| *** britthouser has quit IRC | 00:01 | |
| *** achanda_ has joined #kolla | 00:05 | |
| *** banix has quit IRC | 00:06 | |
| *** achanda has quit IRC | 00:07 | |
| *** diogogmt has quit IRC | 00:09 | |
| *** achanda_ has quit IRC | 00:09 | |
| *** banix has joined #kolla | 00:17 | |
| *** dims has joined #kolla | 00:34 | |
| SamYaple | sdake_: yea it is the v3 modules | 00:39 |
|---|---|---|
| *** daneyon has joined #kolla | 00:39 | |
| *** banix has quit IRC | 00:40 | |
| *** bmace has joined #kolla | 00:45 | |
| *** daneyon_ has joined #kolla | 00:46 | |
| *** daneyon has quit IRC | 00:49 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove Ansible abstraction for Nova https://review.openstack.org/211445 | 00:52 |
| *** weiyu_ has joined #kolla | 00:55 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix group name in vagrant docs https://review.openstack.org/213950 | 00:58 |
| *** banix has joined #kolla | 01:13 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove Ansible abstraction for Neutron https://review.openstack.org/211442 | 01:19 |
| *** erkules has joined #kolla | 01:24 | |
| *** erkules_ has quit IRC | 01:26 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove start.yml and bootstrap.yml https://review.openstack.org/213958 | 01:36 |
| openstackgerrit | Merged stackforge/kolla: Update Fedora magnum images to use fedora repo https://review.openstack.org/212972 | 01:45 |
| SamYaple | ping sdake_ | 01:47 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Move defaults.yml to group_vars/all.yml https://review.openstack.org/213961 | 01:52 |
| *** erkules has quit IRC | 01:52 | |
| *** erkules has joined #kolla | 01:54 | |
| openstackgerrit | XiaBing Yao proposed stackforge/kolla: update config-external.sh in galera,keystone,rabbitmq,swift https://review.openstack.org/213787 | 01:55 |
| *** Kennan2 has joined #kolla | 02:02 | |
| *** Kennan has quit IRC | 02:03 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add Dockerfile template for kolla-ansible https://review.openstack.org/213394 | 02:19 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Update kolla-ansible to include new modules https://review.openstack.org/213977 | 02:19 |
| *** jasonsb_ has quit IRC | 02:22 | |
| *** achanda has joined #kolla | 02:23 | |
| *** dims has quit IRC | 02:30 | |
| *** klint has joined #kolla | 02:54 | |
| *** tfukushima has quit IRC | 02:58 | |
| *** jasonsb has joined #kolla | 03:07 | |
| *** diogogmt has joined #kolla | 03:13 | |
| *** unicell1 has quit IRC | 03:13 | |
| *** Haomeng has quit IRC | 03:17 | |
| *** Haomeng has joined #kolla | 03:18 | |
| *** thumpba has joined #kolla | 03:25 | |
| *** diogogmt has quit IRC | 03:26 | |
| *** sdake has joined #kolla | 03:32 | |
| *** sdake_ has quit IRC | 03:36 | |
| *** unicell has joined #kolla | 03:40 | |
| *** sdake_ has joined #kolla | 03:42 | |
| *** sdake has quit IRC | 03:45 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Prepare for gating by distro, type, docker type https://review.openstack.org/213537 | 03:49 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Use one data container for all data volumes https://review.openstack.org/212804 | 03:55 |
| *** coolsvap|away is now known as coolsvap | 04:10 | |
| sdake_ | hey coolsvap | 04:16 |
| coolsvap | sdake_, hey hi | 04:16 |
| sdake_ | hey re ceilometer, you planning to use mongodb right? | 04:16 |
| coolsvap | sdake_, yes i did not start on that yet, | 04:17 |
| sdake_ | didn't start on ceilometer yet you mean? | 04:17 |
| coolsvap | the ansible part, i will start by this weekend, the configuration part and testing is remaining | 04:18 |
| coolsvap | i still need to get the configuration bits in place | 04:18 |
| SamYaple | sdake_: didnt you say heat needed ceilomter t owork? | 04:18 |
| sdake_ | of which mongo or ceilometer? | 04:18 |
| sdake_ | zaqar and mongo depend on ceilometer afiak | 04:19 |
| coolsvap | ceilometer | 04:19 |
| sdake_ | we are not doing zaqar | 04:19 |
| sdake_ | until mitaka | 04:19 |
| sdake_ | is mongo + ceilometer going to make aug 31? | 04:19 |
| sdake_ | i pushed out all not started blueprints today to rc1 | 04:20 |
| sdake_ | rc1 = sept 25 | 04:20 |
| sdake_ | sept 25 = drop dead for blueprints | 04:20 |
| coolsvap | yes i can do that, i will be back this saturday, i will have one week to do the remaining work | 04:20 |
| sdake_ | after that we only have 2 weeks of bugfix time left | 04:20 |
| sdake_ | mind i ak if your working on any other kolla related things? | 04:21 |
| coolsvap | not now | 04:21 |
| sdake_ | cool | 04:21 |
| *** sdake_ is now known as sdake | 04:22 | |
| sdake | so mongo + ceiloemter is possible | 04:22 |
| coolsvap | i stopeed trove since its not in current milestone | 04:22 |
| sdake | if you want to have a look at liberty-3 vs liberty-rc1 you can see how I think things are prioirtized | 04:22 |
| sdake | notice liberty-rc1 = not started | 04:22 |
| coolsvap | sure i will do that | 04:22 |
| coolsvap | btw i am doing bof on kolla at containercon on wed | 04:23 |
| sdake | nice | 04:23 |
| sdake | interesting to see how that pans out | 04:23 |
| coolsvap | i will let you know how it goes | 04:25 |
| SamYaple | congrats coolsvap | 04:25 |
| SamYaple | btw | 04:25 |
| coolsvap | do you have the link for the first presentation we did during mid-cycle | 04:25 |
| coolsvap | sdake, ^ | 04:26 |
| coolsvap | SamYaple, thx :) | 04:26 |
| sdake | the one done by daneyon? | 04:26 |
| coolsvap | yes, i think it was on dropbox but i am not able to find the link is it on etherpad? | 04:27 |
| *** weiyu_ has quit IRC | 04:27 | |
| coolsvap | yes it is but the file does not exist | 04:27 |
| sdake | let me check my inbox | 04:28 |
| coolsvap | sdake, please send if you find, i will be back after dinner | 04:29 |
| sdake | email is what? | 04:29 |
| coolsvap | me@coolsvap.net | 04:30 |
| sdake | looks deleted | 04:30 |
| sdake | i'll ask daneyon for a copy | 04:30 |
| coolsvap | thx! | 04:32 |
| *** thumpba has quit IRC | 04:33 | |
| openstackgerrit | XiaBing Yao proposed stackforge/kolla: update config-external.sh in galera,keystone,rabbitmq,swift https://review.openstack.org/213787 | 04:33 |
| *** thumpba has joined #kolla | 04:34 | |
| sdake | samyaple your nova review works yay :) | 04:35 |
| SamYaple | sdake: yea i wasnt able to -1 workflow last night due to gerrit sucking | 04:35 |
| SamYaple | i hadn't tested it yet | 04:35 |
| sdake | roger | 04:35 |
| sdake | say samyaple had 1 technical question re keepalived and haproxy | 05:01 |
| *** weiyu_ has joined #kolla | 05:01 | |
| sdake | you haven't heard this one yet :) | 05:01 |
| *** banix has quit IRC | 05:02 | |
| sdake | what would it take for load balancing to roll in front of haproxy? | 05:02 |
| sdake | isn't that what haproxy does? | 05:02 |
| SamYaple | haproxy is the loadbalancer | 05:03 |
| sdake | does it roundrobin? | 05:04 |
| SamYaple | dude not again tonight | 05:04 |
| sdake | or is it like master slave | 05:04 |
| SamYaple | we just went over this yesterday | 05:04 |
| SamYaple | these exact questions | 05:04 |
| sdake | i remember thinking everything went through one haproxy | 05:04 |
| SamYaple | it does | 05:04 |
| SamYaple | whichever one the keepalived vip is on | 05:05 |
| sdake | that doesn't sound like load balancing | 05:05 |
| SamYaple | that is exactly what loadbalncing is | 05:05 |
| sdake | in a loadbalancer evertyhing always goes through one process? | 05:05 |
| sdake | I thought it roundrobined to the different processes | 05:05 |
| sdake | i get keepalived doesn't move the vif around, you said that yesterday | 05:06 |
| sdake | only on faiure | 05:06 |
| sdake | i've always thought load balancers move the traffic aroudn to "balance load" | 05:07 |
| *** thumpba has quit IRC | 05:07 | |
| sdake | is it that they only load balance outgoing traffic and not incoming traffic? | 05:07 |
| sdake | they being haproxy | 05:07 |
| *** thumpba has joined #kolla | 05:08 | |
| *** weiyu_ has quit IRC | 05:08 | |
| *** thumpba has quit IRC | 05:08 | |
