| kfox1111 | jascott1: the global api has a mechanism to provide multiple levels of value lookups, so that setting values that are general with one value, and refining them with possibly more values is possible. | 00:00 |
|---|---|---|
| jascott1 | yes | 00:00 |
| kfox1111 | it would solve the "I want to set one value and get all of mariadb-pv, mariadb-svc, mariadb-statefulset" etc, conditionally on. | 00:00 |
| kfox1111 | but it does require a macro to run. | 00:00 |
| kfox1111 | what if the code in helm was changed to rather then a value, called a gotl template and used the result? | 00:01 |
| kfox1111 | then the values could be processed through the template to provide much richer logic? | 00:01 |
| jascott1 | templating values? | 00:01 |
| jascott1 | templating the values file itself? | 00:01 |
| kfox1111 | no, templating the subpackage selection. | 00:01 |
| kfox1111 | rather then just reading the raw values. | 00:01 |
| jascott1 | "values processed through the template" | 00:02 |
| kfox1111 | instad of the 2 pass processing, make it 3: | 00:02 |
| kfox1111 | 1. process all non conditional templates and load them for processing. | 00:03 |
| kfox1111 | 2. run macro's in that environment specified in the conditinoal subchart's conditions line | 00:03 |
| kfox1111 | if enabled, load those charts into environment too. | 00:03 |
| kfox1111 | 3 process all the loaded templates. | 00:03 |
| jascott1 | oh you meant template the condition or put logic in it | 00:03 |
| kfox1111 | yeah. instead of the condition line having values to look for booleans for enabled, | 00:04 |
| kfox1111 | call the privided macro isntead and use the output to determine enabled or not. | 00:04 |
| kfox1111 | the provided macro can be the kolla_get_val_str method in our case, | 00:04 |
| kfox1111 | and we can use the global api then to do the enabled lookups. | 00:05 |
| portdirect | man that sounds complex | 00:05 |
| jascott1 | brb | 00:05 |
| kfox1111 | thats the drawback. but would handle a lot more cases then the current proposal. | 00:06 |
| kfox1111 | as the conditional logic comes from gotl with all its support for logic. | 00:06 |
| kfox1111 | while the values solution only supports flat bools. | 00:06 |
| portdirect | could you not do this just through addint he logic into each template? | 00:07 |
| openstackgerrit | sayantani proposed openstack/kolla-ansible: Update bifrost playbooks https://review.openstack.org/424384 | 00:07 |
| kfox1111 | thats the other way to do it. | 00:07 |
| jascott1 | but what does kolla_get_val_str get you? | 00:08 |
| kfox1111 | we have prototyped that in memcached already. | 00:08 |
| portdirect | where? | 00:08 |
| * portdirect goes to github | 00:08 | |
| * jascott1 goes home | 00:08 | |
| kfox1111 | jascott1: a tree structure where you can hang off values that affect multiple packages. | 00:08 |
| kfox1111 | portdirect: helm/microservice/memcached-deployment/templates/memcached-deployment.yaml | 00:09 |
| jascott1 | we need the inverse as well tho, this rabbit needs different settings than that | 00:09 |
| kfox1111 | portdirect: there's an enabled flag there that will enable/disable that template. | 00:09 |
| kfox1111 | jascott1: the global api provides that too. | 00:09 |
| jascott1 | using condition field i can already access global api | 00:09 |
| kfox1111 | the global value space in helm becomes a tree. | 00:09 |
| *** schwicht has quit IRC | 00:09 | |
| jascott1 | imeant global values | 00:09 |
| kfox1111 | the levels of the tree get more and more specific to individual packages. | 00:10 |
| kfox1111 | so you can target a value between all packages, and just one, or somewhere in between. | 00:10 |
| *** schwicht has joined #openstack-kolla | 00:10 | |
| kfox1111 | like, all nova services. | 00:10 |
| jascott1 | yes | 00:10 |
| *** portdirect is now known as portdirect_awayz | 00:10 | |
| kfox1111 | jascott1: hmm... | 00:10 |
| kfox1111 | can you put a list in conditionals, | 00:11 |
| kfox1111 | and have it honor the first defined one? | 00:11 |
| jascott1 | would be easy | 00:11 |
| kfox1111 | that may do the trick. | 00:11 |
| jascott1 | but some would want compounds or other logic | 00:11 |
| jascott1 | kfox1111 think I see what you mean tho | 00:11 |
| kfox1111 | yeah. I can see them wanting logic. which was where the gotl idea was for. | 00:12 |
| *** schwicht has quit IRC | 00:12 | |
| *** schwicht has joined #openstack-kolla | 00:12 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Openvswitch Service package with kuberentes entrypoint https://review.openstack.org/424182 | 00:12 |
| kfox1111 | jascott1: so, have a look at this. | 00:12 |
| kfox1111 | helm/microservice/memcached-deployment/templates/memcached-deployment.yaml | 00:12 |
| jascott1 | alright I will add it to the existing percolations bubbling in my head | 00:13 |
| kfox1111 | see search path: | 00:13 |
| kfox1111 | {{- $searchPath := printf ":global.kolla.%s.deployment:global.kolla.%s.all:global.kolla.memcached.deployment:global.ko | 00:13 |
| kfox1111 | lla.memcached.all:global.kolla.all" $elementName $elementName }} | 00:13 |
| jascott1 | ya | 00:13 |
| kfox1111 | what that means, is, if the user defines 'enabled' at: | 00:13 |
| jascott1 | be a lot easier in helm since its go and not just gotpl | 00:13 |
| kfox1111 | globals.kolla.mymariadb.deployment.enabled=true, it will honor that. | 00:14 |
| *** schwicht has quit IRC | 00:14 | |
| kfox1111 | but they also can go up higher and do it really generically, globals.kolla.memcached.all.enabled=false | 00:14 |
| kfox1111 | turn off all memcached's everywhere by default. | 00:14 |
| kfox1111 | if those two entries were placed in a users values.yaml, | 00:15 |
| kfox1111 | they could laucjn several servcie pakcages, and get only a mariadb on the one service. | 00:15 |
| *** schwicht has joined #openstack-kolla | 00:15 | |
| kfox1111 | the problem with that package htough, is the search path is dynamic... | 00:16 |
| kfox1111 | so wouldn't exactly work with a static value path in dependencies.yaml. | 00:17 |
| kfox1111 | but if the service package didn't allow the user to rename the embeded package, then it would be possible to statically define it I guess. | 00:17 |
| openstackgerrit | sayantani proposed openstack/kolla-ansible: Update bifrost playbooks https://review.openstack.org/424384 | 00:17 |
| kfox1111 | I don't know. tradeoffs there. | 00:17 |
| kfox1111 | but, values lists could work. | 00:18 |
| kfox1111 | so could macro based. | 00:18 |
| kfox1111 | so, up to you I guess which one? | 00:18 |
| jascott1 | i will keep digin | 00:18 |
| *** schwicht_ has joined #openstack-kolla | 00:19 | |
| kfox1111 | k. | 00:19 |
| kfox1111 | if you get the values thoug as a list of values where the first defined is used, and maybe even in the format defined in the other charts, Ithink that would work though. | 00:19 |
| *** inc0 has quit IRC | 00:20 | |
| *** schwicht has quit IRC | 00:20 | |
| kfox1111 | thought you couldn't say, do logic that said, enable pvc/disable pvc, and enable/pv if pvc is enabled and storage_class not defined. | 00:20 |
| kfox1111 | which we may want? | 00:21 |
| *** schwicht_ has quit IRC | 00:21 | |
| jascott1 | right thats getting into the logic | 00:21 |
| jascott1 | none of the stuff in this area sounds hard to implement in helm | 00:21 |
| kfox1111 | cool. | 00:21 |
| *** schwicht has joined #openstack-kolla | 00:21 | |
| jascott1 | cool thanks gonna bounce home but bbiab | 00:22 |
| kfox1111 | k. thanks for talking it out. :) | 00:22 |
| *** sayantani01 has joined #openstack-kolla | 00:25 | |
| *** schwicht has quit IRC | 00:26 | |
| *** schwicht has joined #openstack-kolla | 00:26 | |
| *** schwicht has quit IRC | 00:29 | |
| *** schwicht_at_work has joined #openstack-kolla | 00:29 | |
| *** schwicht_at_work has quit IRC | 00:32 | |
| *** schwicht_at_work has joined #openstack-kolla | 00:32 | |
| *** schwicht_at_work has quit IRC | 00:34 | |
| *** schwicht_at_work has joined #openstack-kolla | 00:35 | |
| *** adrian_otto has joined #openstack-kolla | 00:36 | |
| *** duonghq has joined #openstack-kolla | 00:37 | |
| duonghq | morning guys | 00:40 |
| duonghq | anybody got trouble with helm tiler (cannot pull from grc.io) | 00:41 |
| portdirect_awayz | yeah | 00:41 |
| duonghq | oh, docker 1.13.0 issue | 00:42 |
| * duonghq sad | 00:42 | |
| portdirect_awayz | kubectl delete –namespace=kube-system deploy tiller-deploy; helm init –tiller-image docker.io/port/tiller:v2.1.3 | 00:42 |
| portdirect_awayz | is a dodgy hack that works in the short term | 00:42 |
| portdirect_awayz | i promise that that image is clean :) | 00:42 |
| duonghq | thanks portdirect_awayz | 00:43 |
| *** portdirect_awayz is now known as authz | 00:43 | |
| duonghq | why do you transform into authz ? | 00:44 |
| duonghq | Will you go to PTG? | 00:44 |
| authz | to hide from people who dont notice :) | 00:44 |
| authz | yes - I got an email late tonight | 00:45 |
| authz | have you had feeback | 00:45 |
| *** huzhengchuan has joined #openstack-kolla | 00:45 | |
| duonghq | ya, no luck for me this time | 00:45 |
| duonghq | how about you? | 00:45 |
| authz | :( that sucks | 00:45 |
| authz | yeah | 00:45 |
| *** authz is now known as portdirect | 00:45 | |
| duonghq | nice | 00:45 |
| duonghq | looking for livestream | 00:46 |
| portdirect | was surprised as i thought they were all going out on friday | 00:46 |
| duonghq | ya | 00:46 |
| portdirect | I promise that I'll do everything I can to get you there (via stream) man | 00:47 |
| portdirect | though i think inc0 is working on that | 00:47 |
| portdirect | as we were in the same boat its only fair i try to help you :) | 00:47 |
| openstackgerrit | sayantani proposed openstack/kolla-ansible: Fix MySQL DB error in bifrost container https://review.openstack.org/424391 | 00:48 |
| duonghq | portdirect, thanks | 00:48 |
| * portdirect back to lurking | 00:48 | |
| duonghq | last meeting, iirc inc0 and sdake is doing their best | 00:48 |
| *** zhurong has joined #openstack-kolla | 00:48 | |
| *** portdirect is now known as pd_ | 00:48 | |
| *** jtriley has joined #openstack-kolla | 00:49 | |
| duonghq | nice to know you guys | 00:49 |
| *** goldyfruit has joined #openstack-kolla | 00:54 | |
| *** cuongnv has joined #openstack-kolla | 00:57 | |
| openstackgerrit | sayantani proposed openstack/kolla: Update Bifrost documentation https://review.openstack.org/417114 | 00:59 |
| duonghq | pd_, I don't know why but your image is CrashBackLoop | 01:04 |
| duonghq | https://github.com/kubernetes/helm/issues/1838 painful issue here | 01:05 |
| *** saneax is now known as saneax-_-|AFK | 01:06 | |
| *** sayantani01 has quit IRC | 01:08 | |
| *** zhubingbing has joined #openstack-kolla | 01:08 | |
| *** sayantani01 has joined #openstack-kolla | 01:09 | |
| *** nihilifer has quit IRC | 01:09 | |
| *** nihilifer has joined #openstack-kolla | 01:10 | |
| *** sayantan_ has joined #openstack-kolla | 01:11 | |
| *** sayantani01 has quit IRC | 01:13 | |
| *** sayantan_ has quit IRC | 01:16 | |
| *** goldyfruit has quit IRC | 01:18 | |
| *** sdake has joined #openstack-kolla | 01:20 | |
| sdake | sup peeps | 01:21 |
| jascott1 | yo | 01:21 |
| duonghq | sup sdake | 01:21 |
| *** schwicht_at_work has quit IRC | 01:21 | |
| sdake | duonghq just finished eating, going to work on getting the dev env going | 01:22 |
| sdake | and documented | 01:22 |
| sdake | rather the deploy | 01:22 |
| sdake | kfox1111 and portdirect made a patch for it | 01:22 |
| duonghq | roger | 01:22 |
| sdake | i tested it earlier this morning and it failed | 01:22 |
| sdake | it needs some attention | 01:22 |
| sdake | the dev env is going and documented | 01:23 |
| duonghq | today you have dinner sooner than usual? | 01:23 |
| sdake | it wasn't that hard, portdirect did a mjaority of the work | 01:23 |
| sdake | 6pm is pretty normal for me | 01:23 |
| sdake | brb, need power cable | 01:23 |
| kfox1111 | sorry. was fighting a fire. just got back. | 01:24 |
| sdake | kfox1111 - during the deploy operation i ran genconfig, and it failed with a reconfigure on memcachew | 01:27 |
| sdake | what is odd about it is it failed with an error about 0.4.0-1 as the container version | 01:27 |
| sdake | i only spent 1 hour looking at it before my eye appointment which cratered my day | 01:27 |
| jascott1 | kfox1111 threw a preliminary 'would anyone freak if we made conditions do logic etc' over to helm peeps and it was not shot down so thats promising :) | 01:28 |
| sdake | sup jascott1 | 01:29 |
| jascott1 | hey mang! | 01:29 |
| kfox1111 | sdake: I think I saw, while reviewing the ps, that the gate is pinning kolla-ansible back a bit. | 01:29 |
| kfox1111 | maybe that is related? | 01:29 |
| sdake | back a bit? | 01:29 |
| sdake | you mean an old version? | 01:29 |
| kfox1111 | I think sbezverk put in the patch. but I don't remeber why off hand. | 01:29 |
| sdake | right - so maybe master is broken | 01:30 |
| kfox1111 | jascott1: awesome. :) | 01:30 |
| kfox1111 | sdake: yeah. grep setup_gate.sh for 4.0.0.0b2 | 01:30 |
| kfox1111 | looks like it is pulling back a bit. | 01:30 |
| sdake | i see thanks :) | 01:30 |
| kfox1111 | np. | 01:31 |
| sdake | i found other problems in the script as well | 01:31 |
| kfox1111 | I kind of forgot we did that. we should figure out why and unpin. | 01:31 |
| sdake | i'll update the review with the corrections | 01:31 |
| sdake | say - where does the iscsi device come from in the iscsi dev deploy workflow script? | 01:32 |
| kfox1111 | setup_gate_iscsi I think. | 01:33 |
| kfox1111 | or something similar.... | 01:34 |
| kfox1111 | no, tests/bin/setup_gate_loopback_lvm.sh it looks like. | 01:34 |
| kfox1111 | ah.. here's the docker issue's root: https://github.com/docker/docker/issues/30083 | 01:35 |
| bmace_ | sdake: talking about the dev container scripts sdake? i was running through trying to get that working today based on the latest bits. seems like maybe the recent memcache change is causing an issue in it. | 01:36 |
| sdake | bmace_ agreed its busted | 01:36 |
| sdake | also kolla needs to be renamed kolla_ansible in the script | 01:36 |
| bmace_ | sdake: happy to hear it wasn't just me :) | 01:37 |
| kfox1111 | seems like other repos then google are affected too. :/ | 01:37 |
| kfox1111 | bmace_: yeah. the script is untested, but should be closeish. just gotta find the remaining bugs. | 01:37 |
| bmace_ | yeah, i didn't like the whole flow of that.. i was working with my own version that handled a lot more of the dependency stuff during build, and also has kolla-kubernetes and kolla-ansible both right in /opt with the right names | 01:37 |
| kfox1111 | feel free to make changes. :) | 01:38 |
| kfox1111 | In my much earlier version, I prechecked out the stuff into the container during build too. | 01:39 |
| bmace_ | if i had the whole thing actually working and i was confident in it i definitely would. | 01:39 |
| kfox1111 | then only need to do git pulls during udpate. | 01:39 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 01:39 |
| *** schwicht has joined #openstack-kolla | 01:39 | |
| bmace_ | at this point things keep changing and so it keeps breaking in new ways :) | 01:40 |
| sdake | bmace_ think of that as a joint community review | 01:40 |
| kfox1111 | yeah. I'm hoping the microservices stay relatively stable from here on out. | 01:40 |
| sdake | bmace_ and please push what you learn to it - this is blocking a whole slew of people | 01:40 |
| kfox1111 | before it was changing rapidly. | 01:40 |
| sdake | kfox1111 you mean the microcharts | 01:40 |
| sdake | ? | 01:40 |
| kfox1111 | the more we can gate on too, the less likely that it will break in the future. | 01:41 |
| kfox1111 | sdake: yeah. | 01:41 |
| sdake | kfox1111 yup that is correct. as wel las the more developers looking at it :) | 01:41 |
| kfox1111 | they should be mostly stable from here, with maybe some api additions rather then changes. | 01:41 |
| kfox1111 | +1 | 01:41 |
| bmace_ | at this point i am definitely happy just getting it working. i reset a lot of my changes so most of my local changes now are just to get it to work behind our pretty strict firewall. once it is completely working i'll go back and make some of my more drastic changes that people can chat / disagree about :) | 01:42 |
| kfox1111 | cool. | 01:42 |
