla | 17:01 | |
| *** crushil has joined #openstack-kolla | 17:02 | |
| * krtaylor looks for andymaier for z | 17:02 | |
| jmccarthy | Hmm anyone familiar with barbican and this rpm uwsgi-plugin-python ? I notice I need to yum install it for my build | 17:02 |
| openstackgerrit | Dan Prince proposed openstack/kolla master: Add heat-all container https://review.openstack.org/433299 | 17:02 |
| *** penfold has quit IRC | 17:02 | |
| *** fragatina has quit IRC | 17:03 | |
| *** jascott1_ has quit IRC | 17:03 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 17:05 |
| hrw | jmccarthy: barbican is something I will have to look at as it ftfbs on debian-source | 17:05 |
| des_consolado | well now I'm getting this, which is similar http://termbin.com/4g7n | 17:06 |
| des_consolado | it doesn't understand this neutron_services thing | 17:07 |
| *** srwilkers has joined #openstack-kolla | 17:08 | |
| jmccarthy | hrw: Ok trying to work out from where I need to try and get this rpm built from | 17:08 |
| des_consolado | but the neutron_services are defined in defaults/main.yml D: | 17:08 |
| *** Serlex has quit IRC | 17:09 | |
| hrw | jmccarthy: I built barbican-* for centos, debian/jessie-backports, ubuntu in both binary and source builds. debian/stretch only in binary one | 17:09 |
| jmccarthy | hrw: I'm trying to build barbican images for oraclelinux at the moment, which is in with the conditional for centos and rhel in the Dockerfile.j2 | 17:10 |
| hrw | jmccarthy: oraclelinux7? | 17:11 |
| jmccarthy | hrw: Yep | 17:11 |
| hrw | jmccarthy: it is in epel7 | 17:11 |
| jmccarthy | hrw: Hmm and source rpms ? I haven't looked there | 17:12 |
| hrw | jmccarthy: should be there as well | 17:12 |
| jmccarthy | hrw: Ok will take a look :) ! | 17:12 |
| *** egonzalez has quit IRC | 17:12 | |
| hrw | jmccarthy: https://koji.fedoraproject.org/koji/buildinfo?buildID=833779 | 17:13 |
| jmccarthy | hrw: Nice - thanks ! | 17:15 |
| hrw | jmccarthy: I spent far too mcuh time around koji during last 3 years ;D | 17:16 |
| des_consolado | it's definitely to do with this host_in_groups stuff | 17:17 |
| *** jascott1_ has joined #openstack-kolla | 17:24 | |
| *** jascott1_ has quit IRC | 17:24 | |
| hrw | krtaylor: I looked at original ppc64le patch on top of mine. it is fugly | 17:27 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix chrony genconfig https://review.openstack.org/434412 | 17:28 |
| *** duonghq has quit IRC | 17:31 | |
| hrw | krtaylor: and it's part which touch docker/ directory needs to be rebased on top of my 'enable all images for debian' patch | 17:34 |
| *** rwallner has quit IRC | 17:35 | |
| *** rwallner has joined #openstack-kolla | 17:35 | |
| *** rwallner has quit IRC | 17:35 | |
| openstackgerrit | OpenStack Release Bot proposed openstack/kolla-ansible master: Update reno for stable/ocata https://review.openstack.org/434419 | 17:36 |
| *** sudipto_ has quit IRC | 17:38 | |
| *** sudipto has quit IRC | 17:38 | |
| *** unicell has joined #openstack-kolla | 17:40 | |
| *** rwallner has joined #openstack-kolla | 17:40 | |
| *** spiette has joined #openstack-kolla | 17:41 | |
| *** rwallner has quit IRC | 17:41 | |
| *** athomas has quit IRC | 17:42 | |
| bpattewar_ | I am trying to setup my dev environement. I could deploy rabbitmq, memcached, mariadb, keystone. But after keystone I am hitting following error + /opt/kolla-kubernetes/tests/bin/endpoint_test.sh Discovering versions from the identity service failed when creating the password plugin. Attempting to determine version from URL. Unable to establish connection to http://172.16.35.11:5000/v3/auth/tokens: HTTPConnectionPool(ho | 17:43 |
| bpattewar_ | Anyone faced this issue ? | 17:43 |
| *** unicell has quit IRC | 17:47 | |
