| *** diogogmt has joined #openstack-kolla | 00:00 | |
| *** Pavo has joined #openstack-kolla | 00:03 | |
| *** slagle has joined #openstack-kolla | 00:03 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: ovs setup bridge, fix /run, add missing hostipc https://review.openstack.org/369785 | 00:04 |
|---|---|---|
| *** Pavo has quit IRC | 00:04 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: ovs setup bridge, fix /run, add missing hostipc https://review.openstack.org/369785 | 00:05 |
| *** Pavo has joined #openstack-kolla | 00:05 | |
| *** hieulq has joined #openstack-kolla | 00:13 | |
| *** huikang has quit IRC | 00:16 | |
| *** fguillot is now known as fguillot_afk | 00:20 | |
| *** fguillot_afk has quit IRC | 00:20 | |
| openstackgerrit | Takashi Sogabe proposed openstack/kolla: Change the behavior of waiting process for wsrep_cluster_status. https://review.openstack.org/370491 | 00:21 |
| *** eaguilar has joined #openstack-kolla | 00:23 | |
| *** tonanhngo has quit IRC | 00:24 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Fixes to get Nova to work https://review.openstack.org/370459 | 00:25 |
| *** Marcellin__ has quit IRC | 00:27 | |
| *** Pavo has quit IRC | 00:30 | |
| *** williamcaban has joined #openstack-kolla | 00:31 | |
| *** williamcaban has quit IRC | 00:35 | |
| *** tonanhngo has joined #openstack-kolla | 00:36 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Use "type" to take place of "which" and update the output https://review.openstack.org/369551 | 00:37 |
| openstackgerrit | caoyuan proposed openstack/kolla: Use "type" command to take place of "which" https://review.openstack.org/370497 | 00:37 |
| *** tonanhngo has quit IRC | 00:38 | |
| *** tonanhngo has joined #openstack-kolla | 00:39 | |
| *** awiddersheim has joined #openstack-kolla | 00:40 | |
| awiddersheim | trying to integrate kolla into a larger set of playbooks | 00:41 |
| awiddersheim | trying to avoid just calling 'kolla-ansble' using something like 'command' | 00:41 |
| awiddersheim | and would rather just 'include' kolla's playbooks | 00:41 |
| openstackgerrit | caoyuan proposed openstack/kolla: Upate "which" command to "type" https://review.openstack.org/370497 | 00:42 |
| *** tonanhngo has quit IRC | 00:43 | |
| awiddersheim | but I also want to be able to make the version of kolla used dynamic | 00:44 |
| awiddersheim | seems very difficult with ansible since kolla's playbooks can only be included at a top level and not in a task | 00:44 |
| awiddersheim | wondering if anyone has done something like this before and has any recommendations | 00:44 |
| openstackgerrit | caoyuan proposed openstack/kolla: Update "which" command to "type" https://review.openstack.org/370497 | 00:45 |
| *** rkrum has joined #openstack-kolla | 00:46 | |
| *** fguillot has joined #openstack-kolla | 00:46 | |
| *** tonanhngo has joined #openstack-kolla | 00:46 | |
| *** tonanhngo_ has joined #openstack-kolla | 00:49 | |
| *** phuongnh has joined #openstack-kolla | 00:49 | |
| *** JoseMello has joined #openstack-kolla | 00:50 | |
| *** tonanhngo has quit IRC | 00:51 | |
| *** tonanhngo_ has quit IRC | 00:53 | |
| *** duonghq has joined #openstack-kolla | 00:57 | |
| duonghq | morning | 00:57 |
| *** rkrum has quit IRC | 01:02 | |
| britthouser | morning duonghq | 01:06 |
| duonghq | morning britthouser | 01:07 |
| *** tonanhngo has joined #openstack-kolla | 01:16 | |
| *** tonanhngo has quit IRC | 01:20 | |
| *** tonanhngo has joined #openstack-kolla | 01:21 | |
| *** tonanhngo has quit IRC | 01:25 | |
| *** Pavo has joined #openstack-kolla | 01:31 | |
| *** mliima has joined #openstack-kolla | 01:32 | |
| *** tonanhngo has joined #openstack-kolla | 01:33 | |
| *** mliima_ has quit IRC | 01:34 | |
| *** tonanhngo has quit IRC | 01:38 | |
| *** salv-orlando has joined #openstack-kolla | 01:41 | |
| *** ssurana has quit IRC | 01:43 | |
| *** salv-orl_ has quit IRC | 01:43 | |
| *** rbergero1 has joined #openstack-kolla | 01:44 | |
| *** tonanhngo has joined #openstack-kolla | 01:46 | |
| *** tonanhngo has quit IRC | 01:50 | |
| openstackgerrit | Takashi Sogabe proposed openstack/kolla: Change the behavior of waiting process for wsrep_cluster_status. https://review.openstack.org/370491 | 01:52 |
| *** zhubingbing has joined #openstack-kolla | 02:03 | |
| *** unicell has quit IRC | 02:06 | |
| *** tonanhngo has joined #openstack-kolla | 02:08 | |
| *** huikang has joined #openstack-kolla | 02:09 | |
| *** diogogmt has quit IRC | 02:09 | |
| *** tonanhngo has quit IRC | 02:13 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: add renote monitoring https://review.openstack.org/370345 | 02:16 |
| *** fguillot is now known as fguillot_afk | 02:19 | |
| *** williamcaban has joined #openstack-kolla | 02:19 | |
| *** fguillot_afk has quit IRC | 02:19 | |
| *** tonanhngo has joined #openstack-kolla | 02:19 | |
| *** awiddersheim has quit IRC | 02:21 | |
| *** fguillot has joined #openstack-kolla | 02:23 | |
| *** williamcaban has quit IRC | 02:23 | |
| *** tonanhngo has quit IRC | 02:24 | |
| *** zhubingbing has quit IRC | 02:27 | |
| *** huikang has quit IRC | 02:30 | |
| *** zhubingbing has joined #openstack-kolla | 02:30 | |
| *** huikang has joined #openstack-kolla | 02:30 | |
| *** absubram has quit IRC | 02:31 | |
| *** huikang has quit IRC | 02:35 | |
| *** ssurana has joined #openstack-kolla | 02:37 | |
| *** huikang has joined #openstack-kolla | 02:37 | |
| *** inc0 has joined #openstack-kolla | 02:43 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add etcd ansible role https://review.openstack.org/364662 | 02:43 |
| openstackgerrit | Merged openstack/kolla: Add ansible role for vmtp container https://review.openstack.org/358305 | 02:44 |
| *** zhubingbing has quit IRC | 02:45 | |
| inc0 | hello | 02:46 |
| *** yuanying has quit IRC | 02:47 | |
| duonghq | hello inc0 | 02:49 |
| *** daneyon has joined #openstack-kolla | 03:00 | |
| *** rbergero1 is now known as rbergeron | 03:00 | |
| *** rbergeron has joined #openstack-kolla | 03:00 | |
| *** sdake has joined #openstack-kolla | 03:01 | |
| sdake | evening folks | 03:02 |
| *** daneyon has quit IRC | 03:04 | |
| duonghq | evening sdake | 03:06 |
| sdake | hey | 03:06 |
| *** lyncos has quit IRC | 03:09 | |
| duonghq | sdake: I see some wishlist bug is targated to rc1 | 03:15 |
| sdake | just woek up | 03:16 |
| sdake | eating | 03:16 |
| sdake | be around in a bit | 03:16 |
| duonghq | iirc, there is a special name for this sleep in English or in Arizone? | 03:17 |
| openstackgerrit | Takashi Sogabe proposed openstack/kolla: Change the behavior of waiting process for wsrep_cluster_status. https://review.openstack.org/370491 | 03:18 |
| sbezverk | sdake: when you have a chance please check this ps: https://review.openstack.org/#/c/370440/ | 03:18 |
| *** ssurana has quit IRC | 03:22 | |
| *** ssurana has joined #openstack-kolla | 03:24 | |
| *** zhubingbing has joined #openstack-kolla | 03:27 | |
| *** ssurana has quit IRC | 03:31 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla: Generate full /etc/hosts on every node https://review.openstack.org/370525 | 03:32 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add etcd ansible role https://review.openstack.org/364662 | 03:32 |
| *** fguillot has quit IRC | 03:33 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add etcd ansible role https://review.openstack.org/364662 | 03:34 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla: Generate full /etc/hosts on every node https://review.openstack.org/370525 | 03:36 |
| *** tonanhngo has joined #openstack-kolla | 03:39 | |
| *** zhubingbing has quit IRC | 03:41 | |
| *** tonanhngo has quit IRC | 03:43 | |
| *** inc0 has quit IRC | 03:45 | |
| *** yuanying has joined #openstack-kolla | 03:51 | |
| *** tonanhngo has joined #openstack-kolla | 04:04 | |
| *** eaguilar has quit IRC | 04:07 | |
| *** williamcaban has joined #openstack-kolla | 04:07 | |
| *** eaguilar has joined #openstack-kolla | 04:09 | |
| *** tonanhngo has quit IRC | 04:12 | |
| *** williamcaban has quit IRC | 04:12 | |
| *** tonanhngo has joined #openstack-kolla | 04:13 | |
| *** mtaylor22 has joined #openstack-kolla | 04:14 | |
| mtaylor22 | hello | 04:14 |
| mtaylor22 | Jeffrey4l_: sorry no chance to work on that RTNETLINK issue, been on-site for the last 2 days | 04:15 |
| *** ssurana has joined #openstack-kolla | 04:17 | |
| sdake | duonghq ya - in arizona we call it a ciesta | 04:17 |
| sdake | siesta. A Spanish word meaning sixth (hour), used to mean a sleep taken in the afternoon, especially after the midday meal. | 04:18 |
| sdake | apprently ican't spell spanish :) | 04:18 |
| duonghq | sdake: is midday meal same as lunch? | 04:19 |
| sdake | for me its a 3pm meal | 04:19 |
| sdake | but ya lunch | 04:19 |
| sdake | duonghq the main objective is to dodge the heat | 04:19 |
| sdake | to be productive when the wheather is cool | 04:20 |
| sdake | I operate differently during the winter | 04:20 |
| duonghq | differenet timetable? | 04:20 |
| sdake | not sure when spanish folks eat lucnh | 04:20 |
| sdake | mexico lacks air conditioning for hte most part | 04:21 |
| sdake | so it really makes a big impaact on their society | 04:21 |
| sdake | mexico's culture slipped into hte us a bit - we are a bit of a melting pot of culturess | 04:21 |
| duonghq | understood | 04:21 |
| sdake | and arizona in particular | 04:22 |
| sdake | nothing ever goes wrong in arizona, but its hotter then hell | 04:22 |
| sdake | thatis why most of the datacenter peering happens in Phoenix | 04:22 |
| duonghq | so you still have 3 meals/day? | 04:22 |
| sdake | god i wish | 04:23 |
| sdake | i struggle with getting breakfast | 04:23 |
| sdake | but iwth kids | 04:23 |
| sdake | its hard to have a proper breakfast | 04:23 |
| sdake | usually busy playing rill seargent in the morning | 04:23 |
| sdake | my breakfast used to be 3 sodas,but I've nearly cut soda out of my diet | 04:24 |
| sdake | so my mornings are pretty low energy because of lack of carbs | 04:24 |
| duonghq | you mean soda drink? | 04:25 |
| sdake | yup | 04:25 |
| sdake | i was on 6-9 a day | 04:25 |
| sdake | no more | 04:25 |
| sdake | feel much better | 04:25 |
| sdake | i typically hae one soda a day with dinner | 04:25 |
| sdake | i made that change at the beginning of the new year | 04:25 |
| sdake | it was a rough adjustment | 04:26 |
| duonghq | American folks consume really much soda | 04:26 |
| sdake | duonghq you got that right | 04:26 |
| sdake | now i drink gataorade or water | 04:26 |
| sdake | gatorade is sugarflavored water | 04:26 |
| sdake | more or leess | 04:27 |
| sdake | now that caffeine is gone from my died | 04:27 |
| sdake | diet | 04:27 |
| duonghq | ya, much more healthy | 04:27 |
| sdake | i plan to cut sugar via drinks from it as wel | 04:27 |
| sdake | (.e. ditchthe gatorage) | 04:27 |
| sdake | the way the body metabolises sugaris nearlly as bad for you as drinking beer | 04:28 |
| sdake | atleasat the crap they put in sodas | 04:28 |
| sdake | real sugaris used in gatorage | 04:28 |
| sdake | rather then the chemical corn byproduct thatis used in sugar | 04:29 |
| sdake | rther used in sodas | 04:29 |
| *** fragatin_ has joined #openstack-kolla | 04:29 | |
| sdake | high fructose corn syrup is sweeterthe n sugar (120% vs 100% for sugar which is the baseline for sweetness) | 04:30 |
| sdake | but it is hell on your liver | 04:30 |
| duonghq | ya | 04:30 |
| sdake | and is used all over the place in us food products | 04:30 |
| sdake | imo it shuld be banned | 04:30 |
| sdake | and real sugar or xylitol should be used | 04:31 |
| sdake | corn syrup is superdangerous - that is whyI think I feel so much better, not the cafeeine | 04:31 |
| duonghq | do you take fastfood by daily manner? | 04:31 |
| sdake | i'd say often | 04:32 |
| sdake | not happy about it | 04:32 |
| sdake | havinga properly cooked meal during daytime is difficult to execute | 04:32 |
| *** fragatina has quit IRC | 04:32 | |
| duonghq | sadly but it's had to prevent fastfood before retire | 04:33 |
| *** fragatin_ has quit IRC | 04:33 | |
| *** coolsvap has joined #openstack-kolla | 04:34 | |
| duonghq | sdake: for the wishlist bugs targeted for rc1: as I understand, it's not needed to be fix in this timeframe? | 04:37 |
| *** zhubingbing has joined #openstack-kolla | 04:39 | |
| sdake | if you think you can fix a bug - fix it | 04:42 |
| sdake | wishlist bugs can sometimes be fetures | 04:42 |
| *** daneyon has joined #openstack-kolla | 04:48 | |
| *** aswadr_ has joined #openstack-kolla | 04:49 | |
| *** unicell has joined #openstack-kolla | 04:51 | |
| *** eaguilar has quit IRC | 04:52 | |
| *** daneyon has quit IRC | 04:53 | |
| *** unicell has quit IRC | 04:57 | |
| *** unicell has joined #openstack-kolla | 04:57 | |
| *** Jeffrey4l__ has joined #openstack-kolla | 05:00 | |
| *** unicell1 has joined #openstack-kolla | 05:01 | |
| *** unicell has quit IRC | 05:02 | |
| *** sdake has quit IRC | 05:03 | |
| *** Jeffrey4l_ has quit IRC | 05:04 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: Integrate neutron-vpnaas-agent into the neutron Ansible role https://review.openstack.org/335502 | 05:08 |
| openstackgerrit | zhubingbing proposed openstack/kolla: add renote monitoring https://review.openstack.org/370345 | 05:14 |
| *** zhubingbing has quit IRC | 05:14 | |
| *** huikang has quit IRC | 05:19 | |
| *** sdake has joined #openstack-kolla | 05:22 | |
| *** fragatina has joined #openstack-kolla | 05:32 | |
| *** fragatina has quit IRC | 05:37 | |
| *** rstarmer has joined #openstack-kolla | 05:48 | |
| rstarmer | does anyone know why defining the --tag parameter to kolla-build doesn't work? I want to build 2.0.3 from binaries, but I only get 2.0.2 | 05:49 |
| *** williamcaban has joined #openstack-kolla | 05:55 | |
| *** msimonin has quit IRC | 05:58 | |
| *** williamcaban has quit IRC | 06:00 | |
| *** egonzalez90 has joined #openstack-kolla | 06:01 | |
| bdaca | rstarmer, can you provide more info about it? | 06:05 |
| bdaca | When you use --tag parameter in kolla-build, it only create docker images with this tag (you can specify anything in it) | 06:06 |
| bdaca | What are you trying to achieve? | 06:06 |