| sdake | bmace_ if you push a review tht works for me, I'd +14 million it :) | 01:42 |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Add fluentd role https://review.openstack.org/407392 | 01:42 |
| *** cuongnv is now known as cuong | 01:42 | |
| bmace_ | lol, well, need to figure out the memcache issue for now. you said you were hitting the same issue? | 01:43 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 01:43 |
| sdake | i wonder if ther e is a gerrit counter | 01:43 |
| sdake | besides stackalytics | 01:43 |
| *** cuong is now known as cuongnv | 01:43 | |
| kfox1111 | gotta head out for a while. l8r. | 01:43 |
| sdake | i've only done 2414 +2's for the lifetime of oenstack | 01:44 |
| bmace_ | i hit the issue during config gen.. looks like it tries to "check memcached container" and talk to docker to do that.. but during config gen it seems like it really shouldn't be doing all of that. | 01:44 |
| sdake | later kfox1111 enjoy your evening | 01:44 |
| sdake | bmace_ have exact same issue | 01:44 |
| bmace_ | back before the memcache change i had config gen working like a champ | 01:44 |
| sdake | bmace_ and agree with your assertion | 01:44 |
| sdake | which memcache change are you tlaking about? | 01:45 |
| bmace_ | i think i saw one go in over the weekend, about using memcache rather than a db to do something or rather? perhaps not causation but that is the only change related to memcache i noticed recently | 01:45 |
| sdake | bmace_ the bug is in kolla-ansible not kolla-kubernetes | 01:45 |
| sdake | genconfig rusn out of kolla-ansible | 01:46 |
| sdake | bmace_ do you recall the deliverable the ptach went into so I can hunt it down and revert it in my local repo? | 01:46 |
| bmace_ | sure, but was memcache enabled before for kubernetes? there are still some enable / disable settings done into the ansible environment for certain services | 01:46 |
| sdake | i see what your saying | 01:46 |
| sdake | that memcache was enabled in the default kubernetes config | 01:47 |
| sdake | and this exposed a bug in genconfig | 01:47 |
| bmace_ | yeah, that is my expectation.. something like that. | 01:47 |
| sdake | brb connecting to vpn | 01:48 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Support rolling upgrades of horizon https://review.openstack.org/333996 | 01:50 |
| *** sdake has quit IRC | 01:53 | |
| *** sdake has joined #openstack-kolla | 01:55 | |
| *** sdake has joined #openstack-kolla | 01:55 | |
| *** adrian_otto has quit IRC | 01:56 | |
| *** goldyfruit has joined #openstack-kolla | 01:59 | |
| *** Marcellin__ has quit IRC | 01:59 | |
| *** goldyfruit has quit IRC | 02:02 | |
| *** cuongnv_ has joined #openstack-kolla | 02:04 | |
| *** cuongnv has quit IRC | 02:06 | |
| bmace_ | i'm not great at debugging ansible issues, but it looks like for some reason it isn't properly ignoring the "Check memcached container" when we are doing a genconfig action.. do you know, does it always follow those in the right order? | 02:08 |
| *** schwicht has quit IRC | 02:08 | |
| bmace_ | in ansible/roles/memcached/tasks/config.yml seems like if the genconfig check is first and it always follows the order maybe it will fix it? we might be hitting the error when it goes to check to see if the service is enabled.. which is before it in the when flow right now | 02:09 |
| bmace_ | another kind of annoying part of the current dev container thing, it always re-pulls kolla-ansible on run from upstream where i think it would be nicer if it just used both a local kolla-kubernetes and kolla-ansible so you can tweak things in both and test fixes, etc. | 02:11 |
| bmace_ | afk for a few | 02:12 |
| *** schwicht has joined #openstack-kolla | 02:24 | |
| sbezverk | kfox1111: ping | 02:25 |
| *** cuongnv has joined #openstack-kolla | 02:26 | |
| *** cuongnv_ has quit IRC | 02:29 | |
| sdake | sbezverk kfox1111 went to bed i think :) | 02:34 |
| sbezverk | sdake: ok thanks | 02:35 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for ceilometer https://review.openstack.org/423606 | 02:36 |
| bmace_ | sdake: odd.. tweaked that line in place in my container and tried again. same error. maybe for some reason it isn't picking up the action is genconfig in there. | 02:36 |
| bmace_ | hrm.. looks like action=config is what it is called when passed into the playbook, not genconfig | 02:38 |
| jascott1 | kfox1111 in you pv/pvc example earlier storage_class is a values key right? | 02:38 |
| bmace_ | sdake: fyi, that was the fix.. someone somewhere changed either config or genconfig and didn't catch all the places it effected the other. any idea which way it should be? | 02:39 |
| bmace_ | maybe genconfig is the old name? | 02:39 |
| bmace_ | i only found the term genconfig used in keystone, glance, memcached, searchlight and senlin | 02:40 |
| Pavo | evening everyone | 02:43 |
| sdake | bmace_ can you paste a diff of he fix so i can see what your talking bout? | 02:43 |
| bmace_ | sure | 02:44 |
| *** harlowja has quit IRC | 02:44 | |
| sdake | pavo sup | 02:44 |
| Pavo | notta, in Baltimore for the week | 02:44 |
| Pavo | bored | 02:44 |
| sdake | sorry to hear it | 02:44 |
| *** tovin07_ has joined #openstack-kolla | 02:44 | |
| kfox1111 | sbezverk: ping. | 02:45 |
| bmace_ | diff --git a/ansible/roles/memcached/tasks/config.yml b/ansible/roles/memcached/tasks/config.yml | 02:45 |
| bmace_ | index 6b00816..63438c6 100644 | 02:45 |
| bmace_ | --- a/ansible/roles/memcached/tasks/config.yml | 02:45 |
| bmace_ | +++ b/ansible/roles/memcached/tasks/config.yml | 02:45 |
| bmace_ | @@ -29,5 +29,5 @@ | 02:45 |
| bmace_ | when: | 02:45 |
| *** goldyfruit has joined #openstack-kolla | 02:45 | |
| bmace_ | - inventory_hostname in groups[service.group] | 02:45 |
| bmace_ | - service.enabled | bool | 02:45 |
| bmace_ | - - action != "genconfig" | 02:45 |
| bmace_ | + - action != "config" | 02:45 |
| bmace_ | notify: Restart memcached container | 02:45 |
| bmace_ | that is really it.. if you look at the genconfig command it passed the action as "config" not "genconfig" | 02:45 |
| kfox1111 | jascott1: if pvc is enabled, and storage_class (or whatever its called), isn't defined, then they may want to create the pv right away) | 02:45 |
| sdake | bmace_ moment - at parents watching tv at 15 million decibels bekuz they are hard of hearing | 02:46 |
| bmace_ | there are 5 or 6 containers that look to see if the action is "genconfig" to avoid doing certain actions, like checking container state, which will fail in a config generation only scenario | 02:46 |
| kfox1111 | though, I've never been really convinced anythone but the cloud admin should create pv's. as there's no real way to tie a pv to a pvc except by not having any but 1. | 02:46 |
| Pavo | got a quick question sdake, if I wanted to make some changes to different openstack services using ansible, things that couldn't be changed using a reconfigure, would I need to use only the docker exec module of ansible to achieve it? | 02:46 |
| sdake | pavo i dont have a good answer for that question | 02:47 |
| bmace_ | sdake: if you make that change i posted above you will get past the memcached failure and then blow up in keystone, because it has a similar genconfig check, which I think should also be just "config" | 02:48 |
| sdake | pavo the concept if reconfigure is it can change anything | 02:48 |
| sdake | bmace_ trying to get to the repo to look at the git blame logs :) | 02:48 |
| Pavo | well reconfigure can't add new services that wasn't enabled during deploy | 02:48 |
| sdake | pavo the reality of reconfigure is it doesnt change everything | 02:48 |
| sdake | pavo for that, I think you can do a kolla-ansible deploy | 02:48 |
| jascott1 | kfox1111 thanks. im updating proposal with complex use case | 02:49 |
| sdake | pavo or a kolla-ansible upgrade | 02:49 |
| *** prameswar has joined #openstack-kolla | 02:49 | |
| Pavo | but doing a deploy would make me lose everything that I had deploied already correct? | 02:49 |
| sdake | pavo no it would not | 02:49 |
| sdake | pavo deploy is idempotent | 02:49 |
| kfox1111 | jascott1: k. thanks. | 02:49 |
| sdake | pavo personally i'd use the upgrade faatue to be safe | 02:49 |
| sdake | pavo also depends on your version of kolla | 02:50 |
| Pavo | hmmm thought it would, I did try the upgrade feature but doesn't do anything unless you retag all images to something later than what you deployed with | 02:51 |
| sdake | bmace_ see https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/memcached/tasks/config.yml#L30-L32 | 02:52 |
| sbezverk | kfox1111: all green https://review.openstack.org/#/c/424182/ | 02:52 |
| sdake | bmace_ to me, that ansible looks correct | 02:53 |
| sdake | i noticed the deploy script pins to an old version of ansible | 02:53 |
| sdake | 2.1 or less | 02:53 |
| sdake | 2.1 has a slew of bugs - many related to how when works | 02:53 |
| bmace_ | but there is no such thing as a "genconfig" action | 02:53 |
| bmace_ | the kolla-ansible genconfig command passes in a "config" action | 02:53 |
| sdake | bmace_ let me look at that - moment | 02:53 |
| sdake | bmace_ https://github.com/openstack/kolla-ansible/blob/master/tools/kolla-ansible#L220 | 02:55 |
| bmace_ | exactly. config != genconfig | 02:55 |
| bmace_ | which is why when genconfig is changed to config in the task it starts to work like a charm | 02:56 |
| sdake | https://github.com/openstack/kolla-ansible/blame/master/tools/kolla-ansible#L220 | 02:56 |
| sdake | the correct solution is to change the action to be genconfig in kolla-ansible | 02:57 |
| sdake | and change all parts where "config" is used in current master to genconfig | 02:57 |
| sdake | the api name should match the action name | 02:57 |
| sdake | i think inc0 fixed it the wrong way :) | 02:57 |
| bmace_ | yeah, those genconfig action checks could have never done anything the way kolla-ansible is written | 02:58 |
| sdake | bmace_ could you be more specific with a github pointer | 02:58 |
| bmace_ | well, you posted them all already. the added config action was done 8 months ago. the genconfig checks were all newer, but since the action looks like it was always "config" and not "genconfig" i don't see how any of those != genconfig checks would have ever done anything | 02:59 |
| sdake | bmace_ yup - untested gerrit review looks like | 03:00 |
| kfox1111 | sbezverk: looks good to me. | 03:00 |
| bmace_ | the lesser of two evils, since the config / kolla-ansible code is older, is just swap the 6 or so places where genconfig is checked in the config tasks, imo | 03:01 |
| sdake | bmace_ https://github.com/openstack/kolla-ansible/commit/34ce7aad5acaa04bf5940c6eb87f7788ad13c4fd | 03:02 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 03:02 |
| bmace_ | i'm fine making the fixes if you want, if we go the route of just changing genconfig to config in the ansible config tasks. | 03:03 |
| sdake | bmace_ that doesn't look like a regression to me, it looks like Jeffrey4l_ broke the code ;) | 03:03 |
| kfox1111 | sdake: can I get a quick rereview of https://review.openstack.org/#/c/422950/ ? | 03:04 |
| *** tovin07 has joined #openstack-kolla | 03:07 | |
| sdake | kfox1111 done | 03:07 |
| sdake | bmace_ Jeffrey4l_ has led the optimization of reconfigure | 03:07 |
| sdake | bmace_ which probably broke genconfig in some way | 03:07 |
| bmace_ | well, i know the way :) genconfig just needs to be config in 5 files and it will at least work better :) | 03:08 |
| sdake | bmace_ i guess since kolla-ansible used "config" in newton, we are stuck with it | 03:08 |
| kfox1111 | sdake: thx. | 03:08 |
| sdake | bmace_ cool fix away :) | 03:08 |
| bmace_ | kk | 03:08 |
| sdake | we may want to change all the config to genconfig at some other time | 03:09 |
| sdake | although in the medium term we are removing genconfig entirely | 03:09 |
| sdake | so it may not matter | 03:09 |
| bmace_ | yeah, i figured in the short term, with the change set so small, this is the path of least disruption | 03:11 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for ceilometer https://review.openstack.org/423606 | 03:12 |
| Jeffrey4l_ | bmace_, is reconfig still break the k8s? | 03:13 |
| bmace_ | genconfig is broken in the k8s dev container work, and sort of in general.. i'll have a review up shortly | 03:13 |
| Jeffrey4l_ | cool. | 03:15 |
| openstackgerrit | BorneMace proposed openstack/kolla-ansible: Fixed config generation to check correct actions https://review.openstack.org/424421 | 03:15 |
| sdake | bmace_ does that get you to a working openstack? | 03:15 |
| sdake | (on kubernetes) via the shell script? | 03:16 |
| bmace_ | i don't love the mix of double and single quotes in that code :/ | 03:16 |
| bmace_ | hrm.. fully working openstack? lets not hope for too much ;). i'm just happy to have genconfig complete successfully | 03:16 |
| Jeffrey4l_ | oops. i know what is happening. | 03:16 |
| sdake | bmace_ fwiw I did manage to get genconfig to work properly by installing newotn version of kolla and managed to get helm to launch all the packages against my vagrant environment | 03:17 |
| Jeffrey4l_ | but, are u are you patch works? i am afraid that handler will be trigger still, because config file change. bmace_ | 03:17 |
| bmace_ | i need to make the rest of those changes in my container and run it, and i'll let you know how far i make it. | 03:17 |
| sdake | Jeffrey4l_ if you hve a better version plesae push it up - you know this code as well as anyone | 03:18 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Fix old container https://review.openstack.org/424422 | 03:18 |
| sdake | Jeffrey4l_ this is blocking kolla-kubernetes for alot of folks | 03:18 |
| bmace_ | i made the change in my local environment, ran genconfig and it correctly did not process the container check | 03:18 |
| Jeffrey4l_ | i am not sure now. need make some test. | 03:18 |
| sdake | bmace_ what Jeffrey4l_ just said there is reconfigure will break with that change | 03:18 |
| sdake | bmace_ which is not acceptable either :) | 03:19 |
| Jeffrey4l_ | no. genconfig may still not work. | 03:19 |
| sdake | atlast i think that is what he said | 03:19 |
| Jeffrey4l_ | let me test. | 03:19 |
| sdake | ok | 03:19 |
| bmace_ | kk, have at it, and yeah, if there are better fixes go for it. these at least get me past where i was stuck | 03:19 |
| bmace_ | but the genconfig string there seemed blatantly wrong either way | 03:19 |
| sdake | i think it should be "" not '' tho | 03:20 |
| sdake | consistently used | 03:20 |
| sdake | my kids are running me ragged | 03:20 |
| bmace_ | other than in tox the string genconfig isn't use anywhere else in the entire kolla-ansible codebase, so i don't think reconfigure or anything else does anything magical to make action="genconfig" valid | 03:20 |
| sdake | bmace_ cool - i think the same | 03:21 |
| openstackgerrit | QiLiang proposed openstack/kolla-kubernetes: Typo fix: mysql => memcached https://review.openstack.org/424423 | 03:22 |
| kfox1111 | hmm.... I think the docker hub is acting up again. :/ | 03:23 |
| *** prameswar has quit IRC | 03:24 | |
| kfox1111 | we really need to finish up the container caching. :/ | 03:24 |
| *** qiliang28 has joined #openstack-kolla | 03:27 | |
| bmace_ | sdake: still hitting some issues. looks like the when ordering of that line does matter, at least in glance, and i'm hitting a similar issue in heat so i'm adding in a check there | 03:29 |
| *** qiliang27 has quit IRC | 03:29 | |
| *** qiliang28 is now known as qiliang27 | 03:29 | |
| *** schwicht has quit IRC | 03:30 | |
| *** prameswar has joined #openstack-kolla | 03:30 | |
| bmace_ | sdake: fixes heat issue also.. i'll post all the updates. | 03:31 |
| bmace_ | just blowing up now talking to my kube environment so getting closer ;) | 03:31 |
| *** schwicht has joined #openstack-kolla | 03:31 | |
| kfox1111 | oh.. so close.... | 03:32 |
| kfox1111 | http://logs.openstack.org/50/418550/43/check/gate-kolla-kubernetes-deploy-centos-binary-t-ceph-multi-nv/71958f7/console.html | 03:32 |
| kfox1111 | neutron seems to have changed their behavior, again. :/ | 03:32 |
| openstackgerrit | BorneMace proposed openstack/kolla-ansible: Fixed config generation to check correct actions https://review.openstack.org/424421 | 03:33 |
| *** sp_ has quit IRC | 03:33 | |
| Jeffrey4l_ | bmace_, need condition in handlers/main.yml file too. | 03:34 |