| sdake | samyaple if you have an answer to that last q, i'd apprecate to hear it | 05:16 |
| *** dims has joined #kolla | 05:20 | |
| *** dims has quit IRC | 05:26 | |
| SamYaple | loadbalancers everything goes through one place, yes | 05:28 |
| SamYaple | haproxy is the singular point that it goes through sdake | 05:28 |
| SamYaple | haproxy sends it to whatever it is configured to send it to | 05:28 |
| sdake | right so haproxy will load balance across 3 control nodes for exampel | 05:28 |
| sdake | how does google get all their requests to go through one machine? | 05:29 |
| SamYaple | yea but the other 2 haproxy instances are doing nothing | 05:29 |
| sdake | i got that | 05:29 |
| SamYaple | dude | 05:29 |
| SamYaple | youre the one that had basic netorking questions, dont give me that "how does google send it all to one place" | 05:29 |
| SamYaple | i can tell you how they do stuff, but thats not going to help you here | 05:29 |
| SamYaple | in the end its the same concept | 05:30 |
| sdake | so i know they use dns roundrobin to loadbalance | 05:30 |
| SamYaple | thats one of the things they do, yes | 05:31 |
| sdake | is there any other technique or is that it? | 05:31 |
| SamYaple | those roundrobin'd vips that they hit are held by a _single_ loadbalancer | 05:31 |
| SamYaple | that then spreads it to other nodes | 05:31 |
| sdake | so multilayer | 05:32 |
| SamYaple | thats how they do it yes | 05:32 |
| SamYaple | we dont have dns roundrobin or geodns obviously | 05:32 |
| SamYaple | but in the end an ip is only going to exist in one location | 05:32 |
| SamYaple | hence all connections go through one loadbalncer | 05:32 |
| sdake | do you know if haproxy is threaded? | 05:33 |
| SamYaple | yes | 05:33 |
| SamYaple | haproxy is a really solid loadbalancer | 05:33 |
| SamYaple | for software anyway | 05:33 |
| sdake | does it read a request from network and write to network or use some other advanced zero-copy technique? | 05:34 |
| *** tfukushima has joined #kolla | 05:34 | |
| SamYaple | #haproxy | 05:34 |
| sdake | lool ok | 05:34 |
| SamYaple | youve now stopped talking about architecture | 05:34 |
| sdake | thanks | 05:34 |
| sdake | i have unmeasured concerns haproxy will be a bottleneck | 05:35 |
| SamYaple | it wont be, but even if it was | 05:35 |
| sdake | pure speculation | 05:35 |
| SamYaple | its our only option | 05:35 |
| SamYaple | its hands down the best software loadbalancer | 05:35 |
| sdake | i'm just sort of curious if a load balancer for our load balancer would make sense for more scale | 05:36 |
| sdake | but sounds like not unless it was dns based | 05:36 |
| SamYaple | i dont think you understand how this stuff works to be honest | 05:36 |
| SamYaple | youd still have a single point | 05:37 |
| sdake | i'm getting an idea | 05:37 |
| sdake | what I mean is spin up 3 haproxy with 3 vifs | 05:37 |
| SamYaple | why use vips then :/ | 05:37 |
| SamYaple | how would each service know who to talk to | 05:37 |
| sdake | ok spin up 3 haproxy | 05:37 |
| SamYaple | how is that ha | 05:37 |
| SamYaple | now an ip goes away and bam no one can talk | 05:37 |
| sdake | ok spin up 9 haproxy 3 vifs :) | 05:38 |
| sdake | with dns round robin | 05:38 |
| SamYaple | vip | 05:38 |
| SamYaple | not vif | 05:38 |
| SamYaple | vip | 05:38 |
| sdake | sorry vip | 05:38 |
| SamYaple | and no thats not how this works | 05:38 |
| *** tfukushima has quit IRC | 05:39 | |
| sdake | say change of topic have you had a look at bluebox's implementation? | 05:40 |
| SamYaple | who | 05:40 |
| SamYaple | oh you mean openstack implementation | 05:41 |
| sdake | yup | 05:42 |
| sdake | they have some ansible work kicking around in github | 05:42 |
| sdake | were aquired by ibm recently | 05:43 |
| SamYaple | i didnt think they ere doig kolla | 05:43 |
| *** achanda has quit IRC | 05:43 | |
| sdake | no they have their own, curious on deltas | 05:43 |
| sdake | https://github.com/blueboxgroup/ursula | 05:44 |
| SamYaple | but... they arent using docker.... | 05:45 |
| SamYaple | so how is there even a comparison | 05:45 |
| sdake | i get that delta obviously :) | 05:45 |
| sdake | looks like ubuntu only | 05:47 |
| *** achanda has joined #kolla | 05:47 | |
| sdake | from bluebox docs: | 05:51 |
| sdake | This will have installed ursula-cli, the various openstack clients, and our patched fork of Ansible. | 05:51 |
| SamYaple | just another 'do it our way' deployment tool | 05:52 |
| SamYaple | with the same conf file issues as all the other ones | 05:52 |
| sdake | can you point me at the conf file for this project- i can't seem to find it and have been searching for 15 minutes | 05:56 |
| sdake | is everything hard coded? | 05:56 |
| sdake | a patched fork of ansible is cracksauce | 05:57 |
| sdake | straight up cracksauce | 05:57 |
| sdake | its not like they added a couple of modules | 05:57 |
| sdake | they rewrote major portions of it :) | 05:57 |
| SamYaple | devs gonna dev | 05:57 |
| sdake | is this an opinionated installer or does it offer config options in some way? | 05:58 |
| SamYaple | same way OSAD does but to a far lesser extent than OSAD | 05:59 |
| sdake | so templating the options | 05:59 |
| SamYaple | and they appear to be using hte playbooks to fix past mistakes too | 05:59 |
| SamYaple | so its a rolling thing | 05:59 |
| SamYaple | which i dont think is a good idea | 05:59 |
| sdake | what do you mean fix past mistakes/ example? | 06:01 |
| SamYaple | if you add a file in v2, but dont want it in v3 they have a task to remove the file | 06:02 |
| SamYaple | rather than an upgrade script | 06:02 |
| SamYaple | as in one time script | 06:02 |
| sdake | you mean add a file to the host? | 06:02 |
| SamYaple | sure | 06:03 |
| sdake | that is their migration strategy i guess | 06:04 |
| sdake | that is where the question about ports came from yeseterday ;) | 06:04 |
| sdake | which i see you nicely resolved ;) | 06:04 |
| SamYaple | aint no listen 0.0.0.0 in my stuff | 06:05 |
| SamYaple | thats just laziness | 06:05 |
| *** weiyu_ has joined #kolla | 06:07 | |
| sdake | i like the internal external address, happy i finally understand that | 06:07 |
| sdake | that is a killer feature | 06:08 |
| SamYaple | pretty sure everyone has that | 06:08 |
| sdake | probably | 06:08 |
| sdake | can this be in some other file: https://github.com/stackforge/kolla/blob/master/etc/kolla/globals.yml#L22 | 06:09 |
| SamYaple | depends on how much abstraction we want | 06:10 |
| *** juggler has quit IRC | 06:10 | |
| SamYaple | we can move all but the 4 required options to the defaults.yml | 06:10 |
| sdake | you said in the review for ports there is a way to override | 06:12 |
| sdake | how precisely? | 06:12 |
| SamYaple | what do you mean? | 06:12 |
| sdake | in the commit log you indicated the values could be overridden | 06:12 |
| SamYaple | the same way all values are overridden | 06:13 |
| SamYaple | you put a new value for the variable | 06:13 |
| sdake | ya i dont know how that is done, just define again? | 06:13 |
| sdake | that is more like overwritten ;) | 06:14 |
| SamYaple | overridden would be correct | 06:16 |
| sdake | i'd like all the various things in globals.conf to be undefined except those that are minimum nececessity with the abilityt o change directly in thatone file | 06:17 |
| sdake | so sane defaults happen but overrides happen in one file rather then hunting around in etc | 06:19 |
| sdake | (for migration) | 06:19 |
| sdake | every time master changes i have to recopy my etc dir and reconfigure | 06:19 |
| sdake | i dont like the idea of making a tool to merge the configs from old to new | 06:19 |
| sdake | for example the ports are in ansible directory not etc directory with your port change | 06:20 |
| sdake | to override would i have to go hunt the value in the ansible directory? | 06:20 |
| SamYaple | sdake: thats not how that works | 06:21 |
| SamYaple | you cant have them defined | 06:21 |
| SamYaple | you have to not have them defined for the override to work | 06:21 |