| rstarmer | Yeah, the default listed in /etc/kolla/globals.yml is 2.0.3, and I've built other systems using centos/binary that are tagged 2.0.3. But I've got an environment that builds only 2.0.2 tagged images. Even if I do kolla-build --tag 2.0.3 | 06:06 |
| rstarmer | I am scratching my head as to what is driving that if not the command line parameter! | 06:07 |
| *** v1k0d3n has quit IRC | 06:11 | |
| bdaca | what version of kolla do you use? | 06:12 |
| *** mickt has quit IRC | 06:13 | |
| rstarmer | stable/mitaka is what I pulled from git | 06:14 |
| rstarmer | ah... kolla-build --version => 2.0.2 | 06:15 |
| rstarmer | the "other" system, has vesion of 2.0.3 | 06:16 |
| *** ssurana has quit IRC | 06:16 | |
| *** absubram has joined #openstack-kolla | 06:26 | |
| *** absubram_ has joined #openstack-kolla | 06:28 | |
| *** absubram has quit IRC | 06:31 | |
| *** absubram_ has quit IRC | 06:34 | |
| *** sdake has quit IRC | 06:36 | |
| *** absubram has joined #openstack-kolla | 06:38 | |
| *** tonanhngo has quit IRC | 06:42 | |
| *** tonanhngo has joined #openstack-kolla | 06:43 | |
| *** tonanhngo has quit IRC | 06:43 | |
| *** berendt has joined #openstack-kolla | 06:45 | |
| *** berendt has quit IRC | 06:45 | |
| *** berendt has joined #openstack-kolla | 06:46 | |
| berendt | morning guys | 06:46 |
| duonghq | morning berendt | 06:46 |
| *** jascott1_ has joined #openstack-kolla | 06:47 | |
| *** jascott1 has quit IRC | 06:49 | |
| bdaca | hello | 06:49 |
| *** salv-orlando has quit IRC | 06:50 | |
| *** salv-orlando has joined #openstack-kolla | 06:51 | |
| *** absubram has quit IRC | 06:58 | |
| *** JoseMello has quit IRC | 07:01 | |
| *** papacz has quit IRC | 07:05 | |
| bjolo_ | morning all | 07:08 |
| *** bjolo_ is now known as bjolo | 07:08 | |
| *** berendt has quit IRC | 07:08 | |
| *** berendt has joined #openstack-kolla | 07:09 | |
| *** msimonin has joined #openstack-kolla | 07:10 | |
| *** shardy has joined #openstack-kolla | 07:20 | |
| *** zhurong has joined #openstack-kolla | 07:22 | |
| *** matrohon has joined #openstack-kolla | 07:24 | |
| egonzalez90 | request for release 3.0.0 tag was not made yet, right? still can merge without bug id? | 07:27 |
| *** NMI has joined #openstack-kolla | 07:32 | |
| *** NMI has left #openstack-kolla | 07:33 | |
| *** maggro has joined #openstack-kolla | 07:35 | |
| *** salv-orl_ has joined #openstack-kolla | 07:40 | |
| *** salv-orlando has quit IRC | 07:43 | |
| *** jmccarthy has left #openstack-kolla | 07:57 | |
| *** jascott1_ has quit IRC | 08:05 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Bump version of Kibana from 4.4 to 4.6 https://review.openstack.org/369076 | 08:11 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Allow a custom keepalived configuration file https://review.openstack.org/368956 | 08:13 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Allow a custom haproxy configuration file https://review.openstack.org/368959 | 08:13 |
| *** nradojevic has joined #openstack-kolla | 08:16 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Add doc with notes about making reviews https://review.openstack.org/369037 | 08:18 |
| *** tonanhngo has joined #openstack-kolla | 08:19 | |
| *** tonanhngo has quit IRC | 08:19 | |
| *** mbound has joined #openstack-kolla | 08:20 | |
| *** gfhellma has joined #openstack-kolla | 08:21 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Add doc with notes about making reviews https://review.openstack.org/369037 | 08:21 |
| *** daneyon has joined #openstack-kolla | 08:25 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Document how to add log rotation for a service https://review.openstack.org/369037 | 08:26 |
| *** mewald has joined #openstack-kolla | 08:29 | |
| *** daneyon has quit IRC | 08:29 | |
| nradojevic | Morning, I got the following error message while trying to start an instance: internal error: process exited while connecting to monitor: Could not access KVM kernel module: Permission denied. I'm running trusty containers on xenal hosts. Any advice? | 08:30 |
| bjolo | nradojevic, sounds like KVM might not work. Has it worked before on this node? | 08:32 |
| bjolo | docker exec nova_libvirt virt-host-validate | 08:32 |
| *** gfhellma has quit IRC | 08:32 | |
| nradojevic | bjolo: yes, it has worked before. I deployed yesterday and now I'm getting this error | 08:36 |
| bjolo | where do you get the error message? | 08:37 |
| bjolo | log? horizon? | 08:37 |
| bjolo | cli? | 08:37 |
| nradojevic | nova_compute logs | 08:39 |
| mewald | bjolo: we are also seeing this when running "virsh start <name>" from within the nova_libvirt container | 08:40 |
| mewald | bjolo: the validate command seems to succeed: https://gist.github.com/mewald1/1da854f1416c43dd1500737451db4f61 | 08:41 |
| *** dwalsh has joined #openstack-kolla | 08:42 | |
| bjolo | what kolla version is it? | 08:43 |
| bjolo | lsmod | grep kvm | 08:43 |
| nradojevic | 3.0.0 | 08:45 |
| nradojevic | kvm_intel 172032 0 | 08:45 |
| nradojevic | kvm 536576 1 kvm_intel | 08:45 |
| nradojevic | irqbypass 16384 1 kvm | 08:45 |
| *** williamcaban has joined #openstack-kolla | 08:45 | |
| bjolo | for redhat, but... https://bugzilla.redhat.com/show_bug.cgi?id=950436 | 08:47 |
| openstack | bugzilla.redhat.com bug 950436 in qemu "Could not access KVM kernel module: Permission denied" [Unspecified,Closed: errata] - Assigned to virt-maint | 08:47 |
| bjolo | selinux? | 08:47 |
| bjolo | getenforce? | 08:48 |
| *** mtaylor22 has quit IRC | 08:48 | |
| bjolo | ah right, ubuntu uses apparmor still even on 1604? | 08:48 |
| *** williamcaban has quit IRC | 08:49 | |
| mewald | bjolo: we are checking apparmor right now | 08:53 |
| bjolo | as a temp fix/workaround, try rmmod kvm_intel; rmmod_kvm; modprobe kvm; modprobe kvm_intel | 08:55 |
| rstarmer | anyone running ironic with Kolla? any pointers? | 08:57 |
| bjolo | rstarmer, not yet, but it is on my list to try out shortly | 08:57 |
| bjolo | like in the next few weeks, so i might pick your brain :) | 08:58 |
| rstarmer | roger that. It clearly starts, headed down the RDO-manager/triple-O like path with this, so it could be quite interesting (assuming I get it to work :) | 08:58 |
| mewald | bjolo: it's a bit weird: I moved all profiles to complain mode in apparmor. After that it still didnt work. After a reboot the "docker-default" profile was back in enforce mode again but now instances are starting ... | 09:07 |
| bjolo | ? | 09:08 |
| mewald | bjolo: exactly my reaction :D | 09:08 |
| egonzalez90 | Jeffrey4l__, ping | 09:08 |
| *** dwalsh has quit IRC | 09:10 | |
| *** jmccarthy has joined #openstack-kolla | 09:11 | |
| *** mbound has quit IRC | 09:12 | |
| bjolo | newbie question. i have a suggestion/question about design. how do i raise that? | 09:16 |
| bjolo | the concern is rabbitmq-clusterer config | 09:17 |
| berendt | bjolo just ask it? | 09:17 |
| bjolo | the default config in jinja template is to use the ansible_hostname | 09:17 |
| bjolo | as an operator i dont like that since it make rabbit dependable on DNS working | 09:18 |
| bjolo | i would like that to be IP based instead | 09:18 |
| bjolo | i.e. if DNS stops, the whole cloud stops | 09:18 |
| berendt | bjolo we use local hosts file on the nodes, this way it does not depend on dns | 09:18 |
| berendt | bjolo https://review.openstack.org/#/c/370525/ | 09:19 |
| bjolo | berendt, you mean you do on the hosts, but not kolla | 09:19 |
| berendt | the pre install tasks of kolla handle this | 09:19 |
| bjolo | well the preinstall task checks that it is resolved, but not that it is from /etc/hosts right | 09:20 |
| berendt | no the pre install task are generating a /etc/hosts file containing all hosts defined in the inventory file of kolla | 09:21 |
| *** DanyC has joined #openstack-kolla | 09:21 | |
| bjolo | ? | 09:22 |
| bjolo | where is that hosts file? | 09:22 |
| *** hieulq has quit IRC | 09:22 | |
| jmccarthy | Morning all ! | 09:22 |
| *** hieulq has joined #openstack-kolla | 09:22 | |
| bjolo | i only have localhost and local hostname in my /etc/hosts on all nodes | 09:23 |
| berendt | you used the pre install tasks? | 09:23 |
| bjolo | yes | 09:23 |
| jmccarthy | I was wondering about the globals.yml, is there a special reason there is stuff uncommented in it (like tempest_floating_network_name) which are not found in the all.yml ? Should *everything in globals be commented out ? | 09:23 |
| bjolo | i quite frankly i dont know if i would like for kolla to update the local /etc/hosts file on the nodes | 09:23 |
| bjolo | we have other playbooks/cm systems outside of kolla that handles that | 09:24 |
| berendt | bjolo kolla only adds the inventory hosts to the hosts file, it does not overwrite/replace the file | 09:24 |
| berendt | sorry, "generate" was the wrong wording | 09:24 |
| *** zhurong has quit IRC | 09:24 | |
| bjolo | no but next time i run my playbooks outside of kolla, they will revert the changes done by kolla to /etc/hosts | 09:25 |
| bjolo | again breaking the cloud | 09:25 |
| berendt | then the playbook handing the /etc/hosts file is broken | 09:25 |
| bjolo | IMO we should not have one :) | 09:25 |
| bjolo | all configs should be IP based only | 09:26 |
| berendt | no | 09:26 |
| bjolo | and not touch the local /etc/hosts | 09:26 |
| bjolo | ok why not? >) | 09:26 |
| bjolo | :) | 09:26 |
| berendt | i hate it to write ip addresses | 09:26 |
| bjolo | why? | 09:26 |
| berendt | feel free to open a bug report to propose to use ip addresses, but i do not think that it makes sense | 09:26 |
| berendt | i do not want to discuss this in detail, i prefer hostnames and i do not get the point why we should not modify the /etc/hosts file | 09:27 |
| bjolo | why not. hostnames can only resolve to one IP right. Kolla multinode uses many different networks like api, storage, tunnel, etc | 09:28 |
| rstarmer | I personally prefer hostnames over IP addresses. Clever naming can actually let you use equivalent config files rather than forcing a change on IP based schemas. | 09:28 |
| rstarmer | IPs are also singular. :} | 09:29 |
| rstarmer | but you should be free to use IP addresses instead of hostnames | 09:29 |
| berendt | i think so | 09:29 |
| *** DanyC has left #openstack-kolla | 09:32 | |
| *** phuongnh has quit IRC | 09:34 | |
| *** zhubingbing has joined #openstack-kolla | 09:34 | |
| *** phuongnh has joined #openstack-kolla | 09:34 | |
| openstackgerrit | Prithiv proposed openstack/kolla: Add networking guide for Neutron extensions https://review.openstack.org/355386 | 09:34 |
| bjolo | further more, /etc/hosts might not be the authorative answer. depends on what you have in nsswitch.conf | 09:39 |
| openstackgerrit | zhubingbing proposed openstack/kolla: add renote monitoring https://review.openstack.org/370345 | 09:40 |
| bjolo | rstarmer, please elaborate on the config files and clever naming? | 09:40 |
| bjolo | IMO if kolla wants to use hosts file, it should be a seperate one that is copied into the containers and not touch the external one for the reasons i have listed | 09:43 |
| berendt | bjolo please open a bug report and motivate you POV there | 09:44 |
| bjolo | ok | 09:44 |
| berendt | i think we simply assume that only our playbooks are touching the used nodes | 09:44 |
| berendt | in my environment i do not use the pre tasks at all, i use a separate playbook to prepare the nodes | 09:45 |
| openstackgerrit | zhubingbing proposed openstack/kolla: add renote monitoring https://review.openstack.org/370345 | 09:48 |
| openstackgerrit | Prithiv proposed openstack/kolla: Add networking guide for Neutron extensions https://review.openstack.org/355386 | 09:51 |
| bjolo | berendt, please dont get me wrong, dont mean to come across unhappy and crumpy :) i like kolla a lot | 09:52 |
| berendt | i think it makes sense to think about container specific hosts files, but i think it is better to discuss it in a bug report | 09:53 |
| openstackgerrit | zhubingbing proposed openstack/kolla: add renote monitoring https://review.openstack.org/370345 | 09:53 |
| bjolo | in my experience when running large critial envs, we always cut away as many dependencies as possible to ensure stability. hostname resolution is one of them | 09:53 |
| jmccarthy | I was wondering about the globals.yml, is there a special reason there is stuff uncommented in it (like tempest_floating_network_name) which are not found in the all.yml ? Shouldn't *everything in globals be commented out initially ? | 09:56 |
| berendt | it depends on the environment, local hosts files on the nodes works for me | 09:57 |
| *** gfidente has joined #openstack-kolla | 09:57 | |
| berendt | jmccarthy the variable is defined in ansible/roles/tempest/defaults/main.yml, i think it is safe to comment those variables in globals.yml | 09:58 |
| *** phuongnh has quit IRC | 09:58 | |
| openstackgerrit | Prithiv proposed openstack/kolla: Add networking guide for Neutron extensions https://review.openstack.org/355386 | 09:59 |
| *** phuongnh has joined #openstack-kolla | 09:59 | |
| *** gfidente has quit IRC | 10:00 | |
| jmccarthy | Oh hmm ok is that a trend that is happening, stuff defined within the service vs in all.yml ? | 10:01 |
| *** salv-orl_ has quit IRC | 10:01 | |
| berendt | i do not know. but i think it makes sense to define role specific variables inside the role. | 10:01 |
| *** phuongnh has quit IRC | 10:04 | |
| *** phuongnh has joined #openstack-kolla | 10:05 | |
| jmccarthy | Ok, I just didn't notice globals had uncommented stuff by default until just now | 10:06 |
| berendt | jmccarthy i think it is fine to comment it | 10:06 |
| jmccarthy | Ok cool | 10:06 |
| berendt | the whole file has to be cleaned up | 10:06 |
| openstackgerrit | Prithiv proposed openstack/kolla: Modify collectd permissions https://review.openstack.org/370663 | 10:06 |
| *** mewald has left #openstack-kolla | 10:07 | |
| openstackgerrit | Prithiv proposed openstack/kolla: Modify collectd permissions https://review.openstack.org/370663 | 10:11 |
| *** gfidente has joined #openstack-kolla | 10:12 | |
| *** gfidente has joined #openstack-kolla | 10:12 | |
| *** rstarmer has quit IRC | 10:19 | |
| openstackgerrit | Prithiv proposed openstack/kolla: Modify collectd permissions https://review.openstack.org/370663 | 10:22 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Fix ceilometer.conf template https://review.openstack.org/370677 | 10:27 |