| Jeffrey4l_ | bmace_, i will update your patch. ;) | 03:34 |
| bmace_ | sure.. not sure why i need it, but you know better than I about the config gen.. i know with my change i was able to make a clean / full genconfig run with no access to docker / etc to do the container checks, which is where it was blowing up before. | 03:35 |
| sdake | Jeffrey4l_ while your at it can yo ufix the ' to " | 03:36 |
| sdake | bmace_ can you test out Jeffrey4l_ 's patch | 03:36 |
| Jeffrey4l_ | sdake, np | 03:36 |
| sdake | i am at my parents at present and don't have access to my personal lab | 03:36 |
| sdake | as i got a new eero and don't know my ip address | 03:36 |
| sdake | and eero sadly doesn't support dyndns | 03:37 |
| sdake | i did manage to take care of myself today and get a new pair of glasses so i can see straight | 03:38 |
| sdake | altough tht cratered the whole day | 03:39 |
| *** zhurong has quit IRC | 03:40 | |
| sdake | i can't believe how thourough my eye doctor was | 03:41 |
| sdake | he spent like 1 hr with me | 03:41 |
| sdake | and the exam was 72 bucks | 03:41 |
| sdake | of couse the glasses wer e700 bucks | 03:41 |
| sdake | which is cheap compare dto what i've paid in teh past | 03:41 |
| bmace_ | once he pushes them up i can give it a go in my dev container, sure | 03:42 |
| sdake | bmace_ fwiw I am doing this on bare metal as well | 03:42 |
| sdake | bmace_ i'd like both to work well (for th eoperator docs) | 03:43 |
| bmace_ | absolutely | 03:43 |
| sdake | genconfig has some wierd thing where eht0 doesn't work | 03:43 |
| sdake | and it hsa to be enp0s6 or something | 03:43 |
| sdake | for network_interface | 03:43 |
| *** schwicht has quit IRC | 03:43 | |
| kfox1111 | redhat changed the network naming in el7. | 03:43 |
| kfox1111 | (it was supported in el6 too but defaulted off) | 03:43 |
| bmace_ | i remember hitting an ethX related issue.. i recall working around it somehow but don't recall exactly how at this moment. | 03:44 |
| kfox1111 | in general, its a good improvement. | 03:44 |
| bmace_ | I don't have any eth1 / eth0 in my dev container, it is eno1, and genconfig still seems ok | 03:44 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Fixed config generation to check correct actions https://review.openstack.org/424421 | 03:45 |
| Jeffrey4l_ | bmace_, please check it ^^ | 03:45 |
| *** sp_ has joined #openstack-kolla | 03:45 | |
| *** sudipto has joined #openstack-kolla | 03:46 | |
| *** sudipto_ has joined #openstack-kolla | 03:46 | |
| Jeffrey4l_ | bmace_, the issue is: the container is not started up after run "reconfig" | 03:46 |
| bmace_ | Jeffrey4l_: it also seemed like the ordering was important in some places, like glance for example, where some of the "when" checks might actually cause failures when you were in "config" mode, so is it safe to just always put the config check first? | 03:47 |
| bmace_ | it ends up being nicely consistent if we can | 03:48 |
| *** mdnadeem has joined #openstack-kolla | 03:48 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for manila https://review.openstack.org/421365 | 03:48 |
| bmace_ | it is already in most places. there are only a couple of files i think where it isn't first in the when clause | 03:49 |
| Jeffrey4l_ | 1. action is always defined and not empty 2. in geconfig cause, just do not run the check container task and all handlers. | 03:49 |
| Jeffrey4l_ | so i think it is OK. | 03:49 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for manila https://review.openstack.org/421365 | 03:49 |
| sdake | bmace_ i just set network_interface in globals.yml to the name of the management interface | 03:50 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 03:50 |
| bmace_ | ugh, the way the dev container is set up right now getting the changes into there is a pain :/ | 03:51 |
| Jeffrey4l_ | bmace_, sorry? | 03:52 |
| bmace_ | not your fault Jeffrey :) | 03:52 |
| sdake | Jeffrey4l_ there is a patch portdirect and kfox1111 wrote | 03:53 |
| Jeffrey4l_ | ;). tell me whether the patch works. | 03:53 |
| Jeffrey4l_ | roger. | 03:53 |
| Jeffrey4l_ | after you tested it. | 03:53 |
| sdake | hey fellas, i gotta drive home - 20 mins drive - should be bck online | 03:53 |
| sdake | i'll check out the log for results | 03:53 |
| *** rhallisey has quit IRC | 03:54 | |
| *** jrobinson is now known as jrobinson-afk | 03:55 | |
| kfox1111 | bmace_: feel free to tweak it so that its less painful. | 03:58 |
| bmace_ | sdake: Jeffrey4l_: genconfig continues to work fine with Jeffreys additional changes. | 03:58 |
| *** sdake has quit IRC | 03:58 | |
| bmace_ | kfox1111: once it is all working i definitely have some tweaks i would like to continue to make to it, to better support running behind a proxy, using local kolla-ansible code, etc. | 03:59 |
| Jeffrey4l_ | bmace_, no container is running after run "genconfig", right? | 03:59 |
| bmace_ | Jeffrey4l_: that is correct. it couldn't even talk to a docker daemon if it wanted to, so definitely no containers running :) | 04:00 |
| Jeffrey4l_ | cool. we will fix all the under review patch. | 04:00 |
| kfox1111 | bmace_: cool. :) | 04:04 |
| kfox1111 | bmace_: would be cool if we could get a helm package into kubernetes/charts too. | 04:04 |
| kfox1111 | so all you have to do to get going is helm install stable/kollakube or something like that. | 04:04 |
| kfox1111 | gota go. l8r. | 04:05 |
| bmace_ | nite kfox1111 | 04:05 |
| jascott1 | kfox1111 for tomorrow https://github.com/kubernetes/helm/issues/1837 | 04:06 |
| jascott1 | have a good evening | 04:06 |
| *** MasterOfBugs has quit IRC | 04:34 | |
| sp_ | Jeffrey4l_: done necessary changes required after your and pbourk comments. Please check now https://review.openstack.org/#/c/422444/13 | 04:36 |
| *** sayantani01 has joined #openstack-kolla | 04:40 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: remove /var/log/rally in rally dockerfile https://review.openstack.org/424429 | 04:41 |
| *** zhubingbing has quit IRC | 04:42 | |
| *** jrobinson-afk is now known as jrobinson | 04:44 | |
| *** williamcaban has quit IRC | 05:00 | |
| *** adrian_otto has joined #openstack-kolla | 05:08 | |
| *** msimonin has joined #openstack-kolla | 05:08 | |
| *** goldyfruit has quit IRC | 05:09 | |
| Jeffrey4l_ | sp_, roger. sorry busy on better reconfigure bp. this should be done before m3 release, which is Jan 25. | 05:12 |
| sp_ | Jeffrey4l_: yes | 05:14 |
| Jeffrey4l_ | sp_, imo, test koll.summary == True is meaning less. | 05:14 |
| Jeffrey4l_ | it return a dict, you need test whether the return dict is your expected. | 05:14 |
| sp_ | Jeffrey4l_: for now I checked it and focused more coverage | 05:15 |
| Jeffrey4l_ | just increase coverage is useless. we need make sure the test is useful. | 05:15 |
| sp_ | Jeffrey4l_: ok, I will check | 05:16 |
| *** ssurana has joined #openstack-kolla | 05:16 | |
| *** ssurana has quit IRC | 05:16 | |
| sp_ | Jeffrey4l_: I will do this and will let you know | 05:17 |
| Jeffrey4l_ | thanks. | 05:17 |
| sp_ | Jeffrey4l_: thanks | 05:17 |
| *** adrian_otto1 has joined #openstack-kolla | 05:19 | |
| *** severion has joined #openstack-kolla | 05:22 | |
| *** adrian_otto has quit IRC | 05:23 | |
| *** zhangqiankun has quit IRC | 05:25 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for ceilometer https://review.openstack.org/423606 | 05:26 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for aodh https://review.openstack.org/423670 | 05:28 |
| *** williamcaban has joined #openstack-kolla | 05:29 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 05:30 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for rally https://review.openstack.org/421893 | 05:31 |
| *** cuongnv_ has joined #openstack-kolla | 05:32 | |
| *** cuongnv has quit IRC | 05:32 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for mistral https://review.openstack.org/423787 | 05:33 |
| *** williamcaban has quit IRC | 05:33 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for cloudkitty https://review.openstack.org/422516 | 05:34 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for gnocchi https://review.openstack.org/423133 | 05:35 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for trove https://review.openstack.org/423990 | 05:38 |
| *** adrian_otto1 has quit IRC | 05:41 | |
| *** adrian_otto has joined #openstack-kolla | 05:46 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for panko https://review.openstack.org/424164 | 05:46 |
| *** sudipto_ has quit IRC | 05:47 | |
| *** skramaja has joined #openstack-kolla | 05:48 | |
| *** sudipto has quit IRC | 05:49 | |
| openstackgerrit | Sajauddin proposed openstack/kolla: Kolla support for power server https://review.openstack.org/423239 | 05:51 |
| *** adrian_otto has quit IRC | 05:51 | |
| *** adrian_otto has joined #openstack-kolla | 05:52 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add the missing option into globals.yml https://review.openstack.org/424436 | 05:59 |
| *** trinaths has joined #openstack-kolla | 06:05 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Fix the panko_database_port for panko https://review.openstack.org/424442 | 06:10 |
| *** adrian_otto has quit IRC | 06:11 | |
| *** adrian_otto has joined #openstack-kolla | 06:12 | |
| *** satyar has joined #openstack-kolla | 06:16 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Fix a Syntax Error for panko bootstrap.yml https://review.openstack.org/424451 | 06:26 |
| satyar | Hi jeffrey4l_ | 06:27 |
| Jeffrey4l_ | sup satyar | 06:27 |
| *** sacharya_ has quit IRC | 06:28 | |
| satyar | i was looking at the bootstrap.yml for nova | 06:28 |
| satyar | saw few issues over there | 06:28 |
| satyar | i was trying to enable the nova_api_cell0 | 06:28 |
| satyar | and getting an error | 06:29 |
| Jeffrey4l_ | satyar, kolla do not support nova cell v2 now. | 06:29 |
| *** trinaths has quit IRC | 06:29 | |
| Jeffrey4l_ | a patch is trying to fix it. | 06:29 |
| Jeffrey4l_ | https://review.openstack.org/415472 | 06:29 |
| Jeffrey4l_ | but no time on it now. | 06:29 |
| satyar | This patch i know :) | 06:30 |
| *** prameswar has quit IRC | 06:30 | |
| satyar | i have done the necessary changes for cell0 | 06:30 |
| satyar | and the permissions | 06:30 |
| Jeffrey4l_ | satyar, cool. | 06:31 |
| Jeffrey4l_ | what error you got? | 06:31 |
| satyar | pasting | 06:31 |
| *** trinaths has joined #openstack-kolla | 06:32 | |
| Jeffrey4l_ | hrm. are u satroutr? aha, i think so :) | 06:33 |
| *** harlowja has joined #openstack-kolla | 06:34 | |
| *** trinaths1 has joined #openstack-kolla | 06:35 | |
| *** trinaths has quit IRC | 06:35 | |
| *** cuongnv_ is now known as cuongnv | 06:39 | |
| satyar | jeffrey4l_ http://paste.openstack.org/show/596183/ | 06:39 |
| Jeffrey4l_ | it says credential is not correct ;( | 06:40 |
| satyar | jeffrey4l_ http://paste.openstack.org/show/596184/ | 06:41 |
| Jeffrey4l_ | satyar, may nova changed the default cell0 database name. | 06:42 |
| Jeffrey4l_ | but no time to check this. will fix it after m3 is release. | 06:42 |
| satyar | hmm | 06:43 |
| satyar | let me check with nova | 06:43 |
| Jeffrey4l_ | busy on finish some bp ;) | 06:43 |
| satyar | :) | 06:44 |
| *** adrian_otto has quit IRC | 06:50 | |
| satyar | jeffrey4l_ https://review.openstack.org/#/c/420437/ | 07:01 |
| *** msimonin has quit IRC | 07:01 | |
| *** msimonin has joined #openstack-kolla | 07:01 | |
| Jeffrey4l_ | satyar, cool. thanks. | 07:01 |
| satyar | :) | 07:02 |
| *** Pavo has quit IRC | 07:02 | |
| *** Pavo has joined #openstack-kolla | 07:04 | |
| *** williamcaban has joined #openstack-kolla | 07:05 | |
| *** msimonin has quit IRC | 07:05 | |
| *** williamcaban has quit IRC | 07:09 | |
| *** harlowja has quit IRC | 07:26 | |
| *** sacharya has joined #openstack-kolla | 07:28 | |
| *** sayantani01 has quit IRC | 07:30 | |
| *** williamcaban has joined #openstack-kolla | 07:31 | |
| *** sacharya has quit IRC | 07:33 | |
| *** williamcaban has quit IRC | 07:36 | |
| *** zhubingbing has joined #openstack-kolla | 07:39 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for watcher https://review.openstack.org/424105 | 07:42 |
| *** prameswar has joined #openstack-kolla | 07:43 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Add fluentd role https://review.openstack.org/407392 | 07:45 |
| *** bmace_ has quit IRC | 07:47 | |
| *** bmace_ has joined #openstack-kolla | 07:47 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Fix undefine error for watcher bootstrap.yml https://review.openstack.org/424488 | 07:52 |
| *** ingmar23 has joined #openstack-kolla | 07:54 | |
| *** sacharya has joined #openstack-kolla | 07:59 | |
| *** openstackgerrit has quit IRC | 08:03 | |
| *** sacharya has quit IRC | 08:04 | |
| *** bjolo has joined #openstack-kolla | 08:06 | |
| *** openstackgerrit has joined #openstack-kolla | 08:07 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for sahara https://review.openstack.org/424165 | 08:07 |
| *** sayantani01 has joined #openstack-kolla | 08:07 | |
| openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Code coverage improvement for kolla https://review.openstack.org/422444 | 08:08 |
| sp_ | Jeffrey4l_: done | 08:10 |
| *** msimonin has joined #openstack-kolla | 08:11 | |
| *** sayantani01 has quit IRC | 08:11 | |
| *** saneax-_-|AFK is now known as saneax | 08:16 | |
| *** eaguilar has quit IRC | 08:19 | |
| *** eaguilar has joined #openstack-kolla | 08:20 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: change pankor to panko https://review.openstack.org/424499 | 08:20 |
| *** mgoddard has joined #openstack-kolla | 08:21 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimize reconfigure action for neutron https://review.openstack.org/424504 | 08:33 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for magnum https://review.openstack.org/424162 | 08:34 |
| *** egonzalez has joined #openstack-kolla | 08:42 | |
| *** pcaruana has joined #openstack-kolla | 08:45 | |
| openstackgerrit | Zeyu Zhu proposed openstack/kolla: Use required_if in kolla_docker.py https://review.openstack.org/424512 | 08:46 |
| *** zhubingbing has quit IRC | 08:51 | |
| openstackgerrit | Niklas Wik proposed openstack/kolla-ansible: Fixes package check when ansible_python_interpreter config is used in inventory file https://review.openstack.org/424515 | 08:52 |
| *** eaguilar has quit IRC | 08:52 | |
| *** zhubingbing has joined #openstack-kolla | 08:55 | |
| *** sayantani01 has joined #openstack-kolla | 09:01 | |
| *** eaguilar has joined #openstack-kolla | 09:01 | |
| openstackgerrit | Merged openstack/kolla-ansible: Fix a Syntax Error for panko bootstrap.yml https://review.openstack.org/424451 | 09:02 |
| openstackgerrit | Merged openstack/kolla-ansible: Fix undefine error for watcher bootstrap.yml https://review.openstack.org/424488 | 09:03 |
| openstackgerrit | Merged openstack/kolla-ansible: Fix the panko_database_port for panko https://review.openstack.org/424442 | 09:03 |
| *** sayantani01 has quit IRC | 09:05 | |
| *** Serlex has joined #openstack-kolla | 09:10 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add the missing option into globals.yml https://review.openstack.org/424436 | 09:11 |
| openstackgerrit | Niklas Wik proposed openstack/kolla-ansible: Support non standard Python path https://review.openstack.org/424515 | 09:20 |
| *** salkin has joined #openstack-kolla | 09:22 | |
| *** gfidente has joined #openstack-kolla | 09:23 | |
| *** gfidente has quit IRC | 09:23 | |
| *** gfidente has joined #openstack-kolla | 09:23 | |
| openstackgerrit | Cornelio Hopmann proposed openstack/kolla-ansible: Fix magnum [DEFAULT] configuration. https://review.openstack.org/422896 | 09:27 |
| *** jrobinson has quit IRC | 09:31 | |
| *** williamcaban has joined #openstack-kolla | 09:33 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add freezer ansible role https://review.openstack.org/402224 | 09:37 |
| *** williamcaban has quit IRC | 09:38 | |
| *** zhubingbing has quit IRC | 09:41 | |
| trinaths1 | hi | 09:43 |
| *** mgoddard has quit IRC | 09:43 | |
| trinaths1 | while bring up the setup, | 09:43 |
| trinaths1 | I was struck at | 09:43 |
| trinaths1 | TASK [kube-join : join workers to the initialized kubernetes cluster] ********** | 09:43 |
| trinaths1 | skipping: [kube1] | 09:43 |
| trinaths1 | and it was idle there from more than 5 hours. is there some thing wrong to decide ? | 09:44 |