| SamYaple | yes you would have to find the values you need to change | 06:22 |
| SamYaple | such is life | 06:22 |
| *** britthou_ has quit IRC | 06:22 | |
| *** coolsvap is now known as coolsvap|away | 06:22 | |
| *** britthouser has joined #kolla | 06:22 | |
| sdake | there isn't ian if not defined define this? | 06:23 |
| sdake | keep in mind i have less then 50 hours of ansible experience | 06:23 |
| sdake | and I haven't rtfmed anything | 06:23 |
| SamYaple | i dont have time for these questions go do that | 06:23 |
| sdake | overrides in the ansible directory are not ideal, ansible directory would be installed in /usr/share/kolla ideally | 06:25 |
| sdake | which is read only | 06:25 |
| SamYaple | the only file you modify is /etc/kolla/global.yml for the 100th time | 06:25 |
| sdake | https://review.openstack.org/#/c/213961/ | 06:26 |
| sdake | this patch places ports in the ansible directory | 06:26 |
| SamYaple | im aware | 06:26 |
| sdake | can the ports then be set in global.yml? | 06:26 |
| SamYaple | yes | 06:27 |
| SamYaple | as ive said, global.yml overrides all | 06:27 |
| sdake | thanks | 06:27 |
| SamYaple | why do i have to keep saying the same things over and over | 06:27 |
| SamYaple | seriously all the time | 06:27 |
| SamYaple | its the same questions | 06:27 |
| SamYaple | its a waste of my time | 06:27 |
| sdake | it helps me communicate effectively how kolla works | 06:27 |
| sdake | if thats a waste of your itme i dont know what to tell you | 06:27 |
| SamYaple | its a waste of my time to answer the same question from you over and over | 06:28 |
| SamYaple | youre wasting my time | 06:28 |
| SamYaple | not kolla | 06:28 |
| sdake | your answers aren't consistent, which is why i ask multiple times | 06:29 |
| SamYaple | they are indeed consistent | 06:29 |
| SamYaple | this is why i pull logs from weeks ago and paste them | 06:29 |
| sdake | you said earlier if i want to change ports, i had to hunt the ansible directory | 06:29 |
| SamYaple | you asked to find the name of the override variable | 06:30 |
| sdake | then you said if you put in globals.yml it overrides all | 06:30 |
| SamYaple | for that you must hunt | 06:30 |
| SamYaple | to set it you set it in the only file you configure | 06:30 |
| SamYaple | feel free to read the logs | 06:30 |
| sdake | [23:20:34] <sdake>for example the ports are in ansible directory not etc directory with your port change | 06:30 |
| sdake | [23:20:50] <sdake>to override would i have to go hunt the value in the ansible directory? | 06:30 |
| SamYaple | 06:22 < SamYaple> yes you would have to find the values you need to change | 06:31 |
| sdake | i was talking about the port values not the variable names | 06:31 |
| sdake | but i can see how we didn't understand each other | 06:31 |
| SamYaple | if oyu are asking where to set things there is only one answer, globals.yml | 06:31 |
| SamYaple | period | 06:31 |
| SamYaple | as has been since the beginning | 06:31 |
| sdake | good i like that | 06:32 |
| sdake | thanks for helping clear that up | 06:32 |
| sdake | looks like we need to move some variables from globals.yml then | 06:32 |
| SamYaple | like I said, we can move all the variables | 06:33 |
| SamYaple | just depends on how bare we want to make it | 06:33 |
| SamYaple | do we want to have it commented out | 06:33 |
| SamYaple | but listed | 06:33 |
| sdake | yup | 06:33 |
| SamYaple | do we want just the 4 needed | 06:33 |
| SamYaple | its a decision I can't make alone | 06:33 |
| sdake | let me think on that | 06:33 |
| SamYaple | but doesnt affect structure so i dont really care | 06:33 |
| sdake | i'll bring it up in wednesday's team meeting | 06:33 |
| SamYaple | just keep in mind if they are simply commentted they are 'docs' and may get out of date | 06:34 |
| bmace | programatically it isn't super hard to have them spread out / isolated and just overridden in globals. it is even sort of nice from an isolation point of view, instead of everything in one file.. but i can see if you are hand editing it could be less fun to find what you want to change.. though all the changes just go in one place | 06:35 |
| sdake | i am working to improve the docs to answer these questions so we don't have to answer them a milliion times | 06:35 |
| SamYaple | bmace: thats what we currently do | 06:35 |
| SamYaple | bmace: we have variables all over the place | 06:35 |
| bmace | yeah, i know :) | 06:35 |
| SamYaple | i like that way better | 06:35 |
| SamYaple | having only 4 items i have to think about to get started is great | 06:35 |
| bmace | i don't mind it, from a programmatic perspective, though a bit more onerous for hand editing | 06:36 |
| sdake | most cats will hand edit ;) | 06:36 |
| SamYaple | the question then becomes how often do we hand edit those files? | 06:36 |
| bmace | but maybe we can get to the point there is no more hand editing :) | 06:36 |
| sdake | bmace that is what i'm after | 06:36 |
| bmace | working on it :) | 06:36 |
| sdake | i think what i'd like is everything in the ansible directory except mandatory config options | 06:39 |
| sdake | with common overrides in globals.yml commented | 06:39 |
| sdake | with a gate job to check they dont get out of wack ;) | 06:39 |
| sdake | but its 12pm | 06:40 |
| SamYaple | define common | 06:40 |
| sdake | and i've been up since 5am | 06:40 |
| sdake | common = everything in globals.yml now | 06:40 |
| *** openstackgerrit_ has joined #kolla | 06:42 | |
| sdake | i told my wife i was working on using ansible to deploy heat | 06:55 |
| sdake | and she said "your using an orchestration tool to deploy an orchestration tool?" | 06:55 |
| sdake | full of zingers that woman | 06:56 |
| SamYaple | dont tell her about tripleo | 06:56 |
| bmace | is tripleo the expression on someones face when you tell them about tripleo? OoO | 06:56 |
| SamYaple | what happens to thier nose?!? | 06:57 |
| *** mfalatic has quit IRC | 06:57 | |
| bmace | that is their mouth so wide open you don't even notice their nose anymore.. | 06:57 |
| SamYaple | why are thier eyes bigger than thier mouth? | 06:57 |
| bmace | they are that surprised! | 06:57 |
| SamYaple | what do people look like where you come from bmace?! | 06:58 |
| bmace | it is like the Mars effect from total recall | 06:58 |
| bmace | they all look like this? http://www.2dayblog.com/images/2009/april/animeeye.jpg | 06:59 |
| SamYaple | hurrhurrhurr | 07:03 |
| bmace | :P | 07:04 |
| *** bmace is now known as bmace_away | 07:05 | |
| *** jmccarthy has quit IRC | 07:07 | |
| openstackgerrit | Merged stackforge/kolla: update config-external.sh in galera,keystone,rabbitmq,swift https://review.openstack.org/213787 | 07:19 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Prepare for gating by distro, type, docker type https://review.openstack.org/213537 | 07:22 |
| *** akwasnie has joined #kolla | 07:27 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Templatize Murano container https://review.openstack.org/213558 | 07:31 |
| *** achanda has quit IRC | 07:33 | |
| *** shardy has joined #kolla | 07:36 | |
| *** dims has joined #kolla | 07:48 | |
| sdake | samyaple my wife had a zinger about tripleo | 07:51 |
| sdake | it involved the word "singleo" | 07:51 |
| sdake | I dont remember the context, but we laughed for a good 5 minutes | 07:51 |
| sdake | CRS Ftl | 07:51 |
| *** dims has quit IRC | 07:53 | |
| *** bradjones has quit IRC | 07:53 | |
| *** bradjones has joined #kolla | 07:56 | |
| *** bradjones has quit IRC | 07:56 | |
| *** bradjones has joined #kolla | 07:56 | |
| *** athomas has joined #kolla | 08:01 | |
| *** jmccarthy has joined #kolla | 08:01 | |
| *** gfidente has joined #kolla | 08:01 | |
| *** inc0 has joined #kolla | 08:18 | |
| inc0 | good morning | 08:19 |
| SamYaple | morning inc0 | 08:20 |
| inc0 | wassup Sam? | 08:20 |
| openstackgerrit | XiaBing Yao proposed stackforge/kolla: update haproxy/ensure_latest_config.sh https://review.openstack.org/214068 | 08:23 |
