| *** bradjones has quit IRC | 00:22 | |
| *** bradjones has joined #kolla | 00:23 | |
| *** bradjones has joined #kolla | 00:23 | |
| *** jasonsb has quit IRC | 00:26 | |
| *** jasonsb has joined #kolla | 00:27 | |
| *** jasonsb has quit IRC | 00:31 | |
| *** dimsum__ has quit IRC | 00:39 | |
| *** dimsum__ has joined #kolla | 00:41 | |
| *** dims_ has joined #kolla | 00:50 | |
| *** dimsum__ has quit IRC | 00:50 | |
| *** erkules_ has joined #kolla | 01:05 | |
| *** erkules has quit IRC | 01:08 | |
| *** dimsum__ has joined #kolla | 01:17 | |
| *** dims_ has quit IRC | 01:17 | |
| *** dims_ has joined #kolla | 01:40 | |
| *** dimsum__ has quit IRC | 01:41 | |
| *** erkules_ is now known as erkules | 01:54 | |
| *** jtriley has joined #kolla | 02:16 | |
| *** jtriley has quit IRC | 02:20 | |
| *** jtriley has joined #kolla | 02:46 | |
| *** jasonsb has joined #kolla | 02:59 | |
| *** jtriley has quit IRC | 03:00 | |
| SamYaple | sdake_: https://review.openstack.org/#/c/193870/ | 03:09 |
|---|---|---|
| SamYaple | you saw that right? | 03:09 |
| *** sdake_ has quit IRC | 03:10 | |
| *** dims_ has quit IRC | 03:40 | |
| *** sdake has joined #kolla | 04:04 | |
| *** dimsum__ has joined #kolla | 04:41 | |
| *** dimsum__ has quit IRC | 04:46 | |
| *** Haomeng|2 has joined #kolla | 05:11 | |
| *** Haomeng has quit IRC | 05:15 | |
| *** nihilifer has joined #kolla | 05:26 | |
| nihilifer | good morning | 05:27 |
| diga | sdake: Hi | 05:28 |
| SamYaple | morning diga nihilifer | 05:29 |
| diga | morning SamYaple | 05:30 |
| diga | can you check this BP, I have just registerd - https://blueprints.launchpad.net/kolla/+spec/kolla-compose-script | 05:30 |
| diga | SamYaple: you can edit the details if you want | 05:31 |
| SamYaple | diga: it looks like sdake already approved it? | 05:31 |
| SamYaple | oh sorry nvm | 05:32 |
| diga | No, that was kolla-ansible | 05:32 |
| SamYaple | im not certain we can approve this because im not sure what the state of kolla-compose is going to be. As far as I know we are keeping that the default deploy method for single hosts | 05:33 |
| *** nihilifer has quit IRC | 05:46 | |
| diga | okay | 05:52 |
| diga | sdake suggested me to draft a BP on this | 05:54 |
| SamYaple | diga: yea im ok with the change proposed. but i haven't seen the path forward discussed anywhere and I dont want to step on anyones toes | 05:54 |
| diga | ok | 05:55 |
| diga | SamYaple: If it is fitting to our path, then we may reject this BP | 05:55 |
| diga | I dont mind if this not approves | 05:56 |
| *** dasm has joined #kolla | 06:15 | |
| *** inc0 has joined #kolla | 06:22 | |
| inc0 | good morning | 06:22 |
| *** erkules has quit IRC | 06:22 | |
| *** erkules has joined #kolla | 06:22 | |
| dasm | inc0: hello | 06:23 |
| dasm | all: good morning | 06:23 |
| SamYaple | morning diga inc0 | 06:23 |
| SamYaple | inc0: got a minute to talk about templating dockerfiles? | 06:23 |
| inc0 | let me get cup of tee | 06:24 |
| inc0 | brb 10mins | 06:24 |
| inc0 | tea even | 06:24 |
| SamYaple | drinking tee sounds painful | 06:24 |
| inc0 | well, no better way to get you up and awake right? | 06:24 |
| SamYaple | nothing like a good ole suprise trichotomy | 06:25 |
| diga | Morning SamYaple, dasm, inc0 | 06:31 |
| inc0 | allright, we have one less tee in site, what's up SamYaple?;) | 06:41 |
| inc0 | SamYaple, also, look at comment here plz: https://review.openstack.org/#/c/187981/ | 06:48 |
| SamYaple | inc0: sure but i dont think you get the issue at hand with the container | 06:56 |
| inc0 | I know you're reffering to this PID checking stuff | 06:57 |
| SamYaple | if we take down the loadbalancer, the vip needs to come down as well since it wont be going anywhere | 06:57 |
| SamYaple | i understand your concern, but keepalived has all kinds of ways to check who gets the vip | 06:57 |
| SamYaple | it doesnt need a pid check | 06:57 |
| inc0 | but it will even without HAProxy, for example to APIs themselves | 06:57 |
| SamYaple | in its current form, the keepalived container it is not reliable | 06:58 |
| SamYaple | it needs some check script | 06:58 |
| SamYaple | no neccessarialy pids | 06:58 |
| inc0 | yeah, but we don't know what that script will be | 06:59 |
| SamYaple | we should figure that out | 06:59 |
| inc0 | that's admins decision | 06:59 |
| SamYaple | sure, but we need a way to incorporate it in the container | 06:59 |
| SamYaple | right now there is no way t odo so | 06:59 |
| inc0 | I don't like PIDs, that's for sure | 07:00 |
| SamYaple | agreed. plus it wouldnt work due to seperate pid namespaces | 07:01 |
| inc0 | (I'm not even sure if we can get other containers PID without some hacking around) | 07:01 |
| SamYaple | putting haproxy and keepalived in the host pid namespace works, but it has other issue (i know from yaodu) | 07:01 |
| inc0 | we can make keepalived+haproxy combined container | 07:02 |
| SamYaple | i wont say no to that, but that violates the one process per container thing | 07:02 |
| inc0 | yeah, and limits options | 07:02 |
| SamYaple | it may be what we have to do for the sake of reliability, but lets think of other optiosn first | 07:02 |
| SamYaple | the keepalived script can literally be any script | 07:03 |
| SamYaple | so we need a way for keepalived to know haproxy is live | 07:03 |
| SamYaple | I would say a shared socket plus a script that checks into the socket to see if it responds | 07:04 |