| *** duonghq has left #openstack-kolla | 10:30 | |
| *** williamcaban has joined #openstack-kolla | 10:33 | |
| *** williamcaban has quit IRC | 10:37 | |
| *** zhubingbing has quit IRC | 10:38 | |
| *** hieulq has quit IRC | 10:51 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Add kolla_keystone_role.py to kolla-toolbox https://review.openstack.org/370688 | 10:51 |
| *** hieulq has joined #openstack-kolla | 10:51 | |
| *** shardy has quit IRC | 10:57 | |
| *** salv-orlando has joined #openstack-kolla | 11:02 | |
| jmccarthy | Does this keepavlid.conf seem ok in general (test setup) ? http://paste.openstack.org/show/577368/ | 11:02 |
| *** dave-mccowan has joined #openstack-kolla | 11:06 | |
| *** salv-orlando has quit IRC | 11:06 | |
| sbezverk | Jeffrey4l: ping | 11:07 |
| *** daneyon has joined #openstack-kolla | 11:07 | |
| *** tonanhngo has joined #openstack-kolla | 11:09 | |
| *** tonanhngo has quit IRC | 11:11 | |
| *** daneyon has quit IRC | 11:12 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Create the heat_stack_user role in the heat role https://review.openstack.org/351220 | 11:14 |
| *** msimonin has quit IRC | 11:15 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Create the heat_stack_user role in the heat role https://review.openstack.org/351220 | 11:17 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: User os_user_role to create cloudkitty rating role https://review.openstack.org/370702 | 11:18 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: User os_user_role to create cloudkitty rating role https://review.openstack.org/370702 | 11:19 |
| *** phuongnh has quit IRC | 11:23 | |
| *** jmccarthy has left #openstack-kolla | 11:24 | |
| *** salv-orlando has joined #openstack-kolla | 11:28 | |
| *** fragatina has joined #openstack-kolla | 11:30 | |
| *** tonanhngo has joined #openstack-kolla | 11:33 | |
| *** fragatina has quit IRC | 11:34 | |
| *** tonanhngo has quit IRC | 11:34 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: User os_keystone_role to create cloudkitty rating role https://review.openstack.org/370702 | 11:46 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Create the heat_stack_user role in the heat role https://review.openstack.org/351220 | 11:47 |
| mliima | morning | 11:48 |
| egonzalez90 | morning mliima | 11:48 |
| *** bdaca has quit IRC | 11:58 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Add cloudkitty to heka-openstack.toml.j2 https://review.openstack.org/370734 | 12:03 |
| *** dwalsh has joined #openstack-kolla | 12:05 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Document how to add log rotation for a service https://review.openstack.org/369037 | 12:06 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Add cloudkitty to heka-openstack.toml.j2 https://review.openstack.org/370734 | 12:06 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Document how to add logging for a service https://review.openstack.org/369037 | 12:07 |
| *** msimonin has joined #openstack-kolla | 12:13 | |
| *** dwalsh has quit IRC | 12:17 | |
| bjolo | i have an opengrok server setup, but it is public if anyone wants to use it | 12:20 |
| bjolo | http://94.246.117.128/source/ | 12:21 |
| *** williamcaban has joined #openstack-kolla | 12:21 | |
| bjolo | very quick code search tool to look around | 12:22 |
| bjolo | for ex. to see where ansible_hostname is being used in the code | 12:22 |
| bjolo | http://94.246.117.128/source/search?q=ansible_hostname&defs=&refs=&path=&hist=&type=&project=kolla | 12:22 |
| *** williamcaban has quit IRC | 12:25 | |
| *** mgoddard has joined #openstack-kolla | 12:29 | |
| *** lyncos has joined #openstack-kolla | 12:31 | |
| *** rhallisey has joined #openstack-kolla | 12:36 | |
| *** sdake has joined #openstack-kolla | 12:38 | |
| mliima | guys, i launched an instance, but i'm unable to access the internet :/ | 12:38 |
| mliima | someone has gone through this problem? | 12:38 |
| sdake | hey cats | 12:43 |
| sdake | today is tag day | 12:43 |
| sdake | t-12 hours | 12:43 |
| sdake | anyone jammed up? | 12:43 |
| bjolo | morning sdake | 12:45 |
| bjolo | up early today huh? :) | 12:45 |
| sdake | morning fine sir | 12:45 |
| sdake | i wake up at 5-6am typically | 12:45 |
| sdake | early is 3am | 12:45 |
| sdake | sometimes tht happeens becauseee my kids get into it | 12:46 |
| bjolo | 3am is go to bed time | 12:46 |
| sbezverk | sdake: morning can you please ack this ps really important: https://review.openstack.org/#/c/370440/ | 12:46 |
| sdake | and thtrealy makes me angry | 12:46 |
| sdake | sbezverk roger doing right now | 12:46 |
| *** tonanhngo has joined #openstack-kolla | 12:46 | |
| sdake | sbezverk why is mariadb bootstrapping important for you? | 12:46 |
| sdake | sbezverk not that it it issn't an important bug to fix | 12:46 |
| sdake | just want to make sure you linked the correct review | 12:47 |
| sbezverk | it is a show stopper for kube | 12:47 |
| *** tonanhngo has quit IRC | 12:48 | |
| sdake | sbezverk oh nm i just read the commit log | 12:48 |
| sdake | sbezverk sorry dude i just woke up | 12:48 |
| sbezverk | sdake: np | 12:48 |
| *** jamesdenton has joined #openstack-kolla | 12:48 | |
| sbezverk | has been tested and the gate is green | 12:48 |
| kfox1111 | so, I've got kolla-kubernetes just about working in minikube. | 12:48 |
| sdake | sbezverk let me read patch more clolsely- i htought this was huikang's mariadb fix | 12:48 |
| sdake | what is minikube | 12:49 |
| kfox1111 | having a strange issue with qemu though. | 12:49 |
| kfox1111 | sdake: https://github.com/kubernetes/minikube | 12:49 |
| kfox1111 | a very simple way to get a test kubernetes running. | 12:49 |
| rhallisey | it's another kube tool | 12:49 |
| *** shardy has joined #openstack-kolla | 12:50 | |
| kfox1111 | suppoerts macosx/windows/linux. | 12:50 |
| sdake | its a shame our countiners need a differnt bootstrap mechanism for kube and ansible | 12:51 |
| sdake | i was hopeful that would remain the same | 12:51 |
| rhallisey | sdake, that will get fixed | 12:51 |
| rhallisey | I don't like that either | 12:51 |
| kfox1111 | qemu though is failing to launch vm's with an error | 12:52 |
| rhallisey | the only way it won't is if there is no workflow/some other tool in fornt | 12:52 |
| sdake | rhallisey roger gglad you know its a problem and are thinking about it | 12:52 |
| kfox1111 | qemu-system-x86_64: -chardev pty,id=charserial1: Failed to create chardev | 12:52 |
| sdake | kfox1111 i hvae seen tht a mllion times | 12:52 |
| *** mgoddard_ has joined #openstack-kolla | 12:52 | |
| rhallisey | sdake, ya I'm in agreement | 12:52 |
| sdake | kfox1111 let me have a smoke and try to remembeer why it happens | 12:52 |
| kfox1111 | sdake: was hoping someone would. | 12:52 |
| rhallisey | mounting /dev didn't work :( | 12:52 |
| *** mgoddard has quit IRC | 12:52 | |
| rhallisey | neither changing host perms | 12:52 |
| kfox1111 | sdake: cool. have a good smoke. :) | 12:53 |
| openstackgerrit | Merged openstack/kolla: Mariadb bootstrap - no need to check for cluster ready in Kube https://review.openstack.org/370440 | 12:54 |
| sdake | kfox1111 is it ony minikube that causes this chardev problem? | 12:57 |
| *** jtriley has joined #openstack-kolla | 12:57 | |
| sdake | kfox1111 my guess is permissions on /dev/ptymx (or whatever that thing is called) is wrong | 12:57 |
| kfox1111 | I haven't tried i anywhere else yet. | 12:57 |
| *** fguillot has joined #openstack-kolla | 12:57 | |
| kfox1111 | so, I straced it. | 12:58 |
| sdake | docker used to hae a problem where if /dev was bindmounted, /dev/ptymx was set to 000 permisisons | 12:58 |
| *** mgoddard_ has quit IRC | 12:58 | |
| kfox1111 | it opens /dev/pts/ptmx, | 12:58 |
| sdake | thats the file | 12:58 |
| kfox1111 | then tries to chown /dev/pts4 to owner qemu,tty | 12:58 |
| kfox1111 | the chown fails with permission denied | 12:59 |
| sdake | cool so ptmx is working for yuo | 12:59 |
| sdake | ptmx is a control interface | 12:59 |
| sdake | that creates ptys | 12:59 |
| kfox1111 | there was a getuid shortly before and it returned qemu. | 12:59 |
| kfox1111 | yeah. | 12:59 |
| sdake | ls -l /dev/pts/* for me please | 12:59 |
| *** rstarmer has joined #openstack-kolla | 13:00 | |
| sdake | and ls -ld /dev/pts | 13:00 |
| kfox1111 | http://paste.openstack.org/show/Bj0HheHOOsMNQlVJwJrb/ | 13:00 |
| *** v1k0d3n has joined #openstack-kolla | 13:00 | |
| kfox1111 | drwxr-xr-x 2 root root 0 Sep 14 23:14 /dev/pts | 13:00 |
| rhallisey | kfox1111, chmod 666 /dev/pts/ptmx | 13:01 |
| rhallisey | :) | 13:01 |
| sdake | rhallisey yes that would work | 13:01 |
| sdake | but lets get to the bottom of why its happening | 13:01 |
| sdake | or may work | 13:01 |
| kfox1111 | it opened ptm ok. it was /dev/pts/4 that it failed to chown | 13:01 |
| rhallisey | kfox1111, did you try that already? | 13:01 |
| kfox1111 | yes. | 13:01 |
| sdake | remind me libvirt container runs as which user? | 13:02 |
| rhallisey | root | 13:02 |
| kfox1111 | root | 13:02 |
| sdake | are you using qemu.conf to set uid/gidd of libvirt | 13:02 |
| rhallisey | good question | 13:03 |
| kfox1111 | let me look... | 13:03 |
| rhallisey | I don't think so | 13:03 |
| sdake | kolla doe sthis | 13:03 |
| kfox1111 | know where that's defined? | 13:03 |
| sdake | defautls/qemu.conf.j2 | 13:04 |
| sdake | in nova role | 13:04 |
| kfox1111 | this is a stock kolla centos binary nova compute 2.0.1 container from the hub. | 13:04 |
| sdake | kolla-kube generates defaults from the j2 files in ansible no? | 13:05 |
| *** rstarmer has quit IRC | 13:05 | |
| kfox1111 | hmm... there is only ls /etc/kolla/nova-compute/ | 13:05 |
| kfox1111 | config.json nova.conf | 13:05 |
| kfox1111 | should it be generating a qemu.conf too? | 13:05 |
| rhallisey | kfox1111, ya I don't think it's includede | 13:05 |
| rhallisey | check the conifgmap | 13:05 |
| sdake | i dont recall | 13:05 |
| rhallisey | kfox1111, it's for libvirt | 13:05 |
| sdake | let me look in kollatree | 13:05 |
| rhallisey | /etc/kolla/nova-libvirt/ | 13:06 |
| rhallisey | kfox1111, libvirt uses nova's configmap | 13:06 |
| rhallisey | it's needs it's own | 13:06 |
| kfox1111 | oh. yeah. sec | 13:06 |
| rhallisey | s/it's/it/ | 13:06 |
| kfox1111 | more /etc/kolla/nova-libvirt/qemu.conf | 13:06 |
| kfox1111 | stdio_handler = "file" | 13:06 |
| sdake | https://github.com/openstack/kolla/blob/master/ansible/roles/nova/templates/libvirtd.conf.j2 | 13:06 |
| sdake | when we went to qemu2.5, qemu automatically dropped root | 13:07 |
| rhallisey | kfox1111, kubctl get configmaps nova-compute-configmap | 13:07 |
| rhallisey | there will be no qemu.conf in there | 13:07 |
| *** jmccarthy has joined #openstack-kolla | 13:07 | |
| kfox1111 | http://paste.openstack.org/show/WZt3yqUYKx41BXgYzQZQ/ | 13:07 |
| sdake | are you bindmoutning /dev? | 13:08 |
| kfox1111 | yeah. | 13:09 |
| sdake | are you sure its actually bindmoutning | 13:09 |
| sdake | (same files on host as not host) | 13:09 |
| kfox1111 | yup. tried making permission changes on the host in /dev and see it in the container. | 13:09 |
| sdake | kfox1111 can you try chmod 666 /dev/pts | 13:10 |
| sdake | and run again | 13:10 |
| kfox1111 | k | 13:10 |
| kfox1111 | pts is a dir. you mean 777? | 13:10 |
| sdake | sorry 777 | 13:10 |
| kfox1111 | k | 13:10 |
| sdake | i never quite understood how i fixed this problem | 13:11 |
| sdake | but i know it revolved around /dev/pts and /dev/pts/ptmx permissions | 13:11 |
| sdake | let me switch netowrks and see what ivegot on my gear | 13:11 |
| kfox1111 | same. | 13:12 |
| *** fguillot has quit IRC | 13:13 | |
| *** sdake_ has joined #openstack-kolla | 13:14 | |
| sbezverk | kfox1111: it could be minikube thing too | 13:14 |
| sdake_ | my /dev/pts is 755 | 13:15 |
| *** salv-orlando has quit IRC | 13:15 | |
| sdake_ | what doe minikube do | 13:15 |
| sbezverk | kfox1111: I am almost there with all-in-one | 13:15 |
| *** sdake has quit IRC | 13:16 | |
| sdake_ | kfox1111 nice job on the strace idea btw ;) | 13:16 |
| kfox1111 | sdake_: minikube is a docker-machine with kube loaded in. | 13:16 |
| kfox1111 | its basically a small kvm vm with docker and kubelet. | 13:16 |
| sdake_ | docer machine is a bare metal deploy tool no? | 13:16 |
| kfox1111 | Linux minikubeVM 4.4.15-boot2docker #1 SMP Fri Jul 15 18:11:45 UTC 2016 x86_64 GNU/Linux | 13:17 |
| kfox1111 | it can do bare metal, but minikube only does vm's at the moment. | 13:17 |
| sdake_ | its a vm image? | 13:17 |
| sdake_ | how does docker-machine interactwith things? | 13:17 |
| britthouser | morning all! | 13:18 |
| sdake_ | hey britthouser | 13:18 |
| britthouser | How goes sdake_ ? | 13:18 |
| kfox1111 | sdake_: http://paste.openstack.org/show/LvChOrYXMOp06mpVTrVl/ | 13:18 |
| kfox1111 | then you can use kubernetes. | 13:18 |
| sdake_ | britthouser pretty dam nwell britt | 13:20 |
| sdake_ | rc1 is lookingrealy good | 13:20 |
| britthouser | w00t | 13:21 |
| *** shardy has quit IRC | 13:22 | |
| mliima | ERROR 1047 (08S01) at line 1: WSREP has not yet prepared node for application use | 13:25 |
| mliima | mariaDB bootstrap | 13:25 |
| mliima | :/ | 13:25 |
| mliima | this error is intermitent | 13:25 |
| sdake_ | mliima ho wd you trigger it? just rnaodmly? | 13:26 |
| mliima | run kolla-ansible deploy | 13:26 |
| sdake_ | mliima how many controller nodes | 13:27 |
| mliima | aio sdake_ | 13:27 |
| sdake_ | where is that error printed | 13:27 |
| sdake_ | during startup, mariadb logs, where? | 13:28 |
| mliima | bootstrap_mariaDB container | 13:28 |
| sdake_ | in docker logs or maraidb logfiles? | 13:28 |
| mliima | docker | 13:28 |
| sdake_ | annoying | 13:28 |
| sdake_ | ok mliima give mea min to look at the code base | 13:29 |
| mliima | It can be a problem of my environment | 13:29 |
| sdake_ | that code has recenty been touched | 13:29 |
| mliima | the deploy works now | 13:29 |
| sdake_ | if its unreliable part of the time | 13:29 |
| sdake_ | it might as well be unreliable all of the time | 13:29 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Remove public_router_id from tempest configuration template https://review.openstack.org/370825 | 13:30 |