| *** diga has joined #kolla | 08:27 | |
| inc0 | SamYaple, quick question..why do we remove start.yml stuff? | 08:45 |
| inc0 | inheritance is good | 08:45 |
| *** tfukushima has joined #kolla | 08:47 | |
| SamYaple | inc0: i agree. it wasn't an easy decision :( | 08:49 |
| SamYaple | it was breaking ansible | 08:49 |
| SamYaple | it was very non-best practice ansible code too | 08:49 |
| inc0 | hmf...well it will help readibility for sure tho | 08:52 |
| SamYaple | hmf? | 08:53 |
| SamYaple | oh right | 08:53 |
| SamYaple | ive asked that before | 08:53 |
| inc0 | thinking loudly | 08:53 |
| SamYaple | yea | 08:53 |
| SamYaple | hmph | 08:53 |
| inc0 | I tend to emit variety of noises | 08:53 |
| SamYaple | yea I was trying to find a clean way to have all those default options in a dict and do a dict merge for the docker module | 08:53 |
| inc0 | akwasnie can confirm | 08:54 |
| SamYaple | that would mean we would only specify the options that were unique | 08:54 |
| akwasnie | confirm | 08:54 |
| SamYaple | akwasnie: you coming over to work at RAX? | 08:54 |
| akwasnie | I am still thinking... | 08:54 |
| inc0 | I'm trying to convince her to come | 08:55 |
| SamYaple | akwasnie: come on youll like it | 08:55 |
| SamYaple | you can meet my wife | 08:55 |
| SamYaple | she likes new people | 08:55 |
| akwasnie | for sure I will like it.... but I'm afraid xD | 08:56 |
| SamYaple | its alot better here now | 08:56 |
| SamYaple | only like 2 or 3 shooting deaths a week now | 08:56 |
| sdake | I guess I need to figureo ut how to get a docimpact job created | 08:57 |
| * sdake groans | 08:57 | |
| SamYaple | sdake: yea i asked for that a few weeks back | 08:57 |
| sdake | more project config beating | 08:57 |
| sdake | did you ask me or in general? | 08:58 |
| SamYaple | no i asked specficially because infra popped a bug on kolla when i used DocImpact | 08:58 |
| SamYaple | you* | 08:58 |
| SamYaple | i assigned the bug to you too | 08:58 |
| sdake | inc0 the reason is because it created 2x tasks to deal with pid:host | 08:58 |
| inc0 | sdake, that's about tox testing review? | 08:59 |
| sdake | inc0 the reason start.yml is being removed | 08:59 |
| sdake | i dont even know what a doc impact job is - i'll ask around when people are awake | 09:00 |
| sdake | i remember seeing the bug and thinking "not a prioirty" :) | 09:00 |
| sdake | we dont even have doc pushing to openstack.org | 09:01 |
| sdake | which i'd definately like to get to | 09:01 |
| sdake | I think we need to port the md to rsts tho for that to work | 09:01 |
| sdake | but i am not sure | 09:01 |
| sdake | has something to do with sphinx black magic | 09:02 |
| sdake | which i know nothing about | 09:02 |
| sdake | more shit to learn | 09:02 |
| inc0 | I've spent some time on sphinx, but I've never did stuff for openstack | 09:04 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: nova-compute daemon misses libvirt python package on start https://review.openstack.org/214079 | 09:07 |
| *** vbel has joined #kolla | 09:07 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: nova-compute daemon misses libvirt python package on start https://review.openstack.org/214079 | 09:09 |
| pbourke | sdake: https://review.openstack.org/#/c/212804/20/ansible/roles/nova/defaults/main.yml -- this isn't consistent with the other main.ymls | 09:22 |
| pbourke | sdake: but after 20 patchsets I'm hesitant to put a -1 on it ;) | 09:22 |
| pbourke | sdake: so up to you, want to leave it? | 09:22 |
| SamYaple | pbourke: the number of patchsets shouldn't change the consistency | 09:26 |
| pbourke | SamYaple: it shouldn't but often does | 09:26 |
| SamYaple | well thats just bad reviewing | 09:26 |
| SamYaple | it is consistent within the nova.yml though | 09:26 |
| SamYaple | thats fine then | 09:26 |
| SamYaple | we can tweak that in a different patch | 09:26 |
| inc0 | harry router and the mask of network | 09:29 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove Ansible abstraction for Nova https://review.openstack.org/211445 | 09:36 |
| SamYaple | sdake: are you ovs or linux-bridge to test that neutron one? | 09:39 |
| SamYaple | ovs works, but i see a problem with linux-bridge | 09:39 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove Ansible abstraction for Neutron https://review.openstack.org/211442 | 09:42 |
| *** athomas has quit IRC | 09:44 | |
| *** athomas has joined #kolla | 09:48 | |
| *** tfukushima has quit IRC | 09:51 | |
| *** weiyu_ has quit IRC | 09:52 | |
| *** tfukushima has joined #kolla | 09:58 | |
| *** tfukushima has quit IRC | 09:58 | |
| *** tfukushima has joined #kolla | 09:58 | |
| openstackgerrit | Merged stackforge/kolla: nova-compute daemon misses libvirt python package on start https://review.openstack.org/214079 | 10:01 |
| *** weiyu_ has joined #kolla | 10:02 | |
| SamYaple | hey python pep8 people | 10:07 |
| SamYaple | how to format a line thats longer than 80 chars | 10:07 |
| SamYaple | endpoint_kwargs['service_name_or_id'] = endpoint_kwargs.pop('service_id') | 10:08 |
| SamYaple | ^^ the line | 10:08 |
| SamYaple | where to break | 10:08 |
| SamYaple | coolsvap|away inc0 ? | 10:08 |
| pbourke | SamYaple: I would prob break after the = | 10:09 |
| SamYaple | pbourke: does that require a \ | 10:09 |
| pbourke | yes | 10:09 |
| inc0 | I'd do this either after [ or ( | 10:09 |
| inc0 | and key with indent | 10:09 |
| *** weiyu_ has quit IRC | 10:10 | |
| *** pbourke has quit IRC | 10:11 | |
| inc0 | can you show more code tho? it seems very deeply nested | 10:12 |
| *** pdb has joined #kolla | 10:14 | |
| SamYaple | its standard ansible stuff | 10:15 |
| SamYaple | ie yea its nested | 10:15 |
| SamYaple | not much i can do there | 10:15 |
| *** sdake has quit IRC | 10:15 | |
| inc0 | SamYaple, http://paste.openstack.org/show/420272/ | 10:16 |
| *** britthouser has quit IRC | 10:17 | |
| inc0 | I don't like line breaks with \ | 10:18 |
| SamYaple | yea im definetely not doing that though | 10:19 |
| SamYaple | thats really bad | 10:19 |
| inc0 | actually its pretty pep8ish | 10:19 |
| inc0 | that's why I said about nesting | 10:19 |
| SamYaple | ansible modules will fail pep8 anyway | 10:20 |
| inc0 | it seems this stuff would need refactoring anyway | 10:20 |
| SamYaple | ill stick with existing coding style before i do that | 10:20 |
| inc0 | probably good idea | 10:21 |
| inc0 | one other way to do this, probably even better | 10:22 |
| inc0 | SamYaple, http://paste.openstack.org/show/420284/ | 10:23 |
| inc0 | also if there are more of such operations, you can squash it all there | 10:24 |
| SamYaple | that I can live with | 10:25 |
| SamYaple | that looks slick | 10:25 |
| *** diga has quit IRC | 10:25 | |
| inc0 | 80 chars in such thins is always somewhat riddle how to do it best | 10:26 |
| inc0 | but long if statements are worse | 10:26 |
| SamYaple | i79 | 10:27 |
| SamYaple | not 80 | 10:27 |
| inc0 | yeah, true | 10:27 |
| inc0 | trying to build our containers from source from kilo-stable | 10:27 |
| inc0 | I wonder what'll break | 10:27 |
| SamYaple | at least neutron | 10:28 |
| inc0 | neutron is always broken | 10:28 |
| *** britthouser has joined #kolla | 10:36 | |
| vbel | hi everyone. Why neutron is broken? | 10:43 |
| SamYaple | vbel: neutron is fine | 10:43 |
| SamYaple | its broken with kilo because of changes in master | 10:43 |
| vbel | SamYaple: ok, thanks! | 10:43 |
| inc0 | neutron is never fine, it can be less broken if anything | 10:47 |
| inc0 | btw SamYaple what would be this breaking change? | 10:47 |
| inc0 | if I'll use config from old kilo | 10:47 |
| SamYaple | i tested it and it was broken | 10:47 |
| SamYaple | i think | 10:47 |
| SamYaple | it was afeter we talked a few days ago | 10:47 |
| inc0 | I remember there were problems with sudoers not present in container | 10:48 |
| inc0 | but that can be solved by either config change or installation of sudo | 10:48 |
| inc0 | anyway, I'll give it a shot | 10:48 |
| SamYaple | that was a long time ago | 10:48 |