| SamYaple | so haproxy provides the socket and responds, the script access the socket and lets keepalived know its active | 07:04 |
| inc0 | uhh, but that means we can't use container x without container y | 07:04 |
| SamYaple | correct | 07:04 |
| inc0 | on same host moreover | 07:05 |
| SamYaple | i am aware. we have no rule against that | 07:05 |
| inc0 | I'd rather go for haproxy+keepalived container in this case | 07:05 |
| inc0 | for sake of usability | 07:06 |
| SamYaple | we have a rule against that | 07:06 |
| SamYaple | so we have agreed to thin containers. if we are going to break that rule, it has to be decided that we will break it | 07:06 |
| SamYaple | the reason I would vote against the fat container is that when the haproxy config needs to be reloaded, we would have t oreload the keepalived service as well which is very disruptive | 07:07 |
| *** mstachow has joined #kolla | 07:12 | |
| inc0 | yeah, I guess | 07:14 |
| inc0 | but I also dislike having HAProxy hardcoded to keepalived | 07:14 |
| inc0 | this pretty much says "yeah, you have keepalived for example for Rabbit, but no, you can't use it" | 07:14 |
| mstachow | o/ all | 07:15 |
| SamYaple | inc0: rabbit goes through haproxy | 07:15 |
| inc0 | I think this is counterintuitive | 07:15 |
| SamYaple | mstachow: hey | 07:15 |
| inc0 | it can it doesn't have tpo | 07:15 |
| *** athomas has joined #kolla | 07:15 | |
| SamYaple | but why would you want keepalived for rabbit | 07:15 |
| SamYaple | openstack can talk to all the rabbit servers directly | 07:16 |
| inc0 | active-passive rabbit isn't that bad of an idea | 07:16 |
| SamYaple | but it wouldnt be | 07:16 |
| SamYaple | it would still need to be clustered | 07:16 |
| SamYaple | and partitions would still occur | 07:16 |
| SamYaple | keepalived wouldnt know about that stuff | 07:16 |
| inc0 | I'm not even considering partitions, because keepalived has no support for there | 07:17 |
| inc0 | (see my comments in HA spec) | 07:17 |
| inc0 | I guess I just dislike idea of binding one container to another | 07:17 |
| inc0 | implicitly | 07:17 |
| SamYaple | well it wouldn't be | 07:18 |
| SamYaple | it would be a script keepalived calls | 07:18 |
| SamYaple | _what_ script it calls can be variable | 07:18 |
| inc0 | I guess we can add check_live.sh to keepalived container | 07:18 |
| inc0 | and call this one | 07:18 |
| inc0 | and let people meddle with it should they want to | 07:19 |
| SamYaple | yes this is what i have been proposing | 07:19 |
| SamYaple | its not a lock in | 07:19 |
| inc0 | and what would be default? | 07:19 |
| SamYaple | haproxy | 07:19 |
| inc0 | shared socket stuff? | 07:20 |
| inc0 | that requires a bit more than a script | 07:20 |
| inc0 | and we get to square one | 07:20 |
| SamYaple | how d oyo ufigure square one? | 07:20 |
| inc0 | keepalived needs haproxy | 07:21 |
| inc0 | whether you like it or not | 07:21 |
| SamYaple | configurably. | 07:21 |
| SamYaple | where is the disconnect here? | 07:21 |
| inc0 | (btw, mstachow, Sam, have you seen this one? https://www.youtube.com/watch?v=Nzq9epS2b1A) | 07:22 |
| inc0 | hold on, let me read about socker sharing in Docker | 07:22 |
| inc0 | (I'm still Docker newbie) | 07:22 |
| SamYaple | inc0: just share the /run folder | 07:22 |
| SamYaple | inc0: those SWTOR videos always look so good | 07:22 |
| SamYaple | game was a bit of a disappointment | 07:23 |
| inc0 | yeah, game itse;f not-so | 07:23 |
| inc0 | but trailers and music are 10/10 | 07:23 |
| *** diga has quit IRC | 07:24 | |
| *** nihilifer has joined #kolla | 07:24 | |
| inc0 | I'll go gram some food and think of this as I go | 07:27 |
| inc0 | brb | 07:27 |
| nihilifer | sdake: are you here? | 07:32 |
| SamYaple | he has teeth issue | 07:34 |
| inc0 | SamYaple, allright, exposing /run and adding new check_alive.sh script which will, by default, check HAProxy socket makes sense | 07:37 |
| inc0 | or...even better! | 07:37 |
| inc0 | check if something is listening on :5000 | 07:38 |
| SamYaple | that might work better. lets think about it | 07:39 |
| inc0 | that will create dependency on keystone | 07:39 |
| inc0 | or haproxy, or any lb you'd want | 07:39 |
| inc0 | and we could expose ENV variable to override default port | 07:40 |
| SamYaple | if yo uwanted to go that route then oyu would have to be able make multiple keepalived containers on the same host | 07:40 |
| inc0 | not unless we'd make possible to look for several ports | 07:41 |
| *** akwasnie has joined #kolla | 07:41 | |
| inc0 | and unless all of them are being listened on, don't migrate | 07:41 |
| SamYaple | thats getting pretty ugly | 07:41 |
| *** gfidente has joined #kolla | 07:41 | |
| *** gfidente has joined #kolla | 07:41 | |
| SamYaple | there is no way to dynamically update that list of listening | 07:42 |
| inc0 | you'll have to restart service | 07:42 |
| inc0 | even redeploy container | 07:42 |
| inc0 | default would be just 5000 | 07:43 |
| *** bradjones has quit IRC | 07:43 | |
| inc0 | you'd be able to deploy keepalived per service if you'll change IP per service | 07:43 |
| SamYaple | running multiple keepalived per host requires more work than you think | 07:44 |
| inc0 | also, remember that we keep script outside, so if someone want's to change it, he's free to do so | 07:44 |