| sdake_ | mliima can you dig into this - i dont havetiem today | 13:30 |
| mliima | ok sdake_ | 13:31 |
| sdake_ | mlima please revert c7c879095fffee711479c40f0bc95b1829a2b069 and try deploy several times and see if the problem resolves | 13:31 |
| sdake_ | you may have to revert other patches on top | 13:31 |
| sdake_ | you can figure that part out (sbezverk recently chnged this code as well) | 13:31 |
| sbezverk | sdake_: the gate was all green | 13:32 |
| sdake_ | sbezverk undretood - intermittent problem | 13:32 |
| sbezverk | it means deployment went ok | 13:32 |
| sdake_ | sbezverk not blaming you- my guess is huikang's implementation isn't quite right | 13:32 |
| sdake_ | so lets either validate or eliminate that as a theory | 13:32 |
| mliima | sdake_, i can't found c7c879095fffee711479c40f0bc95b1829a2b069 | 13:33 |
| mliima | :/ | 13:33 |
| sdake_ | oh- try a git pull --rebase then | 13:35 |
| sdake_ | and dont revert | 13:35 |
| sdake_ | so theory #1 shot down - it is not a regressionfrom hui's patch | 13:36 |
| sdake_ | theory #2 is huikang's ptch resolves that partuclar problem | 13:36 |
| sdake_ | lets either validate that or elimte it as a theory | 13:37 |
| sdake_ | mliima if you haven't rebased, if youcan tell me the current commit your on (git status) i'dapprecate it | 13:38 |
| mliima | ok | 13:38 |
| mliima | i'm working on master | 13:39 |
| sdake_ | yes but which commit ofmaster i need to know | 13:39 |
| mliima | commit 78f1564ed808d5aeaba8fa5f320e5dc95242483b | 13:40 |
| *** inc0 has joined #openstack-kolla | 13:46 | |
| *** huikang has joined #openstack-kolla | 13:47 | |
| *** daneyon has joined #openstack-kolla | 13:49 | |
| inc0 | good morning | 13:51 |
| huikang | morning/evening | 13:53 |
| huikang | Happy Chinese mid-autumn to everyone! | 13:54 |
| *** daneyon has quit IRC | 13:54 | |
| *** nradojevic has quit IRC | 13:54 | |
| sbezverk | kfox1111: still here? | 13:54 |
| kfox1111 | yes | 13:56 |
| kfox1111 | whats up? | 13:57 |
| inc0 | huikang, mid-autumn? how does that work? | 13:58 |
| inc0 | when does Chinese autumn starts? | 13:58 |
| *** maggro has quit IRC | 13:58 | |
| huikang | Hi, inc0. the 15th day of the 8th month according to the Chinese lunar calendar, the Mid-Autumn Festival is the second grandest festival after the Spring Festival in China | 13:59 |
| *** NMI has joined #openstack-kolla | 13:59 | |
| huikang | People eat moon cake today | 14:00 |
| mliima | guys, why physical_interface_mappings = physnet1:{{ neutron_external_interface }} and not {{ network_interface }} ? | 14:00 |
| sbezverk | kfox1111: found interesting bug in nova-api | 14:00 |
| sbezverk | curious if you saw it | 14:01 |
| kfox1111 | oh yeah? | 14:01 |
| kfox1111 | no. what did you see? | 14:01 |
| inc0 | huikang, interesting, I'd love to see one of your festivals someday | 14:01 |
| sbezverk | kfox1111: we assign my_ip based on {{ tunnel_interface }} | 14:01 |
| inc0 | also moon cake sounds great;) in European culture autumn starts at equinox - so in a week or so | 14:02 |
| huikang | inc0, you are very welcome | 14:02 |
| sbezverk | kfox1111: but it is not visible to nova-api | 14:02 |
| sbezverk | since it is not running in net-host | 14:02 |
| kfox1111 | oh. cause its in a container. | 14:02 |
| *** tonanhngo has joined #openstack-kolla | 14:02 | |
| sbezverk | yep | 14:02 |
| inc0 | but no festivals:( unless you're into pagan religions and want to dance naked around fireplace | 14:02 |
| kfox1111 | what part of the api actually uses my_ip though? | 14:02 |
| sbezverk | when I had tunnel interface eth0 it just happened to fall back to eht0 | 14:03 |
| sbezverk | but in reality it was different interface | 14:03 |
| kfox1111 | I think that's dead code? | 14:03 |
| sbezverk | when I changed my baremetal server tunnel interface ti bond0.80 | 14:03 |
| *** tonanhngo has quit IRC | 14:03 | |
| sbezverk | nope nova-api crashes when my_ip is not set properly | 14:04 |
| sbezverk | I am going to push a fix once I test it | 14:04 |
| kfox1111 | I could have swarn I never set it on my other clouster... let me go look. | 14:04 |
| sbezverk | I suspect you use eth0 for network interface in your globals.yml | 14:05 |
| bjolo | kfox1111, http://94.246.117.128/source/search?q=my_ip&defs=&refs=&path=&hist=&type=&project=kolla | 14:05 |
| sbezverk | then it would | 14:05 |
| sbezverk | skip this issue | 14:05 |
| kfox1111 | sbezverk: no, I mean on a production non kolla cluster. | 14:06 |
| sbezverk | no I mean in your kube environment | 14:06 |
| openstackgerrit | Prithiv proposed openstack/kolla: Modify collectd permissions https://review.openstack.org/370663 | 14:07 |
| kfox1111 | in my qube, I'm using eth0. | 14:07 |
| kfox1111 | kube | 14:07 |
| sbezverk | that is why it worked :-) | 14:08 |
| sbezverk | you change your net interface for some thing else nova-api will stop working | 14:08 |
| kfox1111 | yeah. but I think we can probably just remove the code out of the template? | 14:09 |
| sbezverk | kfox1111: if we remove this code, who is going to setup my ip? | 14:09 |
| kfox1111 | yeah. on my production system, my_ip isn't set at all. | 14:09 |
| kfox1111 | I don't think nova-api uses it. it may validate it though. | 14:10 |
| sbezverk | kfox1111: well what I see is if it is not set it fails | 14:10 |
| sbezverk | kfox1111: man I have nova-api traceback complaining about it and you are saying it is not used?!?!?! | 14:10 |
| kfox1111 | I'm saying, I have a cloud that doesn't have it set at all, and nova-api works. | 14:11 |
| kfox1111 | I'm guessing, if you have the var set, and its set to an ip not on the host, nova-api complains. | 14:11 |
| kfox1111 | but if not set, it doesn't care. | 14:11 |
| sbezverk | kfox1111: does it run master nova bits? | 14:11 |
| kfox1111 | mitaka | 14:12 |
| sbezverk | kfox1111: in this case I suspect nova folks added validation now | 14:12 |
| kfox1111 | it never has mattered from essex all the way to mitaka. I doubt it was added only in newton. | 14:12 |
| sbezverk | kfox1111: see for yourself http://paste.openstack.org/show/577452/ | 14:14 |
| kfox1111 | sec... | 14:15 |
| kfox1111 | ah. the metadata stuff | 14:15 |
| kfox1111 | metadata_host = 172.18.0.60 | 14:15 |
| kfox1111 | I do set that var explicitly. | 14:15 |
| kfox1111 | but setting it to the pod's ip would be wrong. | 14:16 |
| kfox1111 | its set to the metadata server ip. | 14:16 |
| *** NMI has quit IRC | 14:16 | |
| *** lrensing has joined #openstack-kolla | 14:16 | |
| *** fguillot has joined #openstack-kolla | 14:17 | |
| kfox1111 | metadata_host should be kolla_external_vip I think. | 14:19 |
| kfox1111 | then my_ip doesn't need to be set. | 14:19 |
| sbezverk | kfox1111: testing now I have working nova.conf and none working the difference I see in my_ip sp far | 14:20 |
| kfox1111 | I think its wrong though. :/ | 14:21 |
| kfox1111 | nova-api's going to give out that ip to metadata requests, | 14:21 |
| kfox1111 | and its the local metadata server, not the pool of metadata servers. | 14:21 |
| kfox1111 | it will cause breakages with rolling upgrades. | 14:21 |
| mliima | guys, why physical_interface_mappings = physnet1:{{ neutron_external_interface }} and not {{ network_interface }} ? | 14:22 |
| kfox1111 | it will apear to work until you hit some edge cases. | 14:22 |
| openstackgerrit | Ryan Hallisey proposed openstack/kolla: Shrink the quickstart massively and break it into 2 https://review.openstack.org/343224 | 14:24 |
| rhallisey | massive rebase O.o | 14:25 |
| *** tonanhngo has joined #openstack-kolla | 14:25 | |
| *** tonanhngo has quit IRC | 14:25 | |
| *** jheroux has joined #openstack-kolla | 14:27 | |
| kfox1111 | bbiab | 14:27 |
| *** jtriley has quit IRC | 14:30 | |
| *** eaguilar has joined #openstack-kolla | 14:36 | |
| kfox1111 | back. | 14:39 |
| *** egonzalez90 has quit IRC | 14:39 | |
| *** jtriley has joined #openstack-kolla | 14:42 | |
| *** zhubingbing has joined #openstack-kolla | 14:42 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Use "type" to take place of "which" and update the output https://review.openstack.org/369551 | 14:43 |
| *** eaguilar_ has joined #openstack-kolla | 14:44 | |
| *** eaguilar has quit IRC | 14:44 | |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Add a tip at the end of tools/init-runonce https://review.openstack.org/370903 | 14:49 |
| *** mbound has joined #openstack-kolla | 14:50 | |
| openstackgerrit | caoyuan proposed openstack/kolla: Use "type" to take place of "which" and update the output https://review.openstack.org/369551 | 14:52 |
| *** absubram has joined #openstack-kolla | 14:52 | |
| *** absubram_ has joined #openstack-kolla | 14:53 | |
| openstackgerrit | prameswar proposed openstack/kolla: kolla config file path corrected for ubuntu Closes-Bug: #1623967 Change-Id: I214d9ab42b57e94bd13448773432580861aee32f https://review.openstack.org/370907 | 14:54 |
| openstack | bug 1623967 in kolla "path /usr/share/kolla/etc_examples/kolla not exist in ubuntu " [Undecided,New] https://launchpad.net/bugs/1623967 - Assigned to prameswar (prameswar) | 14:54 |
| *** tonanhngo has joined #openstack-kolla | 14:55 | |
| *** tonanhngo has quit IRC | 14:56 | |
| *** absubram has quit IRC | 14:57 | |
| *** absubram_ is now known as absubram | 14:57 | |
| *** berendt has quit IRC | 15:03 | |
| *** unicell1 has quit IRC | 15:03 | |
| *** diogogmt has joined #openstack-kolla | 15:03 | |
| *** rstarmer has joined #openstack-kolla | 15:05 | |
| *** unicell has joined #openstack-kolla | 15:07 | |
| *** egonzalez90 has joined #openstack-kolla | 15:09 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: First stab at minikube doc https://review.openstack.org/370930 | 15:10 |
| *** yingjun has joined #openstack-kolla | 15:12 | |
| *** rstarmer has quit IRC | 15:14 | |
| *** pbourke has quit IRC | 15:16 | |
| *** pbourke has joined #openstack-kolla | 15:16 | |
| *** lyncos has quit IRC | 15:16 | |
| *** salv-orlando has joined #openstack-kolla | 15:19 | |
| *** diogogmt has quit IRC | 15:20 | |
| *** salv-orlando has quit IRC | 15:22 | |
| *** ssurana has joined #openstack-kolla | 15:26 | |
| *** daneyon has joined #openstack-kolla | 15:38 | |
| *** msimonin has quit IRC | 15:40 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: First stab at minikube doc https://review.openstack.org/370930 | 15:42 |
| *** daneyon has quit IRC | 15:43 | |
| openstackgerrit | prameswar proposed openstack/kolla: kolla config file path corrected for ubuntu https://review.openstack.org/370907 | 15:43 |
| *** eaguilar_ has quit IRC | 15:43 | |
| *** britthouser has quit IRC | 15:45 | |
| inc0 | sdake_, ping | 15:45 |
| openstackgerrit | Merged openstack/kolla: Add cloudkitty to heka-openstack.toml.j2 https://review.openstack.org/370734 | 15:46 |
| openstackgerrit | Merged openstack/kolla: Minor fix in manila guide https://review.openstack.org/370104 | 15:46 |
| kfox1111 | sdake_: did you get a chance to look at your other system yet? | 15:48 |
| sdake_ | inc0 shoot | 15:48 |
| inc0 | sdake_, when do we release rc1? | 15:48 |
| sdake_ | kfox1111 which part? /dev/pty is 700 | 15:48 |
| sdake_ | inc0 today at 5pm pst | 15:48 |
| inc0 | wfm | 15:48 |
| *** salv-orlando has joined #openstack-kolla | 15:49 | |
| sdake_ | which should be 00:00 | 15:49 |
| sdake_ | but maybe its aug 16 00:00 | 15:49 |
| inc0 | I think we have 2 more bps to merge | 15:49 |
| inc0 | 3* | 15:49 |
| inc0 | kuryr, vpnaas and sfc | 15:49 |
| inc0 | I'd suggest guys focusing our collective review power on these reviews please | 15:50 |
| inc0 | both cores and non-cores alike | 15:50 |
| *** jtriley has quit IRC | 15:51 | |
| kfox1111 | sdake_: k. so even with 777 its not working, so its not that... | 15:51 |
| sdake_ | kfox1111 can i get bac to you | 15:51 |
| kfox1111 | sure. | 15:52 |
| apuimedo | inc0: let me know if there is anything we have to do on the kuryr side | 15:52 |
| sdake_ | huikang is kuryr going to make it | 15:52 |
| kfox1111 | any ohter ideas you have though, please let me know. | 15:52 |
| inc0 | apuimedo, will do, I'm reviewing it now | 15:52 |
| apuimedo | inc0: thanks | 15:52 |
| *** jgriffith has quit IRC | 15:52 | |
| inc0 | sdake_, I'm not sure if all of reviews will make it, but most of it looks good | 15:53 |
| *** eaguilar has joined #openstack-kolla | 15:54 | |
| apuimedo | inc0: do you have a link to the topic branch? | 15:54 |
| inc0 | apuimedo, this one is correct? https://review.openstack.org/#/q/topic:bp/kuryr-docker-plugin | 15:55 |
| apuimedo | I'll check the patches ;-) | 15:56 |
| *** williamcaban has joined #openstack-kolla | 15:57 | |
| inc0 | apuimedo, zhubingbing huikang https://review.openstack.org/#/c/298894/21/ansible/roles/kuryr/templates/kuryr.conf.j2 question | 15:59 |
| apuimedo | I don't see any question there | 16:00 |
| inc0 | this auth looks wrong, how do we configure it in kuryr side? | 16:00 |
| inc0 | ^ the question;) | 16:00 |
| apuimedo | it IS wrong :P | 16:00 |
| *** thumpba has joined #openstack-kolla | 16:00 | |
| apuimedo | inc0: good eyesight | 16:00 |
| apuimedo | well, now we have a script just like Neutron | 16:00 |
| inc0 | I'll comment | 16:00 |
| *** britthouser has joined #openstack-kolla | 16:00 | |
| apuimedo | that generates a sample config | 16:00 |
| apuimedo | but basically, for the [neutron] section | 16:01 |
| apuimedo | you just put what you'd put for any other service in auth_token section | 16:01 |
| inc0 | yeah I guess we need same type of auth creds as any other openstack service | 16:01 |
| inc0 | yes, but auth_token was deprecated in keystone right | 16:01 |
| inc0 | ? | 16:01 |
| apuimedo | inc0: https://paste.fedoraproject.org/427103/47367050/ | 16:02 |
| sdake_ | we broke 5k commits in the repo ! | 16:02 |
| inc0 | woot for kolla! | 16:02 |
| *** williamcaban has quit IRC | 16:02 | |
| inc0 | apuimedo, so I see just usual keystone user/password auth | 16:02 |
| sdake_ | apuimedo you have to read all the reviews not jusst the last one | 16:02 |
| apuimedo | here you have an example with devstack | 16:03 |
| *** egonzalez90 has quit IRC | 16:03 | |
| inc0 | which means we'll need to register kuryr in ks | 16:03 |
| apuimedo | sdake_: I'll check them all | 16:03 |
| inc0 | let me check if it's in patchset | 16:03 |