| inc0 | I need kilo based kolla, possibly deployed with ansible | 10:48 |
| vbel | the same for me :) | 10:48 |
| inc0 | vbel, I'm building source installs as we speak | 10:49 |
| inc0 | I sed'ed .buildconfs to change master to kilo stable | 10:49 |
| vbel | in source based centos we are missing openvswitch folder | 10:49 |
| inc0 | so I'll know what's broken soon enough | 10:49 |
| inc0 | but ovs container should work from binary as well | 10:50 |
| inc0 | I mean, we don't build ovs from source at all | 10:50 |
| inc0 | and it's not bound to openstack releases | 10:50 |
| vbel | I mean in docker/ ... but we switching to templates | 10:50 |
| *** dims has joined #kolla | 10:51 | |
| vbel | yes, I take it from binary | 10:51 |
| inc0 | jpeeler is us based right? | 10:52 |
| SamYaple | yea | 10:54 |
| inc0 | damn, one good thing about reloc is that I'll actually get to talk to more people | 10:54 |
| *** tfukushima has quit IRC | 11:10 | |
| *** britthouser has quit IRC | 11:11 | |
| *** britthouser has joined #kolla | 11:12 | |
| *** tfukushima has joined #kolla | 11:15 | |
| *** tfukushima has quit IRC | 11:22 | |
| *** diga has joined #kolla | 11:23 | |
| *** dims_ has joined #kolla | 11:28 | |
| *** dims has quit IRC | 11:30 | |
| *** diga has quit IRC | 11:32 | |
| *** akwasnie has quit IRC | 11:39 | |
| *** britthouser has quit IRC | 11:43 | |
| *** dwalsh has joined #kolla | 11:45 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Move defaults.yml to group_vars/all.yml https://review.openstack.org/213961 | 11:56 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove defaults.yml from docs https://review.openstack.org/214147 | 11:56 |
| *** dims has joined #kolla | 12:00 | |
| *** dims has quit IRC | 12:00 | |
| *** dims_ has quit IRC | 12:00 | |
| *** dims has joined #kolla | 12:01 | |
| openstackgerrit | Merged stackforge/kolla: Templatize Murano container https://review.openstack.org/213558 | 12:01 |
| *** dims has quit IRC | 12:05 | |
| *** banix has joined #kolla | 12:13 | |
| *** dims has joined #kolla | 12:15 | |
| *** rhallisey has joined #kolla | 12:16 | |
| *** tfukushima has joined #kolla | 12:22 | |
| *** tfukushima has quit IRC | 12:26 | |
| *** banix has quit IRC | 12:30 | |
| *** diogogmt has joined #kolla | 12:39 | |
| *** shardy_ has joined #kolla | 12:39 | |
| *** dims_ has joined #kolla | 12:41 | |
| *** shardy has quit IRC | 12:41 | |
| *** dims has quit IRC | 12:41 | |
| *** tfukushima has joined #kolla | 12:44 | |
| *** shardy_ has quit IRC | 12:45 | |
| *** shardy has joined #kolla | 12:46 | |
| *** rhallisey has quit IRC | 12:56 | |
| *** diogogmt has quit IRC | 12:56 | |
| *** britthouser has joined #kolla | 12:59 | |
| vbel | I am checking latest ansible code. Can we be more db agnostic? I don't like mixing database_host/user with mariadb_port, e.g. here - https://github.com/stackforge/kolla/blob/master/ansible/roles/bootstrap.yml | 12:59 |
| *** inc0 has quit IRC | 13:00 | |
| SamYaple | vbel: how do you mean? | 13:00 |
| vbel | I suggest using database_port in globals.yml (together with user/host) | 13:00 |
| SamYaple | we must specify the port | 13:00 |
| SamYaple | oh i see | 13:00 |
| vbel | -a "login_host='{{ database_address }}' | 13:00 |
| vbel | login_port='{{ mariadb_port }}' | 13:00 |
| vbel | login_user='{{ database_user }}' | 13:00 |
| vbel | login_password='{{ database_password }}' | 13:00 |
| SamYaple | there are lots of changes happening with where options are specified | 13:00 |
| SamYaple | but the most of those options will be moving to the group_vars/all.yml | 13:00 |
| *** britthou_ has joined #kolla | 13:01 | |
| *** dims has joined #kolla | 13:01 | |
| *** dims_ has quit IRC | 13:02 | |
| *** rhallisey has joined #kolla | 13:04 | |
| *** britthouser has quit IRC | 13:04 | |
| rhallisey | SamYaple, around? | 13:04 |
| SamYaple | rhallisey: yes | 13:04 |
| SamYaple | is this about the enable_ceph thing? | 13:05 |
| rhallisey | no | 13:05 |
| SamYaple | ohg | 13:05 |
| SamYaple | wel lthats a thing | 13:05 |
| rhallisey | KOLLA_BOOTSTRAP | 13:05 |
| SamYaple | :) | 13:05 |
| SamYaple | yea? | 13:05 |
| rhallisey | I'm impartial to the enable_ceph thing | 13:05 |
| SamYaple | I think sdake will be fine im moving the defaults.yml to group_vars where it should be | 13:05 |
| SamYaple | it wasnt there initially for other reason that I have fixed | 13:06 |
| rhallisey | Using KOLLA_BOOTSTAP* gets syntax error | 13:06 |
| rhallisey | what were you intending for that | 13:07 |
| rhallisey | what happens atm is cinder_bootstrap doesn't run the db sync | 13:07 |
| SamYaple | KOLLA_BOOTSTRAP: | 13:07 |
| SamYaple | just like that | 13:07 |
| rhallisey | didn't work for me :/ | 13:07 |
| SamYaple | if oy ulook you are missing a 'T' | 13:07 |
| SamYaple | copy the other services | 13:07 |
| rhallisey | omg I see it | 13:08 |
| rhallisey | haha | 13:08 |
| SamYaple | lol ;) | 13:08 |
| rhallisey | yes finally! | 13:08 |
| rhallisey | boostrap fml | 13:08 |
| SamYaple | thats a copypasta error from a long while back | 13:08 |
| vbel | yes, mariadb_port is perfect for all.yml, what about database_port which can be something else. I don't like it hardcoded in bootstrap.yml in roles/ and keystone/ | 13:09 |
| rhallisey | when you put the * at the end I thought you literally meant add a star ugh | 13:09 |
| SamYaple | vbel: database_port is mariadb_port | 13:09 |
| SamYaple | rhallisey: sorry, i meant it like "correction" | 13:10 |
| rhallisey | ya I just realized | 13:10 |
| SamYaple | BOOSTRAP | 13:10 |
| SamYaple | BOOOOO | 13:10 |
| rhallisey | read that word to quickly | 13:10 |
| SamYaple | vbel: i dont really see what oyu mean by hardcoded, its reading the mariadb_port vairbale | 13:10 |
| vbel | yes, why not to name it like database_user/database_password ... and somewhere in global.yml set database_port: "{{ mariadb_port }}" | 13:10 |
| SamYaple | the mariadb_port variable determines what port the database will listen on | 13:11 |
| vbel | we just don't use mariadb :) | 13:11 |
| SamYaple | is this just a naming thing? | 13:11 |
| SamYaple | like you dont care about the fucntionality yo ujust care about the name? | 13:11 |
| vbel | :) It can be confusing later if someone ports it to postgres | 13:11 |
| vbel | I want roles to be db agnostic as much as possible | 13:12 |
| SamYaple | I doubt well have more than one database solution in the kolla base | 13:12 |
| SamYaple | but you are only refering to the naming of the variable | 13:12 |
| SamYaple | that we can accomidate I am sure | 13:12 |
| SamYaple | I have structured it so it isn't reliant on deploying the database through Kolla (though i still have a few tweaks for that) | 13:13 |
| vbel | yes, I just thought it would be more consistent to refer to database user/password and port as database_ everywhere but mariadb tasks | 13:13 |
| *** thumpba has joined #kolla | 13:13 | |
| SamYaple | like I need ot not have the haproxy config listen on mariabd if enable_mariadb == no, i have some patches for that locally its just with the templating things have been difficult | 13:13 |
| *** dims_ has joined #kolla | 13:15 | |
| vbel | it's not a big deal if mariadb is the only db backend now. Just to grow it someday further :) | 13:15 |
| SamYaple | I dont think we will is my point | 13:15 |
| vbel | ok | 13:16 |
| *** dims has quit IRC | 13:17 | |
| *** akwasnie has joined #kolla | 13:18 | |
| *** thumpba has quit IRC | 13:22 | |
| *** thumpba has joined #kolla | 13:29 | |
| *** diogogmt has joined #kolla | 13:30 | |
| *** dims_ has quit IRC | 13:31 | |
| *** dims has joined #kolla | 13:31 | |
| *** vbel has quit IRC | 13:34 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Cinder bootstrap container is not initializing the DB https://review.openstack.org/213845 | 13:35 |
| rhallisey | ok that gets cinder up | 13:35 |
| rhallisey | thanks SamYaple | 13:35 |
| rhallisey | glad you pointed that out. I was going crazy | 13:35 |
| *** jtriley has joined #kolla | 13:39 | |
| *** dims has quit IRC | 13:40 | |
| *** dwalsh has quit IRC | 13:42 | |