| inc0 | yeah, I'm not too concerned about that because that's pretty hardcore | 07:45 |
| inc0 | if someone want's to do this, he'd have to support it | 07:45 |
| *** bradjones has joined #kolla | 07:45 | |
| *** bradjones has joined #kolla | 07:45 | |
| SamYaple | i dont like the port thing for haproxy because the haproxy could have issues but still be listening. If the socket responds, all is good | 07:46 |
| SamYaple | lets go that route for the default script | 07:46 |
| SamYaple | as you said you can add in whatever you want for other scripts | 07:46 |
| SamYaple | the other scripts can listen on a port | 07:46 |
| inc0 | yeah | 07:46 |
| inc0 | and single port or multiple? | 07:46 |
| SamYaple | thats the scripts job | 07:47 |
| SamYaple | the script just needs to return positive | 07:47 |
| inc0 | I'll do netstat -nalp | <some nifty grep> anyway | 07:47 |
| inc0 | yeah, as default | 07:47 |
| inc0 | I'll expose env to override port anyway | 07:47 |
| *** nihilifer has quit IRC | 07:48 | |
| inc0 | aaand do single port for now | 07:48 |
| inc0 | let's not go too far ahead, if someone want's I'll +1 on patch with multiple ports | 07:49 |
| SamYaple | the haproxy socket script will need to be there as per the HA spec | 07:50 |
| *** nihilifer has joined #kolla | 07:50 | |
| inc0 | but maybe as part of haproxy container? | 07:50 |
| SamYaple | it has to exist in the keepalived container | 07:51 |
| inc0 | ehh, I don't want 2 default scripts tho | 07:51 |
| SamYaple | haproxy is the default script | 07:51 |
| inc0 | but socket exposure? | 07:51 |
| inc0 | I think we agreed that this is some wizadry which might go sideways | 07:52 |
| SamYaple | im ok with other options, but not port checks | 07:52 |
| SamYaple | the socket check? no | 07:52 |
| SamYaple | the port check, yes | 07:52 |
| SamYaple | 07:46 < SamYaple> i dont like the port thing for haproxy because the haproxy could have issues but still be listening. If the socket responds, all is good | 07:52 |
| inc0 | also, we already nearly have keepalived container, and no haproxy container yet | 07:54 |
| inc0 | do we want not-working-container-unless-we-land-another-bp thing upstream? | 07:54 |
| inc0 | how about add check later? | 07:55 |
| SamYaple | as it so the keepalived container has no use at the moment. we need the check to make it useful. the haproxy container is not required to write the check | 07:56 |
| SamYaple | stands* | 07:56 |
| inc0 | why? you'll have full active-passive | 07:57 |
| inc0 | but you'll have failover | 07:57 |
| inc0 | api's doesn't need LB to work | 07:57 |
| SamYaple | assuming all services run on all nodes. but what happens when a service fails? | 07:57 |
| SamYaple | the vip doesnt move | 07:57 |
| inc0 | yeah | 07:57 |
| SamYaple | we need a check script. | 07:57 |
| inc0 | but you'll get *something* right now | 07:57 |
| SamYaple | the haproxy check script is required as per the HA spec, that needs to be in the container | 07:57 |
| inc0 | full node failure | 07:58 |
| inc0 | it will, let's just do this in another patch | 07:58 |
| SamYaple | why would we merge half a container? | 07:59 |
| inc0 | it's functional as it is | 07:59 |
| SamYaple | i disagree that is is functional | 07:59 |
| inc0 | not fully functional | 07:59 |
| inc0 | well, full node failure will get vrrp moved right? | 07:59 |
| *** nihilifer has quit IRC | 07:59 | |
| SamYaple | all service failures will not get the vrrp moved, right? | 08:00 |
| inc0 | hence not fully functional | 08:00 |
| SamYaple | hence no merge | 08:00 |
| inc0 | and we can solve this one by port checking | 08:00 |
| SamYaple | we need it to check the services | 08:00 |
| *** diga has joined #kolla | 08:00 | |
| SamYaple | haproxy handles those checks per the ha spec | 08:01 |
| SamYaple | so it needs to check haproxy | 08:01 |
| SamYaple | if you dont want to write it, I can finish the patch. but we do need the check | 08:02 |
| *** bradjones has quit IRC | 08:02 | |
| inc0 | so we drop this patchset until HAProxy lands... | 08:02 |
| inc0 | that's kinda sux, because it already had +W | 08:03 |
| inc0 | and just rebase didn't make this happen | 08:03 |
| SamYaple | if thats the way you want to interpt it, but thats not what I said | 08:03 |
| inc0 | right now checking of HAProxy is pointless, because there is no HAProxy | 08:03 |
| inc0 | container | 08:03 |
| inc0 | so we don't assume people will install it on host | 08:03 |
| inc0 | so default check won't work | 08:04 |
| inc0 | and there is no alternative check which works, as you've pointed out | 08:04 |
| *** shardy has joined #kolla | 08:05 | |
| SamYaple | you havent even registered a blueprint for this | 08:05 |
| inc0 | https://blueprints.launchpad.net/kolla/+spec/keepalived-container | 08:07 |
| inc0 | it's set to implemented because as I said, only rebase prevented it to get into tree | 08:07 |
| SamYaple | ah its marked as implemented | 08:07 |
| SamYaple | well sdkae marked it as implemetned which it is not | 08:07 |
| SamYaple | just because this was missed in previous reviews does not make the issue go away | 08:07 |
| SamYaple | we do need a check script | 08:07 |
| inc0 | but we can't do check script you want right now | 08:08 |
| inc0 | I'm ok to make it, but not binding it to HAProxy right now | 08:08 |
| inc0 | because we don't have it at all | 08:08 |
| SamYaple | you can implement the check script right now. you dont have to use it. | 08:08 |