| *** eaguilar has quit IRC | 17:49 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: support non-x86 architectures (aarch64, ppc64le) https://review.openstack.org/430940 | 17:49 |
| *** mannamne has quit IRC | 17:50 | |
| *** rwellum has quit IRC | 17:51 | |
| *** crushil has quit IRC | 17:51 | |
| *** crushil has joined #openstack-kolla | 17:51 | |
| *** vhosakot has joined #openstack-kolla | 17:53 | |
| *** penfold has joined #openstack-kolla | 17:55 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: install gnupg and dirmngr needed for apt-key https://review.openstack.org/434431 | 17:55 |
| *** crushil has quit IRC | 17:56 | |
| *** crushil has joined #openstack-kolla | 17:56 | |
| *** saneax-_-|AFK is now known as saneax | 17:57 | |
| *** rwallner has joined #openstack-kolla | 17:59 | |
| *** mgoddard has joined #openstack-kolla | 17:59 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 17:59 |
| *** l4yerffeJ has quit IRC | 18:00 | |
| *** srwilkers has quit IRC | 18:01 | |
| *** hieulq_ has quit IRC | 18:02 | |
| *** ccrouch has quit IRC | 18:03 | |
| *** sambetts is now known as sambetts|afk | 18:04 | |
| *** gfidente is now known as gfidente|afk | 18:10 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 18:11 |
| *** jeffrey4l_ircc has joined #openstack-kolla | 18:11 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: enable all images enabled for Ubuntu https://review.openstack.org/432787 | 18:11 |
| *** tonanhngo has joined #openstack-kolla | 18:12 | |
| hrw | hm. need to make branch with 3 patches one after another and then send them again for review | 18:13 |
| *** JoseMello has quit IRC | 18:15 | |
| hrw | ok, let gerrit handle it now | 18:15 |
| *** tonanhngo has quit IRC | 18:16 | |
| *** tonanhngo has joined #openstack-kolla | 18:16 | |
| *** unicell has joined #openstack-kolla | 18:17 | |
| *** adrian_otto has quit IRC | 18:17 | |
| *** dtturner has joined #openstack-kolla | 18:20 | |
| *** tonanhngo has quit IRC | 18:21 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: enable all images enabled for Ubuntu https://review.openstack.org/432787 | 18:21 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: Add support for non-x86 architectures (aarch64, ppc64le) https://review.openstack.org/430940 | 18:21 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: debian: move to stretch https://review.openstack.org/434453 | 18:21 |
| des_consolado | Jeffrey4l: I don't get it D: I'm still having this issue and nobody else is | 18:22 |
| *** tyrola has quit IRC | 18:24 | |
| *** tyrola has joined #openstack-kolla | 18:25 | |
| *** jheroux has quit IRC | 18:27 | |
| hrw | have a nice day | 18:27 |
| krtaylor | hrw, I agree with your "fugly" comment, but I also do not want to hold this up waiting on my setup for testing | 18:28 |
| krtaylor | maybe split aarch64 and ppc64le support patches? | 18:29 |
| *** fragatina has joined #openstack-kolla | 18:34 | |
| *** ipsecguy has joined #openstack-kolla | 18:35 | |
| *** adrian_otto has joined #openstack-kolla | 18:35 | |
| *** adrian_otto has quit IRC | 18:40 | |
| *** tonanhngo has joined #openstack-kolla | 18:43 | |
| *** salv-orlando has quit IRC | 18:44 | |
| *** MasterOfBugs has joined #openstack-kolla | 18:45 | |
| jascott1 | ugh anyone know how to tweak a mac so networking does get foobar'd when you lock it? | 18:46 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 18:48 |
| *** tonanhngo has quit IRC | 18:48 | |
| *** saneax is now known as saneax-_-|AFK | 18:48 | |
| *** tonanhngo has joined #openstack-kolla | 18:48 | |
| *** rwallner has quit IRC | 18:52 | |
| *** tonanhngo has quit IRC | 18:53 | |
| *** srwilkers has joined #openstack-kolla | 18:54 | |
| *** rwallner has joined #openstack-kolla | 18:56 | |
| *** mgoddard has quit IRC | 18:59 | |