| *** dims has joined #kolla | 13:43 | |
| *** klint has quit IRC | 13:44 | |
| SamYaple | rhallisey: glab to help | 13:48 |
| SamYaple | all i can say is, 'been there' | 13:48 |
| rhallisey | this was the 2nd misspelling I found when getting this working | 13:48 |
| rhallisey | both were nearly nvisible | 13:48 |
| rhallisey | :( | 13:49 |
| rhallisey | happens | 13:49 |
| SamYaple | do you know how to test properly with ansible? | 13:49 |
| SamYaple | with --step? | 13:49 |
| SamYaple | -vvvv | 13:49 |
| SamYaple | tags? | 13:49 |
| SamYaple | you can even use pdb to debug | 13:49 |
| SamYaple | I should do a brown bag for oyu guys | 13:49 |
| pdb | :/ | 13:49 |
| SamYaple | not you pdb | 13:50 |
| *** pdb is now known as pbourke | 13:50 | |
| pbourke | why wont my client remember the change! | 13:50 |
| *** pbourke has quit IRC | 13:51 | |
| *** pbourke has joined #kolla | 13:51 | |
| pbourke | that should do it | 13:51 |
| rhallisey | SamYaple, no I don't know how to properly test | 13:54 |
| rhallisey | I'm getting the hang of the code base but I'm totally making it up as I go | 13:54 |
| SamYaple | a big one is --step (combine it will --tags) | 13:55 |
| SamYaple | it will ask you if you wish to execute each tasks | 13:55 |
| SamYaple | to debug with pdb is a bit more complicated | 13:55 |
| rhallisey | oh that would've been useful | 13:57 |
| rhallisey | I was hacking at the bootstrap contianer forcing to to stop | 13:57 |
| *** dwalsh has joined #kolla | 13:58 | |
| SamYaple | yea that is not the best way to do the stuff :) | 14:00 |
| *** dims has quit IRC | 14:04 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Move defaults.yml to group_vars/all.yml https://review.openstack.org/213961 | 14:05 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Move lots of global.yml options https://review.openstack.org/214200 | 14:05 |
| *** dims has joined #kolla | 14:05 | |
| SamYaple | rhallisey: https://review.openstack.org/#/c/214200/1/etc/kolla/globals.yml | 14:05 |
| SamYaple | the new globals.yml | 14:06 |
| SamYaple | think that will make sdake happy :/ | 14:06 |
| SamYaple | 3 options | 14:06 |
| SamYaple | full deploy | 14:06 |
| SamYaple | (assuming we have DockerHub images) | 14:06 |
| *** banix has joined #kolla | 14:11 | |
| *** diogogmt has quit IRC | 14:14 | |
| *** akwasnie_ has joined #kolla | 14:22 | |
| *** dwalsh has quit IRC | 14:23 | |
| *** akwasnie has quit IRC | 14:23 | |
| *** akwasnie_ is now known as akwasnie | 14:23 | |
| *** jasonsb has quit IRC | 14:26 | |
| *** vbel has joined #kolla | 14:29 | |
| *** akwasnie_ has joined #kolla | 14:30 | |
| *** akwasnie_ has quit IRC | 14:31 | |
| *** sdake has joined #kolla | 14:31 | |
| sdake | morning | 14:34 |
| pbourke | question all | 14:35 |
| vbel | morning sdake | 14:35 |
| pbourke | is it ok to assume a certain number of nodes, say 2 in the reference implementation, and then allow people to override? | 14:35 |
| *** dims_ has joined #kolla | 14:35 | |
| pbourke | this is around generating swift rings, you need to know certain info about each node in the cluster, and Im finding it hard to generate this dynamically | 14:36 |
| *** dims has quit IRC | 14:37 | |
| *** sdake_ has joined #kolla | 14:37 | |
| vbel | I would assume 2 as minimum for such implementation. Otherwise it makes no sense | 14:37 |
| vbel | assuming reference implementation != demo | 14:38 |
| pbourke | it would be nice to generate defaults for any number of hosts | 14:40 |
| *** sdake has quit IRC | 14:40 | |
| pbourke | ah I think I can use hostvars | 14:42 |
| vbel | http://docs.openstack.org/developer/swift/deployment_guide.html as default? :) up to 500,000 partitions / 5000 disks | 14:46 |
| akwasnie | jpeeler, are u there? | 14:50 |
| jpeeler | akwasnie: what's up | 14:55 |
| *** pbourke_ has joined #kolla | 14:57 | |
| *** rhallisey has quit IRC | 14:59 | |
| *** pbourke has quit IRC | 14:59 | |
| *** Ti-mo has quit IRC | 14:59 | |
| *** rhallisey has joined #kolla | 14:59 | |
| *** dims has joined #kolla | 15:01 | |
| *** achanda has joined #kolla | 15:03 | |
| *** dims_ has quit IRC | 15:03 | |
| *** achanda has quit IRC | 15:05 | |
| akwasnie | hi, are u working on this https://blueprints.launchpad.net/kolla/+spec/ironic-container? | 15:06 |
| jpeeler | it's on my to do, had a lot of things come before it | 15:09 |
| *** blahRus has joined #kolla | 15:11 | |
| *** Ti-mo has joined #kolla | 15:12 | |
| *** diogogmt has joined #kolla | 15:14 | |
| *** bmace_away is now known as bmace | 15:18 | |
| akwasnie | do u need help with this? | 15:19 |
| jpeeler | akwasnie: we could share some of it if you'd like. do you have special interest in ironic or just looking for something to work on? | 15:20 |
| akwasnie | this is needed for tripleo | 15:23 |
| akwasnie | we can discuss it later with inc0 | 15:25 |
| jpeeler | i started working on it many months ago, need to find that work (today will do so) and see if any of it can be used | 15:25 |
| akwasnie | ok great thanks | 15:26 |
| pbourke_ | vbel: any idea how I might do this for each host in hostvars? http://paste.openstack.org/show/420626/ | 15:27 |
| pbourke_ | SamYaple: ^ | 15:27 |
| bmace | pssst pbourke, see im :) | 15:28 |
| *** dims has quit IRC | 15:29 | |
| *** dims has joined #kolla | 15:30 | |
| *** dwalsh has joined #kolla | 15:30 | |
| vbel | pbourke: you already do this :) | 15:32 |
| pbourke_ | vbel: well, Im trying, but doesn't work :( | 15:33 |
| pbourke_ | vbel: you can try it out with 'ansible-playbook -i "localhost," -c local test.yml -v' | 15:34 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Libvirt is moving around the wrong config file https://review.openstack.org/214241 | 15:34 |
| pbourke_ | looking at docs it would need to become item.value.inventory_hostname. ... | 15:34 |
| pbourke_ | but haven't found a form that works yet | 15:34 |
| jpeeler | can docker pull an entire namespace of images? | 15:36 |
| *** akwasnie has quit IRC | 15:36 | |
| vbel | I could play that ansible | 15:36 |
| *** jasonsb has joined #kolla | 15:42 | |
| *** dims has quit IRC | 15:45 | |
| *** dims has joined #kolla | 15:47 | |
| *** dwalsh has quit IRC | 15:56 | |
| *** mfalatic has joined #kolla | 15:58 | |
| *** vbel has quit IRC | 16:02 | |
| *** mfalatic_ has joined #kolla | 16:03 | |
| *** diga has joined #kolla | 16:04 | |
| sdake_ | morning again | 16:05 |
| sdake_ | apparently I fell asleep | 16:05 |
| sdake_ | jpeeler not that I am aware of - I asked around for this question for mirroring purposes | 16:05 |
| sdake_ | pbourke morning fine sir | 16:06 |
| sdake_ | pbourke there? | 16:06 |
| pbourke_ | sdake_: hi | 16:06 |
| sdake_ | your signed up for swift for l3 | 16:06 |
| sdake_ | the ansible part | 16:06 |
| sdake_ | sept 3rd is the absolute last day :) | 16:06 |
| sdake_ | is that work started? | 16:06 |
| *** mfalatic has quit IRC | 16:07 | |
| sdake_ | and forgive m e if i've already asked | 16:07 |
| sdake_ | i have trouble keeping track of everyone's status | 16:07 |
| pbourke_ | sdake_: sure, I was actually hoping to have it ready today but the ring generation is turning out to be a bit tricky | 16:07 |
| *** mfalatic_ is now known as mfalatic | 16:07 | |
| sdake_ | nice | 16:08 |
| sdake_ | so its "good progress" ? | 16:08 |
| pbourke_ | yes | 16:08 |
| sdake_ | can you set the tracker to that pls | 16:08 |
| pbourke_ | its questionable whether we should generate the rings at all | 16:08 |
| pbourke_ | or have the operator supply them | 16:08 |
| sdake_ | i just woke up so i'll have to let my brain boot before i can offer advice there | 16:08 |
| pbourke_ | sure | 16:08 |
| pbourke_ | will update tracker | 16:09 |
| sdake_ | bluebox i think generates them | 16:09 |
| sdake_ | license is compatible with stealing good ideas if there is any :) | 16:09 |
| sdake_ | so happy to hear that about swift ;) | 16:12 |
| sdake_ | so ring generation, i think it makes sense to put the storage nodes in the ring file | 16:15 |
| sdake_ | the ones from the ansible playbook | 16:15 |