| SamYaple | you can have the check-haproxy.sh and a check-null.sh | 08:09 |
| inc0 | all I'm saying, let's add check-haproxy when we'll have it | 08:09 |
| inc0 | it doesn't have to be one patchset | 08:10 |
| SamYaple | this blueprint isnt complete without it | 08:10 |
| inc0 | no bp is ever complete in software world | 08:10 |
| SamYaple | if you have no futher things to add, we can end the discussion | 08:11 |
| inc0 | let's just finish off - you don't agree with no haproxy check? | 08:12 |
| inc0 | and just adding noop check right now? | 08:12 |
| SamYaple | there is no infrastructure for a check in place at all. a noop check is fine, but a check needs to occur | 08:13 |
| inc0 | I'm ok with adding check ifra | 08:13 |
| inc0 | I agree on this one | 08:13 |
| inc0 | but I don't agree to add haproxy check right now | 08:13 |
| SamYaple | then this blueprint isnt implemented | 08:14 |
| inc0 | I'd rather add noop or port check now and file a bug to add haproxy later | 08:14 |
| SamYaple | no | 08:14 |
| SamYaple | we arent filing a bug for somethign tahts part of a bp | 08:14 |
| inc0 | I didn't mark it as such;) | 08:14 |
| inc0 | we can do this without bug | 08:14 |
| inc0 | but having a bug is good way to keep this in mind | 08:14 |
| *** bradjones has joined #kolla | 08:14 | |
| *** bradjones has joined #kolla | 08:14 | |
| SamYaple | not when the conatienr hasnt even been commited yet | 08:14 |
| inc0 | Sam, then what do you really want? to implemnt useless check? | 08:15 |
| SamYaple | if you want ot implement as-is, add the noop check to just return true | 08:15 |
| inc0 | then this container is useless until HAProxy | 08:15 |
| inc0 | so let's not merge it | 08:15 |
| SamYaple | ill implemetn the haproxy check and we can close this bp | 08:15 |
| inc0 | but it won't work | 08:16 |
| inc0 | anyway, I'll implemnt noop | 08:16 |
| SamYaple | make sure to update the commit to partially implements blueprint | 08:16 |
| inc0 | or port (I'd rather add port check, however faulty, that's better than nothing) | 08:16 |
| SamYaple | no make it a noop | 08:16 |
| SamYaple | port can be optional it you wish | 08:17 |
| *** nihilifer has joined #kolla | 08:17 | |
| inc0 | noop it is then | 08:17 |
| SamYaple | on second though, we can keep it as implemented and close the BP once the noop checks are in place. We can replace the checks on the haproxy BP, but we do need the check infrastructure in place in the keepalived container | 08:22 |
| inc0 | yes, thank you | 08:22 |
| inc0 | let's do that | 08:22 |
| *** nihilifer has quit IRC | 08:23 | |
| SamYaple | yes, but the check infratrcutre still needs to exist. | 08:23 |
| inc0 | I'll do this, I agree with you on this one | 08:23 |
| SamYaple | the haproxy script would only be 2-3 lines large you realize | 08:23 |
| inc0 | I know, but it's just timing | 08:24 |
| inc0 | there is no haproxy yet. that's all | 08:24 |
| SamYaple | sure but you just fought me for 40 minutes on 2-3 lines | 08:24 |
| SamYaple | we both need to communicate better | 08:24 |
| inc0 | I guess | 08:24 |
| inc0 | we need to get a drink at midcycle | 08:24 |
| SamYaple | or two | 08:24 |
| inc0 | and get wasted. This helps with future communication | 08:25 |
| SamYaple | first rounds on me how about that | 08:25 |
| inc0 | drink in my book means "generally undefined amount of alkohol in any number of carriers" | 08:25 |
| inc0 | I won't say no to that | 08:25 |
| SamYaple | inc0: updated the review with our conversation | 08:27 |
| inc0 | cool, thank you, I'll get to implementation | 08:27 |
| *** shardy has quit IRC | 08:35 | |
| *** shardy has joined #kolla | 08:36 | |
| *** bradjones has quit IRC | 08:38 | |
| *** bradjones has joined #kolla | 08:40 | |
| *** bradjones has joined #kolla | 08:40 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Minor doc fixups in dev-quickstart.md https://review.openstack.org/190543 | 08:57 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Move ansible/README.md under docs/ https://review.openstack.org/190544 | 08:59 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add base image for oraclelinux https://review.openstack.org/191013 | 09:00 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Fix broken cinder symlinks for fedora https://review.openstack.org/193625 | 09:00 |
| *** inc0 has quit IRC | 09:02 | |
| *** inc0 has joined #kolla | 09:03 | |
| *** shardy has quit IRC | 09:08 | |
| *** shardy has joined #kolla | 09:08 | |
| *** pdb is now known as pbourke | 09:08 | |
| *** inc0 has quit IRC | 09:23 | |
| *** inc0_ has joined #kolla | 09:23 | |
| *** shardy has quit IRC | 09:36 | |
| *** shardy has joined #kolla | 09:37 | |
| inc0_ | SamYaple, btw you wanted to talk about config templates? | 10:00 |
| SamYaple | not anymore | 10:00 |
| SamYaple | ;) | 10:00 |
| SamYaple | yea give me a minute | 10:00 |
| SamYaple | ok inc0_ | 10:06 |
| SamYaple | so I want to rewrite the build script to use python queues and multithreading for performance | 10:07 |
| SamYaple | I rewrote it in bash to be parrallel but it just go more complicated | 10:07 |
| SamYaple | during the rewrite would be the perfect time to do the templating | 10:07 |
| SamYaple | I was thinking of using jinja2 | 10:07 |
| inc0_ | do we need that performance? don't forget multithreading 1. isn't very good in cpython and 2. is much harder in debugging | 10:08 |