| apuimedo | gotta finish a meeting first ;-) | 16:03 |
| apuimedo | inc0: you can register it to the ks, I do it in devstack | 16:03 |
| apuimedo | or you can use admin or neutron credentials | 16:04 |
| inc0 | yeah, what I mean is we have task to register in our plays | 16:04 |
| apuimedo | cool | 16:04 |
| inc0 | which I don't see present in this patchset | 16:04 |
| inc0 | which means it needs adding for it to work | 16:04 |
| apuimedo | well, originally we just used admin credentials | 16:04 |
| apuimedo | so that's probably where this is coming from | 16:05 |
| inc0 | https://github.com/openstack/kolla/blob/master/ansible/roles/heat/tasks/register.yml | 16:05 |
| inc0 | yeah, we can't do this:) | 16:05 |
| apuimedo | inc0: hence why added the registering in devstack | 16:05 |
| apuimedo | but we do not register endpoints | 16:06 |
| inc0 | second task | 16:06 |
| inc0 | is to create user and role | 16:06 |
| inc0 | that one you need | 16:06 |
| apuimedo | role needs to be admin to be able to bind with ovs agent IIRC | 16:07 |
| inc0 | well, you can just use same approach as neutron, we don't use admin creds for services anywhere | 16:07 |
| apuimedo | which is? | 16:08 |
| inc0 | https://github.com/openstack/kolla/blob/master/ansible/roles/neutron/tasks/register.yml#L25 pretty much the same;) | 16:08 |
| inc0 | also I don't think perms in ks has anything to do with OVS, anyway, just duplicate behavior of neutron register | 16:09 |
| inc0 | also I think what you really talk to is ovs-db | 16:09 |
| apuimedo | yes, role admin | 16:09 |
| apuimedo | as I said ;-) | 16:09 |
| apuimedo | this should work | 16:09 |
| apuimedo | thanks inc0 | 16:09 |
| apuimedo | I'll try to sync with huikang in the morning | 16:10 |
| apuimedo | our timezones do not overlap much | 16:10 |
| inc0 | well, we tag in few hrs | 16:10 |
| inc0 | so if you won't do this in next 8 or so hrs, it bumps to O | 16:10 |
| openstackgerrit | caoyuan proposed openstack/kolla: Remove the redundant operation https://review.openstack.org/370977 | 16:10 |
| *** jgriffith has joined #openstack-kolla | 16:11 | |
| inc0 | let me get through rest of reviews and I can help you guys, I'd love to have it merged in N, so close | 16:11 |
| openstackgerrit | caoyuan proposed openstack/kolla: Remove the redundant operation doc/quickstart.rst https://review.openstack.org/370977 | 16:12 |
| openstackgerrit | caoyuan proposed openstack/kolla: Use "type" to take place of "which" and update the output https://review.openstack.org/369551 | 16:12 |
| apuimedo | inc0: agreed | 16:13 |
| apuimedo | I'm not very knowledgable with kolla's brand of jinja over Dockerfiles | 16:13 |
| inc0 | apuimedo, so I see potential to merge kuryr service ansible part | 16:13 |
| huikang | hi, <apuimedo> | 16:13 |
| apuimedo | but otherwise I'll do as much as I can | 16:13 |
| jmccarthy | Hmm does an error like this ring any bells ? Host: control02_ru5, AnsibleUndefinedVariable: 'dict object' has no attribute u'ansible_bond0' I do have a bond0, and I see it in the ansible facts ? http://paste.openstack.org/show/577665/ | 16:13 |
| apuimedo | hi huikang! | 16:13 |
| huikang | https://review.openstack.org/#/c/370382/ | 16:13 |
| huikang | apuimedo, for kolla, I split the container image to kuryr and kuryr-libnetwork | 16:14 |
| *** mewald1 has joined #openstack-kolla | 16:14 | |
| *** absubram has quit IRC | 16:14 | |
| apuimedo | what's on kuryr image? | 16:14 |
| huikang | because later we probabaly need kuryr-kubernente for kolla-kubernenet | 16:14 |
| mewald1 | guys, is stuff like "adding a controller node" somehow tested? | 16:14 |
| openstackgerrit | caoyuan proposed openstack/kolla: Remove the redundant operation in doc/quickstart.rst https://review.openstack.org/370977 | 16:14 |
| inc0 | jmccarthy, which task? | 16:15 |
| pbourke | mewald1: yes, add new one to your inventory and re-run deploy | 16:15 |
| inc0 | mewald1, uhh... | 16:15 |
| inc0 | pbourke, not that simple | 16:15 |
| huikang | apuimedo, the kuryr image contains kuryr, while kuryr-libnetwork has kuryr-libnetwork | 16:15 |
| jmccarthy | inc0, TASK [haproxy : Copying over haproxy.cfg] | 16:15 |
| mewald1 | pbourke: we did so and crashed miserably :D | 16:15 |
| huikang | apuimedo, does that make sense? | 16:15 |
| pbourke | inc0: this does not bode well for woot for kolla | 16:15 |
| apuimedo | huikang: kuryr is now a dependency of kuryr-libnetwork | 16:16 |
| apuimedo | so I think a single image is enough | 16:16 |
| inc0 | pbourke, issue is, databases aren't exactly happy with just messing with their cluster | 16:16 |
| apuimedo | kuryr is just a library | 16:16 |
| apuimedo | like neutron-lib | 16:16 |
| inc0 | galera being most edgy | 16:16 |
| pbourke | inc0: it's worked for me in the past but maybe there are issues | 16:16 |
| huikang | apuimedo, in kolla, there is a base image, e.g, glance-base. Then its children include glance-api and glance-registry | 16:16 |
| mewald1 | inc0: also our experience. we had trouble with rabbitmq but found a solution (bug report incoming) but mariadb went completely out of control | 16:17 |
| pbourke | that needs to be looked at high priority | 16:17 |
| pbourke | mewald1: could you log a bug? | 16:17 |
| inc0 | so with rabbitmq *I hope* that if you run deploy + reconfigure to change clusterer configs on existing nodes...it might rebalance itself? might not? | 16:17 |
| jmccarthy | Hmm I do see config.json on both nodes in /etc/kolla/haproxy where it should be | 16:17 |
| huikang | apuimedo, I think the same principle is applicalbe to kuryr: kuryr-base, kuryr-libnetwork | 16:17 |
| apuimedo | huikang: then it can make sense, if it's not too much work | 16:17 |
| huikang | apuimedo, actually the commit has that implemented | 16:17 |
| apuimedo | then we can re-use the base when we add the kuryr-kubernetes image in the next cycle | 16:17 |
| huikang | sorry, the PS | 16:17 |
| inc0 | pbourke, tbh I don't really know how to do it safely even manually | 16:17 |
| inc0 | kolla or not kolla | 16:17 |
| huikang | apuimedo, yes, exactly | 16:18 |
| inc0 | adding services is easy, it's databases that suck | 16:18 |
| huikang | https://review.openstack.org/#/c/370382/ apuimedo, this is the PS | 16:18 |
| *** Marcellin__ has joined #openstack-kolla | 16:18 | |
| pbourke | inc0: well, we tend to run our databases on a separate nodes. rabbitmq should work | 16:18 |
| apuimedo | huikang: did you check the new README I contributed? | 16:18 |
| apuimedo | about how we generate config now | 16:18 |
| huikang | apuimedo, could you send me the link | 16:18 |
| apuimedo | sure | 16:18 |
| inc0 | pbourke, that's why you're good;) | 16:18 |
| inc0 | mewald1, if you'd pin mariadb and rabbitmq nodes to specific hosts | 16:19 |
| inc0 | and not extend these two | 16:19 |
| inc0 | you might be good | 16:19 |
| huikang | apuimedo, in kolla, we will address the config in kuryr ansible role once the container image file gets merged | 16:19 |
| inc0 | should be even | 16:19 |
| mewald1 | inc0: ok will try | 16:19 |
| apuimedo | huikang: https://github.com/openstack/kuryr-libnetwork/commit/068e3ad5a19c4c08106d9f5b2964a2a9b402b6af | 16:19 |
| inc0 | as far as scaling up databases go...there is quite a few problems to solve | 16:19 |
| apuimedo | see the changes in README.rst in the commit | 16:19 |
| *** matrohon has quit IRC | 16:19 | |
| inc0 | mewald1, also you'll also want to run reconfigure after you're done with deployment | 16:19 |
| apuimedo | this is the patch that also contributes the upstream Dockerfile | 16:20 |
| inc0 | not sure about ceph-mons too | 16:20 |
| inc0 | there will be strange period of 3 mons thinking they're cluster of 3 and 2 more will think this is cluster of 5 | 16:20 |
| inc0 | so you might want to turn off these things;) | 16:20 |
| huikang | apuimedo, this is cool; we will start the kolla kuryr container like this | 16:21 |
| inc0 | or remove containers of databases/ceph-mons/rabbitmq but keep data? | 16:21 |
| inc0 | meh, this is fun | 16:21 |
| inc0 | good luck man:P | 16:21 |
| apuimedo | thanks huikang | 16:22 |
| apuimedo | let me know anything you need me to do to get this merged | 16:22 |
| inc0 | for what is worth, I probably will do these experiments in following months | 16:22 |
| inc0 | to figure out good way to scale up controllers with minimal downtime | 16:22 |
| apuimedo | I can probably spare two hours in the next 8h | 16:22 |
| inc0 | but there is science to be done! | 16:23 |
| mewald1 | inc0, pbourke: Do you see any problems with adding another network node? | 16:23 |
| huikang | apuimedo, let me try the kuryr-libnetwork in kolla now | 16:23 |
| inc0 | mewald1, I don't think so | 16:23 |
| pbourke | mewald1: not that I know of | 16:23 |
| apuimedo | huikang: cool | 16:23 |
| inc0 | only service that might be affected really is keepalived/neutron vrrp which is keepalived | 16:24 |
| inc0 | and afair keepalived know how to scale up | 16:24 |
| inc0 | btw mewald1 monitoring is merged \o/ | 16:25 |
| inc0 | thanks man:) | 16:25 |
| *** ssurana has quit IRC | 16:25 | |
| inc0 | jmccarthy, sorry, I lost the track | 16:25 |
| mewald1 | inc0: nice! and thanks for taking care of it over the last days :) | 16:25 |
| inc0 | hmm...seems like ansible fact caching? | 16:26 |
| jmccarthy | inc0 it's np :) | 16:26 |
| inc0 | brb changing vpns | 16:26 |
| *** inc0 has quit IRC | 16:26 | |
| *** inc0 has joined #openstack-kolla | 16:27 | |
| *** jascott1 has joined #openstack-kolla | 16:27 | |
| inc0 | back | 16:27 |
| jmccarthy | Yea the config.json on the controller is there in /etc/kolla/haproxy as the task before succeeds just fine, but then the TASK [haproxy : Copying over haproxy.cfg] tasks fails | 16:27 |
| *** jtriley has joined #openstack-kolla | 16:27 | |
| inc0 | because I think haproxy.cfg is first taks when we actually use this | 16:28 |
| jmccarthy | Use which ? | 16:28 |
| inc0 | also, all of the nodes has bond0? | 16:28 |
| inc0 | of haproxy? | 16:28 |
| inc0 | because it will fail like this even if problem is lack of bond0 in different node | 16:28 |
| jmccarthy | They do yep (have the bond0) , I manually check the facts like this also: http://paste.openstack.org/show/577665/ | 16:29 |
| inc0 | try ansible -m setup -i << your kolla inventory >> network | 16:30 |
| jmccarthy | You mean to confirm all nodes have a bond0 ? Ok I'll double check | 16:30 |
| inc0 | yeah | 16:30 |
| *** jascott1 has quit IRC | 16:30 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: Splitting Keystone bootstrap into three steps https://review.openstack.org/369774 | 16:31 |
| *** absubram has joined #openstack-kolla | 16:31 | |
| *** jascott1 has joined #openstack-kolla | 16:31 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Changing keystone-public service to use generic service https://review.openstack.org/369767 | 16:33 |
| *** gfhellma has joined #openstack-kolla | 16:34 | |
| jmccarthy | Just looped through them all, they all do have bond0 | 16:35 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Changing keystone-public service to use generic service https://review.openstack.org/369767 | 16:36 |
| inc0 | jmccarthy, hmm...strange...did you add bond0 now? | 16:41 |
| *** fguillot has quit IRC | 16:41 | |
| inc0 | do you have ansible fact caching enabled? | 16:41 |
| jmccarthy | They were there, the vms are all created with mgmt addresses on bond0 | 16:43 |
| jmccarthy | inc0: Oh where do I check that ? | 16:44 |
| inc0 | ansible condig | 16:47 |
| inc0 | config | 16:47 |
| inc0 | you can also add debug task and see what's there | 16:47 |
| jmccarthy | Hmm I do see =memory in /etc/ansible/ansible.cfg | 16:48 |
| jmccarthy | I mean it is enabled there | 16:48 |
| *** thumpba has quit IRC | 16:48 | |
| inc0 | - debug: msg=" {{ hostvars['yourhosname'] }}" | 16:48 |
| inc0 | memory shouldn't affect it tbh | 16:48 |
| *** thumpba has joined #openstack-kolla | 16:49 | |
| *** vhosakot has joined #openstack-kolla | 16:49 | |
| *** senk has joined #openstack-kolla | 16:51 | |
| jmccarthy | inc0, in the debug, I find ..."bitmq_cluster_cookie': None, u'ansible_bond0': {u'lacp_r" ... | 16:52 |
| inc0 | oO | 16:52 |
| inc0 | on all the network nodes? | 16:52 |
| jmccarthy | I'm not sure what is the deal, yep it's there for both (in my case I have haproxy setup across 2 control nodes) | 16:53 |
| jmccarthy | I did check them all, it's weird ? :/ | 16:54 |
| *** thumpba has quit IRC | 16:54 | |
| inc0 | and all the APIs out there have bond0? | 16:54 |
| inc0 | yeah, wierd | 16:54 |
| sdake_ | mewald1 thanks zhubingbing | 16:55 |
| sdake_ | mewald1 he did the work :) | 16:55 |
| inc0 | https://github.com/openstack/kolla/blob/master/ansible/roles/haproxy/templates/haproxy.cfg.j2 all the api_interfaces there has to be correct | 16:55 |
| inc0 | it's not just haproxy nodes | 16:55 |
| inc0 | it's all of the nodes really | 16:55 |
| inc0 | where we have APIs | 16:55 |
| zhubingbing | hi | 16:55 |
| zhubingbing | i am here | 16:55 |
| sdake_ | NOTE - we have seen evidence that ansible fact caching is broken | 16:55 |
| zhubingbing | hi inc | 16:55 |
| zhubingbing | inc0 | 16:56 |
| inc0 | good morning zhubingbing :) | 16:56 |
| zhubingbing | i am back | 16:56 |
| sdake_ | NOTE continued: the workaround to that is to remove the fact cache | 16:56 |
| jmccarthy | They all do though, it's part of the template I use to set them up, I couldn't ssh into them if they didn't all have their addresses on bond0 | 16:56 |
| sdake_ | NOTE the behavior typically seen in this scenario is variables unresolvable | 16:56 |
| zhubingbing | hikang etcd role or kuyr have problem ？ | 16:56 |
| sdake_ | that obviously should be resolvable | 16:56 |
| sdake_ | kfox1111 back to your problem | 16:57 |
| jmccarthy | sdake_: Oh ? Ok so disable 'fact_caching = memory' I'll try it | 16:57 |
| sdake_ | jmccarthy - you can just delete he on disk fact cache | 16:58 |
| sdake_ | the problem appears that the ondisk fact cache does not garbage collect | 16:58 |
| sdake_ | its supposed to last 5 minutes | 16:58 |
| sdake_ | instead it doesn't get removed | 16:58 |
| jmccarthy | sdake_: Oh ok sorry, delete what now ? :) | 16:58 |
| *** fragatina has joined #openstack-kolla | 16:59 | |
| sdake_ | i dont know where it is precisely | 17:00 |
| sdake_ | i think .ansible | 17:00 |