| *** sambetts|afk is now known as sambetts | 09:47 | |
| *** williamcaban has joined #openstack-kolla | 09:50 | |
| *** zhubingbing has joined #openstack-kolla | 09:50 | |
| *** athomas has joined #openstack-kolla | 09:52 | |
| *** williamcaban has quit IRC | 09:54 | |
| *** sayantani01 has joined #openstack-kolla | 09:55 | |
| *** mdnadeem has quit IRC | 09:58 | |
| *** sayantani01 has quit IRC | 09:59 | |
| *** trinaths1 has quit IRC | 10:00 | |
| *** cuongnv has quit IRC | 10:03 | |
| *** tovin07_ has quit IRC | 10:04 | |
| *** satyar has quit IRC | 10:06 | |
| *** strigazi_AFK is now known as strigazi | 10:09 | |
| *** mgoddard has joined #openstack-kolla | 10:14 | |
| *** mdnadeem has joined #openstack-kolla | 10:15 | |
| *** eaguilar has quit IRC | 10:24 | |
| *** trinaths has joined #openstack-kolla | 10:24 | |
| *** eaguilar has joined #openstack-kolla | 10:26 | |
| *** haplo37 has quit IRC | 10:27 | |
| *** pbourke has quit IRC | 10:29 | |
| *** pbourke has joined #openstack-kolla | 10:29 | |
| *** trinaths has quit IRC | 10:30 | |
| *** trinaths has joined #openstack-kolla | 10:31 | |
| des_consolado | morning all | 10:34 |
| des_consolado | https://ask.openstack.org/en/question/101828/kolla-ansible-deploy-error-run-not-a-shared-mount/ anyone have any idea what I should do with this one? | 10:35 |
| *** bjolo has quit IRC | 10:35 | |
| *** ccesario has joined #openstack-kolla | 10:36 | |
| *** haplo37 has joined #openstack-kolla | 10:36 | |
| des_consolado | gah I think it might be some SELinux shit | 10:43 |
| des_consolado | Ah yes SELinux was still enabled on the target machine, silly me... | 10:48 |
| sp_ | egonzalez: pbourke It would be very helpful if review it. Some more changes done for it. https://review.openstack.org/#/c/422444/14 | 10:48 |
| *** shardy has joined #openstack-kolla | 11:05 | |
| *** bjolo has joined #openstack-kolla | 11:06 | |
| *** hfu has joined #openstack-kolla | 11:10 | |
| des_consolado | hm I'm still getting the problem actually. Destroyed everything on compute01, deleted kolla-toolbox image from my registry and re-built it, no SELinux anywhere at this point, rebooted, tried deploying again and still the same error D: | 11:17 |
| des_consolado | maybe I need to delete my /var/lib/docker | 11:19 |
| *** rhallisey has joined #openstack-kolla | 11:23 | |
| des_consolado | nah, issue persists... wtf... | 11:25 |
| mliima | morning guys | 11:42 |
| openstackgerrit | Merged openstack/kolla-ansible: Fix ironic inspector https://review.openstack.org/420122 | 11:42 |
| *** pd_ is now known as portdirect | 11:47 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Refactor register.yml files https://review.openstack.org/416046 | 11:50 |
| *** williamcaban has joined #openstack-kolla | 11:51 | |
| *** williamcaban has quit IRC | 11:56 | |
| *** rmart04 has joined #openstack-kolla | 11:57 | |
| *** athomas has quit IRC | 11:59 | |
| *** athomas has joined #openstack-kolla | 12:00 | |
| *** bjolo has quit IRC | 12:04 | |
| *** eaguilar has quit IRC | 12:08 | |
| *** simon-AS559 has joined #openstack-kolla | 12:09 | |
| *** zioproto has joined #openstack-kolla | 12:10 | |
| zioproto | hello all | 12:10 |
| simon-AS559 | Hi there :-) | 12:11 |
| *** shardy is now known as shardy_lunch | 12:11 | |
| portdirect | o/ | 12:11 |
| simon-AS559 | zioproto and I have started playing with Kolla in view of deploying it in production for our multi-region IaaS (SWITCHengines). | 12:12 |
| *** sdake has joined #openstack-kolla | 12:12 | |
| portdirect | swiss invasion :) | 12:12 |
| portdirect | I'm work on the kolla-k8s side of things, but there are loads of kolla-ansible peeps here - though this channel is fairly quite until the United States wakes up | 12:13 |
| portdirect | hows it going for you? | 12:14 |
| zioproto | just starting, we are on day 2 of the project | 12:14 |
| simon-AS559 | So-so… I'm trying to deploy the multi-node installation (only controller nodes) on a set of 3 OpenStack instances, and of course the HAProxy/VIP stuff breaks. I'll recreate those instances in a private network with port-security off. | 12:14 |
| simon-AS559 | In production we won't have *that* problem, but another (even worse) one: Our network is pure Layer-3 with no shared subnet between servers. | 12:15 |
| simon-AS559 | So for the VIP stuff we'll probably need to write something to inject those addresses using a routing protocol (we use OSPF/OSPFv3) rather than the usual L2/ARP way. | 12:16 |
| *** sdake_ has joined #openstack-kolla | 12:16 | |
| *** jtriley has quit IRC | 12:17 | |
| sdake_ | sup folks | 12:17 |
| sdake_ | hey portdirect | 12:17 |
| portdirect | hey sdake_ o/ | 12:17 |
| *** sdake has quit IRC | 12:17 | |
| zioproto | anyone coming to the milano OPS-midcycle ? | 12:18 |
| * portdirect laughs at sdakes morning vpn ritual... | 12:18 | |
| sdake_ | zioproto up in the air for me personally | 12:18 |
| sdake_ | portdirect ya i know rediculou | 12:18 |
| sdake_ | someone sid I can use an irc bouncer | 12:18 |
| sdake_ | i'm going to investigat that | 12:18 |
| sdake_ | although I prefer my irc go over TLS | 12:19 |
| *** jtriley has joined #openstack-kolla | 12:19 | |
| sdake_ | ZNC is the tool | 12:19 |
| sdake_ | I really like textual (mac irc client) | 12:19 |
| portdirect | zioproto: not sure yet - going to try and make it but have a lot on that month :/ | 12:19 |
| sdake_ | with znc, you connect to a specific machine | 12:19 |
| zioproto | I am probably gonna be there | 12:19 |
| zioproto | with a lot a of questions | 12:20 |
| zioproto | so I am going back to my inventory files :) | 12:20 |
| portdirect | sdake_: I just bit the bullet and went for irccloud.com - I highly recommend it | 12:20 |
| sdake_ | its a shame the midops midcycle isn't colocated with the ptg or visaversa | 12:20 |
| sdake_ | ya i guess i can check it out - i just really like textual | 12:21 |
| sdake_ | portdirect i've been using irc since i was 15 | 12:21 |
| sdake_ | 42 now | 12:21 |
| sdake_ | so do the math - and this is the first irc cleint that doesnt' suck :) | 12:21 |
| *** saneax is now known as saneax-_-|AFK | 12:22 | |
| sdake_ | does irccloud have a real client or is it entirely web bsaed? | 12:22 |
| portdirect | its pretty much entirely web based - so requires some mental buy in for sure | 12:24 |
| *** jtriley has quit IRC | 12:24 | |
| portdirect | there are a few client apps - bet they are just thin wrappers for the web UI | 12:25 |
| portdirect | *but | 12:25 |
| *** trinaths1 has joined #openstack-kolla | 12:27 | |
| portdirect | took me some convincing - used hexchat for ages - but I'm pretty sold on it now. The upside is always on and the mobile app is awesome (push notifications). The primary downside for me is pretty similar, always on and push notifications... :) | 12:27 |
| sdake_ | the problem with web apps is I often close chrome to reset it | 12:28 |
| sdake_ | because I end up opening 100 tabs or something | 12:28 |
| sdake_ | and i'm not joking with teh 100tabs | 12:28 |
| sdake_ | then to clean it all out i just force close chrome | 12:28 |
| portdirect | most of the channels I'm on don't have logging - so the always on aspect is a big bonus. and I'm too lazy to set up my own bouncer. | 12:28 |
| sdake_ | i'm going the bouncer route if i can find a machine to run it on that isn't a fortune | 12:29 |
| sdake_ | rackspce clouds are super expensive | 12:29 |
| sdake_ | i suspect there is irc bouncer as a service | 12:29 |
| portdirect | i hear you with the chrome tabs thing | 12:29 |
| sdake_ | at one point i was paying 90$/mo for 1 rackspace vm to host a mailing list | 12:29 |
| sdake_ | so i cleared it out and my life is alot better :) | 12:30 |
| *** hfu has quit IRC | 12:30 | |
| portdirect | why not run a bouncer out of your house? | 12:30 |
| *** trinaths has quit IRC | 12:30 | |
| *** hfu has joined #openstack-kolla | 12:30 | |
| *** hfu has quit IRC | 12:31 | |
| sdake_ | because sometimes i am connected to a vpn | 12:31 |
| sdake_ | and the vpn doesn't allow split routing | 12:31 |
| sdake_ | because of rediculous policies | 12:31 |
| sdake_ | i can't even print to my home printers when connected to the vpn | 12:31 |
| *** hfu has joined #openstack-kolla | 12:31 | |
| sdake_ | everyone at cisco compalins about it | 12:31 |
| *** hfu has quit IRC | 12:31 | |
| sdake_ | especially those people that are 100% telecommute | 12:32 |
| *** hfu has joined #openstack-kolla | 12:32 | |
| *** hfu has quit IRC | 12:32 | |
| openstackgerrit | Merged openstack/kolla-ansible: Add the missing option into globals.yml https://review.openstack.org/424436 | 12:40 |
| openstackgerrit | Merged openstack/kolla-ansible: Fix magnum [DEFAULT] configuration. https://review.openstack.org/422896 | 12:41 |
| *** williamcaban has joined #openstack-kolla | 12:42 | |
| *** bjolo has joined #openstack-kolla | 12:43 | |
| *** williamcaban has quit IRC | 12:47 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Optimize reconfigure action for nova https://review.openstack.org/407416 | 12:47 |
| *** rhallisey has quit IRC | 12:47 | |
| *** rhallisey has joined #openstack-kolla | 12:48 | |
| *** sdake_ is now known as sdake | 12:49 | |
| zioproto | dumb questions | 12:49 |
| zioproto | looking at this page | 12:49 |
| zioproto | http://docs.openstack.org/developer/kolla/multinode.html | 12:49 |
| zioproto | it is not exactly clear what is tools/start-registry | 12:49 |
| zioproto | also | 12:49 |
| *** zhubingbing has joined #openstack-kolla | 12:49 | |
| zhubingbing | hello guys | 12:50 |
| zioproto | if I have to install kolla for production, given that there are no ubuntu packages, what is the right installation procedure ? with pip ? or I have to git clone stuff and checkout a stable branch ? | 12:50 |
| *** krtaylor has quit IRC | 12:51 | |
| *** shardy_lunch is now known as shardy | 12:52 | |
| mliima | hello zhubingbing | 12:53 |
| *** schwicht has joined #openstack-kolla | 12:53 | |
| zhubingbing | hi mliima | 12:53 |
| openstackgerrit | Merged openstack/kolla-ansible: Add fluentd role https://review.openstack.org/407392 | 12:54 |
| *** schwicht has quit IRC | 12:55 | |
| berendt | my issue with packaging 16.7 was solved by https://github.com/openstack/requirements/commit/6e673baf41b4d5c1f247357683f6d4974436fa08 | 12:56 |
| berendt | packaging upper constraint is now 16.8 | 12:56 |
| berendt | zioproto: git clone and checkout stable branch | 12:57 |
| zioproto | berendt: ok, but for kolla-ansible there is only master | 12:57 |
| *** huzhengchuan has quit IRC | 12:57 | |
| berendt | zioproto: kolla-ansible was created after the current stable branch | 12:57 |
| berendt | zioproto: when working with stable/newton (3.0.2 release) than everything is included in kolla repository | 12:58 |
| berendt | zioproto: after newton release we moved the ansible roles into the kolla-ansible repository | 12:58 |
| zioproto | okay ! | 12:58 |
| zioproto | so for stable/newton | 12:58 |
| berendt | zioproto: because of that there is not yet a stable branch in kolla-ansible | 12:58 |
| zioproto | I just need https://github.com/openstack/kolla | 12:59 |
| zioproto | and I just need to clone this repo on the operator node | 12:59 |
| zioproto | also, I should let Kolla manage the installation of Docker on all nodes, right ? | 12:59 |
| zioproto | I see that Kolla pins docker to 1.11 | 12:59 |
| zioproto | working on Ubuntu xenial there is any apt repo that I should add ? | 12:59 |
| zioproto | or Kolla with his playbooks knows how to install the right version of Docker on xenial ? | 13:00 |
| berendt | zioproto: yes, checkout kolla repository on the seed/operator node and install it there | 13:00 |
| berendt | zioproto: i do not know how to install docker with kolla on the other nodes because i use a custom ansible role for this part of the deployment | 13:00 |
| berendt | zioproto: but i think the baremetal role will handle this | 13:01 |
| zioproto | I tried already the 1 node installation and | 13:01 |
| zioproto | kolla-ansible -i <<inventory file>> bootstrap-servers | 13:01 |
| zioproto | this playbook | 13:01 |
| zioproto | will install Docker | 13:01 |
| berendt | https://github.com/openstack/kolla/blob/stable/newton/ansible/roles/baremetal/tasks/pre-install.yml#L86-L91 | 13:01 |
| berendt | yes, bootstrap-servers will run the baremetal role to bootstrap everything | 13:02 |
| berendt | and this role adds the required docker repository | 13:02 |
| zioproto | deb https://apt.dockerproject.org/repo ubuntu-xenial main | 13:04 |
| zioproto | great | 13:04 |
| zioproto | is the same one I already use | 13:04 |
| openstackgerrit | harsh proposed openstack/kolla-ansible: Resolves parsing issue with kolla-ansible CLI https://review.openstack.org/424651 | 13:04 |
| zioproto | for the docker stuff | 13:04 |
| zioproto | ok, so I move on | 13:04 |
| zioproto | thanks for clarification | 13:04 |
| *** krtaylor has joined #openstack-kolla | 13:04 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Add zun ansible role https://review.openstack.org/416179 | 13:05 |
| *** apuimedo|flu is now known as apuimedo | 13:06 | |
| openstackgerrit | Merged openstack/kolla-ansible: Allow cinder-volume to be configured to use HNAS nfs https://review.openstack.org/417094 | 13:08 |
| *** huzhengchuan has joined #openstack-kolla | 13:10 | |
| *** jtriley has joined #openstack-kolla | 13:13 | |
| *** williamcaban has joined #openstack-kolla | 13:14 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Split Openvswitch into own role https://review.openstack.org/416793 | 13:15 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Add OpenDaylight role https://review.openstack.org/416367 | 13:15 |
| *** williamcaban has quit IRC | 13:18 | |
| *** jtriley has quit IRC | 13:18 | |
| *** dave-mccowan has joined #openstack-kolla | 13:18 | |
| *** jtriley has joined #openstack-kolla | 13:19 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for panko https://review.openstack.org/424164 | 13:21 |
| Jeffrey4l_ | zhubingbing, mliima please review this https://review.openstack.org/#/c/424421/ | 13:22 |
| zhubingbing | ok | 13:22 |
| mliima | done Jeffrey4l_ | 13:22 |
| Jeffrey4l_ | thanks. | 13:22 |
| zhubingbing | done | 13:24 |
| *** jtriley has quit IRC | 13:24 | |
| Jeffrey4l_ | thanks. | 13:24 |
| zhubingbing | thanks your commit | 13:24 |
| zhubingbing | ;) | 13:24 |
| *** huzhengchuan has quit IRC | 13:25 | |
| Jeffrey4l_ | now need add - action != "config" in all handlers task. If not, it will break k8s genconfig. so you may need update your patch about reconfigure mliima zhubingbing | 13:25 |
| zhubingbing | emm | 13:25 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for magnum https://review.openstack.org/424162 | 13:25 |
| mliima | ok Jeffrey4l_ | 13:25 |
| zhubingbing | ok | 13:25 |
| mliima | caoyuan is helping :) | 13:25 |
| mliima | he is a commit machine haha | 13:26 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Split Openvswitch into own role https://review.openstack.org/416793 | 13:28 |
| openstackgerrit | Merged openstack/kolla-ansible: Fixed config generation to check correct actions https://review.openstack.org/424421 | 13:29 |
| *** williamcaban has joined #openstack-kolla | 13:29 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: install murano dashboard package in horizon https://review.openstack.org/395957 | 13:31 |
| *** williamcaban has quit IRC | 13:33 | |
| *** jmontleon has quit IRC | 13:38 | |
| Jeffrey4l_ | mliima, egonzalez could u review this one? https://review.openstack.org/424504 | 13:39 |
| *** severion has quit IRC | 13:39 | |
| *** v1k0d3n has quit IRC | 13:39 | |
| *** mdnadeem has quit IRC | 13:39 | |
| *** jmontleon has joined #openstack-kolla | 13:40 | |
| *** v1k0d3n has joined #openstack-kolla | 13:40 | |
| mliima | Jeffrey4l_, any notify in Waiting the openvswitch_db service to be ready? | 13:41 |
| mliima | https://review.openstack.org/#/c/424504/1/ansible/roles/neutron/handlers/main.yml | 13:41 |
| Jeffrey4l_ | mliima, yes. check the first task in handlers/main.yml. | 13:42 |
| Jeffrey4l_ | one handler trigger another one. ;) | 13:42 |
| *** Jeffrey4l_ is now known as Jeffrey4l | 13:42 | |
| mliima | oh, hehe | 13:46 |
| mliima | Ok Jeffrey4l | 13:46 |
| mliima | Jeffrey4l, how I create a softlink? | 13:47 |