| inc0_ | jinja2 is cool, I used this a lot for configs | 10:08 |
| SamYaple | inc0_: multithreaded building, as in this thread does this build, this one does that build | 10:08 |
| SamYaple | im not rewriting docker | 10:08 |
| inc0_ | Ansible uses jinja as well? | 10:08 |
| SamYaple | inc0_: it does, hence why i think its a good choice | 10:08 |
| inc0_ | and other stuff can consume these tpls | 10:09 |
| inc0_ | which is nice as well | 10:09 |
| SamYaple | Basically we just want these builds to happen in parallel. It takes 40 minutes to build, in parrallel it took me 10 minutes | 10:09 |
| inc0_ | yeah, makes sense | 10:10 |
| inc0_ | network will be more utilized | 10:10 |
| SamYaple | everything will be | 10:10 |
| SamYaple | anyway. That shouldn't be too hard to do, but the templating needs some thought | 10:11 |
| SamYaple | it would be easy for different distros to share common sections wihtout symlink speghetti | 10:11 |
| SamYaple | but it would make the files more comlicated to read | 10:11 |
| SamYaple | I was thinking large if sections. {% if base == centos|fedora|rhel } do %} | 10:12 |
| inc0_ | or subsections like | 10:12 |
| * SamYaple says syntax is not garunteed accurate | 10:12 | |
| inc0_ | if centos => include centos | 10:13 |
| inc0_ | centos.j2 | 10:13 |
| inc0_ | we can make this readable | 10:13 |
| inc0_ | even for larger sessions | 10:13 |
| SamYaple | the issue I have with that is now its a bunch of files, harder to read | 10:13 |
| inc0_ | for smaller, just plain if | 10:13 |
| SamYaple | but also centos and fedora share alot of sections | 10:13 |
| SamYaple | so now the file centos.j2 needs to be generic | 10:14 |
| inc0_ | I wouldn't think too much about that tho | 10:14 |
| inc0_ | it should be solver per-case | 10:14 |
| *** dimsum__ has joined #kolla | 10:14 | |
| SamYaple | But jinja2 you think would be a good choice? it sounds like we have options | 10:14 |
| inc0_ | jinja2 is great imho. I have good experience with using it as config templating tool | 10:15 |
| *** nihilifer has joined #kolla | 10:15 | |
| SamYaple | same | 10:15 |
| inc0_ | also, we'll just use basic stuff | 10:15 |
| SamYaple | shouldnt be hard to convince others either since well be doing it in ansible | 10:15 |
| inc0_ | forloops and such, and that's good in jinja as well | 10:15 |
| inc0_ | good == easy | 10:15 |
| SamYaple | i cant see why we would need loops in the docker files, but its good to have for sure | 10:16 |
| inc0_ | you know what I mean | 10:16 |
| SamYaple | We can implement the templating first, but if we want it before liberty we have to move fast | 10:16 |
| SamYaple | this overrides discussions about what we will support as well | 10:17 |
| inc0_ | yeah | 10:17 |
| SamYaple | we can support whatever wihtout having a multiplying number of container files | 10:17 |
| inc0_ | it was me who wanted generated Dockerfiles remember?;) | 10:17 |
| SamYaple | thats why I am talking to you | 10:17 |
| SamYaple | though at the time you didnt offer up jinja2, i did ask what that would look like, remember :) | 10:18 |
| SamYaple | I think jinja2 can make generated files clean and readable | 10:18 |
| inc0_ | well, I haven't paid much thought about technologies themselves, but I | 10:18 |
| inc0_ | I'd go for jinja as first choice | 10:18 |
| SamYaple | now you say it | 10:19 |
| * SamYaple rolls eyes | 10:19 | |
| SamYaple | some of my ansible workload is being removed since there is work upstream | 10:19 |
| inc0_ | well...communication...drinks | 10:19 |
| SamYaple | im talking to the folks driving it and probably going to move that along | 10:19 |
| SamYaple | specifically the keystone modules and other OS modules | 10:19 |
| SamYaple | the docker module will need some work, but that can happen later | 10:20 |
| inc0_ | let's not solve all problems right now | 10:22 |
| inc0_ | we'll fail | 10:22 |
| SamYaple | the templating can be done without anything buildwise changing to much | 10:26 |
| SamYaple | im worried about duplicate work with all the source containers were building | 10:26 |
| inc0_ | with templating there won't be too much of it | 10:27 |
| inc0_ | not as much as without it anyway | 10:27 |
| SamYaple | merging 5 containers into one x 30-40 different services gets hefty work wise | 10:27 |
| SamYaple | If you are not interested in pushing this forward right now, it probably wont make it before Liberty since ive got stuff on my plate. If you are interested in driving it I will be able to contribute in the container merging | 10:29 |
| inc0_ | I can't say how much time I'll have for that | 10:31 |
| inc0_ | we need to land production-ready tripleo | 10:31 |
| inc0_ | even if it won't land fully in L, that's not that bad | 10:31 |
| inc0_ | as long as mechanism will be there | 10:31 |
| inc0_ | also, let | 10:32 |
| inc0_ | lets move this on meeting | 10:32 |
| inc0_ | maybe we'll get additional people | 10:32 |
| SamYaple | might be the case there | 10:32 |
| *** nihilifer has quit IRC | 10:33 | |
| *** nihilifer has joined #kolla | 10:37 | |
| *** sdake has quit IRC | 10:43 | |
| *** britthou_ has quit IRC | 10:43 | |
| *** apmelton has quit IRC | 10:43 | |
| *** apmelton has joined #kolla | 10:43 | |
| *** britthouser has joined #kolla | 10:43 | |
| *** sdake has joined #kolla | 10:44 | |
| *** dimsum__ is now known as dims | 10:48 | |