| sdake_ | is that not possible in some ansiblie-ized way? | 16:15 |
| sdake_ | pbourke_ | 16:15 |
| pbourke_ | I think its very possible, probably just my inexperience with ansible is making things painful | 16:16 |
| pbourke_ | will get there in the end | 16:16 |
| sdake_ | it may require an ansible module | 16:16 |
| sdake_ | i am not super keen on ansible modules | 16:17 |
| sdake_ | but in this case i see no harm | 16:17 |
| sdake_ | to get the data in the right format | 16:17 |
| sdake_ | what format is the ring file - have an example | 16:17 |
| pbourke_ | right now Im using a task to exec swift-ring-builder using the swift-proxy container | 16:17 |
| pbourke_ | so my challenge is to loop this in such a way that we can build a ring adding all hosts, while still leaving all values configurable | 16:18 |
| pbourke_ | I'll pastebin something in a min | 16:18 |
| pbourke_ | osad seem to have a playbook around this but there's a lot in it | 16:18 |
| pbourke_ | im not sure how much hacking it would take to make it work in our framework | 16:18 |
| *** bmace has quit IRC | 16:19 | |
| pbourke_ | http://paste.openstack.org/show/420680/ | 16:20 |
| pbourke_ | clearly this is garbage, just to give an idea of what im trying :) | 16:20 |
| sdake_ | osad same story as bluebox, if there is anything worth stealing - license is compatible ;) | 16:20 |
| pbourke_ | can you link me to bluebox whatever that is? | 16:21 |
| *** pbourke_ is now known as pbourke | 16:21 | |
| sdake_ | yup moment | 16:22 |
| sdake_ | https://github.com/blueboxgroup/ursula | 16:23 |
| sdake_ | ibm liked them enough to aquire their company | 16:23 |
| sdake_ | and they have real customers - similiar to cisco's metacloud | 16:23 |
| sdake_ | not as many customers as metacloud from what I understadn ;-) | 16:23 |
| sdake_ | their solution rleies on a super-patches forked version of ansible | 16:23 |
| sdake_ | so its clearly a disaster waiting to happen on that front | 16:24 |
| sdake_ | but they may have something worth referencing | 16:24 |
| sdake_ | sam said last night one of the problems with their code is their migration migrates from v1 to v2 to v3 and fixes errors along the way | 16:24 |
| sdake_ | which is not how we want to handle things in kolla obviously | 16:24 |
| sdake_ | pbourke just looking at your example paybook it seems like your introducing alot of config variables | 16:25 |
| pbourke | sdake_: they gotta go somewhere | 16:25 |
| sdake_ | not to armchair design, but ideally everything should work out the box | 16:25 |
| pbourke | sdake_: that's the goal | 16:25 |
| pbourke | sdake_: all will have working defaults | 16:25 |
| pbourke | its just so ops can customise the rings if they want | 16:26 |
| sdake_ | i'm good with customization | 16:26 |
| *** unicell has quit IRC | 16:26 | |
| sdake_ | as long as it really doesn't end up in globals.yml ;) | 16:26 |
| sdake_ | we tryt o customize via augemntation files | 16:26 |
| sdake_ | keep that in mind | 16:26 |
| sdake_ | we dont want to template much if anything | 16:26 |
| sdake_ | by templae i mean {{ }} | 16:27 |
| pbourke | hmm | 16:27 |
| sdake_ | if you ahve a full patch ready to git review, might help to get some early feedback :) | 16:27 |
| pbourke | sdake_: that actually kind of brings me back round to my initial question | 16:27 |
| pbourke | sdake_: is it ok to provide defaults that assume a certain architecture, i.e. 2 storage nodes, 2 controllers, etc | 16:27 |
| pbourke | sdake_: or does it need to flex to whatever is in the inventory | 16:28 |
| sdake_ | i'd like it to deal with what is in the inventory | 16:28 |
| pbourke | yeah me too :( | 16:28 |
| sdake_ | I know you can do the job with a module | 16:28 |
| sdake_ | if it is only teh swift ring that matters in this case | 16:28 |
| pbourke | from initial reads it looks like bluebox are using a static definition file | 16:28 |
| sdake_ | you may be able to do a for loop in asnbiel as well not sure | 16:28 |
| *** akwasnie has joined #kolla | 16:29 | |
| pbourke | sdake_: yup you're thoughts match my past few hours of hacking | 16:29 |
| pbourke | *your | 16:29 |
| sdake_ | before implementing a module suggest speaking with samyaple via irc | 16:29 |
| sdake_ | (if you get stuck on the ring buiding) | 16:29 |
| sdake_ | he may have some magic way to do a for loop with ansible that is not documented | 16:30 |
| sdake_ | ansible documentation is absolutely shit | 16:30 |
| jpeeler | this could be useful - https://github.com/CenturyLinkLabs/watchtower | 16:30 |
| jpeeler | is there anything else that restarts the container upon a new image being pushed? | 16:31 |
| sdake_ | jpeeler a deploy operation | 16:32 |
| sdake_ | watchtower is propreitary software so we can't directly integrate with it | 16:32 |
| sdake_ | it is only 99$ a node | 16:32 |
| sdake_ | so its not like its super exepsnive | 16:32 |
| sdake_ | like most proprietary software is | 16:32 |
| *** bmace has joined #kolla | 16:33 | |
| sdake_ | and their purchase process is easy (web purchase) vs hard (sales teams) | 16:33 |
| sdake_ | oh jpeeler nevermind | 16:33 |
| jpeeler | yeah it's on github | 16:34 |
| pbourke | one thing that upsets me with yaml is its not very flexible for multiline stuff | 16:34 |
| sdake_ | jpeeler i am confused with some other product from ansible related to watching stuff | 16:34 |
| pbourke | seem to end up cramming everything into one line a lot of the time | 16:34 |
| *** sdake_ is now known as sdake | 16:34 | |
| *** dims has quit IRC | 16:36 | |
| *** dwalsh has joined #kolla | 16:37 | |
| *** dims has joined #kolla | 16:39 | |
| *** diogogmt_ has joined #kolla | 16:41 | |
| *** diogogmt has quit IRC | 16:43 | |
| *** diogogmt_ is now known as diogogmt | 16:43 | |
| *** dwalsh has quit IRC | 16:44 | |
| sdake | pbourke different topic, oracle linux build from source gating | 16:46 |
| sdake | pbourke I'll do the gating patch for it and the patch to test_build | 16:47 |
| pbourke | sdake: its not building right? | 16:47 |
| sdake | I dont know - binary wasn't last i tried | 16:47 |
| sdake | not sure I tried from surce | 16:47 |
| pbourke | sdake: yeah its not going to. I want to add epel/rdo into our base to get it working, just literally haven't had the time :( Im not sure what the usual procedure is here, I assume you should lay down an ultimatum with a date for us to sort it by | 16:48 |
| sdake | pbourke on the binary front is there any opposition to buiding oracle linux with rdo? | 16:49 |
| pbourke | sdake: otherwise we may have to drop it | 16:49 |
| sdake | whats not going to? binary or source ? | 16:49 |
| pbourke | neither binary or source will build without epel/rdo | 16:49 |
| pbourke | as the necessary packages aren't available in the default sources | 16:50 |
| sdake | i keenly want building of oraclelinux before we ship liberty | 16:50 |
| pbourke | me too | 16:50 |
| sdake | earlier is better | 16:50 |
| sdake | what does it take to do the job? | 16:50 |
| sdake | I can do it if you can throw me a bone | 16:50 |
| pbourke | that would be very generous :) | 16:51 |
| pbourke | off the top of my head we just need something like: | 16:51 |
| sdake | where I get stuck is there is no epel-release package for oraclelinux | 16:51 |
| pbourke | {% if "oraclelinux" %} | 16:51 |
| pbourke | ... | 16:51 |
| pbourke | add epel / rdo | 16:51 |
| pbourke | <% endif %} | 16:51 |
| sdake | does the epel repo file need to be changed in some way? | 16:52 |
| pbourke | bah, terrible psedo code :) | 16:52 |
| pbourke | no it should be compatible | 16:52 |
| sdake | we already have that if in the base fie | 16:52 |
| pbourke | oh | 16:52 |
| pbourke | what error was binary throwing then? | 16:52 |
| sdake | its tehe adding epel repo - yum -y install epel-releease fails | 16:52 |
| pbourke | ok let me check that | 16:53 |
| sdake | epel-release is part of centos base packaging | 16:53 |
| sdake | is it part of oraclelinux? | 16:53 |