| *** pcaruana has quit IRC | 19:03 | |
| *** Serlex has joined #openstack-kolla | 19:05 | |
| *** MasterOfBugs has quit IRC | 19:07 | |
| *** MasterOfBugs has joined #openstack-kolla | 19:07 | |
| *** adrian_otto has joined #openstack-kolla | 19:08 | |
| *** shardy has quit IRC | 19:10 | |
| jascott1 | proxies make baby koala's cry | 19:13 |
| jgriffith | anybody around that knows how the iscsid containers work? I'm trying to figure out how to use Kolla to depoly just cinder so I can test | 19:14 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 19:14 |
| jgriffith | inparticular the bug that sdake and Serguei logged a long time ago | 19:14 |
| jgriffith | I have no idea how/where the actual docker build and run commands are hidden | 19:15 |
| *** sdake_ has joined #openstack-kolla | 19:22 | |
| *** sdake has quit IRC | 19:26 | |
| *** sdake has joined #openstack-kolla | 19:27 | |
| *** tonanhngo has joined #openstack-kolla | 19:29 | |
| *** sdake_ has quit IRC | 19:30 | |
| *** srwilkers has quit IRC | 19:32 | |
| *** tonanhngo has quit IRC | 19:34 | |
| *** ccrouch has joined #openstack-kolla | 19:34 | |
| *** tonanhngo has joined #openstack-kolla | 19:36 | |
| *** fragatina has quit IRC | 19:37 | |
| *** fragatina has joined #openstack-kolla | 19:37 | |
| jascott1 | jgriffith i do not. you might check the gate script | 19:40 |
| *** tonanhngo has quit IRC | 19:40 | |
| *** tonanhngo has joined #openstack-kolla | 19:42 | |
| *** tonanhngo has quit IRC | 19:47 | |
| *** tonanhngo has joined #openstack-kolla | 19:48 | |
| jgriffith | jascott1 good idea | 19:49 |
| jascott1 | random complaint: the quickstart guide should really include kolla as a prereq | 19:49 |
| jascott1 | imo biggest pain in docs 6 months ago when I started sleeping on the floor around here is the inter linked "go do everything in this doc" and then "do this" | 19:51 |
| jascott1 | kolla-k8s needs a hit list of steps | 19:51 |
| *** tonanhngo has quit IRC | 19:52 | |
| jgriffith | jascott1 ha! that answers that, just don't run said services :) | 19:53 |
| jgriffith | nova/glance/neutron only it looks like (and required projects for them) | 19:53 |
| inc0 | jgriffith: sbezverk will know:) | 19:56 |
| jgriffith | sbezverk oh, you're online now :) | 19:56 |
| inc0 | but from my experience, ceph is better in general;) | 19:57 |
| jgriffith | inc0 until you actually have to run it | 19:57 |
| jgriffith | inc0 but i'm not here to get in that ridiculous argument | 19:57 |
| inc0 | ironically it's easier for kolla than iscsid | 19:57 |
| jgriffith | sure | 19:58 |
| inc0 | however I get that, we support both | 19:58 |
| inc0 | I just run ceph whenever I need cinder | 19:58 |
| inc0 | (in fact I have one running now for k8s) | 19:58 |
| jgriffith | yeah, as long as we ignore everything else that works out pretty well | 19:59 |
| inc0 | well, anyway, what is exactly your issue with iscsi there? | 19:59 |
| *** tonanhngo has joined #openstack-kolla | 20:00 | |
| jgriffith | inc0 well, iscsi works by interacting with the devices on the node iteself of course, and doing things like putting targets in from of them and mounts etc | 20:00 |
| jgriffith | inc0 my issue is, doing that in containers has a number of challenges, there are basically 3 ways to do it that I know of | 20:00 |
| jgriffith | Kolla inparticular figured out what I have determined is probably the most elegant, but also the most difficult way | 20:01 |
| jgriffith | running tgtd and iscsid in a container | 20:01 |
| inc0 | yeah, we run it all inside containers | 20:01 |
| inc0 | https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/iscsi/tasks/start.yml | 20:01 |
| jgriffith | what I have never been able to figure out when one does that is HOW the nova container for example accesses the iscsiadm cmds that are in some other container? | 20:02 |