| mliima | in reconfigure file | 13:47 |
| mliima | :/ | 13:47 |
| Jeffrey4l | the really concern i care is: host_in_groups variable in neutron/default/main.yml. it is a tricker. | 13:47 |
| Jeffrey4l | mliima, include is fine. | 13:47 |
| mliima | ok :) | 13:47 |
| Jeffrey4l | if u want to create softlink: just use "ln -s deploy.yml reconfigure.yml" | 13:48 |
| *** jmontleon has quit IRC | 13:48 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for designate https://review.openstack.org/424155 | 13:50 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for manila https://review.openstack.org/421365 | 13:50 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for influxdb https://review.openstack.org/424159 | 13:51 |
| *** ccesario has quit IRC | 13:51 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for etcd https://review.openstack.org/424157 | 13:51 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for gnocchi https://review.openstack.org/423133 | 13:52 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for grafana https://review.openstack.org/424158 | 13:52 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for mistral https://review.openstack.org/423787 | 13:52 |
| *** saneax-_-|AFK is now known as saneax | 13:52 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for watcher https://review.openstack.org/424105 | 13:53 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for kuryr https://review.openstack.org/424161 | 13:53 |
| *** jmontleon has joined #openstack-kolla | 13:53 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for sahara https://review.openstack.org/424165 | 13:53 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for karbor https://review.openstack.org/424160 | 13:54 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for solum https://review.openstack.org/424167 | 13:54 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for aodh https://review.openstack.org/423670 | 13:55 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for elasticsearch https://review.openstack.org/424139 | 13:55 |
| *** sp___ has joined #openstack-kolla | 13:55 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for collectd https://review.openstack.org/424133 | 13:56 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for barbican https://review.openstack.org/423728 | 13:56 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for rally https://review.openstack.org/421893 | 13:56 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfiguration for ceilometer https://review.openstack.org/423606 | 13:57 |
| openstackgerrit | harsh proposed openstack/kolla-ansible: Resolves parsing issue with kolla-ansible CLI https://review.openstack.org/424651 | 13:58 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible: Optimize reconfigure action for cloudkitty https://review.openstack.org/422516 | 13:58 |
| *** lamt has joined #openstack-kolla | 14:01 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfiguration for octavia https://review.openstack.org/423846 | 14:02 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfigure action for trove https://review.openstack.org/423990 | 14:04 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla-ansible: Optimize reconfigure action for congress https://review.openstack.org/422690 | 14:05 |
| *** huzhengchuan has joined #openstack-kolla | 14:06 | |
| *** sp_ has quit IRC | 14:06 | |
| *** v1k0d3n has quit IRC | 14:08 | |
| *** v1k0d3n has joined #openstack-kolla | 14:09 | |
| *** williamcaban has joined #openstack-kolla | 14:10 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Add freezer ansible role https://review.openstack.org/402224 | 14:10 |
| *** matrohon has joined #openstack-kolla | 14:12 | |
| *** rmart04 has quit IRC | 14:12 | |
| *** eaguilar has joined #openstack-kolla | 14:13 | |
| *** williamcaban has quit IRC | 14:14 | |
| *** goldyfruit has joined #openstack-kolla | 14:15 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Add freezer ansible role https://review.openstack.org/402224 | 14:16 |
| zioproto | hey I am confused about this tools/start-registry | 14:17 |
| zioproto | installing stable kolla with pip | 14:17 |
| zioproto | where I am supposed to find this script ? | 14:17 |
| *** krtaylor has quit IRC | 14:18 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible: Add freezer ansible role https://review.openstack.org/402224 | 14:18 |
| pbourke | zioproto: its in the kolla repo | 14:18 |
| *** sp_ has joined #openstack-kolla | 14:18 | |
| *** srwilkers has joined #openstack-kolla | 14:19 | |
| *** srwilkers has quit IRC | 14:19 | |
| *** srwilkers has joined #openstack-kolla | 14:20 | |
| *** simon-AS559 has quit IRC | 14:20 | |
| srwilkers | morning | 14:21 |
| *** schwicht has joined #openstack-kolla | 14:21 | |
| zioproto | pbourke: but if I install with pip should I also anyway clone the repo ? | 14:22 |
| *** williamcaban has joined #openstack-kolla | 14:23 | |
| *** flaper87 has quit IRC | 14:24 | |
| openstackgerrit | Merged openstack/kolla-kubernetes: Fix old container https://review.openstack.org/424422 | 14:24 |
| zioproto | pbourke: ok I guess I was using too new docs. Now I will stick to http://docs.openstack.org/developer/kolla/newton/quickstart.html | 14:24 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Openvswitch Service package with kuberentes entrypoint https://review.openstack.org/424182 | 14:25 |
| *** gema has joined #openstack-kolla | 14:26 | |
| gema | berendt: ping | 14:26 |
| *** tovin07_ has joined #openstack-kolla | 14:26 | |
| berendt | gema: pong | 14:26 |
| gema | berendt: I am reading your email about debian deprecation and wanted to ask about this | 14:27 |
| berendt | gema: feel free.. :) | 14:27 |
| gema | berendt: what are the supported os when you deprecate debian? | 14:27 |
| berendt | gema: we actually support ubuntu and centos | 14:27 |
| gema | ok | 14:27 |
| berendt | last cycle we dropped the fedora images, because nobody worked on it | 14:28 |
| gema | the reason I am asking is because I am gearing my team towards starting to contribute images for ARM64 | 14:28 |
| berendt | this is the same for the debian images | 14:28 |
| gema | and they are most likely going to be debian b ased | 14:28 |
| openstackgerrit | Merged openstack/kolla-ansible: Optimize reconfiguration for magnum https://review.openstack.org/424162 | 14:28 |
| berendt | the problem is that actually nobody is working/using the debian images | 14:28 |
| *** satyar has joined #openstack-kolla | 14:28 | |
| berendt | is it not possible to use ubuntu based arm images? | 14:28 |
| gema | well, we are not supposed to be favoring any vendor | 14:29 |
| gema | that's why we like debian | 14:29 |
| openstackgerrit | Merged openstack/kolla-ansible: Optimize reconfiguration for panko https://review.openstack.org/424164 | 14:29 |
| gema | our work has to do with enablement | 14:29 |
| gema | rather than being vendor specific | 14:29 |
| *** tovin07_ has quit IRC | 14:29 | |
| berendt | the problem is that the current debian images need a lot of love and gate jobs | 14:29 |
| berendt | are you willing to work on them? | 14:30 |
| *** breitz has quit IRC | 14:30 | |
| gema | yes we will have to for ARM | 14:30 |
| *** mattmceuen has joined #openstack-kolla | 14:30 | |
| gema | whether we'll be able to do the same for x86 is a question mark | 14:30 |
| gema | or whether the work we do is also usable there | 14:30 |
| berendt | i think it is a good idea to discuss this in our team meeting tomorrow | 14:30 |
| gema | I have hardware to set up 3rd party CI for ARM | 14:31 |
| gema | but I don't have hardware for x86 | 14:31 |
| berendt | can you join? then i will add it to the agenda | 14:31 |
| gema | sure , I will be there | 14:31 |
| *** krtaylor has joined #openstack-kolla | 14:31 | |
| gema | I try to be active but normally get caught between the interop meeting and yours | 14:31 |
| gema | tomorrow I will be mostly in yours :D | 14:31 |
| gema | berendt: thanks | 14:31 |
| berendt | i will add 2 topics: new architecture and debian iamges | 14:32 |
| berendt | will you join the ptg in feb? | 14:32 |
| berendt | gema: can you please reply on the deprecation vote on openstack-dev and wrote down a summary there? | 14:34 |
| berendt | i think this makes sense for the logs | 14:34 |
| gema | berendt: sure will do | 14:35 |
| gema | berendt: wanted to check with you first so that it doesn't come across as this strange person trying to stop this decision :) | 14:35 |
| gema | berendt: I will also be at the ptg in feb, but again I have the conflict with interop WG | 14:36 |
| berendt | for me it is fine to keep the debian images as long as anybody is working with them | 14:36 |
| gema | berendt: so I will have to work out at what times I can be where | 14:36 |
| berendt | yes but maybe we can have a quick floor discussion about this topic there | 14:36 |
| gema | berendt: yes, I understand taht | 14:36 |
| gema | sounds like a plan | 14:36 |
| gema | will reply to your email | 14:36 |
| *** krtaylor has quit IRC | 14:37 | |
| *** tovin07_ has joined #openstack-kolla | 14:38 | |
| *** jtriley has joined #openstack-kolla | 14:38 | |
| *** tovin07_ has quit IRC | 14:38 | |
| *** v1k0d3n has quit IRC | 14:39 | |
| *** severion has joined #openstack-kolla | 14:39 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Integrate designate with neutron https://review.openstack.org/424679 | 14:43 |
| *** lrensing has joined #openstack-kolla | 14:44 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Integrate designate with neutron https://review.openstack.org/424679 | 14:46 |
| *** rmart04 has joined #openstack-kolla | 14:50 | |
| *** tovin07_ has joined #openstack-kolla | 14:50 | |
| obre | :q | 14:50 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Custom init-runonce values https://review.openstack.org/424682 | 14:52 |
| *** severion has quit IRC | 14:52 | |
| *** v1k0d3n has joined #openstack-kolla | 14:52 | |
| *** lamt has quit IRC | 14:52 | |
| openstackgerrit | Merged openstack/kolla: remove /var/log/rally in rally dockerfile https://review.openstack.org/424429 | 14:53 |
| mliima | zhubingbing, egonzalez https://review.openstack.org/#/c/416046/ | 14:53 |
| *** rmart04 has quit IRC | 14:57 | |
| *** krtaylor has joined #openstack-kolla | 14:58 | |
| *** rmart04 has joined #openstack-kolla | 14:58 | |
| *** zhubingbing has quit IRC | 14:58 | |
| *** tovin07_ has quit IRC | 15:01 | |
| openstackgerrit | Merged openstack/kolla: install murano dashboard package in horizon https://review.openstack.org/395957 | 15:04 |
| *** tovin07_ has joined #openstack-kolla | 15:04 | |
| *** fguillot has joined #openstack-kolla | 15:08 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Fix security group creation https://review.openstack.org/424690 | 15:09 |
| *** mgiles has joined #openstack-kolla | 15:11 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Optimize reconfiguration for tacker https://review.openstack.org/422036 | 15:12 |
| sdake | berendt uses the term support loosly there gema :) | 15:16 |
| sdake | berendt the other reason we dropped fedor aimages is because fedora couldn't be madee to work (it was technically impossible as RDO stopped maintenance for fedora) | 15:16 |
| *** rmart04 has quit IRC | 15:16 | |
| *** rmart04 has joined #openstack-kolla | 15:17 | |
| sdake | gema - agree with berendt's conclusion - we don't have it out for debian - instead nobody has stepped up to do any work on them | 15:18 |
| sdake | gema fwiw debian has *never* worked with kolla | 15:18 |
| sdake | gema so your in for a bit of pain making it work :) | 15:19 |
| *** skramaja has quit IRC | 15:19 | |
| sdake | gema if you c an get the images to work its easy enough to add x86_64 gate jobs for debian | 15:19 |
| sdake | gema probably 4-8 hours of work | 15:19 |
| gema | sdake: I am enlisting some expert help | 15:19 |
| gema | from debian :D | 15:19 |
| sdake | although I am 4-8 hours of work out of blocks :) | 15:19 |
| sdake | berendt this exact situation is why we have the deprecatoin policy in place that we do :) | 15:20 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Add cinder control package with kubernetes entrypoint https://review.openstack.org/424697 | 15:22 |
| gema | sdake: 4-8 hours of work per container or in total? | 15:22 |
| gema | sdake: you are making it sound almost too easy | 15:22 |
| gema | sdake: it will be harder to get the lab properly configured and up and running | 15:23 |
| *** msimonin has quit IRC | 15:23 | |
| gema | :D | 15:23 |
| sdake | gema for the gate for someone experienced in gate work | 15:24 |
| sdake | the containers themselves will be a bigger challenge | 15:24 |
| sdake | gema i'd recommend bringup on x86-64 first then arm second | 15:24 |
| sdake | but its your choice as to the order | 15:24 |
| sdake | note kolla does not support nor implement i386 | 15:24 |
| gema | sdake: ack | 15:24 |
| sdake | the reason i recommend that order is x86_64 is in good shape | 15:25 |
| sdake | although maybe arm is too | 15:25 |
| sdake | nicolas pitre (a dude I used to work with at montavista) did a great job of beating the arm kernel into submission years ago | 15:25 |
| sdake | not sure if he is still doing that or not | 15:25 |
| sdake | gema i'd focus on getting it to work locally - then adding a gate job for x86-64, then adding third party ci if you want to do that | 15:26 |
| sdake | i assume debian compiles from the same sources for x86-64 and arm64 | 15:26 |
| gema | sdake: yes | 15:26 |
| gema | sdake: although some packages are called slightly different | 15:27 |
| sdake | if they are named differnetly - that is where the challenge will be | 15:27 |
| sdake | in sorting out which packages to install in which conditions | 15:27 |
| sdake | kolla at this point isn't multiarch | 15:27 |
| sdake | never has been | 15:27 |
| sdake | although i dont' see harm in it if someone is willing to maintain it | 15:28 |
| sdake | we will need to come up with a way to tag images properly for the arm case as well | 15:28 |
| sdake | right now we assume x86_64 for everything | 15:28 |
| sdake | so changes to build.py will be required | 15:28 |
| sdake | gema after the team discussion on wed (which I will miss, I have a conflicting appointment) you will probably want to file a blueprint to track the work items related to this | 15:29 |
| gema | sdake: ok, will do | 15:32 |
| *** huzhengchuan has quit IRC | 15:34 | |
| *** sayantani01 has joined #openstack-kolla | 15:35 | |
| *** sayantani01 has quit IRC | 15:35 | |
| *** aj_power has joined #openstack-kolla | 15:36 | |
| *** sayantani01 has joined #openstack-kolla | 15:36 | |
| *** tonanhngo has joined #openstack-kolla | 15:40 | |
| *** trinaths1 has quit IRC | 15:45 | |
| *** sacharya has joined #openstack-kolla | 15:46 | |
| *** adrian_otto has joined #openstack-kolla | 15:48 | |
| *** hieulq_ has joined #openstack-kolla | 15:51 | |
| *** breitz has joined #openstack-kolla | 15:51 | |
| *** schwicht has quit IRC | 15:57 | |
| *** simon-AS559 has joined #openstack-kolla | 16:00 | |
| *** srwilkers has quit IRC | 16:01 | |
| *** schwicht has joined #openstack-kolla | 16:01 | |
| *** sp_ has quit IRC | 16:02 | |
| *** srwilkers has joined #openstack-kolla | 16:04 | |
| *** sayantani01 has quit IRC | 16:05 | |
| *** sayantani01 has joined #openstack-kolla | 16:06 | |
| *** schwicht_at_work has joined #openstack-kolla | 16:06 | |
| *** schwicht has quit IRC | 16:06 | |
| *** breitz has quit IRC | 16:07 | |
| *** breitz has joined #openstack-kolla | 16:08 | |
| *** rmart04 has quit IRC | 16:11 | |
| *** pcaruana has quit IRC | 16:16 | |
| *** sp_ has joined #openstack-kolla | 16:16 | |
| *** lamt has joined #openstack-kolla | 16:20 | |
| *** TxGirlGeek has joined #openstack-kolla | 16:29 | |
| *** TxGirlGeek has quit IRC | 16:32 | |
| *** TxGirlGeek has joined #openstack-kolla | 16:32 | |
| *** bjolo has quit IRC | 16:33 | |
| *** inc0 has joined #openstack-kolla | 16:33 | |
| *** eaguilar has quit IRC | 16:33 | |
| openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Code coverage improvement for kolla https://review.openstack.org/422444 | 16:34 |
| *** prameswar has quit IRC | 16:34 | |
| inc0 | another day of meetings \\o o// | 16:35 |
| inc0 | \o/ -o- /o\ | 16:35 |
| *** tovin07_ has quit IRC | 16:36 | |
| kfox1111 | morning. | 16:37 |