| sdake_ | in your directory | 17:00 |
| inc0 | sdake_, I don't think memory cache creates files | 17:00 |
| sdake_ | but id not delete evereything there | 17:00 |
| sdake_ | the default is not a memory cache | 17:00 |
| *** msimonin has joined #openstack-kolla | 17:00 | |
| inc0 | anyway, try to remove cache and rerun | 17:00 |
| *** fragatina has quit IRC | 17:00 | |
| zhubingbing | hi inc0 | 17:00 |
| sdake_ | my conceren with specifying memory is you have an on disk cache | 17:01 |
| zhubingbing | https://review.openstack.org/#/c/335502/ | 17:01 |
| *** fragatina has joined #openstack-kolla | 17:01 | |
| zhubingbing | you can review it yesterday, i have fix gate error | 17:01 |
| jmccarthy | sdake_ inc0 Ok cool I'll have a look there ! | 17:01 |
| sdake_ | so that may or may not take precednt no matter what your fact caching says | 17:01 |
| sdake_ | jmccarthy i missed the original problem | 17:01 |
| sdake_ | jmccarthy just letting you know fact caching is suspect in ansible at this time | 17:01 |
| *** unicell has quit IRC | 17:02 | |
| *** msimonin has quit IRC | 17:03 | |
| *** gfhellma has quit IRC | 17:03 | |
| jmccarthy | sdake_ Ok no probs, yea it's a weird behaviour right now, will poke around some more in a bit, have to go sort something else out - thanks guys ! | 17:03 |
| *** huikang_ has joined #openstack-kolla | 17:04 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: add reno note monitoring https://review.openstack.org/370345 | 17:06 |
| *** eaguilar has quit IRC | 17:06 | |
| *** eaguilar has joined #openstack-kolla | 17:08 | |
| *** zhubingbing has quit IRC | 17:08 | |
| *** thumpba has joined #openstack-kolla | 17:17 | |
| *** senk has quit IRC | 17:18 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Additional nova changes required to launch vm in all-in-one https://review.openstack.org/371013 | 17:23 |
| *** msimonin has joined #openstack-kolla | 17:25 | |
| *** fguillot has joined #openstack-kolla | 17:26 | |
| *** msimonin has quit IRC | 17:26 | |
| *** msimonin has joined #openstack-kolla | 17:27 | |
| *** jascott1 has quit IRC | 17:29 | |
| *** jascott1 has joined #openstack-kolla | 17:29 | |
| *** huikang_ has quit IRC | 17:32 | |
| *** jascott1 has quit IRC | 17:34 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: net=host + resolv.conf workaround. https://review.openstack.org/369807 | 17:35 |
| *** msimonin has quit IRC | 17:35 | |
| *** fragatina has quit IRC | 17:36 | |
| openstackgerrit | Merged openstack/kolla: Switch arguments to usermod https://review.openstack.org/370450 | 17:37 |
| *** ssurana has joined #openstack-kolla | 17:38 | |
| *** yingjun has quit IRC | 17:39 | |
| *** inc0 has quit IRC | 17:41 | |
| *** inc0 has joined #openstack-kolla | 17:41 | |
| *** gfhellma has joined #openstack-kolla | 17:41 | |
| inc0 | guys, can this bp be closed? | 17:42 |
| inc0 | https://blueprints.launchpad.net/kolla/+spec/enable-networking-sfc-support | 17:42 |
| *** gfidente has quit IRC | 17:44 | |
| *** jascott1 has joined #openstack-kolla | 17:46 | |
| *** williamcaban has joined #openstack-kolla | 17:46 | |
| *** msimonin has joined #openstack-kolla | 17:48 | |
| *** bdaca has joined #openstack-kolla | 17:49 | |
| *** williamcaban has quit IRC | 17:51 | |
| *** unicell has joined #openstack-kolla | 17:52 | |
| *** eaguilar has quit IRC | 17:54 | |
| *** jascott1 has quit IRC | 17:58 | |
| *** jascott1 has joined #openstack-kolla | 17:58 | |
| *** jascott1 has quit IRC | 17:59 | |
| *** jascott1 has joined #openstack-kolla | 17:59 | |
| *** jascott1_ has joined #openstack-kolla | 17:59 | |
| *** jascott1 has quit IRC | 18:01 | |
| *** huikang_ has joined #openstack-kolla | 18:03 | |
| *** fragatina has joined #openstack-kolla | 18:04 | |
| *** tonanhngo has joined #openstack-kolla | 18:06 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: ovs setup bridge, fix /run, add missing hostipc https://review.openstack.org/369785 | 18:11 |
| *** huikang_ has quit IRC | 18:13 | |
| *** huikang has quit IRC | 18:17 | |
| *** eaguilar has joined #openstack-kolla | 18:19 | |
| *** daneyon has joined #openstack-kolla | 18:20 | |
| *** jascott1_ has quit IRC | 18:21 | |
| *** jascott1 has joined #openstack-kolla | 18:22 | |
| *** daneyon has quit IRC | 18:25 | |
| *** sdake_ is now known as sdake | 18:26 | |
| *** dupengfei has quit IRC | 18:30 | |
| *** eaguilar_ has joined #openstack-kolla | 18:33 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: ovs setup bridge, fix /run, add missing hostipc https://review.openstack.org/369785 | 18:36 |
| kfox1111 | sdake: rhallisey: sbezverk: I switched just the nova-compute and nova-libvirt containers to ubuntu and the vm fired up. | 18:36 |
| *** eaguilar has quit IRC | 18:36 | |
| kfox1111 | so its something specific to the minikube kernel and the centos libvirt. | 18:37 |
| rhallisey | oo | 18:37 |
| rhallisey | interesting | 18:37 |
| sdake | SE LINUX | 18:37 |
| kfox1111 | no selinux on the host. | 18:37 |
| kfox1111 | I checked. | 18:37 |
| sdake | oops ssorry about caps | 18:37 |
| kfox1111 | and no app armor either. | 18:37 |
| kfox1111 | :) | 18:37 |
| sdake | I meant SElinux | 18:37 |
| kfox1111 | no worries. selinux has that effect on people. ;) | 18:38 |
| sdake | makes me hold onto my shift key longer then neessary:) | 18:38 |
| kfox1111 | that was the first thing I checked when I saw weirdness. | 18:38 |
| sdake | fwiw - there is a huge delta between ubuntu and centos on the qemu front | 18:38 |
| kfox1111 | yeah. I could believe that. | 18:39 |
| kfox1111 | though I've built a version of centos's from scratch. | 18:39 |
| kfox1111 | like the linux kernel, there is a bazillion patches on top. | 18:39 |
| kfox1111 | like hundreds. | 18:39 |
| sdake | we use the virtualiation sig version | 18:39 |
| sdake | of 2.5 - not many carried paches in that version iiuc | 18:40 |
| kfox1111 | hmm... I saw the -ev version being used, which I thougth was the one from overt. | 18:40 |
| kfox1111 | ovirt. | 18:40 |
| sdake | its from the virtualiation sig not the ovirt sig | 18:40 |
| kfox1111 | hmm. k. | 18:40 |
| kfox1111 | I'll have to take a closer look at some point then. | 18:41 |
| sdake | the ovirt sig and rdo use the virtualization sig's work product | 18:41 |
| kfox1111 | I'm using an ovirt version on one of my clouds. but want to go to a more stock version. | 18:41 |
| sdake | ovirt sig intends then to use :) | 18:41 |
| sdake | my info is wrong | 18:41 |
| *** fragatina has quit IRC | 18:43 | |
| *** gfhellma has quit IRC | 18:43 | |
| *** fragatina has joined #openstack-kolla | 18:43 | |
| *** fguillot has quit IRC | 18:44 | |
| kfox1111 | ok. so with the ubuntu libvirt/nova-compute, the network's working execpt for the nova metadata server... that next. | 18:47 |
| openstackgerrit | Merged openstack/kolla: Bump version of Kibana from 4.4 to 4.6 https://review.openstack.org/369076 | 18:48 |
| openstackgerrit | Merged openstack/kolla: Remove public_router_id from tempest configuration template https://review.openstack.org/370825 | 18:49 |
| *** lrensing has quit IRC | 18:50 | |
| openstackgerrit | Merged openstack/kolla: Document how to add logging for a service https://review.openstack.org/369037 | 18:51 |
| *** fguillot has joined #openstack-kolla | 18:51 | |
| *** fguillot has quit IRC | 18:51 | |
| *** fguillot has joined #openstack-kolla | 18:52 | |
| rhallisey | sdake, https://review.openstack.org/343224 | 18:55 |
| *** gfhellma has joined #openstack-kolla | 18:55 | |
| rhallisey | rebased | 18:55 |
| sdake | rhallisey thanks | 18:56 |
| sdake | reading for correctness nwo | 18:56 |
| sdake | moment | 18:56 |
| rhallisey | sure | 18:56 |
| openstackgerrit | Merged openstack/kolla-kubernetes: First stab at minikube doc https://review.openstack.org/370930 | 18:57 |
| openstackgerrit | prameswar proposed openstack/kolla: kolla config file path corrected for ubuntu https://review.openstack.org/370907 | 18:59 |
| *** lyncos has joined #openstack-kolla | 19:00 | |
| *** jascott1_ has joined #openstack-kolla | 19:00 | |
| *** absubram has quit IRC | 19:01 | |
| *** jascott1 has quit IRC | 19:04 | |
| *** salv-orlando has quit IRC | 19:07 | |
| *** matrohon has joined #openstack-kolla | 19:07 | |
| *** rstarmer has joined #openstack-kolla | 19:11 | |
| mliima | guys | 19:12 |
| mliima | guys, why physical_interface_mappings = physnet1:{{ neutron_external_interface }} and not {{ network_interface }} ? | 19:12 |
| *** lyncos has quit IRC | 19:13 | |
| *** msimonin has quit IRC | 19:13 | |
| *** rstarmer has quit IRC | 19:15 | |
| *** huikang has joined #openstack-kolla | 19:15 | |
| *** fragatina has quit IRC | 19:17 | |
| *** fragatina has joined #openstack-kolla | 19:18 | |
| *** fragatina has quit IRC | 19:19 | |
| *** fragatina has joined #openstack-kolla | 19:19 | |
| *** eaguilar_ has quit IRC | 19:21 | |
| *** bjolo_ has joined #openstack-kolla | 19:21 | |
| bjolo_ | goodevening | 19:23 |
| inc0 | mliima, because you don't want neutron messing your api_interface and such;) | 19:30 |
| mliima | hmm, i'm using linuxbridget and my instances were unable to access the internet | 19:31 |
| *** aswadr_ has quit IRC | 19:32 | |
| mliima | so i change physical_interface_mappings and gotcha! i have internet | 19:32 |
| mliima | :/ | 19:32 |
| *** fragatina has quit IRC | 19:32 | |
| mliima | is there any network configuration that needs to be made so that the instances have access to internet? inc0 | 19:33 |
| *** williamcaban has joined #openstack-kolla | 19:34 | |
| inc0 | mliima, well, quite some | 19:35 |
| mliima | oh god, inc0 :( | 19:37 |
| *** fragatina has joined #openstack-kolla | 19:38 | |
| *** williamcaban has quit IRC | 19:39 | |
| *** fragatin_ has joined #openstack-kolla | 19:40 | |
| *** fragatina has quit IRC | 19:40 | |
| *** david-lyle has quit IRC | 19:40 | |
| *** david-lyle has joined #openstack-kolla | 19:40 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Additional nova changes required to launch vm in all-in-one https://review.openstack.org/371013 | 19:41 |
| *** eaguilar has joined #openstack-kolla | 19:41 | |
| inc0 | mliima, sooo generally if you create flat network | 19:41 |
| inc0 | in this interface | 19:41 |
| inc0 | network from this interface is routable to internet | 19:42 |
| inc0 | it should be ok | 19:42 |
| inc0 | basically just networking mliima ;) | 19:43 |
| inc0 | this is quite a field to get to know | 19:43 |
| *** fragatin_ has quit IRC | 19:44 | |
| mliima | I think it might be a bug in kolla, i can access the instances when not using the kolla. | 19:46 |
| mliima | or it might be a bug in me hahaha | 19:47 |
| inc0 | it's not kolla issue;) in fact we'd break stuff if we put neutron on network_interface | 19:47 |
| inc0 | and I mean really bad breakage;) | 19:47 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Additional nova changes required to launch vm in all-in-one https://review.openstack.org/371013 | 19:48 |
| mliima | i can't put neutron on network_interface | 19:48 |
| *** sdake has quit IRC | 19:49 | |
| *** sdake_ has joined #openstack-kolla | 19:49 | |
| inc0 | no, you can't and shouldnt | 19:49 |
| mliima | hehe | 19:49 |
| mliima | i know | 19:49 |
| mliima | :) | 19:49 |
| *** jascott1_ has quit IRC | 19:51 | |
| *** jascott1 has joined #openstack-kolla | 19:52 | |
| *** jascott1 has quit IRC | 19:52 | |
| mliima | the question here is more simple, why when i'm not using kolla i can access the internet and using kolla my instance are unable to access it? there is some additional configuration that kolla does not do? | 19:52 |
| *** jascott1 has joined #openstack-kolla | 19:52 | |
| *** jascott1 has quit IRC | 19:53 | |
| *** jascott1 has joined #openstack-kolla | 19:54 | |
| sdake_ | mlima that takes a rocket scientist understanding of networking to sort out | 19:54 |
| sdake_ | not to discourage you :) | 19:54 |
| *** jascott1 has quit IRC | 19:57 | |
| mliima | np sdake, but it's an issue that I want and need resolve. :) | 19:58 |
| mliima | it'll work | 19:58 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Changing keystone-public service to use generic service https://review.openstack.org/369767 | 20:00 |
| *** fragatina has joined #openstack-kolla | 20:01 | |
| *** bdaca has quit IRC | 20:03 | |
| *** awiddersheim has joined #openstack-kolla | 20:04 | |
| *** fragatina has quit IRC | 20:05 | |
| *** mliima has quit IRC | 20:08 | |
| *** absubram has joined #openstack-kolla | 20:09 | |
| *** tonanhngo has quit IRC | 20:10 | |
| *** fragatina has joined #openstack-kolla | 20:10 | |
| *** fragatina has quit IRC | 20:15 | |
| *** rhallisey has quit IRC | 20:16 | |
| *** lyncos has joined #openstack-kolla | 20:19 | |
| *** msimonin has joined #openstack-kolla | 20:23 | |
| *** msimonin has quit IRC | 20:26 | |
| *** fguillot has quit IRC | 20:28 | |
| *** fguillot has joined #openstack-kolla | 20:28 | |
| *** jascott1 has joined #openstack-kolla | 20:29 | |
| *** fragatina has joined #openstack-kolla | 20:29 | |
| *** ccesario has quit IRC | 20:30 | |
| *** tonanhngo has joined #openstack-kolla | 20:30 | |
| *** lyncos has quit IRC | 20:31 | |
| sdake_ | huikang where are you with kuryr | 20:32 |
| sdake_ | huikang doable today or not doable? | 20:33 |
| apuimedo | sdake_: he told me he got it working with https://review.openstack.org/#/c/298894/22/ | 20:33 |
| sdake_ | theanks apuimedo - i'll review now | 20:33 |
| apuimedo | thanks for the wait and review | 20:33 |
| sdake_ | rhi have reviewed partof your docs changes, still working on it - feel there are many changes needed | 20:34 |
| sdake_ | rhallisey i'l get em in the review today aftr fires are out | 20:34 |
| *** tonanhngo has quit IRC | 20:35 | |
| huikang | sdake_, the PS are tested by following the guide | 20:39 |
| huikang | but the first patch you need to merge is the kuryr-libnetwork one | 20:40 |
| sdake_ | huikang rviewing atm - i'll be back ;) | 20:40 |
| huikang | sure; i need some rest as well | 20:40 |
| *** jtriley has quit IRC | 20:40 | |
| sdake_ | huikang lets get a link up | 20:40 |
| sdake_ | i'll have someone take over th review from hear if you need to hit the rack | 20:40 |
| sdake_ | hear/here | 20:41 |