| inc0 | if you look at volumes, this is what we mount from host | 20:02 |
| inc0 | it won't be nova, it will be libvirt | 20:02 |
| jgriffith | Unless you just link the entrypoint of the iscsid container to a directory that you then pass in to the containers that you want to use it | 20:02 |
| inc0 | so i'd guess that libvirt uses socket | 20:02 |
| jgriffith | ? | 20:03 |
| inc0 | so, we mount /dev and /run from host to container | 20:03 |
| jgriffith | inc0 ahh... this helps | 20:04 |
| jgriffith | inc0 so yes, it's passing in the local mount points for things like dev etc | 20:04 |
| jgriffith | inc0 now that you've pointed me in the right direction maybe I can figure out how nova is actually consuming this :) | 20:04 |
| inc0 | yeah, and we pass both /dev and /run to both containers | 20:04 |
| inc0 | so I'd guess it's libvirt rather than nova really | 20:05 |
| *** tonanhngo has quit IRC | 20:05 | |
| jgriffith | inc0 yeah.. so what's in /run ? | 20:05 |
| inc0 | https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/nova/defaults/main.yml#L13-L23 | 20:05 |
| jgriffith | DErP :) | 20:05 |
| *** tonanhngo has joined #openstack-kolla | 20:06 | |
| inc0 | jgriffith: /run is usually for shared memory, pids and sockets | 20:06 |
| inc0 | so I'd look there | 20:06 |
| jgriffith | inc0 ok, so yeah if you look down at L#123 the /etc/iscsi dir is passed in from the host | 20:07 |
| jgriffith | so you have everything in the nova container, /dev and /etc/iscsi... the only piece I'm still missing is the actual iscsiadm cmd | 20:07 |
| inc0 | well libvirt might not use it | 20:08 |
| jgriffith | inc0 BTW, what I'm trying to do is break all of this down to deploy just cinder by itself, with LVM iscsi support | 20:08 |
| inc0 | it might communicate over raw socket | 20:08 |
| jgriffith | inc0 well that's not how nova works though | 20:09 |
| inc0 | rly depends on libvirt/qemu internals | 20:09 |
| jgriffith | inc0 nova calls cinder, asks for connection info, make an iscsi connection and then passes the device into libvirt | 20:09 |
| jgriffith | inc0 for nova libvirt doesn't talk directly to the iscsi pieces, it relies on user space to set everything up | 20:10 |
| *** tonanhngo has quit IRC | 20:10 | |
| jgriffith | inc0 one nice thing with RBD is it doesn't need the user space efforts, can just use the kernel | 20:10 |
| inc0 | jgriffith: https://github.com/openstack/kolla/blob/master/docker/nova/nova-compute/Dockerfile.j2#L17 | 20:11 |
| jgriffith | inc0 HA! Looks like I'm going to have to buy you a beer! | 20:12 |
| inc0 | jgriffith: looks like it | 20:12 |
| jgriffith | so that's the piece I was looking for | 20:12 |
| *** tonanhngo has joined #openstack-kolla | 20:12 | |
| inc0 | there are variants in ubutnu too, al in this file tho | 20:12 |
| jgriffith | although now the question is WTF is with the jinja syntax in the dockerfiles :) | 20:12 |
| inc0 | well, that's how we roll | 20:12 |
| jgriffith | inc0 yeah, that's just because the package names are different | 20:12 |
| inc0 | that's how you can have multiple distros and lots of nifty features | 20:13 |
| inc0 | like customization allowing you to install your own cinder driver stuff if you need it;) | 20:13 |
| jgriffith | although, kinda confusing... the iscsi packages are being installed in teh compute container, but there's also that iscsid container out there | 20:13 |
| jgriffith | inc0 haha... touchet | 20:13 |
| Serlex | bmace ping | 20:13 |
| inc0 | well you still need iscsid and tgtd running on host | 20:14 |
| inc0 | (or in container) | 20:14 |
| jgriffith | inc0 well, what I'm getting at is that kolla has seperate containers for those two things | 20:14 |
| inc0 | ofc | 20:15 |