| srwilkers | inc0, such is my life today | 16:37 |
| inc0 | how are y'all doing? | 16:37 |
| kfox1111 | alive. you? :) | 16:38 |
| pbourke | inc0: would like to propose a vote on ML that we gut the kolla docs to only info related to building images | 16:39 |
| pbourke | inc0: and keep the current docs under kolla-ansible for improvement | 16:40 |
| pbourke | inc0: right now there's two copies of the same thing and its getting really confusing | 16:40 |
| kfox1111 | pbourke: +1 | 16:40 |
| openstackgerrit | Merged openstack/kolla-ansible: Ansible config for nova-serialproxy console https://review.openstack.org/415870 | 16:40 |
| kfox1111 | well, +1 to the vote, and +1 my vote. :) | 16:41 |
| pbourke | I know there are some people looking at improving the docs in general so will give them a day or so to chime in. But I think this can be done now and will help everyone | 16:41 |
| kfox1111 | it will just get more confusing too, as kolla-kubernetes matures. | 16:41 |
| kfox1111 | arg.... really need to get this local repo ps finished so we can get local repo's going for the other versions. :/ | 16:44 |
| *** aj_power has quit IRC | 16:45 | |
| *** saneax is now known as saneax-_-|AFK | 16:46 | |
| kfox1111 | hmm... interesting problem... | 16:46 |
| kfox1111 | its failing because the keystone-admin service is internal only. | 16:47 |
| *** schwicht_at_work has quit IRC | 16:47 | |
| *** schwicht has joined #openstack-kolla | 16:47 | |
| kfox1111 | we need a way to make the service external conditionally. | 16:47 |
| *** saneax-_-|AFK is now known as saneax | 16:47 | |
| *** dmsimard is now known as dmsimard|off | 16:48 | |
| *** klindgren has quit IRC | 16:48 | |
| berendt | pbourke: can you add the docs topic to the agenda for tomorrow? | 16:49 |
| pbourke | berendt: ok | 16:49 |
| openstackgerrit | Merged openstack/kolla-ansible: Optimize reconfigure action for rally https://review.openstack.org/421893 | 16:50 |
| inc0 | pbourke, I don't think we need voting | 16:54 |
| pbourke | inc0: fair enough, if the general consenus is positive in the meeting I'll do up a patch | 16:55 |
| inc0 | there is no reason for keeping deploy guide in kolla other than noone wrote good guide that will allow people to navigate across deliverables | 16:55 |
| sbezverk | kfox1111: ping | 16:55 |
| inc0 | that's my requirement <- people being able to find kolla-ansible from kolla docs | 16:55 |
| inc0 | and kolla-k8s | 16:55 |
| pbourke | can we not just link to kolla-[ansible,k8s,*] from the kolla docs? | 16:55 |
| *** sudipto has joined #openstack-kolla | 16:55 | |
| inc0 | maybe | 16:55 |
| *** sudipto_ has joined #openstack-kolla | 16:56 | |
| berendt | i think it makes sense to have a central landing page for all documents and do maintain the documents itself in each repository | 16:56 |
| inc0 | I dunno, I'm sure we can figure that one out | 16:56 |
| inc0 | berendt, agree, or architectural guide which will also explain use-cases, maturity and such of each deliverable | 16:56 |
| pbourke | tbh, I dont want a massive discussion on our final vision for where the docs need to be | 16:56 |
| pbourke | I just want to get rid of the duplication | 16:56 |
| pbourke | because right now its horrible | 16:56 |
| inc0 | +2 pbourke | 16:57 |
| berendt | confirmed | 16:57 |
| inc0 | btw, we got consensus on how to add new deliverable | 16:57 |
| pbourke | oh? | 16:57 |
| inc0 | 2*+2 + PTL vote | 16:57 |
| kfox1111 | sbezverk: ping. | 16:57 |
| berendt | great | 16:57 |
| inc0 | I just closed poll | 16:57 |
| inc0 | yeah, I hoped this will be the case | 16:57 |
| pbourke | isn't there a wider discussion in that area though | 16:57 |
| *** egonzalez has quit IRC | 16:57 | |
| inc0 | 12 out of 17 people voted;) | 16:57 |
| berendt | do you send the result to the mailing list? | 16:58 |
| pbourke | about whether they need to be under the kolla umbrella | 16:58 |
| *** QuentinM has quit IRC | 16:58 | |
| *** zioproto has quit IRC | 16:58 | |
| sbezverk | kfox1111: I think I captured a log confirming that entrypoint is not checking if there is actual process behind a service | 16:58 |
| inc0 | pbourke, we'll deal with it later | 16:58 |
| kfox1111 | ugg. :/ | 16:58 |
| pbourke | inc0: check | 16:58 |
| berendt | inc0: can you see who has not voted? maybe it makes sense to ask them why | 16:58 |
| inc0 | issue I'm having with making full breakup is mess it's going to cause | 16:58 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla: Use install-pip macro in most source images https://review.openstack.org/413861 | 16:58 |
| inc0 | we barely recovered from kolla-ansible split | 16:58 |
| inc0 | I can't :( or I don't see how | 16:58 |
| *** ingmar23 has quit IRC | 16:59 | |
| kfox1111 | sbezverk: can you file a bug with them? last time they were quite responsive. | 16:59 |
| berendt | maybe it is not possible at all | 16:59 |
| sbezverk | I have dependency on cinder-api service, cinder-api is in init state, but cinder-scheduler which has cinder-api as a dependency is alredy in running state | 16:59 |
| qwang | good morning | 16:59 |
| kfox1111 | without that getting fixed, entrypoint's far less useful. :/ | 16:59 |
| kfox1111 | sbezverk: can you double check the dep container is in the pod? | 16:59 |
| sbezverk | kfox1111: maybe race conditions we keep seeing are related to it.. | 16:59 |
| kfox1111 | we missed one before. | 16:59 |
| *** aj_power has joined #openstack-kolla | 16:59 | |
| kfox1111 | I thought I hit all of them, but managed to miss one. so its possible I missed 2. | 17:00 |
| sbezverk | kfox1111: will do that before.. | 17:00 |
| kfox1111 | or more. | 17:00 |
| kfox1111 | thx. | 17:00 |
| *** simon-AS559 has quit IRC | 17:00 | |
| *** sudipto has quit IRC | 17:01 | |
| openstackgerrit | Merged openstack/kolla: Add ipxe packages to Ironic PXE container https://review.openstack.org/424313 | 17:01 |
| *** sudipto_ has quit IRC | 17:01 | |
| sbezverk | kfox1111: not this case, as it uses common stateful template | 17:01 |
| sbezverk | and log tells that dependency got resolved | 17:01 |
| kfox1111 | ok. thanks for double checking. | 17:02 |
| kfox1111 | that does sound like a bug. | 17:02 |
| kfox1111 | they might not have considered that pods can be in init state. | 17:02 |
| *** schwicht has quit IRC | 17:02 | |
| kfox1111 | only pending/not ready. | 17:02 |
| *** schwicht_at_work has joined #openstack-kolla | 17:03 | |
| kfox1111 | I don't know if they are using init containers yet. | 17:03 |
| *** sudipto has joined #openstack-kolla | 17:04 | |
| *** schwicht_at_work has quit IRC | 17:04 | |
| *** sudipto_ has joined #openstack-kolla | 17:04 | |
| *** schwicht has joined #openstack-kolla | 17:05 | |
| sbezverk | kfox1111: filed the issue and gave them link for logs. do you know when PS' logs get trashed? | 17:07 |
| sbezverk | is it on the next run or that link stays alive for a while? | 17:07 |
| *** markmcclain has quit IRC | 17:09 | |
| *** markmcclain has joined #openstack-kolla | 17:11 | |
| *** dave-mccowan has quit IRC | 17:11 | |
| kfox1111 | sbezverk: haven't figured that out. they seem to stay alive a long time. | 17:13 |
| *** sudipto_ has quit IRC | 17:14 | |
| *** sudipto has quit IRC | 17:14 | |
| *** sp___ has quit IRC | 17:14 | |
| sbezverk | kfox1111: cool thanks, hopefully sufficient enough for entrypoint people to check this problem | 17:15 |
| *** klindgren has joined #openstack-kolla | 17:15 | |
| kfox1111 | thanks for filing the issue. | 17:15 |
| kfox1111 | do you have a link handy so I can follow it? | 17:15 |
| openstackgerrit | Francisco Edigleison Barbosa proposed openstack/kolla: Add how configure backend to cinder HNAS iSCSI driver guide https://review.openstack.org/422157 | 17:16 |
| sbezverk | kfox1111: https://github.com/stackanetes/kubernetes-entrypoint/issues/19 | 17:17 |
| kfox1111 | thx. | 17:17 |
| qwang | sbezverk: I can still open the logs from PS that's merged a couple of months ago. | 17:18 |
| *** lrensing has quit IRC | 17:19 | |
| openstackgerrit | Francisco Edigleison Barbosa proposed openstack/kolla: Add how configure backend to cinder HNAS iSCSI driver guide https://review.openstack.org/422157 | 17:19 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Add cinder control package with kubernetes entrypoint https://review.openstack.org/424697 | 17:20 |
| sbezverk | qwang: nice, thanks for confirming it | 17:21 |
| *** lrensing has joined #openstack-kolla | 17:23 | |
| qwang | sbezverk: you're welcome. glad to help :) | 17:24 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Typo fix: mysql => memcached https://review.openstack.org/424423 | 17:26 |
| *** saneax is now known as saneax-_-|AFK | 17:28 | |
| kfox1111 | ugg. | 17:29 |
| kfox1111 | the gate really is starting to get really messy with all the forks. :/ | 17:29 |
| kfox1111 | making fixing things harder and harder. :/ | 17:29 |
| *** inc0 has quit IRC | 17:30 | |
| kfox1111 | I was going to do this right in this ps, but its going to take a major refactor to do that. :/ | 17:32 |
| sbezverk | kfox1111: we are almost all done with microservices and services, after that it will be easier for gate refactor | 17:36 |
| kfox1111 | even just doing ceph_workflow.sh and iscsi_workflow.sh woudl really help. :/ | 17:37 |
| kfox1111 | the problem is, they aren't using the global api for config, so its getting harder to tweak things. | 17:37 |
| kfox1111 | I'm going to lay a bit of groundwork in this ps to do that, | 17:37 |
| kfox1111 | then we can fix it for real in a follow up ps as soon as it merges. | 17:38 |
| sdake | kfox1111 was it you that told me that openstack had sorted out a way to add new language choices? | 17:38 |
| kfox1111 | no. but I heard they were working on it. | 17:39 |
| kfox1111 | didn't hear they were done though. | 17:39 |
| sdake | i thought someon told m e they were done | 17:39 |
| sdake | not sure who it was | 17:39 |
| sdake | was looking for a pointer on the ml for the thread | 17:39 |
| sdake | as jascott1 and srwilkers are taking a look at it for kolla-kubneretes | 17:40 |
| kfox1111 | yeah. I can understand that. | 17:40 |
| sbezverk | kfox1111: sounds like a plan, it will also give us a bit of time to finish services.. | 17:40 |
| sdake | srwilkers i'll ask a few folks on the tc on that point | 17:41 |
| *** Serlex has quit IRC | 17:41 | |
| sdake | the lsat time i asked it was blocked on a variety of things - however - I am pretty sure its unbloked now | 17:41 |
| sdake | unblocked | 17:41 |
| kfox1111 | sbezverk: yeah. services is in much better shape already, as it was written much more recently. | 17:41 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 17:42 |
| kfox1111 | sdake: unblocked doesn't meen imminent though. | 17:42 |
| kfox1111 | I think they decided a little while ago, that they would come up with a process by which they could certify a language was ok to use. | 17:42 |
| kfox1111 | which unblocked the discussion. | 17:42 |
| sdake | kfox1111 cool - i haven't been following the tc meetings - too busy | 17:43 |
| kfox1111 | but that doesn't mean they nessisarily could come up with the criteria very quickly. | 17:43 |
| sdake | kfox1111 i'll ask around | 17:43 |
| *** adrian_otto has quit IRC | 17:43 | |
| kfox1111 | yeah. let me know what you find out. :) | 17:43 |
| sdake | jascott1 srwilkers does that semi-answer your q, or do you need m0ar minerals? :) | 17:43 |
| sbezverk | kfox1111: could you please ack nova control ps? | 17:43 |
| kfox1111 | in a few. meeting time. | 17:44 |
| sbezverk | k | 17:44 |
| srwilkers | sdake, kfox1111 an email came across last week saying the new requirements for new languages had been approved by the TC | 17:44 |
| sdake | srwilkers sounds like you know more aobut it then i do :) | 17:44 |
| sdake | srwilkers got alink to hte mail - i'll parse it if you like | 17:44 |
| srwilkers | yeah, sec | 17:45 |
| pbourke | does anyone know if oslo push pypi packages for stable branches? | 17:45 |
| sdake | pbourke i'd assume so | 17:45 |
| sdake | pbourke #openstack-oslo is the place to ask - btw your up late ;) | 17:45 |
| pbourke | sdake: nope only 17:44 here ;) | 17:45 |
| sdake | oh cool | 17:45 |
| srwilkers | sdake: https://governance.openstack.org/tc/reference/new-language-requirements.html | 17:45 |
| sdake | sbezverk do yu ahe to have a helm server active to do helm prebuild? | 17:46 |
| *** hieulq_ has quit IRC | 17:46 | |
| sbezverk | sdake: only to prebuild services | 17:47 |
| sdake | srwilkers where the existing, already accepted, languages have been proven to not meet the technical requirements | 17:47 |
| sbezverk | srwilkers: man that is a lot of requirements ;) | 17:47 |
| sdake | have to make a technical argument (which people are doing for go) not a religious one :) | 17:47 |
| *** eaguilar has joined #openstack-kolla | 17:48 | |
| sdake | Define how the deliverables are distributed¶ | 17:48 |
| srwilkers | yep, its not just religious for me. makes sense from a performance standpoint | 17:48 |
| sdake | golang has like 5 ways to do builds | 17:48 |
| sdake | this one is going to be challenging for go specifically | 17:48 |
| jascott1 | he preference for us is | 17:48 |
| jascott1 | to produce Debian based containers | 17:48 |
| jascott1 | stupid paste | 17:49 |
| sdake | we could just take over the universe and produce containers ;) | 17:49 |
| sdake | jascott1 right - a deliverable of go in a container makes alot of sense | 17:49 |
| *** eaguilar_ has joined #openstack-kolla | 17:49 | |
| jascott1 | sorry was trying to quote the ML about debian | 17:49 |
| sdake | since go is targeted at cloud native apps, and building a container in go is somethign that can be delivered vs the 5 build tools | 17:49 |
| sdake | jascott1 oh - well I was serious | 17:50 |
| jascott1 | sdake i agree that go in containers does make sense :) | 17:50 |
| sdake | dep managemeent | 17:50 |
| sdake | thats going to be a pain in the rear | 17:50 |
| *** dave-mccowan has joined #openstack-kolla | 17:50 | |
| sdake | whoever manages to thread the needle on golang for openstack will be my hero for sorting out those sets of problems :) | 17:51 |
| sdake | jascott1 srwilkers it will take a team - I'd suggest posting to the ml and seeing who is interested and join forces | 17:51 |
| sdake | the more people interested in bringing go to openstack, the more likely it is to happen | 17:51 |
| srwilkers | sdake, yep. i agree. jascott and i have been chatting for about a week or so in regards to getting the discussion started | 17:52 |
| sdake | could probably form a technical working group around i t - however - I am overloaded and can't do that myself | 17:52 |
| bmace_ | unless things have improved recently i found golang dependency management amazingly obnoxious.. i know there were some third party tools for version pinning, but most projects just sat on the bleeding edge and therefore broke all the time | 17:52 |
| sdake | bmace_ agreed | 17:52 |
| *** eaguilar has quit IRC | 17:52 | |
| portdirect | i think some of the swift peeps may be interested? | 17:52 |
| sdake | srwilkers there are a whole slew of people in the community interested in the problem | 17:52 |
| sdake | including what portdirect just pointed out | 17:52 |
| bmace_ | unless there are real performance needs i'm not sure i see a lot of upside. i find python a lot more readable also. | 17:52 |
| sdake | my recommendation is to start talking to them | 17:53 |
| sdake | bmace_ ya it is clear go is a clear performance winner | 17:53 |
| sdake | bmace_ although go also has significant downside | 17:53 |
| srwilkers | sdake, yep. jascott1 showed me some of the swift work, and we had planned to reach out to them | 17:53 |
| sdake | (dep managmenet, interationalization, docs generation, etc) | 17:53 |
| bmace_ | there is that new grumpy thing. just write your code in python and compile it down to Go :) | 17:53 |
| sdake | i think it would be fantastic if go were a language choice in openstack | 17:53 |
| sdake | lots ofpeople want it | 17:53 |
| jascott1 | yes we do | 17:54 |
| sdake | nobody really knows how to do it yet :) | 17:54 |
| sdake | bmace_ that sounds interesting :) | 17:54 |
| sdake | bmace_ its a preprocessor? | 17:54 |
| bmace_ | you hadn't heared about that? | 17:54 |
| sdake | bmace_ dude i hear abou tnew tech all day - i can't keep it all straight | 17:54 |
| *** srwilkers is now known as srwilkers_brb | 17:54 | |