| * sdake_ learns grammar today | 20:41 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: More test stuff and create nova-metadata svc. https://review.openstack.org/371148 | 20:41 |
| *** tonanhngo has joined #openstack-kolla | 20:41 | |
| sdake_ | huikang beforey ou depart | 20:41 |
| sdake_ | why is /usrlib/docker needeed as a bindmount? | 20:41 |
| *** mewald1 has quit IRC | 20:43 | |
| huikang | Becasue docker needs access that directory on the host | 20:43 |
| huikang | apuimedo, can you answer that? | 20:43 |
| apuimedo | I can! | 20:44 |
| apuimedo | when you request a driver in a docker operation, such as docker network create | 20:44 |
| apuimedo | docker searches /usr/lib/docker or /etc/docker subdirs for network/storage plugin specs or json definitions | 20:45 |
| apuimedo | (on the host, of course) | 20:45 |
| openstackgerrit | Hui Kang proposed openstack/kolla: Add Kuryr ansible role https://review.openstack.org/298894 | 20:45 |
| apuimedo | so it's either have ansible place the file there, or volume mount it and let the container place the file there | 20:45 |
| bjolo_ | trying to setup up multinode and make use of all network types like storage, cluster, tunnel. api is still using network_interface tho. But i get a lot of errors that interfaces are not there but they are | 20:47 |
| bjolo_ | FAILED! => {"failed": true, "msg": "'dict object' has no attribute u'ansible_bond0.822'"} | 20:47 |
| bjolo_ | anyone done a multinode multinetwork setup? | 20:48 |
| bjolo_ | storage_interface: "bond0.822" | 20:49 |
| *** inc0 has quit IRC | 20:50 | |
| sdake_ | sbezverk - pleae look at just this one change: https://review.openstack.org/#/c/298894/22/etc/kolla/globals.yml | 20:50 |
| sdake_ | apuimedo makes sense - i have suggested in the review to put that in the commit message | 20:51 |
| sdake_ | huikang my review is up | 20:52 |
| sdake_ | i think its nearly ready for merge | 20:52 |
| sdake_ | any other core reviews that can take a look at https://review.openstack.org/#/c/298894/23 | 20:53 |
| sdake_ | bjolo_ - today is tag day - we are super busy atm - can we get back to you? | 20:53 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Fix svc to point to nova-api pod housing the metadata server https://review.openstack.org/371153 | 20:53 |
| bjolo_ | i.e. can anyone confirm that multinetwork should work and that the problem is on my end | 20:53 |
| sdake_ | bjolo_ - actually I know what your problem is | 20:53 |
| sdake_ | bjolo_ ansible fact caching | 20:54 |
| sdake_ | delete your ansible fact cache | 20:54 |
| huikang | sdake_, I will address your review now. Meanwhile, could you look at this https://review.openstack.org/#/c/370382/ | 20:54 |
| bjolo_ | aahhhh | 20:54 |
| sdake_ | hui you got it | 20:54 |
| sdake_ | bjolo_ just a guess | 20:54 |
| sdake_ | bjolo_ but try it | 20:54 |
| bjolo_ | sdake_, think you are right, but let me try first | 20:54 |
| sdake_ | we ran into this on osic | 20:54 |
| *** egonzalez90 has joined #openstack-kolla | 20:58 | |
| sdake_ | huikang merged that one | 20:58 |
| *** Pavo has quit IRC | 20:59 | |
| *** thumpba has quit IRC | 20:59 | |
| *** thumpba has joined #openstack-kolla | 20:59 | |
| openstackgerrit | Merged openstack/kolla: Add kuryr-libnetwork Dockerfile https://review.openstack.org/370382 | 21:00 |
| sdake_ | zomg vnapps long review | 21:03 |
| *** matrohon has quit IRC | 21:05 | |
| *** gfhellma has quit IRC | 21:09 | |
| *** fguillot has quit IRC | 21:09 | |
| *** awiddersheim has quit IRC | 21:09 | |
| sdake_ | is networing-sfc comlete? | 21:09 |
| sdake_ | all outstanding reviews have merged | 21:09 |
| sdake_ | anyone know if more is coming? | 21:09 |
| sdake_ | hui one last thing - I dont know if you had a releaenote on that change or not | 21:10 |
| sdake_ | but if not, it needs one | 21:10 |
| *** fguillot has joined #openstack-kolla | 21:10 | |
| *** thumpba has quit IRC | 21:10 | |
| *** salv-orlando has joined #openstack-kolla | 21:11 | |
| openstackgerrit | Merged openstack/kolla: Integrate neutron-vpnaas-agent into the neutron Ansible role https://review.openstack.org/335502 | 21:11 |
| *** gfhellma has joined #openstack-kolla | 21:12 | |
| *** huikang has quit IRC | 21:12 | |
| bjolo_ | sdake_, fact_caching = memory should be ok right? | 21:12 |
| *** huikang has joined #openstack-kolla | 21:13 | |
| apuimedo | huikang: In case you missed the message (I saw your connection drop) "sdake_ : hui one last thing - I dont know if you had a releaenote on that change or not but if not, it needs one" | 21:13 |
| sdake_ | bjolo_ no clue - but if you try it we will document it ;) | 21:14 |
| sdake_ | its an ansible bug | 21:15 |
| sdake_ | the efacts cache doesn't time out as it should | 21:15 |
| sdake_ | i don't know if ansible knows yetor not | 21:15 |
| sdake_ | fact cache only supposed to lsat 5 minuts | 21:15 |
| bjolo_ | still error | 21:17 |
| *** huikang has quit IRC | 21:17 | |
| sdake_ | bjolo_ try deleting the fact cache | 21:17 |
| sdake_ | bjolo_ Jeffrey4l__ hass resolved this problem in the past | 21:17 |
| sdake_ | Jeffrey4l__ he has a better understanding o the problem then I do | 21:18 |
| bjolo_ | i would if i could find it :) | 21:18 |
| sdake_ | its in ~root/.ansible | 21:18 |
| sdake_ | or if your running as non-sudo it may be in your home dir under .ansible | 21:18 |
| sdake_ | that much I know :) | 21:18 |
| *** salv-orlando has quit IRC | 21:18 | |
| *** absubram has quit IRC | 21:19 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Fix svc to point to nova-api pod housing the metadata server https://review.openstack.org/371153 | 21:19 |
| sdake_ | kfox1111 need any merges before tag in kolla? | 21:19 |
| kfox1111 | As far as I can tell, we're good now. | 21:19 |
| sdake_ | kfox1111 thanks | 21:20 |
| kfox1111 | got nova/neutron/keystone/horion and support services working from trunk. | 21:20 |
| sdake_ | fantastic | 21:20 |
| sdake_ | compute kit ftw | 21:20 |
| kfox1111 | :) | 21:20 |
| *** fguillot has quit IRC | 21:21 | |
| *** jascott1 has quit IRC | 21:21 | |
| *** salv-orlando has joined #openstack-kolla | 21:22 | |
| *** absubram has joined #openstack-kolla | 21:24 | |
| *** egonzalez90 has quit IRC | 21:25 | |
| *** huikang has joined #openstack-kolla | 21:27 | |
| *** fguillot has joined #openstack-kolla | 21:27 | |
| *** root____1 has joined #openstack-kolla | 21:28 | |
| huikang | sdake_, in your comment, do you mean adding "admin, internal" endpoints for kuryr, although they may not be used? | 21:30 |
| sdake_ | huikang right | 21:30 |
| sdake_ | it is standard for all apis to havve 3 endpoints | 21:30 |
| huikang | no problem | 21:30 |
| sdake_ | intrnal, external, and admin | 21:30 |
| sdake_ | and in olla we definately need internal and external | 21:31 |
| sdake_ | because they mean different things | 21:31 |
| huikang | I think they are admin_endpoint, internal_endpoint, and public_endpoint | 21:31 |
| sdake_ | and interact with tls | 21:31 |
| sdake_ | sound right | 21:31 |
| sdake_ | copy paste should do it ;) | 21:31 |
| huikang | cool | 21:32 |
| sdake_ | as for the interface to bind to, bind to the in3rnal_endpoint | 21:32 |
| sdake_ | which is what your doing | 21:32 |
| sdake_ | thinking of htis triggered a thought on haproxy | 21:32 |
| sdake_ | i didn't see haproxy integration in the code | 21:32 |
| sdake_ | has that already merged? | 21:32 |
| *** jascott1 has joined #openstack-kolla | 21:34 | |
| *** jascott1 has quit IRC | 21:39 | |
| *** jascott1 has joined #openstack-kolla | 21:39 | |
| *** dwalsh has joined #openstack-kolla | 21:41 | |
| *** jascott1 has quit IRC | 21:42 | |
| *** eaguilar has quit IRC | 21:43 | |
| apuimedo | sdake_: haproxy with kuryr-libnetwork? | 21:44 |
| sdake_ | apuimedo kuryr has a control plane right? | 21:44 |
| sdake_ | like some type of api | 21:44 |
| apuimedo | REST, but only accessible through docker | 21:45 |
| sdake_ | hrm | 21:45 |
| *** ntpttr_ has quit IRC | 21:45 | |
| apuimedo | with kuryr-kubernetes we do use LBaaSv2 | 21:45 |
| sdake_ | ok - i'll have to think about that | 21:45 |
| sdake_ | this is for kolla ha purposes | 21:45 |
| sdake_ | we use haproxy to ensure we can run 3 copies sof each control node | 21:45 |
| *** wirehead_ has quit IRC | 21:46 | |
| apuimedo | sdake_: no ha, you're supposed to only listen to the local docker daemon requests | 21:46 |
| apuimedo | with kuryr-libnetwork | 21:46 |
| sdake_ | apuimedo is this only on networking nodes? | 21:46 |
| apuimedo | in any node you want to 'docker run' containers that use neutron networks | 21:47 |
| apuimedo | so typically you have controller nodes where there is Neutron but you don't run workloads. Kuryr-libnetwork doesn't go there | 21:47 |
| apuimedo | it goes to where you'd put nova-compute if you are interested in running containers instead (or alongside) VMs | 21:48 |
| sdake_ | i see - part of the compute role | 21:48 |
| sdake_ | roel wrongword | 21:48 |
| sdake_ | anyway part of the compute infrastructure | 21:48 |
| apuimedo | I didn't say it explicitly because you may want just container and not VMs, which compute has come to imply | 21:49 |
| apuimedo | sdake_: I foresee that at some point you may use it in controller nodes | 21:49 |
| *** jascott1 has joined #openstack-kolla | 21:49 | |
| apuimedo | for example if you want an agent running on the overlay, you could run it with docker in the controller node | 21:49 |
| sdake_ | huikang i noticed one problem with your patch just now after discussing apuimedo | 21:49 |
| huikang | sure | 21:49 |
| sdake_ | yourmissing the inventory exampless | 21:49 |
| apuimedo | but, like Octavia VMs, probably better to have it close to the "useful" containers | 21:50 |
| sdake_ | what is ocavia :) | 21:50 |
| huikang | kuryr runs on ompute node | 21:50 |
| *** jascott1 has quit IRC | 21:50 | |
| sdake_ | huikang right but doesn't it need inventory file content? | 21:50 |
| *** wirehead_ has joined #openstack-kolla | 21:51 | |
| sdake_ | sup wirehead_ | 21:51 |
| huikang | sdake_, do you mean [kuryr] compute ? | 21:52 |
| huikang | now I am doing https://review.openstack.org/#/c/298894/22/ansible/site.yml | 21:52 |
| sdake_ | huikang not sure | 21:52 |
| sdake_ | huikang https://github.com/openstack/kolla/blob/master/ansible/inventory/all-in-one | 21:54 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Fixes to get Nova to work https://review.openstack.org/370459 | 21:54 |
| openstackgerrit | Hui Kang proposed openstack/kolla: Add Kuryr ansible role https://review.openstack.org/298894 | 21:54 |
| sdake_ | huikang i htink its ok - libvirt is not specified in the inventory file | 21:55 |
| huikang | ok, cool | 21:55 |
| sdake_ | and since you instruct kuryr to deploy on all compute nodes | 21:55 |
| huikang | for kuryr to work, please also review the etcd PS | 21:57 |
| *** daneyon has joined #openstack-kolla | 21:57 | |
| *** jheroux has quit IRC | 21:57 | |
| huikang | sdake_, link: https://review.openstack.org/#/c/364662/ | 21:57 |
| *** jascott1 has joined #openstack-kolla | 22:01 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: More test stuff and create nova-metadata svc. https://review.openstack.org/371148 | 22:01 |
| *** daneyon has quit IRC | 22:01 | |
| *** absubram has quit IRC | 22:02 | |
| *** absubram has joined #openstack-kolla | 22:08 | |
| *** dwalsh has quit IRC | 22:12 | |
| sdake_ | huikang on it | 22:13 |
| *** huikang has quit IRC | 22:14 | |
| sdake_ | juietcd not merged yet? | 22:14 |
| *** huikang has joined #openstack-kolla | 22:14 | |
| apuimedo | I don't see it merged | 22:14 |
| apuimedo | only has inc0's +2 | 22:15 |
| apuimedo | https://review.openstack.org/#/c/364662/ | 22:15 |
| *** unicell has quit IRC | 22:16 | |
| *** fguillot has quit IRC | 22:16 | |
| sdake_ | huikang doesn't etcd dneed some magic conig to work in ha mode | 22:17 |
| sdake_ | ok merged | 22:19 |
| *** huikang has quit IRC | 22:19 | |
| sdake_ | what else do we need for kuryr? | 22:19 |
| *** vhosakot has quit IRC | 22:20 | |
| apuimedo | I only see this one https://review.openstack.org/#/c/298894/ | 22:20 |
| sdake_ | damn - im behind :) | 22:21 |
| openstackgerrit | Merged openstack/kolla: Add etcd ansible role https://review.openstack.org/364662 | 22:21 |
| sdake_ | just curious - what is kuryr.spec? | 22:22 |
| apuimedo | sdake_: the file that docker looks for when trying to find network/storage drivers | 22:22 |
| *** absubram has quit IRC | 22:23 | |
| apuimedo | it looks for either .spec or .json | 22:23 |
| sdake_ | got it - just wnted to make sure it wasn't an rpm spec file | 22:23 |
| apuimedo | different formats too | 22:23 |
| sdake_ | (accidental inclusion) | 22:23 |
| apuimedo | no, no. No binary build yet ;-) | 22:23 |
| apuimedo | I'll be doing the packages soon | 22:23 |
| apuimedo | but not yet | 22:23 |
| *** absubram has joined #openstack-kolla | 22:23 | |
| *** williamcaban has joined #openstack-kolla | 22:24 | |
| sdake_ | the reason i asked is yuo mentoined you were doing that firsst | 22:28 |
| *** williamcaban has quit IRC | 22:28 | |
| sdake_ | (some time ago) | 22:28 |
| sdake_ | ok - so need another cor ereviewer | 22:28 |
| sdake_ | anyone around | 22:28 |
| apuimedo | :-) | 22:29 |
| sdake_ | Jeffrey4l__ mandre pbourke - need review on https://review.openstack.org/#/c/298894/ | 22:30 |
| apuimedo | sdake_: but then you cleverly (we wouldn't have had a bit of chance of making it otherwise) suggested first go for source | 22:30 |
| sdake_ | right :) | 22:30 |
| sdake_ | thanks for that | 22:30 |
| *** ntpttr has quit IRC | 22:31 | |
| *** salv-orlando has quit IRC | 22:32 | |
| apuimedo | thanks to you | 22:33 |
| root____1 | hi all | 22:35 |
| *** ntpttr has joined #openstack-kolla | 22:35 | |
| root____1 | I realise, a total noob question, but if I may ask? | 22:38 |
| *** eaguilar has joined #openstack-kolla | 22:38 | |
| sdake_ | groan kuryr is in merge conflict | 22:39 |
| apuimedo | :O | 22:39 |
| sdake_ | i'll fix that in a moment | 22:39 |
| sdake_ | i think huikang went for some well deserved sleep :) | 22:39 |
| *** eaguilar has quit IRC | 22:40 | |
| *** root____1 has quit IRC | 22:40 | |
| sdake_ | RootServ nnob questions ar ewelcome | 22:41 |
| *** thumpba has joined #openstack-kolla | 22:41 | |
| *** ntpttr_ has joined #openstack-kolla | 22:41 | |
| *** sjors_ has joined #openstack-kolla | 22:42 | |