| inc0 | well these services runs on different hosts really | 20:15 |
| jgriffith | inc0 so if theyr'e in the container that uses them, why are they external as well.. and how are those linked | 20:15 |
| inc0 | but kolla has 1 process per container across the board | 20:15 |
| jgriffith | inc0 right, which is AWESOME! | 20:15 |
| inc0 | well it's kinda funny when you consider that vms also runs in container, but that's how we roll;) | 20:16 |
| jgriffith | that's why I'm saying it's not actually running iscsid in the compute container, it's in that iscsid container... but I can't figure out how it executes that cmd in the iscsid container | 20:16 |
| jgriffith | inc0 haha... there's that :) | 20:16 |
| inc0 | that being said, if you kill docker, vms stay alive | 20:16 |
| jgriffith | inc0 my workstation is set up that way with just kvm.. running kvm in a container, works gerat | 20:16 |
| *** tonanhngo has quit IRC | 20:17 | |
| bmace | Serlex: hey, whats up? | 20:17 |
| inc0 | well you don't run kvm in container;) but you do run qemu | 20:17 |
| jgriffith | inc0 yeah, because you are probably privelleged then and running on the node itself, just executing things in the conatiner I'm assuming ? | 20:17 |
| inc0 | it's a bit of wizardry that is | 20:17 |
| inc0 | qemu + kvm is kernel-owned process | 20:17 |
| inc0 | so docker, since it doesn't meddle with kernel | 20:17 |
| jgriffith | inc0 right, and that's shared, so... there ya have it | 20:18 |
| inc0 | doesn't kill it, even tho it would like to | 20:18 |
| inc0 | fun fact, if you run qemu without kvm | 20:18 |
| inc0 | it will actually kill a vm | 20:18 |
| *** tonanhngo has joined #openstack-kolla | 20:18 | |
| inc0 | so thats a bit scary, however nobody who cares about vms runs qemu without kvm | 20:18 |
| Serlex | bmace Im running the second script, "run_dev_image.sh" "/root/.kube/config: permission denied" | 20:18 |
| Serlex | didnt retrieve the k8s credentials properly? | 20:19 |
| inc0 | so to the topic of iscid, I can't help you about internals, but sbezverk is person to talk to | 20:19 |
| bmace | Serlex: if you look at the script it mounts in your .kube/config from the home dir of the user you are running the container as. what are the permissions on your .kube and .kube/config outside the container? | 20:20 |
| bmace | Serlex: for me, my .kube is 755 and .kube/config is 644 | 20:21 |
| *** kbyrne has quit IRC | 20:22 | |
| *** salv-orlando has joined #openstack-kolla | 20:22 | |
| *** tonanhngo has quit IRC | 20:23 | |
| *** cliles_ has quit IRC | 20:23 | |
| Serlex | same here, i'm running everything as root thought | 20:23 |
| Serlex | trying again | 20:24 |
| *** cliles has joined #openstack-kolla | 20:24 | |
| *** salv-orlando has quit IRC | 20:26 | |
| bmace | Serlex: if it fails you can change the entrypoint to /bin/bash and debug the permissions on that file before you run /opt/kolla-kubernetes/tools/setup_dev_env.sh | 20:27 |
| *** jeffrey4l_ircc has quit IRC | 20:27 | |
| Serlex | ok | 20:30 |
| *** satyar has quit IRC | 20:30 | |
| *** rwallner has quit IRC | 20:31 | |
| *** rwallner has joined #openstack-kolla | 20:32 | |
| *** rwallner has quit IRC | 20:32 | |
| *** rwallner has joined #openstack-kolla | 20:33 | |
| jgriffith | inc0 hey, thanks for the conversation and pointers, appreciate it | 20:33 |
| *** jtriley has quit IRC | 20:39 | |
| *** jtriley has joined #openstack-kolla | 20:39 | |
| *** tonanhngo has joined #openstack-kolla | 20:41 | |
| *** sdake has quit IRC | 20:41 | |
| inc0 | any time jgriffith | 20:42 |
| *** tonanhngo_ has joined #openstack-kolla | 20:43 | |
| *** tonanhngo has quit IRC | 20:45 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-ansible master: Updated from global requirements https://review.openstack.org/431962 | 20:46 |