| sdake | i'm good with not using yum to get the repo file into the system for oraclelinux conditionally | 16:53 |
| sdake | just not totallyf amiliar with oraclelinux :) | 16:53 |
| pbourke | ol is basically repackaged red hat afaik | 16:54 |
| *** dims has quit IRC | 16:54 | |
| pbourke | the whole idea is you can move between the two seamlessly | 16:54 |
| *** unicell has joined #kolla | 16:55 | |
| pbourke | docker run -it oraclelinux yum install -y https://rdoproject.org/repos/openstack-kilo/rdo-release-kilo.rpm | 16:55 |
| pbourke | this seems to work ^ | 16:55 |
| pbourke | i'll try the template now | 16:55 |
| sdake | epel-release is from centos i think - addon package | 16:55 |
| *** shardy_ has joined #kolla | 16:55 | |
| pbourke | wait hang on that was rdo | 16:55 |
| pbourke | sorry getting tired now | 16:55 |
| *** shardy has quit IRC | 16:55 | |
| *** tfukushima has quit IRC | 16:56 | |
| *** dims has joined #kolla | 16:57 | |
| pbourke | ok so the issue is OL doesn't have the epel-release meta package in its base repos | 16:57 |
| sdake | does https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm | 16:57 |
| pbourke | yes the url seems to work | 16:57 |
| sdake | there is the repo file | 16:57 |
| sdake | i'm good with a conditional there | 16:58 |
| pbourke | works for me | 16:58 |
| sdake | is that all it takes to get it building? | 16:58 |
| pbourke | lets try | 16:58 |
| sdake | pbourke I have an agile waste my day meeting now but i'll be watching irc for results :) | 16:58 |
| pbourke | :) | 16:59 |
| pbourke | trying it now | 16:59 |
| *** shardy_ has quit IRC | 17:00 | |
| *** shardy has joined #kolla | 17:01 | |
| *** jmccarthy has quit IRC | 17:01 | |
| *** bmace has quit IRC | 17:02 | |
| *** dims_ has joined #kolla | 17:02 | |
| *** dims has quit IRC | 17:04 | |
| *** diga has quit IRC | 17:05 | |
| *** rhallisey has quit IRC | 17:08 | |
| *** dims has joined #kolla | 17:08 | |
| *** dims_ has quit IRC | 17:09 | |
| *** rhallisey has joined #kolla | 17:11 | |
| *** dims_ has joined #kolla | 17:11 | |
| *** dims has quit IRC | 17:14 | |
| *** dims_ has quit IRC | 17:20 | |
| *** absubram has joined #kolla | 17:20 | |
| *** dims has joined #kolla | 17:20 | |
| *** dims has quit IRC | 17:20 | |
| *** dims has joined #kolla | 17:21 | |
| *** achanda has joined #kolla | 17:27 | |
| pbourke | the base file is so messy | 17:29 |
| *** britthouser has joined #kolla | 17:30 | |
| *** britthou_ has quit IRC | 17:33 | |
| *** loth has joined #kolla | 17:40 | |
| pbourke | sdake: have binary base building now, will have to test the rest in the morning as have to run for my bus! here's the patch: http://paste.openstack.org/show/420786/ | 17:44 |
| *** bmace has joined #kolla | 17:45 | |
| *** bmace has quit IRC | 17:56 | |
| *** banix has quit IRC | 17:57 | |
| *** thumpba has quit IRC | 18:00 | |
| *** thumpba has joined #kolla | 18:04 | |
| *** banix has joined #kolla | 18:05 | |
| *** gfidente has quit IRC | 18:05 | |
| sdake | pbourke thanks! | 18:07 |
| *** SourabhP has joined #kolla | 18:10 | |
| *** thumpba has quit IRC | 18:11 | |
| *** openstackgerrit_ has quit IRC | 18:11 | |
| sdake | rhallisey can you review https://review.openstack.org/#/c/212804/ | 18:25 |
| *** SourabhP has quit IRC | 18:27 | |
| *** thumpba has joined #kolla | 18:27 | |
| *** athomas has quit IRC | 18:28 | |
| *** akwasnie has quit IRC | 18:29 | |
| *** dwalsh has joined #kolla | 18:31 | |
| *** thumpba has quit IRC | 18:32 | |
| sdake | samyaple fantastic work on globals.yml and the new defaults thing - love it ! | 18:34 |
| *** athomas has joined #kolla | 18:34 | |
| *** SourabhP has joined #kolla | 18:35 | |
| *** dwalsh has quit IRC | 18:52 | |
| *** thumpba has joined #kolla | 18:52 | |
| sdake | rhallisey with https://review.openstack.org/#/c/213845/ does cinder work entirely? | 18:55 |
| sdake | thats the padded one with the bootstrap container | 18:55 |
| *** sdake_ has joined #kolla | 18:56 | |
| *** sdake has quit IRC | 18:59 | |
| *** banix has quit IRC | 19:00 | |
| *** banix has joined #kolla | 19:04 | |
| *** thumpba has quit IRC | 19:08 | |
| rhallisey | ya | 19:13 |
| rhallisey | it works | 19:13 |
| sdake_ | nice | 19:13 |
| rhallisey | sam found the spelling error that I was going crazy over | 19:13 |
| sdake_ | make sure to wear a helmet ;) | 19:13 |
| sdake_ | ya I saw that | 19:13 |
| rhallisey | 2nd one man | 19:13 |
| sdake_ | I look at the review queue about 10 times a day | 19:13 |
| rhallisey | took forever to find | 19:13 |
| rhallisey | source source was spelled ource | 19:13 |
| rhallisey | then that one | 19:13 |
| sdake_ | so cinder is gtg now? | 19:14 |
| rhallisey | ya | 19:14 |
| rhallisey | worked for me | 19:14 |
| sdake_ | nice how do you use it | 19:14 |
| rhallisey | db finally syncd | 19:14 |
| sdake_ | where is the storage stored? | 19:14 |
| rhallisey | this is the same old cinder | 19:14 |
| rhallisey | so same docs | 19:14 |
| sdake_ | 4gb storage? | 19:14 |
| rhallisey | ya | 19:14 |
| sdake_ | can you remind me , was that a file or a disk | 19:14 |
| rhallisey | I made the arbitrary value you can make it whatever | 19:14 |
| rhallisey | file | 19:15 |
| sdake_ | cool and that review on the single data container? | 19:15 |
| rhallisey | I haven't review it yet | 19:15 |
| rhallisey | sorry | 19:15 |
| *** sdake_ is now known as sdake | 19:16 | |
| *** thumpba has joined #kolla | 19:19 | |
| *** achanda has quit IRC | 19:22 | |
| *** banix has quit IRC | 19:24 | |
| *** rhallisey has quit IRC | 19:28 | |
| *** banix has joined #kolla | 19:30 | |
| *** jasonsb has quit IRC | 19:47 | |
| *** britthouser has quit IRC | 19:54 | |
| *** britthouser has joined #kolla | 19:54 | |
| *** jasonsb has joined #kolla | 19:56 | |
| *** britthou_ has joined #kolla | 20:12 | |
| *** banix has quit IRC | 20:12 | |
| *** britthouser has quit IRC | 20:15 | |
| *** britthouser has joined #kolla | 20:18 | |
| *** britthou_ has quit IRC | 20:19 | |
| *** achanda has joined #kolla | 20:22 | |
| *** achanda has quit IRC | 20:22 | |
| *** achanda has joined #kolla | 20:23 | |
| openstackgerrit | Merged stackforge/kolla: Remove Ansible abstraction for Nova https://review.openstack.org/211445 | 20:25 |
| *** achanda has quit IRC | 20:28 | |
| *** achanda has joined #kolla | 20:28 | |
| *** thumpba has quit IRC | 20:40 | |
| *** britthou_ has joined #kolla | 20:45 | |
| *** britthouser has quit IRC | 20:47 | |
| *** dwalsh has joined #kolla | 20:51 | |
| *** britthou_ has quit IRC | 20:51 | |
| *** dwalsh has quit IRC | 21:03 | |
| sdake | coolsvap|away you around? | 21:09 |
| *** jtriley has quit IRC | 21:16 | |
| *** dwalsh has joined #kolla | 21:17 | |
| *** rhallisey has joined #kolla | 21:22 | |
| *** rhallisey has quit IRC | 21:22 | |
| *** athomas has quit IRC | 21:27 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Add oslo_versionedobjects dependency for neutron https://review.openstack.org/214362 | 21:35 |
| *** dwalsh has quit IRC | 21:52 | |
| *** dwalsh has joined #kolla | 21:54 | |
| *** absubram has quit IRC | 22:01 | |
| *** Haomeng has quit IRC | 22:23 | |
| sdake | master doesn't work for me | 22:46 |
| sdake | nova boot fails | 22:46 |
| *** dwalsh has quit IRC | 22:48 | |
| *** rhallisey has joined #kolla | 22:53 | |
| *** blahRus has quit IRC | 23:11 | |
| *** dims_ has joined #kolla | 23:20 | |
| *** banix has joined #kolla | 23:21 | |
| *** achanda has quit IRC | 23:22 | |
| *** dims has quit IRC | 23:22 | |
| *** britthouser has joined #kolla | 23:24 | |
| sdake | such a worthless irritating error | 23:24 |
| sdake | message": "No valid host was found. There are not enough hosts available. | 23:24 |
| sdake | hypervisor-uptime shows hypervisors kicking around | 23:24 |
| *** britthou_ has joined #kolla | 23:25 | |
| *** britthouser has quit IRC | 23:28 | |
| *** achanda has joined #kolla | 23:30 | |
| *** achanda has quit IRC | 23:30 | |
| *** achanda has joined #kolla | 23:31 | |
| *** diogogmt has quit IRC | 23:48 | |
| *** britthou_ has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