| bmace_ | sdake: https://github.com/google/grumpy | 17:54 |
| jascott1 | heres the post talking about the python perf wall https://opensource.googleblog.com/2017/01/grumpy-go-running-python.html | 17:55 |
| bmace_ | Google has a TON of python code and they are trying to make it more performant without doing total rewrites. | 17:55 |
| sbezverk | bmace_: that is really cool, thanls for the link! | 17:56 |
| bmace_ | :) | 17:56 |
| *** mgoddard has quit IRC | 17:56 | |
| bmace_ | i thought it was a really interesting idea, and based on the upside of it maybe Google won't abandon this one, lol. | 17:56 |
| srwilkers_brb | lol well | 17:56 |
| sdake | GROAN: Two design choices we made had big consequences. First, we decided to forgo support for C extension modules. | 17:56 |
| srwilkers_brb | thats a gamble | 17:56 |
| srwilkers_brb | brb | 17:57 |
| srwilkers_brb | yep | 17:57 |
| srwilkers_brb | also made me groan when i first read it | 17:57 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Add cinder control package with kubernetes entrypoint https://review.openstack.org/424697 | 17:57 |
| bmace_ | yeah. there is certainly a set of libs that will probably need to be re-written / unsupported by Grumpy. | 17:57 |
| sdake | anyway jascott1 srwilkers_brb if you fellas are keen - psot on the mailing list and ask others if someone would be keen to get going on it | 17:57 |
| sdake | pun intended :) | 17:57 |
| bmace_ | but at least it can compile down all the pure python stuff | 17:57 |
| jascott1 | sdake i think srwilkers and I can conspire todo that | 17:58 |
| *** athomas has quit IRC | 17:59 | |
| sdake | jascott1 where in kolla is there a performance limitation? | 17:59 |
| sdake | or srwilkers_ ^ | 17:59 |
| sdake | we are always limitedd by disk io speed and netowrk bandiwdth | 17:59 |
| sdake | rarely cpu utilizaiton | 17:59 |
| bmace_ | i can see it maybe in base ostk.. but in what we do in kolla / kolla-k8s i don't know where Go rather than python would actually make a bit perf diff | 18:00 |
| sdake | kolla in general is io-bound not cpu-bound | 18:00 |
| *** inc0 has joined #openstack-kolla | 18:00 | |
| sdake | bmace_ especially some of the base libraries like oslo.log for example | 18:00 |
| *** matrohon has quit IRC | 18:00 | |
| sdake | bmace_ also - today got jammed up on iscsi | 18:01 |
| *** inc0 has quit IRC | 18:01 | |
| bmace_ | i'm sure i'll hit it soon sdake :) | 18:02 |
| sdake | bmace_ "there is no iscsi directory availble" when running the dev env (the last script in the chain) | 18:02 |
| jascott1 | sdake swift | 18:02 |
| *** krtaylor has quit IRC | 18:02 | |
| *** adrian_otto has joined #openstack-kolla | 18:03 | |
| jascott1 | i think the swift/hummingbird approach made sense. they hit a performance wall and they just re-wrote the parts in go that they needed to improve | 18:03 |
| *** inc0 has joined #openstack-kolla | 18:04 | |
| *** inc0 has quit IRC | 18:05 | |
| *** strigazi is now known as strigazi_AFK | 18:05 | |
| jascott1 | imo for kolla the advantage would be to take advantage of other tooling like operators that others (coreos etc) have written | 18:07 |
| sdake | jascott1 yareuse is another technicl advantage | 18:07 |
| sdake | reuse | 18:07 |
| jascott1 | yar | 18:07 |
| sdake | groan - hopefully ou parsed that :) | 18:08 |
| jascott1 | :) | 18:08 |
| sdake | i wish i had time to get go into openstack - sadly i am going 18 hrs a day to keep up with kids + work + wife + parents + extended fmailly | 18:08 |
| sdake | I've got nothing left to give for anyone else :) | 18:08 |
| *** inc0 has joined #openstack-kolla | 18:08 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla: Create deliverables.yaml https://review.openstack.org/424784 | 18:16 |
| *** sean-k-mooney has quit IRC | 18:16 | |
| *** inc0 has quit IRC | 18:16 | |
| *** sean-k-mooney has joined #openstack-kolla | 18:17 | |
| *** TxGirlGeek has quit IRC | 18:17 | |
| jascott1 | sdake man you sounding like the giving tree | 18:21 |
| sdake | jascott1 lol | 18:21 |
| sbezverk | jascott1: where is the fun of reusing other people work? it kind of kills innovation spirit, no? | 18:22 |
| *** inc0 has joined #openstack-kolla | 18:22 | |
| jascott1 | i always hated the kid in that story.. greedy b457ard | 18:22 |
| *** lrensing has quit IRC | 18:22 | |
| sdake | sbezverk we reuse the acii character set on a daily basis | 18:22 |
| *** TxGirlGeek has joined #openstack-kolla | 18:22 | |
| sdake | and atleast in the US, we use the QWERTY keyboard | 18:22 |
| *** sambetts is now known as sambetts|afk | 18:22 | |
| sdake | and we use tables, knives, forks, spoons, plates | 18:23 |
| sbezverk | sdake: if you enjoy it, good for you ;) | 18:23 |
| sdake | these are all examples of reuse :) | 18:23 |
| jascott1 | sbezverk agree "the fun" is diminished but saving time and eating with family is nice ;) | 18:23 |
| sdake | sbezverk ya its not fun integrating thats for sure | 18:23 |
| sdake | sbezverk i think - in the oeprator example, we want to have a bit of diversity in opinions about what those things really look like | 18:24 |
| *** hfu has joined #openstack-kolla | 18:24 | |
| *** sp___ has joined #openstack-kolla | 18:24 | |
| sdake | not sure a reuse technical argument makes sense for k8s-oeprators | 18:24 |
| jascott1 | finally got PTG approval email :) | 18:24 |
| sdake | however ther eis reuse of the interface | 18:24 |
| *** inc0 has quit IRC | 18:24 | |
| *** bpattewar has joined #openstack-kolla | 18:25 | |
| jascott1 | sdake yup interface{} | 18:25 |
| sdake | interface to kubernetes that is (the rest api) | 18:25 |
| *** lrensing has joined #openstack-kolla | 18:25 | |
| sdake | persoanlly i'm happy with python calling kubectl or helm via shellout | 18:25 |
| jascott1 | all the libraries for k8s/helm/yaml/templating etc | 18:25 |
| sdake | since it runs in a contianer who cares | 18:25 |
| sdake | that is what rocks about containers- they turn an imperative system into a delcartive system | 18:26 |
| sdake | and remove all the religious battles in the process :) | 18:26 |
| sbezverk | sdake: I think the main re-use argument, is that people do not want to undersatnd details, it is easier to get blackbox and use it hoping it will work as expected.. | 18:27 |
| *** bjolo has joined #openstack-kolla | 18:27 | |
| jascott1 | not to mention with golang you can create a container that is 80% smaller than a typical one if you so desire | 18:27 |
| jascott1 | if the blackbox tests ok, i dont care (usually) | 18:28 |
| *** dave-mccowan has quit IRC | 18:30 | |
| jascott1 | if there was a generic operator that worked well it would be silly to port to python just because its not python | 18:30 |
| *** sp___ has quit IRC | 18:31 | |
| sbezverk | jascott1: do not get me wrong, I am for golang.. the operator I am working on now will be re-written in go, guaranteed.. | 18:32 |
| sbezverk | jascott1: it is just since I enjoy doing what I am doing, I do not like idea of people reducing amount of fun for me.. :) | 18:32 |
| jascott1 | oh sorry wasnt trying to under appreciate your efforts! | 18:33 |
| jascott1 | im happy to be fun adjacent myself | 18:33 |
| *** mgoddard has joined #openstack-kolla | 18:33 | |
| jascott1 | the caffeine must flow | 18:33 |
| jascott1 | brb | 18:34 |
| *** sdake has quit IRC | 18:37 | |
| *** gfidente is now known as gfidente|afk | 18:43 | |
| *** hfu has quit IRC | 18:44 | |
| *** krtaylor has joined #openstack-kolla | 18:44 | |
| *** inc0 has joined #openstack-kolla | 18:44 | |
| *** sdake has joined #openstack-kolla | 18:45 | |
| *** hfu has joined #openstack-kolla | 18:46 | |
| *** hfu has quit IRC | 18:46 | |
| sdake | sbezverk having fun is good - my kids keep me so busy - fun is not a luxury i really have often | 18:54 |
| sdake | sbezverk might as well maek work enjoyable | 18:55 |
| *** inc0 has quit IRC | 18:57 | |
| *** dave-mccowan has joined #openstack-kolla | 18:57 | |
| *** inc0 has joined #openstack-kolla | 19:00 | |
| *** shardy is now known as shardy_afk | 19:04 | |
| openstackgerrit | sayantani proposed openstack/kolla: Update Bifrost documentation https://review.openstack.org/417114 | 19:04 |
| *** l4yerffeJ__ has joined #openstack-kolla | 19:10 | |
| *** l4yerffeJ_ has quit IRC | 19:10 | |
| *** TxGirlGeek has quit IRC | 19:11 | |
| *** TxGirlGeek has joined #openstack-kolla | 19:12 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Add cinder control package with kubernetes entrypoint https://review.openstack.org/424697 | 19:13 |
| *** inc0 has quit IRC | 19:14 | |
| *** MasterOfBugs has joined #openstack-kolla | 19:14 | |
| *** lrensing has quit IRC | 19:16 | |
| *** inc0 has joined #openstack-kolla | 19:20 | |
| *** lrensing has joined #openstack-kolla | 19:20 | |
| openstackgerrit | sayantani proposed openstack/kolla-ansible: Update bifrost playbooks https://review.openstack.org/424384 | 19:21 |
| *** schwicht has quit IRC | 19:26 | |
| *** sayantani01 has quit IRC | 19:27 | |
| *** sayantani01 has joined #openstack-kolla | 19:28 | |
| *** schwicht has joined #openstack-kolla | 19:28 | |
| *** sayantani01 has quit IRC | 19:32 | |
| *** inc0 has quit IRC | 19:41 | |
| *** harlowja has joined #openstack-kolla | 19:41 | |
| *** inc0 has joined #openstack-kolla | 19:45 | |
| *** lrensing has quit IRC | 19:49 | |
| *** TxGirlGeek has quit IRC | 19:49 | |
| *** mgoddard has quit IRC | 19:50 | |
| *** TxGirlGeek has joined #openstack-kolla | 19:51 | |
| *** sdake has quit IRC | 19:52 | |
| *** inc0 has quit IRC | 19:56 | |
| *** sdake has joined #openstack-kolla | 19:57 | |
| *** flaper87 has joined #openstack-kolla | 19:59 | |
| *** eaguilar_ has quit IRC | 20:05 | |
| sdake | sup flaper87 | 20:07 |
| flaper87 | sdake: yo! :D | 20:08 |
| * flaper87 in TC meeting but almost dying | 20:08 | |
| sdake | flaper87 there was alot of q&a earlier on golang as a language choice for openstack - any chance you know who to put jascott1 and srwilkers_ in touch with that is leading that effort? | 20:10 |
| *** sayantani01 has joined #openstack-kolla | 20:11 | |
| kfox1111 | sbezverk: here's an odd one: http://logs.openstack.org/68/424368/1/check/gate-kolla-kubernetes-deploy-centos-binary-3-ceph-multi-nv/8c04334/logs/openstack/kolla-neutron-openvswitch-agent-compute-38zf2.txt | 20:11 |
| sdake | flaper87 tell me aout it - had an eye exam yesterday - doctor was overly-thorough - eyes kiling me | 20:11 |
| flaper87 | sdake: that'd be me, at least as far as the process for introducing a new language goes. Then, I know dtroyer is working on adding testing for golang | 20:11 |
| sdake | flaper87 i think thes eguys want to build some critical mass - but not sure | 20:12 |
| flaper87 | sdake: :( that sucks :( hope you'll feel better soon | 20:12 |
| sdake | flaper87 the more people working on it, i think the more likely it is to happen | 20:12 |
| sdake | i'll be fine :) | 20:12 |
| flaper87 | sdake: yup, I'm happy to chat and answer questions on the topic | 20:12 |
| *** adrian_otto has quit IRC | 20:12 | |
| flaper87 | or point ppl in the right direction | 20:12 |
| sdake | its not as bad as a dentist torture session | 20:12 |
| sdake | flaper87 i don't personally ahve time to get involved, however, i know jascott1 and srwilkers_ have an interest | 20:13 |
| sdake | flaper87 they may be afk at the moment - lunch time on the west coast for jascott1, not certain of srwilkers_ 's timezone | 20:14 |
| *** adrian_otto has joined #openstack-kolla | 20:17 | |
| *** mgoddard has joined #openstack-kolla | 20:21 | |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla: Update dumb-init to version 1.2.0 https://review.openstack.org/424832 | 20:21 |
| *** TxGirlGeek has quit IRC | 20:24 | |
| *** adrian_otto has quit IRC | 20:24 | |
| *** sdake_ has joined #openstack-kolla | 20:24 | |
| *** TxGirlGeek has joined #openstack-kolla | 20:24 | |
| sbezverk | kfox1111: I think ovs datapath is not loaded | 20:25 |
| sbezverk | that is why br-int returns has datapath-ID [] | 20:26 |
| kfox1111 | yeah. kind of seems like it. | 20:26 |
| kfox1111 | but theres a dep there. | 20:26 |
| *** TxGirlGeek has quit IRC | 20:26 | |
| sbezverk | kfox1111: it is not dependency, it is modprobe | 20:26 |
| *** TxGirlGeek has joined #openstack-kolla | 20:27 | |
| kfox1111 | I thought neutron-openvswitch-agent depended on openvswitch socket though? | 20:27 |
| kfox1111 | seems like it came up too early though? | 20:27 |
| *** sdake has quit IRC | 20:28 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP Add cinder control package with kubernetes entrypoint https://review.openstack.org/424697 | 20:29 |
| kfox1111 | is there a case where the socket gets created before the db's initted somehow? | 20:29 |
| sbezverk | kfox1111: ahhhh that is our issue I hit, I knew it looked familiar | 20:29 |
| sbezverk | kfox1111: if neutron agent comes up before vswitchd it happens | 20:30 |
| *** TxGirlGeek has quit IRC | 20:30 | |
| sbezverk | kfox1111: as vswitchd loads ovs module | 20:30 |
| kfox1111 | I thought the vswitchd created the unix socket though? | 20:30 |
| kfox1111 | or is that the db? | 20:30 |
| sbezverk | kfox1111: no onlt ovsdb creates socket | 20:31 |
| kfox1111 | ah. | 20:31 |
| kfox1111 | so, does the db need to load the module, | 20:31 |
| kfox1111 | or must we ensure both are up? | 20:31 |
| sbezverk | kfox1111: vswitchd loads module, never tried to load it in ovsdb though | 20:32 |
| kfox1111 | or we could just do it in both. | 20:32 |
| sbezverk | kfox1111: if we do it in both, second might generate an error and the whole thing fails then.. | 20:35 |
| sbezverk | kfox1111: better solution is to ensure that pods are running, no? | 20:36 |
| *** srwilkers_brb is now known as srwilkers | 20:36 | |
| *** lrensing has joined #openstack-kolla | 20:38 | |
| kfox1111 | I think a modprobe of somethign thats already there is a noop. | 20:44 |
| kfox1111 | if both pods really need to be running though, then yeah, we should build that in somehow. | 20:44 |
| kfox1111 | the problem is non pod support too. | 20:44 |
| kfox1111 | the check to see if the pods are running should work with the case where your running openvswitch on the host too. | 20:45 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 20:46 |
| * sdake_ wtb bed | 20:46 | |
| *** sdake_ is now known as sdake | 20:46 | |
| sbezverk | kfox1111: that is more complicated scenario | 20:47 |
| kfox1111 | ah... interesting. | 20:47 |
| kfox1111 | the arm folks are interesting in kolla with debian. | 20:48 |
| kfox1111 | sbezverk: yes/no. it works today, as openvswitch is always started before docker. | 20:48 |
| kfox1111 | so never really a problem. | 20:48 |
| kfox1111 | but if we add more deps handling it may run afoul of that. | 20:48 |
| kfox1111 | just need to be careful. | 20:49 |
| sbezverk | kfox1111: sure when you control things at microservice level, absolutely. I was talking about more generic service level deploymnet, where you start everything and just wait for it to converge | 20:50 |
| sbezverk | the way neutron ovs agen is written today it does not restart | 20:50 |
| sbezverk | it just dies | 20:50 |
| sbezverk | that is why it is really important for it to make sure that either internal vswitchd loaded module or your host has it already.. | 20:51 |
| sbezverk | kfox1111: do you know a way to disassemble yum package to see what it does exactly during the install? | 20:53 |
| kfox1111 | hmm... k. will have to think on that one... | 20:53 |
| kfox1111 | sbezverk: from file or from installed package? | 20:53 |
| sbezverk | kfox1111: no worries, I will google it | 20:53 |
| sbezverk | kfox1111: does not matter I guess | 20:54 |
| kfox1111 | rpm --scripts -qpf foo.rpm #for rpm | 20:54 |
| kfox1111 | rpm --scripts -q rpmname for installed. | 20:54 |
| kfox1111 | well, rpm --scripts -qp foo.rpm #for rpm | 20:54 |
| kfox1111 | had an extra f there. | 20:54 |
| sbezverk | kfox1111: thanks! | 20:54 |
| kfox1111 | np | 20:54 |
| *** TxGirlGeek has joined #openstack-kolla | 20:55 | |
| openstackgerrit | sayantani proposed openstack/kolla-ansible: Fix bootstrap.yml in Bifrost https://review.openstack.org/424271 | 20:55 |
| *** TxGirlGeek has quit IRC | 21:00 | |
| *** jrobinson has joined #openstack-kolla | 21:04 | |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla: Clean dockerfile output on Ubuntu distro https://review.openstack.org/424849 | 21:05 |