| *** dims has quit IRC | 10:53 | |
| *** inc0_ has quit IRC | 10:54 | |
| *** dims has joined #kolla | 11:05 | |
| *** nihilifer has quit IRC | 11:05 | |
| *** nihilifer has joined #kolla | 11:17 | |
| *** rhallisey has joined #kolla | 11:21 | |
| *** sdake_ has joined #kolla | 11:27 | |
| sdake_ | morning | 11:27 |
| openstackgerrit | Merged stackforge/kolla: Fix broken cinder symlinks for fedora https://review.openstack.org/193625 | 11:28 |
| *** sdake has quit IRC | 11:30 | |
| SamYaple | sdake_: morning | 11:35 |
| nihilifer | morning sdake_ | 11:36 |
| sdake_ | samyaple can you ack this https://review.openstack.org/#/c/193900/ | 11:38 |
| sdake_ | morning nih* | 11:39 |
| SamYaple | sdake_: https://review.openstack.org/#/c/193870/ | 11:39 |
| SamYaple | you ack i was first! | 11:39 |
| SamYaple | :P | 11:39 |
| SamYaple | i dont care which is merged, lets just pick one | 11:40 |
| nihilifer | one general question - when do we plan to approve bp about ubuntu containers? | 11:41 |
| sdake_ | samyaple I approved your change | 11:42 |
| sdake_ | nih* is one written yet? | 11:42 |
| nihilifer | yes, it's written, but not approved | 11:43 |
| nihilifer | https://review.openstack.org/#/c/189853/ | 11:43 |
| nihilifer | aww, not this | 11:43 |
| openstackgerrit | Merged stackforge/kolla: Unbreak the gate https://review.openstack.org/193870 | 11:43 |
| nihilifer | https://blueprints.launchpad.net/kolla/+spec/install-from-ubuntu | 11:43 |
| sdake_ | needs an assigneee to be approved | 11:44 |
| SamYaple | nihilifer: we can approve it. i dont have the cycles to implement it but I have most of the containers in yaodu | 11:44 |
| SamYaple | I was the intended assignee originally | 11:44 |
| nihilifer | yaodu? | 11:46 |
| nihilifer | however, I can help with making ubuntu containers, I need them ;) | 11:48 |
| SamYaple | nihilifer: sorry https://github.com/SamYaple/yaodu/tree/master/ansible/roles/docker_build/templates/ubuntu | 11:49 |
| SamYaple | those are some templates (yaodu was based on source installs) but that should get you started if you want to lead that | 11:49 |
| SamYaple | nihilifer: do you want me to assign the blueprint to you? | 11:51 |
| nihilifer | SamYaple: yes, assign me please | 11:51 |
| SamYaple | ive got two entries for nihilifer in launchpad, nihilifer-old and mrostecki? | 11:52 |
| nihilifer | mrostecki is a correct one | 11:53 |
| SamYaple | assigned | 11:53 |
| nihilifer | thanks | 11:53 |
| SamYaple | the non openstack containers you should be able to pull from Yaodu in the link above. That should get oyu started | 11:53 |
| SamYaple | ive worked out all the deps there | 11:54 |
| *** shardy has quit IRC | 11:57 | |
| *** shardy has joined #kolla | 12:02 | |
| *** inc0 has joined #kolla | 12:13 | |
| *** britthou_ has joined #kolla | 12:16 | |
| *** jtriley has joined #kolla | 12:16 | |
| *** britthouser has quit IRC | 12:17 | |
| pbourke | sdake_: hope the tooth is holding up. Did you have a question for me over the weekend or does it matter now? | 12:36 |
| *** jmccarthy has quit IRC | 12:37 | |
| inc0 | from mail he wrote it seems not too much matters now | 12:38 |
| *** shardy has quit IRC | 12:39 | |
| *** shardy has joined #kolla | 12:39 | |
| *** jmccarthy has joined #kolla | 12:40 | |
| *** dwalsh has joined #kolla | 12:44 | |
| *** jtriley has quit IRC | 12:45 | |
| *** dwalsh has quit IRC | 12:50 | |
| *** bradjones has quit IRC | 12:52 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Keepalived container https://review.openstack.org/187981 | 12:54 |
| *** shardy has quit IRC | 12:54 | |
| *** bradjones has joined #kolla | 12:54 | |
| *** bradjones has joined #kolla | 12:54 | |
| *** shardy has joined #kolla | 12:56 | |
| inc0 | SamYaple, take a look plz | 12:57 |
| SamYaple | inc0: yup looks good | 12:57 |
| inc0 | yay. | 12:57 |
| SamYaple | can you confirm that when the script fails the vip moves? | 12:57 |
| inc0 | hold on, I'll test that | 12:57 |
| *** jtriley has joined #kolla | 13:03 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Move ansible/README.md under docs/ https://review.openstack.org/190544 | 13:04 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Minor doc fixups in ansible-deployment.md https://review.openstack.org/190546 | 13:04 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add base image for oraclelinux https://review.openstack.org/191013 | 13:05 |
| *** dwalsh has joined #kolla | 13:06 | |
| sdake_ | mornn g again | 13:09 |
| inc0 | hi sdake_ you alive? | 13:14 |
| *** jtriley has quit IRC | 13:14 | |
| mstachow | hi sdake how are u? | 13:16 |
| sdake_ | i'm fineonc0 | 13:17 |
| *** pbourke has quit IRC | 13:19 | |
| *** pbourke has joined #kolla | 13:19 | |
| pbourke | sdake_: sourced some heroin? | 13:20 |
| inc0 | I've heard morphine is good for creativity | 13:21 |
| *** shardy has quit IRC | 13:22 | |
| *** shardy has joined #kolla | 13:22 | |
| *** shardy has quit IRC | 13:22 | |
| *** shardy has joined #kolla | 13:22 | |
| *** pradk_ is now known as prad | 13:24 | |
| mstachow | or You can also break your leg. In this case You will have still tooth pain, but leg pain will be greater (: | 13:24 |
| * mstachow watching Happy Tree Friends :) | 13:24 | |
| inc0 | also you'd have more space around your seat in plane | 13:24 |
| *** dasm is now known as dasm|afk | 13:26 | |