| *** sdake has joined #openstack-kolla | 20:47 | |
| *** sdake has quit IRC | 20:47 | |
| *** rwallner has quit IRC | 20:54 | |
| -openstackstatus- NOTICE: We're currently battling an increase in log volume which isn't leaving sufficient space for new jobs to upload logs and results in POST_FAILURE in those cases; recheck if necessary but keep spurious rebasing and rechecking to a minimum until we're in the clear. | 20:56 | |
| *** mannamne has joined #openstack-kolla | 21:01 | |
| *** rwallner has joined #openstack-kolla | 21:08 | |
| kfox1111 | https://review.openstack.org/#/c/410641/ should be ready. | 21:11 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 21:11 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Mon https://review.openstack.org/410660 | 21:17 |
| openstackgerrit | Saverio Proto proposed openstack/kolla master: Fix 404 to documentation in the Readme https://review.openstack.org/434522 | 21:18 |
| openstackgerrit | Merged openstack/kolla-ansible master: Update reno for stable/ocata https://review.openstack.org/434419 | 21:18 |
| *** rwallner has quit IRC | 21:19 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Optimize reconfigure action for congress https://review.openstack.org/422690 | 21:21 |
| *** rwallner has joined #openstack-kolla | 21:22 | |
| sbezverk | kfox1111: have you ever tried to add 2 externalIP to the service? | 21:23 |
| *** rwallner_ has joined #openstack-kolla | 21:24 | |
| *** zioproto has joined #openstack-kolla | 21:25 | |
| zioproto | hello | 21:25 |
| kfox1111 | sbezverk: I tried once on another project, though a while ago. I wana say it worked fine. | 21:26 |
| *** rwallner has quit IRC | 21:26 | |
| *** rwallner_ has quit IRC | 21:29 | |
| *** Pavo has quit IRC | 21:29 | |
| openstackgerrit | Saverio Proto proposed openstack/kolla-ansible master: Fix BASEDIR because of new kolla-ansible project name https://review.openstack.org/434526 | 21:31 |
| sbezverk | kfox1111: great thanks, I need to use it for ironic-api | 21:32 |
| *** Jeffrey4l has quit IRC | 21:34 | |
| *** Jeffrey4l has joined #openstack-kolla | 21:35 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Mon https://review.openstack.org/410660 | 21:40 |
| *** haplo37_ has quit IRC | 21:41 | |
| *** haplo37_ has joined #openstack-kolla | 21:41 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Mon https://review.openstack.org/410660 | 21:44 |
| *** sdake has joined #openstack-kolla | 21:46 | |
| *** zioproto has quit IRC | 21:49 | |
| *** ipsecguy_ has joined #openstack-kolla | 21:51 | |
| *** crushil has quit IRC | 21:51 | |
| *** ipsecguy has quit IRC | 21:52 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 22:07 |
| *** jtriley_ has joined #openstack-kolla | 22:07 | |
| *** jtriley has quit IRC | 22:10 | |
| *** jtriley_ has quit IRC | 22:13 | |
| *** srwilkers has joined #openstack-kolla | 22:16 | |
| jgriffith | sbezverk ping | 22:19 |
| *** srwilkers has quit IRC | 22:20 | |
| *** salv-orlando has joined #openstack-kolla | 22:23 | |
| sbezverk | jgriffith: pong, how are you? long time no talk | 22:27 |
| *** mgiles has quit IRC | 22:27 | |
| jgriffith | sbezverk hey, yes, it's been a VERY long time! | 22:27 |
| jgriffith | sbezverk I was wondering if you'll be at the PTG next week? | 22:27 |
| jgriffith | It would be good to finally catch up if possible | 22:27 |
| sbezverk | jgriffith: unfortunately no :( | 22:28 |
| jgriffith | I'm still trying to work out how you figured out using the external iscsi containers :) | 22:28 |
| jgriffith | sbezverk oh, that's dissappointing! | 22:28 |
| *** salv-orlando has quit IRC | 22:28 | |
| sbezverk | jgriffith: yeah. management thinks travelling 4 times a year is too much | 22:28 |
| jgriffith | sbezverk yeah, management is *funny* about those things lately it seems | 22:29 |