| *** eaguilar has joined #openstack-kolla | 22:42 | |
| *** unicell has joined #openstack-kolla | 22:43 | |
| sjors_ | so, my question, I tried to run kolla on fedora 23 vm, and guess? right, so is it totally a no go? | 22:44 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Add Kuryr ansible role https://review.openstack.org/298894 | 22:44 |
| sdake_ | sjors_ - yup fedora is busted | 22:45 |
| sdake_ | sjors_ may i suggeset centos or ubuntu | 22:45 |
| apuimedo | I also think so | 22:45 |
| apuimedo | I reckon he didn't get any sleep | 22:45 |
| *** salv-orlando has joined #openstack-kolla | 22:46 | |
| *** ChanServ sets mode: +o sdake_ | 22:46 | |
| *** sdake_ changes topic to "RC1 tag blocked on this review: https://review.openstack.org/#/c/298894/ - Kolla IRC meetngs on Wednesdays @ 16:00 UTC - see agenda @ https://goo.gl/OXB0DL - IRC channel is *LOGGED* @ http://goo.gl/3mzZ7b (old logs from #kolla http://goo.gl/VKpPzA)" | 22:47 | |
| *** tonanhngo has quit IRC | 22:47 | |
| *** ChanServ sets mode: -o sdake_ | 22:47 | |
| sjors_ | sdake_, ok well I started with centos, just to see the diff, can't figure it out | 22:47 |
| sjors_ | but centos it is then | 22:48 |
| sdake_ | sjors_ can't figure out which different? | 22:48 |
| sdake_ | centos uses yum, fedora 23 uses dnf | 22:48 |
| sdake_ | we have no dnf implementation | 22:48 |
| sdake_ | and nobody shows much interest in fedora (or debian) | 22:48 |
| sdake_ | so we are liely deprecating these two operating systems following the standard deprecation cycle | 22:49 |
| sjors_ | sdake, the diff in packages, seems all the same, tox however seems to throw errors on hardening in fedora, I guess the unit tests miss a thing | 22:49 |
| sdake_ | we have unit tests | 22:50 |
| sdake_ | fedora doesn't work | 22:51 |
| sdake_ | fedora is deprecated afaik | 22:51 |
| sdake_ | fedora's liecycle is too short to be of interest to operators | 22:51 |
| sdake_ | (our main audience for this project) | 22:51 |
| sjors_ | jup, I tried just to see what breaks, stupid maybe, and yes fedora moves to fast | 22:52 |
| *** gfhellma has quit IRC | 22:53 | |
| sjors_ | anyway, centos building :) interesting way, I have to explore this, so I will have some questions in the future ;) | 22:56 |
| *** tonanhngo has joined #openstack-kolla | 22:56 | |
| *** jamesdenton has quit IRC | 22:56 | |
| *** Pavo has joined #openstack-kolla | 22:56 | |
| *** absubram has quit IRC | 22:59 | |
| *** tonanhngo has quit IRC | 23:00 | |
| *** mtaylor22 has joined #openstack-kolla | 23:02 | |
| *** absubram has joined #openstack-kolla | 23:03 | |
| *** huikang has joined #openstack-kolla | 23:07 | |
| apuimedo | seems we can't get another core at this time of the day :/ | 23:08 |
| apuimedo | I remain here a bit longer in case another one shows up and has some questoin | 23:08 |
| apuimedo | *question | 23:09 |
| *** inc0 has joined #openstack-kolla | 23:13 | |
| sjors_ | thanks all, laters ;) | 23:13 |
| huikang | apuimedo, what is the time in barcelona? | 23:14 |
| *** sjors_ has quit IRC | 23:14 | |
| apuimedo | still early huikang. 1:14am | 23:14 |
| *** PyroMani has quit IRC | 23:14 | |
| apuimedo | it's only gonna get earlier, in fact | 23:14 |
| huikang | nice; just finished my dinner | 23:14 |
| inc0 | sdake_, ping | 23:15 |
| huikang | apuimedo, at lease we have etcd and the kuryr-libnetwork merged | 23:15 |
| inc0 | apuimedo, kuryr creates endpoint type neutron? | 23:15 |
| sdake_ | icn0 shoot | 23:15 |
| sdake_ | inc0 shot | 23:15 |
| apuimedo | inc0: endpoint type neutron?! | 23:15 |
| sdake_ | huikang i rebased your patch | 23:15 |
| inc0 | https://review.openstack.org/#/c/298894/25/ansible/roles/kuryr/tasks/register.yml | 23:15 |
| huikang | sure, thanks sdake_ | 23:15 |
| inc0 | it creates endpoint type neutron, does it even need endpoint? | 23:16 |
| sdake_ | service_type=neutron | 23:16 |
| apuimedo | interesting | 23:16 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Additional nova changes required to launch vm in all-in-one https://review.openstack.org/371013 | 23:16 |
| apuimedo | let me check what I put with devstack | 23:16 |
| huikang | inc0, I am not sure this part | 23:16 |
| inc0 | yeah | 23:16 |
| sdake_ | ya - i missed that | 23:16 |
| apuimedo | oh, right | 23:16 |
| apuimedo | I remember | 23:16 |
| sdake_ | what is there is wrong | 23:16 |
| apuimedo | I create a service, but I don't put endpoints usually, since they are local to each comput host | 23:17 |
| apuimedo | the problem is that in Kolla apparently service and endpoints are tied together | 23:17 |
| *** eaguilar has quit IRC | 23:17 | |
| inc0 | neutron communicates with kuryr over network? | 23:17 |
| apuimedo | nope | 23:17 |
| apuimedo | kuryr talks to neutron | 23:17 |
| apuimedo | not the other way around (except replies) | 23:17 |
| inc0 | depends what do you mean by service | 23:17 |
| apuimedo | so with huikang, we had to decide either to not create a service or to create one with useless endpoints | 23:18 |
| inc0 | this is keystone service | 23:18 |
| *** PyroMani has joined #openstack-kolla | 23:18 | |
| apuimedo | sure | 23:18 |
| apuimedo | I'm talking about keystone service | 23:18 |
| inc0 | do we need to have ks service really? | 23:18 |
| apuimedo | maybe the best would be to drop it | 23:18 |
| apuimedo | I usually create it just to have it listed there | 23:18 |
| apuimedo | not for finding endpoints | 23:18 |
| inc0 | if kuryr doesn't have any endpoints, we shouldn't put them there | 23:19 |
| apuimedo | (kind of service catalog) | 23:19 |
| apuimedo | inc0: it doesn't have region wide endpoints, no | 23:19 |
| inc0 | ok then remove the task please | 23:20 |
| huikang | let me test the kuryr without creating any endpoint, apuimedo and inc0 | 23:20 |
| inc0 | if it doesn't work, we'll treat it as bug | 23:20 |
| sdake_ | yup lets get it merged (working) | 23:20 |
| inc0 | well, it used to if you tested it before;) but sure, re-test it | 23:20 |
| sdake_ | if it has no endpoints its more like a plugin | 23:21 |
| apuimedo | sdake_: inc0: huikang: http://paste.openstack.org/show/577742/ | 23:21 |
| apuimedo | this is how we set it up in devstack | 23:21 |
| apuimedo | (before last week, we didn't even create the service entry) | 23:21 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Fix svc to point to nova-api pod housing the metadata server https://review.openstack.org/371153 | 23:22 |
| sdake_ | so is the problem we need a service, but no endpoint? | 23:22 |
| inc0 | ok, so let's not do this, I'd rather skip it than rewrite our kolla_keystone_service plugin;) | 23:22 |
| inc0 | sdake_, I don't think it needs service at all | 23:22 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Remove nova compute bootstrap https://review.openstack.org/370373 | 23:22 |
| apuimedo | inc0: you are very right | 23:24 |
| apuimedo | it doesn't need it at all | 23:24 |
| sdake_ | huikang i know your running out of steam - any chance youcan fix that up | 23:24 |
| *** eaguilar has joined #openstack-kolla | 23:24 | |
| sdake_ | huikang apologies for not catching it | 23:24 |
| huikang | sdake_, np | 23:24 |
| sdake_ | i spent 30 minutes reviewing it | 23:24 |
| sdake_ | but someitmes sthingsget missed- which is why we have two core reviewers | 23:25 |
| huikang | sdake_, inc0, do you mean removing Creating the Kuryr service and endpoint in the register.yml? | 23:25 |
| inc0 | huikang, yes | 23:26 |
| inc0 | just remove this task alltogether, keep the keystone user creation | 23:26 |
| huikang | sure, inc0 | 23:26 |
| inc0 | question, what's this file? https://review.openstack.org/#/c/298894/25/ansible/roles/kuryr/templates/kuryr.spec.j2 | 23:27 |
| apuimedo | inc0: :-) | 23:27 |
| inc0 | seems...wierd | 23:27 |
| apuimedo | it's what docker uses to find storage or network drivers when it receives a command that specifies drivers | 23:27 |
| apuimedo | like docker network create -d kuryr | 23:28 |
| apuimedo | this prompts it to find .spec or .json files that describe endpoints | 23:28 |
| apuimedo | sees the url | 23:28 |
| apuimedo | sends an activate request | 23:28 |
| inc0 | interesting, I never did stuff like that with docker | 23:28 |
| apuimedo | and receives the types or drivers it implements | 23:28 |
| apuimedo | ;-) | 23:28 |
| apuimedo | it's a funny way to extend it | 23:28 |
| inc0 | here we are, used to net=host | 23:28 |
| apuimedo | inc0: I also use --net host | 23:30 |
| *** thumpba has quit IRC | 23:30 | |
| huikang | still deploying without creating kuryr endpoints | 23:30 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Additional nova changes required to launch vm in all-in-one https://review.openstack.org/371013 | 23:30 |
| apuimedo | for running the infrastructure containers like you, it's the most logical thing | 23:30 |
| apuimedo | (now I'm writing the kuryr-kubernetes devstack, which uses k8s from containers, --net host is what I use) | 23:31 |
| inc0 | apuimedo, by the way, when we merge it, consider using kolla as devstack replacement:) | 23:32 |
| inc0 | at least cleaning up/redeploying is a breeze | 23:32 |
| apuimedo | inc0: I will certainly do so for kuryr-libnetwork | 23:33 |
| apuimedo | bash has trolled me enough | 23:33 |
| apuimedo | inc0: is the infra container registry thing solved? | 23:35 |
| apuimedo | because that would make it a much easier choice | 23:35 |
| apuimedo | having to build the containers on each job is a bit of a drag | 23:36 |
| inc0 | what do you mean by infra container registyr? | 23:36 |
| huikang | FYI docker network create -d kuryr --ipam-driver=kuryr works without endpoint | 23:36 |
| inc0 | I run local registry and I don't rebuild every time;) | 23:36 |
| apuimedo | huikang: thanks for the confirmation | 23:36 |
| apuimedo | inc0: I mean for gates | 23:37 |
| apuimedo | *gate jobs | 23:37 |
| huikang | so I will update the PS by removing that part and the *endpoints | 23:37 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Fixes to get Nova to work https://review.openstack.org/370459 | 23:38 |
| inc0 | well, so far you'd be first to ask for it tbh | 23:38 |
| inc0 | as building containers for every gate is pretty important for us;) | 23:38 |
| inc0 | but if you'd want to use kolla as replacement of devstack in infra, that would be the first, not that I don't agree | 23:39 |
| *** schwicht has joined #openstack-kolla | 23:39 | |
| Jeffrey4l__ | morning guys. | 23:39 |
| apuimedo | inc0: I already asked for official container publishing and registry some time ago | 23:39 |
| apuimedo | :P | 23:40 |
| *** Jeffrey4l__ is now known as Jeffrey4l | 23:40 | |
| inc0 | we do have a dockerhub account.. | 23:40 |
| huikang | jeffrey4l__, you should have vacation today for Chinese mid-autumn festival | 23:40 |
| huikang | :) | 23:40 |
| apuimedo | but the kolla requirements seemed to make infra concerned | 23:40 |
| Jeffrey4l | ya. happy festival day :) | 23:41 |
| inc0 | Jeffrey4l, drink one for Kolla;) | 23:42 |
| huikang | all highway are toll-free today in China | 23:42 |
| Jeffrey4l | huikang, no. not this time. | 23:42 |
| huikang | oh, I see; I will get a toll fine then :) | 23:43 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Revert 'Add neutron-lbaas-agent to exlude image for Ubuntu binary' https://review.openstack.org/370446 | 23:43 |
| Jeffrey4l | where are we now? is kolla ready for the rc1? | 23:44 |
| openstackgerrit | Hui Kang proposed openstack/kolla: Add Kuryr ansible role https://review.openstack.org/298894 | 23:47 |
| huikang | inc0, I have removed creating service and endpoint for kuryr | 23:48 |
| apuimedo | thanks huikang | 23:48 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Fixes to get Nova to work https://review.openstack.org/370459 | 23:49 |
| inc0 | ok, +2 with a string attached, do a lot of testing now guys, let's get it merged asap, but do *a lot* of testing before we release 3.0.0 | 23:49 |
| inc0 | we can fix it bug-style now, just let's make sure it's pristine quality when we release:) | 23:49 |
| inc0 | deal? | 23:50 |
| Jeffrey4l | agree | 23:50 |
| *** unicell has quit IRC | 23:50 | |
| Jeffrey4l | why kurry deploy on all compute in default? | 23:51 |
| huikang | jeffrey4l, containers can be created on all the compute nodes | 23:51 |
| apuimedo | Jeffrey4l: it is disabled by default, isn't it? | 23:52 |
| huikang | so kuryr server needs run on all compute nodes | 23:52 |
| huikang | apuimedo, yes, by default kuryr is disabled | 23:52 |
| Jeffrey4l | it is disabled in default. but i am try to know why it is started on all compute node. it is necessary? | 23:53 |
| *** fguillot has joined #openstack-kolla | 23:54 | |
| Jeffrey4l | when we have hundred compute node. we will run hundred kuryr service instance. | 23:54 |
| apuimedo | right | 23:54 |
| apuimedo | one docker engine, one kuryr driver | 23:54 |
| apuimedo | they need to run in the same machine | 23:54 |
| huikang | jeffrey4l, kuryr is a network plugin for docker engine | 23:55 |
| huikang | so every docker engine needs its kuryr server | 23:55 |
| Jeffrey4l | 1. what the kuryr service for? the host machine docker engine ( running docker? ) | 23:55 |
| apuimedo | Jeffrey4l: it's a driver backend for Docker | 23:56 |
| apuimedo | docker wants networking driver or ipam driver | 23:56 |
| Jeffrey4l | 2. if every docker engine use kurry , why not deploy kurry on all nodes? rather than all compute nodes. | 23:56 |
| apuimedo | it asks kuryr what it provides | 23:56 |
| apuimedo | kuryr replies, ipam and remote driver | 23:56 |
| apuimedo | and Docker knows, from that moment, that `docker network` commands can be sent to the kuryr rest api | 23:57 |
| Jeffrey4l | make sense. | 23:57 |
| huikang | In kolla, I think compute nodes is to run container instances for tenant | 23:57 |
| *** thumpba has joined #openstack-kolla | 23:57 | |
| huikang | so we want to deploy kuryr only on compute nodes | 23:57 |
| apuimedo | Jeffrey4l: it's just so that you have compute nodes (ansible role) be where you run virtual workloads (be them VM or containers) | 23:57 |
| Jeffrey4l | ok. | 23:57 |
| Jeffrey4l | Got it. thanks for the explanation :) | 23:58 |
| apuimedo | Jeffrey4l: you're most welcome | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!