| mstachow | I've got question about rabbit in ha. Is it acceptable to distribute rabbit cookie through whole rabbit cluster via env variable? | 13:33 |
| sdake_ | pbourke i got a friend who has a drug dealer to get me "blues" which are 30mg oxycodone | 13:34 |
| sdake_ | and I split them in 4 pieces | 13:34 |
| sdake_ | my dentist said he would write me some oxycodone when he got back tuesday | 13:34 |
| pbourke | :D | 13:35 |
| pbourke | dentist, drug dealer, whats the difference | 13:35 |
| sdake_ | 40$ a peice | 13:36 |
| sdake_ | rediculous but gets rid of the pain | 13:36 |
| pbourke | akwasnie: hi you around? | 13:38 |
| akwasnie | hi yes | 13:38 |
| * jmccarthy doesn't want to know what a pain level of 8 on the holy f*ck scale is like | 13:39 | |
| inc0 | I have some hardcore painkillers at home, I've got a bottle after my tooth removal, and I keep them just in case | 13:39 |
| *** jtriley has joined #kolla | 13:40 | |
| pbourke | akwasnie: just wondering how you're doing with the keystone source patch. did the review comments sound ok etc? | 13:40 |
| *** absubram has quit IRC | 13:41 | |
| akwasnie | yes, i'm working on them now and try to make it worked | 13:47 |
| SamYaple | hey guys. Looks like most of the modules for ansible have been written and landed upstream thanks to the efforts of a few folks and openstack shade | 13:47 |
| SamYaple | so we shouldn't have to maintain big modules, only a few targeted ones for our needs | 13:47 |
| inc0 | SamYaple, where are these modules? anslible-openstack project? | 13:50 |
| SamYaple | inc0: no | 13:51 |
| SamYaple | https://github.com/ansible/ansible-modules-core/tree/devel/cloud/openstack | 13:51 |
| SamYaple | all the needed modules are expected to land in 2.0 | 13:52 |
| inc0 | wo, cool | 13:52 |
| SamYaple | we may need to pull them in (and a few noncommited ones) if the license permits | 13:52 |
| SamYaple | but we can drop them with 2.0 drops | 13:52 |
| *** kevsi has quit IRC | 13:54 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Seperate check_required_vars one per line for Glance https://review.openstack.org/194152 | 14:02 |
| *** jtriley has quit IRC | 14:02 | |
| inc0 | SamYaple, after some tinkering I've confirmed, check works | 14:03 |
| inc0 | as it is | 14:03 |
| *** jtriley has joined #kolla | 14:04 | |
| inc0 | pbourke, about stub config | 14:05 |
| inc0 | it is usable right now | 14:05 |
| inc0 | just will not confirm anything works underneath, so it will migrate vrrp only when host dies | 14:05 |
| inc0 | but that's something | 14:05 |
| SamYaple | inc0: does it migrate when the script returns false? | 14:05 |
| inc0 | it does | 14:06 |
| SamYaple | cool | 14:06 |
| inc0 | I've changed default script to failing on usual master | 14:06 |
| inc0 | and it works | 14:06 |
| inc0 | I mean second node gets IP | 14:06 |
| *** shardy has quit IRC | 14:07 | |
| *** shardy has joined #kolla | 14:07 | |
| pbourke | inc0: ok sounds good | 14:07 |
| SamYaple | inc0: alright if oy ucan address the most recent comments ill +2 it and we can get this merged | 14:08 |
| *** nihilifer has quit IRC | 14:09 | |
| inc0 | you mean socat? | 14:10 |
| SamYaple | inc0: socat and the check-script documentation. just need a bit more info. and the few nitpicks since youre already doing another patch. | 14:11 |
| SamYaple | the nitpicks wouldnt hold up the merge, but since youre there | 14:11 |
| *** jtriley has quit IRC | 14:11 | |
| *** jtriley has joined #kolla | 14:12 | |
| inc0 | SamYaple, what's socat for here? | 14:12 |
| SamYaple | socat is needed for talking to unix sockets | 14:12 |
| inc0 | ah, for later check | 14:13 |
| SamYaple | yes | 14:13 |
| inc0 | ok, makes sense | 14:13 |
| inc0 | hmm | 14:18 |
| inc0 | something's not right | 14:18 |
| SamYaple | ? | 14:18 |
| inc0 | I've made test check script which does if -f /test then echo foo; exit 1 | 14:19 |
| inc0 | this should move vrrp once the moment I "touch /test" right? | 14:20 |
| SamYaple | if the exit is in the if statement, sounds right | 14:20 |
| inc0 | then I've did something wrong :/ | 14:20 |
| SamYaple | use `docker exec` to change the script live | 14:21 |
| SamYaple | do something simple | 14:21 |
| SamYaple | have it be `exit 0` by default | 14:21 |
| SamYaple | change it to `exit 1` and see i it flips | 14:22 |
| *** thomasem has quit IRC | 14:28 | |
| *** thomasem has joined #kolla | 14:29 | |
| inc0 | ok, that's it, I'm going home | 14:32 |
| inc0 | with a lot of colorful swears coming out of my mouth | 14:32 |
| inc0 | cya tomorrow | 14:32 |
| *** inc0 has quit IRC | 14:37 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add all missing check_required_vars for Glance https://review.openstack.org/194166 | 14:37 |
| *** shardy has quit IRC | 14:41 | |
| *** shardy has joined #kolla | 14:42 | |
| *** jruano has joined #kolla | 14:49 | |
| *** shardy has quit IRC | 14:49 | |
| *** shardy has joined #kolla | 14:49 | |
| *** shardy has quit IRC | 14:49 | |
| *** shardy has joined #kolla | 14:49 | |
| *** absubram has joined #kolla | 14:50 | |
| *** sdake_ has quit IRC | 14:52 | |
| *** thomasem has quit IRC | 14:57 | |
| pbourke | http://docs.openstack.org/infra/shade/ - "However, if what you want is to be able to write an application that talks to clouds no matter what crazy choices the deployer has made in an attempt to be more hipster than their self-entitled narcissist peers, then shade is for you." | 14:58 |