| jgriffith | sbezverk well I'm bummed to hear that | 22:29 |
| sbezverk | jgriffith: but boston is a go :) so we should make sure to hook up | 22:29 |
| jgriffith | what is it two years of IRC and Webex now :) | 22:29 |
| jgriffith | Yes, Boston for sure! | 22:29 |
| jgriffith | In the meantime, I guess I'm wondering if you could help me understand the usage of said containers. Are you just accomplishing this just by sharing the directories across all the containers? | 22:30 |
| sbezverk | jgriffith: now we have iscsi on kubernetes ;) works similarely as in kolla-ansible | 22:30 |
| jgriffith | sbezverk haha! I haven't even caught up on the ansible stuff yet :) | 22:30 |
| sbezverk | jgriffith: I have not worked on ansible for a while, so we are at the same level now ;) | 22:31 |
| jgriffith | sbezverk so IIRC and based on some looking and chatting with inc0 earlier... the way that works is something like: | 22:31 |
| jgriffith | ha!!! | 22:31 |
| sbezverk | jgriffith: but I would be happy to try to help | 22:31 |
| *** rhallisey has quit IRC | 22:32 | |
| jgriffith | run the iscsiadm container and share/mount the iscsiadm cmd and device directories on the node | 22:32 |
| jgriffith | launch the nova compute node, and pass in that shared dir there as well | 22:32 |
| jgriffith | but how does nova actually "call" iscsiadm? | 22:32 |
| jgriffith | the new AllDevices thing in Docker is super slick, but i can't get past how to call the executable when it's in another container? | 22:33 |
| sbezverk | jgriffith: hmm, I remember seeing some references to iscsiadm in nova compute code | 22:33 |
| jgriffith | unless you share it somewhere and modify the path variable in the compute container | 22:33 |
| jgriffith | sbezverk yeah, nova-compute does everything by making calls to iscsiadm to set up the connection etc | 22:33 |
| jgriffith | I'm trying to figure out "how" it's calling that :) | 22:34 |
| sbezverk | jgriffith: iscsiadm is bundled into the container image | 22:34 |
| jgriffith | only thing I can think of is the shared volume thing and an env variable to change the path, or even a symbolic link I guess | 22:34 |
| jgriffith | sbezverk right | 22:34 |
| jgriffith | so how is nova container issuing that command :) | 22:34 |
| jgriffith | containers can't call executables in other containers (yet) :) | 22:34 |
| sbezverk | jgriffith: man, but it is local | 22:35 |
| sbezverk | iscsiadm software IS in nova-compute container | 22:35 |
| jgriffith | sbezverk oh! hurumph | 22:35 |
| jgriffith | so what's the purpose of the iscsid container? | 22:35 |
| sbezverk | when we build nova-compute we add iscsi-initiator-utils | 22:36 |
| jgriffith | sbezverk ahh, yes, I see that! | 22:36 |
| sbezverk | jgriffith: that one runs daemon | 22:36 |
| sbezverk | iscsid | 22:36 |
| jgriffith | haha! | 22:36 |
| hrw | krtaylor: there is nearly nothing ppc64le related in my patch. all needed tweaking is in images. and I would prefer to get info how it went first before working on them. | 22:36 |
| sbezverk | it is required in order tp iscsiadm command to work | 22:36 |
| jgriffith | sbezverk ahaha! I think I get it now | 22:37 |
| sbezverk | cause it check for a process id and socket for iscsid | 22:37 |
| hrw | krtaylor: especially when it means applying it on top of other 3 patches which are required to even start | 22:37 |
| jgriffith | sbezverk so iscsid is in fact JUST the deamon, which then exposes the socket to the nova container | 22:37 |
| jgriffith | which has iscsiutils (iscsiadm) and so that's why it works | 22:37 |
| sbezverk | jgriffith: yep you got it | 22:37 |
| jgriffith | sbezverk well that was easy :) | 22:38 |
| jgriffith | sbezverk thank you!!! | 22:38 |
| sbezverk | jgriffith: anytime | 22:38 |