| *** TxGirlGeek has joined #openstack-kolla | 21:05 | |
| *** gfidente|afk has quit IRC | 21:06 | |
| *** satyar has quit IRC | 21:08 | |
| *** srwilkers has quit IRC | 21:08 | |
| *** dave-mccowan has quit IRC | 21:10 | |
| *** adrian_otto has joined #openstack-kolla | 21:10 | |
| *** sacharya has quit IRC | 21:12 | |
| *** sacharya has joined #openstack-kolla | 21:14 | |
| *** msimonin has joined #openstack-kolla | 21:16 | |
| *** mgoddard has quit IRC | 21:20 | |
| *** shardy_afk is now known as shardy | 21:27 | |
| *** schwicht has quit IRC | 21:27 | |
| *** jrobinson has quit IRC | 21:27 | |
| *** jrobinson has joined #openstack-kolla | 21:28 | |
| *** williamcaban has quit IRC | 21:29 | |
| *** williamcaban has joined #openstack-kolla | 21:29 | |
| *** l4yerffeJ has joined #openstack-kolla | 21:34 | |
| *** Jeffrey4l_ has joined #openstack-kolla | 21:34 | |
| *** l4yerffeJ__ has quit IRC | 21:35 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** schwicht has joined #openstack-kolla | 21:39 | |
| *** zioproto has joined #openstack-kolla | 21:42 | |
| *** portdirect is now known as portdirect_away | 21:46 | |
| *** zioproto has quit IRC | 21:46 | |
| jascott1 | sdake FYI srwilkers is in St Louis | 21:48 |
| *** williamcaban has quit IRC | 21:51 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Add cinder control package with kubernetes entrypoint https://review.openstack.org/424697 | 21:54 |
| kfox1111 | oh... setip is run later.. hmm. | 21:55 |
| *** breitz has quit IRC | 21:59 | |
| *** breitz has joined #openstack-kolla | 22:00 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 22:01 |
| sbezverk | kfox1111: cinder deployment was successful, when you have a chance please review: https://review.openstack.org/424697 | 22:02 |
| *** aj_power has quit IRC | 22:03 | |
| *** jtriley has quit IRC | 22:04 | |
| *** TxGirlGeek has quit IRC | 22:14 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:15 | |
| *** rhallisey has quit IRC | 22:15 | |
| *** l4yerffeJ has quit IRC | 22:16 | |
| *** l4yerffeJ has joined #openstack-kolla | 22:16 | |
| *** inc0 has joined #openstack-kolla | 22:18 | |
| *** sayantani01 has quit IRC | 22:22 | |
| *** sayantani01 has joined #openstack-kolla | 22:22 | |
| *** sayantani01 has quit IRC | 22:23 | |
| *** sayantani01 has joined #openstack-kolla | 22:23 | |
| *** fguillot has quit IRC | 22:23 | |
| *** TxGirlGeek has quit IRC | 22:24 | |
| *** TxGirlGeek has joined #openstack-kolla | 22:24 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Add cinder control package with kubernetes entrypoint https://review.openstack.org/424697 | 22:25 |
| *** sayantani01 has quit IRC | 22:25 | |
| *** sayantani01 has joined #openstack-kolla | 22:26 | |
| *** krtaylor has quit IRC | 22:27 | |
| *** lrensing has quit IRC | 22:27 | |
| *** sayantani01 has quit IRC | 22:30 | |
| inc0 | centos gates are broken/ | 22:32 |
| *** MasterOfBugs has quit IRC | 22:33 | |
| *** pramodrj07 has joined #openstack-kolla | 22:33 | |
| *** matrohon has joined #openstack-kolla | 22:34 | |
| kfox1111 | inc0 more specific please | 22:34 |
| kfox1111 | kolla-ansible, kolla-kubernetes, kolla? | 22:35 |
| inc0 | https://review.openstack.org/#/c/424784/1 lots of red | 22:35 |
| inc0 | kolla | 22:35 |
| kfox1111 | k. thanks. | 22:35 |
| *** mgiles has quit IRC | 22:39 | |
| *** TxGirlGeek has quit IRC | 22:40 | |
| *** adrian_otto has quit IRC | 22:41 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed openstack/kolla: Create deliverables.yaml https://review.openstack.org/424784 | 22:41 |
| inc0 | SamYaple, wanna get kolla-salt into the fold? | 22:42 |
| inc0 | just submit patch to ^ | 22:42 |
| inc0 | :) | 22:42 |
| *** TxGirlGeek has joined #openstack-kolla | 22:43 | |
| inc0 | bbl, moving home | 22:44 |
| *** inc0 has quit IRC | 22:44 | |
| *** huzhengchuan has joined #openstack-kolla | 22:47 | |
| *** TxGirlGeek has quit IRC | 22:48 | |
| *** lamt has quit IRC | 22:49 | |
| *** david-lyle has quit IRC | 22:50 | |
| *** lamt has joined #openstack-kolla | 22:51 | |
| *** david-lyle has joined #openstack-kolla | 22:53 | |
| bmace_ | sdake: try out the latest containerized dev env stuff today? | 22:56 |
| sdake | bmace_ indeed this morning | 22:56 |
| sdake | bmace_ complaint from the devenv script about iscsi not available | 22:57 |
| sdake | bmace_ i had some quesitons about whether iscsi needs to be enabled in teh host os or something similar | 22:57 |
| sdake | bmace_ how far did you get? | 22:57 |
| *** TxGirlGeek has joined #openstack-kolla | 22:57 | |
| bmace_ | got a new error in genconfig when i just re-built mine (to pick up the genconfig fixes from yesterday rather than using local patches). something in fluentd | 22:58 |
| bmace_ | need to chase it down | 22:58 |
| sdake | i did not get a new error | 22:58 |
| sdake | around fulentd | 22:58 |
| bmace_ | and you did a complete container rebuild recently? | 22:58 |
| sdake | bmace_ - i am installing actually on bare metal not using the container part | 22:59 |
| bmace_ | since i was getting well past this point yesterday and the same env i had before it feels likely it is a new issue. | 22:59 |
| kfox1111 | bmace_: maybe just pin to the same version of kolla-ansible the gate's using, until all the issues can be resolved? | 22:59 |
| *** huzhengchuan has quit IRC | 23:00 | |
| bmace_ | kfox1111: yeah, i may need to do just that. never like picking up the bleeding edge for exactly these kinds of reasons. | 23:01 |
| sdake | bmace_ i wiped out my gear entirely and was running the scripts manually rather then in the container | 23:01 |
| sdake | against a self-installed version of kubernetes | 23:02 |
| sdake | via kubeadm | 23:02 |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 23:02 |
| *** eaguilar has joined #openstack-kolla | 23:04 | |
| sdake | bmace_ i think if you wipe out your entire thingy and start with master and the script with a cherrypick it should work for you up to the iscsi part | 23:04 |
| bmace_ | sdake: why not do it in a container? | 23:04 |
| sdake | kfox1111 do you know what the deal is with iscsi and its requirements for the host os? | 23:04 |
| sdake | bmace_ complicates things imo - although container I guess is worth a shot | 23:04 |
| kfox1111 | sdake: no. sbezverk would know. | 23:04 |
| sdake | sbezverk wtb help :) | 23:05 |
| bmace_ | sdake: well, the magic of containers, i did wipe out my entire "thingy" with an image delete and built a new one :) | 23:05 |
| sdake | hogepodge ping | 23:05 |
| kfox1111 | sdake: or the tools/setup_gate_iscsi.sh script | 23:05 |
| hogepodge | sdake: hi | 23:05 |
| *** mliima has quit IRC | 23:05 | |
| sbezverk | sdake: what is up? | 23:05 |
| *** mliima has joined #openstack-kolla | 23:06 | |
| sdake | hogepodge i don't know if you would be interested, however, we have a biweekly unofficial workgroup with helm happening between openstack on kubernetes efforts and the helm upstream community | 23:06 |
| bmace_ | sdake: i think doing it all in a container actually simplifies things a lot. you don't need to have your scripts nearly as complicated, supporting multiple os types, etc.. just support the container base image type. | 23:06 |
| sdake | hogepodge since your involved in interop not sure if its your area or someone elses | 23:06 |
| sdake | hogepodge if you like I can forward you teh meeting invitiation | 23:06 |
| sdake | bmace_ roger - got it | 23:06 |
| hogepodge | sdake: I'm definitely interested | 23:07 |
| sdake | bmace_ ok well not here to argue about methods, I think real operators will want to deploy on bare metla not via a contianer - but could be wrong ;) | 23:07 |
| kfox1111 | bmace_: +1 | 23:07 |
| sdake | bmace_ as a result i was taking that approach | 23:07 |
| kfox1111 | sdake: the ideal is, the users install k8s and helm. | 23:07 |
| sdake | kfox1111 right | 23:07 |
| hogepodge | sdake: I'm going to be getting more involved with cross-community projects between OpenStack and Kubernetes this year, so it's squarely something I'd like to participate in | 23:07 |
| kfox1111 | then we have the dev env, in a helm chart in the stable helm repo. | 23:07 |
| sdake | hogepodge pm your email address and i'll forward you the invite | 23:07 |
| kfox1111 | so the user just has to helm install kolla-dev-env. | 23:08 |
| sdake | hogepodge sweet - if you know others that are interested let me know | 23:08 |
| kfox1111 | and they get all the deps handled. :) | 23:08 |
| *** harlowja has quit IRC | 23:08 | |
| hogepodge | sdake: I will. quite likely dfflanders | 23:08 |
| sdake | w eare working on turning this into an official workgroup with biwekely workhops hosted by someone :) | 23:08 |
| bmace_ | sdake: well, you can use the container to get everything deployed onto bare metal just as easily as getting it deployed into the kube environment, at least that is the goal for me. | 23:08 |
| *** eaguilar has quit IRC | 23:08 | |
| kfox1111 | bmace_: +1 :) | 23:08 |
| sdake | hogepodge where we re jammed up is we dont know how to yet turn that into an official thing and who would do the hosting | 23:09 |
| sdake | hogepodge for the moment we are working on making it "less kolla centric" and more "what does openstack need out of kubernetes and helm" centric | 23:09 |
| hogepodge | sdake: yeah, I've heard a little bit about it. | 23:09 |
| sdake | hogepodge its open to anyone - was on the mailing list last year | 23:09 |
| *** krtaylor has joined #openstack-kolla | 23:09 | |
| *** matrohon has quit IRC | 23:10 | |
| *** sacharya has quit IRC | 23:12 | |
| sbezverk | bmace_: deploying things in container is great when it is working, but for debugging it adds an extra layer of complexity and dependency.. | 23:13 |
| *** msimonin1 has joined #openstack-kolla | 23:13 | |
| *** huzhengchuan has joined #openstack-kolla | 23:13 | |
| sbezverk | if you container engine is screwed up | 23:13 |
| sbezverk | conatiner is not going to hel much ;-) | 23:13 |
| bmace_ | sbezverk: it hasn't been too bad really. i just swapped the on run of the container to a forever sleep and exec -it into it to do all my running of scripts / debugging | 23:13 |
| sbezverk | bmace_: you must have missed people were not happy with docker 1.13 recently | 23:14 |
| sbezverk | that is what I am tlaking about | 23:14 |
| bmace_ | sbezverk: oh, that, yes I saw a lot of discussion around that. | 23:15 |
| SamYaple | sbezverk: what happened with docker 1.13? | 23:15 |
| sbezverk | SamYaple: you'd beter ask portdirect | 23:15 |
| SamYaple | im running it without issue :/ | 23:15 |
| sbezverk | since he put together a workaround. I think it is somehitng with '-' in repo name | 23:16 |
| bmace_ | SamYaple: i believe there were some helm / tiller issues that popped up because of some 1.13 behavior | 23:16 |
| SamYaple | bmace_: gotcha | 23:16 |
| sbezverk | or something like that but I might be wrong | 23:16 |
| *** msimonin has quit IRC | 23:16 | |
| SamYaple | portdirect_away: ping ^ poke ^ | 23:16 |
| *** lrensing has joined #openstack-kolla | 23:16 | |
| *** mliima has quit IRC | 23:17 | |
| *** sayantani01 has joined #openstack-kolla | 23:18 | |
| *** haplo37 has quit IRC | 23:18 | |
| *** mliima has joined #openstack-kolla | 23:18 | |
| kfox1111 | no. | 23:19 |
| kfox1111 | the heart of the issue is: | 23:19 |
| kfox1111 | docker 1.13 became more strict with the format retured from docker-registry servers. | 23:19 |
| kfox1111 | since there wasn't a spec for the protocol, all the third party repo's didn't implement it exatly as docker-registry did, | 23:20 |
| kfox1111 | and 1.13 broke 3rd party repo's. :/ | 23:20 |
| SamYaple | with regard to naming? | 23:20 |
| SamYaple | or something else? | 23:20 |
| sbezverk | kfox1111: was it not with somehitng in repo name? | 23:20 |
| kfox1111 | and aparently, docker never tried to pull from google's repo or some of the others to validate. | 23:20 |
| *** schwicht_at_work has joined #openstack-kolla | 23:20 | |
| kfox1111 | SamYaple: no, it was a server http return field that became more strict. | 23:20 |
| SamYaple | kfox1111: i think they talked about that a while ago | 23:22 |
| SamYaple | it was part of there backwards compat effort | 23:22 |
| kfox1111 | the review I saw seemed to indicate they might have known it was a problem but pushed forward anyway... | 23:22 |
| SamYaple | so now client versions will be a bit looser (older clients with newer daemon and vice versa) | 23:22 |
| kfox1111 | can't find the issue now though. :/ | 23:22 |
| *** sayantani01 has quit IRC | 23:23 | |
| sbezverk | kfox1111: this one https://github.com/docker/docker/issues/30083 | 23:23 |
| sbezverk | ? | 23:23 |
| *** schwicht has quit IRC | 23:23 | |
| kfox1111 | https://github.com/kubernetes/helm/issues/1838 | 23:23 |
| kfox1111 | was the root, but its nested several down.... | 23:24 |
| kfox1111 | sbezverk: yeah. thats the one. | 23:24 |
| *** schwicht has joined #openstack-kolla | 23:25 | |
| *** haplo37 has joined #openstack-kolla | 23:27 | |
| *** schwicht_at_work has quit IRC | 23:28 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Local registry in gate for trunk builds https://review.openstack.org/418550 | 23:31 |
| *** dave-mccowan has joined #openstack-kolla | 23:34 | |
| kfox1111 | sbezverk: there is an issue with the openvswitch service package. | 23:34 |
| *** schwicht has quit IRC | 23:34 | |
| kfox1111 | the dep is vswitchd_network. | 23:34 |
| kfox1111 | but if the user sets type=compute, then it has no deps. | 23:34 |
| sbezverk | kfox1111: values.yaml covers only gate scenario when network is used | 23:36 |
| kfox1111 | yeah, but if it was moved up a level, then it would workf or both type=compute and type=network. | 23:37 |
| kfox1111 | the value is just too specific. | 23:37 |
| sbezverk | kfox1111: user just need to add section for compute | 23:37 |
| kfox1111 | the deps will be the same both ways. seems like a waste to force the user to set deps when just changing the type. | 23:37 |
| sdake | hogepodge its Chris Hoge and David Flanders correct? | 23:38 |
| sbezverk | I do not think itis that easy as it would add dependency to ovsdb | 23:39 |
| sdake | hogepodge I know your name is chris, and I think its spelled hoge :) | 23:39 |
| sbezverk | kfox1111: ^^ | 23:39 |
| sdake | hogepodge jsut want ot make sure so lachie can get the names correctly in the invitation | 23:39 |
| kfox1111 | sbezverk: should be able to set it at global.kolla.openvswitch.vswitchd.all so that it affects all vswitchd's but not db's. | 23:40 |
| sbezverk | kfox1111: hm, I do not think we have it in search path | 23:40 |
| sbezverk | actually you are right, we do have it | 23:41 |
| kfox1111 | I thoguth we did. but if not, we probbaby shoudl add it. | 23:41 |
| sdake | hogepodge well if I get it wrong, sorry about that :) | 23:41 |
| sbezverk | kfox1111: do you want to fix it? | 23:42 |
| kfox1111 | I've got a lot of ps's I'm juggling already. if you have a moment, could you do it please? | 23:42 |
| kfox1111 | this trunk build one is going to be the death of me. :/ | 23:42 |
| *** mliima has quit IRC | 23:42 | |
| sbezverk | kfox1111: sure no worries | 23:42 |
| kfox1111 | thanks. :) | 23:43 |
| *** mliima has joined #openstack-kolla | 23:43 | |
| kfox1111 | trying to get the trunk one in, so I can reuse the infra to cache the 2.x and 3.x containers too, to make the gate more stable. | 23:43 |
| *** trey_ has joined #openstack-kolla | 23:43 | |
| kfox1111 | it was really close until we hit a neutron cli change. :/ | 23:43 |
| kfox1111 | then it went sideways really quick. :/ | 23:44 |
| *** sdake has quit IRC | 23:45 | |
| openstackgerrit | Marcus Williams proposed openstack/kolla: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 23:46 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Fixing location of dependency for vswitchd https://review.openstack.org/424902 | 23:47 |
| hogepodge | Got it, thanks. Was out walking the dog (and myself) | 23:50 |
| kfox1111 | arg.... looks like the openstack cli may be doing something different, the other way, in older versions. :/ | 23:51 |
| kfox1111 | so basic_tests may have to know version now. :/ | 23:51 |
| *** schwicht has joined #openstack-kolla | 23:53 | |
| *** schwicht has quit IRC | 23:58 | |
| *** harlowja has joined #openstack-kolla | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!