| *** thomasem has joined #kolla | 15:02 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Auto generate minimal environment variables doc https://review.openstack.org/185814 | 15:03 |
| *** thomasem has quit IRC | 15:04 | |
| *** thomasem has joined #kolla | 15:04 | |
| *** thomasem has quit IRC | 15:04 | |
| *** jpeeler has joined #kolla | 15:05 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Auto generate minimal environment variables doc https://review.openstack.org/185814 | 15:06 |
| *** thomasem has joined #kolla | 15:06 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Auto generate minimal environment variables doc https://review.openstack.org/185814 | 15:08 |
| *** thomasem has quit IRC | 15:11 | |
| *** thomasem has joined #kolla | 15:11 | |
| *** thomasem has quit IRC | 15:12 | |
| *** thomasem has joined #kolla | 15:12 | |
| *** thomasem has quit IRC | 15:12 | |
| *** thomasem has joined #kolla | 15:13 | |
| *** thomasem has quit IRC | 15:13 | |
| *** daneyon has joined #kolla | 15:17 | |
| *** thomasem has joined #kolla | 15:17 | |
| *** nihilifer has joined #kolla | 15:19 | |
| *** daneyon_ has joined #kolla | 15:26 | |
| *** daneyon has quit IRC | 15:29 | |
| *** pbourke has quit IRC | 15:32 | |
| *** pdb has joined #kolla | 15:38 | |
| *** pdb is now known as pbourke | 15:38 | |
| *** shardy has quit IRC | 15:44 | |
| *** shardy has joined #kolla | 15:50 | |
| *** shardy has quit IRC | 15:57 | |
| openstackgerrit | Michal Rostecki proposed stackforge/kolla: Move Designate-Sink start.sh to docker/common https://review.openstack.org/194205 | 15:59 |
| *** shardy has joined #kolla | 16:02 | |
| *** jasonsb has quit IRC | 16:04 | |
| *** nihilifer has quit IRC | 16:32 | |
| *** nihilifer has joined #kolla | 16:33 | |
| *** daneyon has joined #kolla | 16:46 | |
| *** daneyon_ has quit IRC | 16:47 | |
| *** Slower_ is now known as Slower | 16:49 | |
| *** shardy has quit IRC | 16:50 | |
| *** shardy has joined #kolla | 16:51 | |
| pbourke | away till Friday folks if anyone is looking for me | 16:52 |
| *** daneyon_ has joined #kolla | 16:58 | |
| *** daneyon has quit IRC | 16:58 | |
| *** daneyon has joined #kolla | 16:59 | |
| *** jasonsb has joined #kolla | 17:02 | |
| *** daneyon_ has quit IRC | 17:03 | |
| *** dims has quit IRC | 17:09 | |
| *** dims has joined #kolla | 17:09 | |
| *** sdake_ has joined #kolla | 17:18 | |
| *** daneyon has quit IRC | 17:18 | |
| *** daneyon has joined #kolla | 17:19 | |
| *** daneyon has quit IRC | 17:20 | |
| *** daneyon has joined #kolla | 17:21 | |
| *** daneyon has quit IRC | 17:23 | |
| *** shardy has quit IRC | 17:25 | |
| *** shardy has joined #kolla | 17:26 | |
| *** shardy has quit IRC | 17:26 | |
| *** shardy has joined #kolla | 17:26 | |
| *** daneyon has joined #kolla | 17:30 | |
| *** vinkman has joined #kolla | 17:32 | |
| *** gfidente has quit IRC | 17:35 | |
| *** dwalsh has quit IRC | 17:58 | |
| *** shardy has quit IRC | 18:00 | |
| *** shardy has joined #kolla | 18:00 | |
| *** rhallisey_ has joined #kolla | 18:09 | |
| *** _mstachow has joined #kolla | 18:09 | |
| *** rhallisey has quit IRC | 18:09 | |
| *** dwalsh has joined #kolla | 18:17 | |
| *** rhallisey_ has quit IRC | 18:47 | |
| *** rhallisey has joined #kolla | 18:57 | |
| openstackgerrit | Merged stackforge/kolla: Move Designate-Sink start.sh to docker/common https://review.openstack.org/194205 | 19:05 |
| *** dwalsh has quit IRC | 19:08 | |
| openstackgerrit | Harm Weites proposed stackforge/kolla: Configure Heat service domain https://review.openstack.org/194308 | 19:09 |
| *** akwasnie has quit IRC | 19:11 | |
| *** dwalsh has joined #kolla | 19:24 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add option to not run db sync https://review.openstack.org/194313 | 19:31 |
| *** shardy has quit IRC | 20:00 | |
| *** shardy has joined #kolla | 20:02 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add option to not run db sync https://review.openstack.org/194313 | 20:32 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Remove Keystone's variable defaults https://review.openstack.org/194353 | 20:32 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add option to not run db sync https://review.openstack.org/194313 | 20:37 |
| rhallisey | hmm let's see if that breaks the gate.. | 20:45 |
| *** shardy has quit IRC | 20:47 | |
| *** shardy has joined #kolla | 20:47 | |
| *** shardy has quit IRC | 20:53 | |
| *** shardy has joined #kolla | 20:53 | |
| *** nihilifer has quit IRC | 20:54 | |
| *** sdake_ has quit IRC | 20:56 | |
| *** rhallisey has quit IRC | 20:58 | |
| *** dwalsh has quit IRC | 21:21 | |
| *** juggler has joined #kolla | 21:21 | |
| *** absubram has quit IRC | 21:24 | |
| *** nihilifer has joined #kolla | 21:54 | |
| *** jtriley has quit IRC | 21:59 | |
| *** jasonsb has quit IRC | 22:20 | |
| *** jtriley has joined #kolla | 22:26 | |
| *** daneyon has quit IRC | 22:52 | |
| *** daneyon has joined #kolla | 22:53 | |
| *** jtriley has quit IRC | 23:00 | |
| *** shardy has quit IRC | 23:06 | |
| *** shardy has joined #kolla | 23:07 | |
| *** jtriley has joined #kolla | 23:09 | |
| *** jasonsb has joined #kolla | 23:13 | |
| *** prad has quit IRC | 23:20 | |
| *** jruano has quit IRC | 23:35 | |
| *** jtriley has quit IRC | 23:42 | |
| *** athomas has quit IRC | 23:44 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!