| jgriffith | I've been messing with approaches that use chroot and other tricks, and it's not very nice | 22:38 |
| sbezverk | jgriffith: let's talk about external iscsi driver for cinder!!!! | 22:39 |
| jgriffith | sbezverk YES! | 22:39 |
| sbezverk | that would be super super useful | 22:39 |
| jgriffith | I'm actually getting closer to having that | 22:39 |
| jgriffith | I have an http interface almost done | 22:39 |
| sbezverk | jgriffith: if you need any additional testing, I could try to help. I have external iscsi storage which I use | 22:40 |
| jgriffith | so it will be just like all the other cinder-drivers | 22:40 |
| jgriffith | sbezverk ok, I'll take you up on that! | 22:40 |
| jgriffith | I don't think I'll have it wrapped up and ready until probably Boston though I'm afraid :( | 22:41 |
| jgriffith | oh crap, I have a meeting I'm late for! | 22:41 |
| jgriffith | sorry, let's catch up later tonight or tomorrow | 22:41 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Mon https://review.openstack.org/410660 | 22:43 |
| *** gfidente|afk has quit IRC | 22:45 | |
| sbezverk | jgriffith: sure thing, ping me when you can | 22:46 |
| berendt | i finally got nova and neutron working with kolla and xenserver :) | 22:51 |
| berendt | i opened a blueprint for the required changes, hope that we can add xenserver as supported hypervisor the next cycle | 22:52 |
| *** dave-mccowan has quit IRC | 22:53 | |
| *** adrian_otto1 has joined #openstack-kolla | 22:59 | |
| *** adrian_otto has quit IRC | 23:01 | |
| *** salv-orlando has joined #openstack-kolla | 23:02 | |
| kfox1111 | sbezverk: https://github.com/kubernetes/dashboard/pull/1332 :) | 23:09 |
| *** lamt has quit IRC | 23:09 | |
| kfox1111 | we really gotta get the icons in at the ptg. :) | 23:09 |
| *** lamt has joined #openstack-kolla | 23:11 | |
| *** lamt has quit IRC | 23:15 | |
| *** adrian_otto1 has quit IRC | 23:17 | |
| *** adrian_otto has joined #openstack-kolla | 23:19 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 23:19 |
| *** mannamne has quit IRC | 23:21 | |
| *** Serlex has quit IRC | 23:22 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: Helm Ceph Mon https://review.openstack.org/410660 | 23:36 |
| jascott1 | anyone working on quickstart atm? | 23:41 |
| jascott1 | (quickstart the guide for kolla-k8s) | 23:42 |
| *** sdake has quit IRC | 23:45 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes master: WIP adding Ironic microservices packages https://review.openstack.org/433322 | 23:46 |
| kfox1111 | jascott1: like, for bare metal, or for a dev env? | 23:46 |
| *** jrobinson has joined #openstack-kolla | 23:47 | |
| jascott1 | i was talking about the existing one... are we splitting it ? | 23:47 |
| *** salv-orlando has quit IRC | 23:47 | |
| jascott1 | sorry not up on the plans for docs | 23:47 |
| sbezverk | kfox1111: yeah, would be cool to be able to associate kolla helm packages with a log icon :) | 23:48 |
| kfox1111 | log? | 23:49 |
| sbezverk | kfox1111: sorry meant dashboard | 23:54 |
| jascott1 | so we have a quickstart guide. I am all for also having ones targeted at dev/prod etc but I think we need a hit list of steps to get everything installed | 23:54 |
| sbezverk | I was going through the log so brain got stuck on this word | 23:54 |
| kfox1111 | sbezverk: ah. ok. :) | 23:56 |
| kfox1111 | jascott1: yeah. agreed. we probably need to start with the devenv stuff and grow it/make it more modular for the more generic deployment case. | 23:57 |
| bmace | agreed. getting it to deploy the bits onto bare metal not just halcyon-vagrant would be good and probably not a huge step.. mostly we need ssh access into the env and the ips / network ports stuff for the config. | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
