Sunday, 2017-02-05

sdake_that almost broke the bot lol00:03
sdake_bmace_ you probably could have compressed those into one commit - but its a done deal now00:03
sdake_gerrit pokey because ofdaily 00:00 utc backup00:04
bmace_sdake_: ah, yeah.. i have a couple more to go.. i will try to at least compress into per-service.00:04
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Neutron-dhcp-agent for devenv  https://review.openstack.org/42931500:05
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Neutron-l3-agent for devenv  https://review.openstack.org/42931600:05
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Neutron-metadata-agent for devenv  https://review.openstack.org/42931700:05
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Neutron-openvswitch-agent for devenv  https://review.openstack.org/42931800:05
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Neutron-sever-haproxy for devenv  https://review.openstack.org/42931900:05
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Neutron-server for devenv  https://review.openstack.org/42932000:05
sdake_soundsgood bmace_00:05
bmace_sdake_: going to wait an hour or so to let the gate chill for the last 400:06
bmace_horizon / iscsid / nova and tgtd00:06
sdake_bmace_ might a well get em in, the gate is not busy atm00:07
bmace_kk00:07
*** david-lyle has joined #openstack-kolla00:07
kfox1111sdake_: its like almost a dozen vm's per ps.00:08
sdake_kfox1111 it is too late now00:08
sdake_the patches are submitted00:08
kfox1111well, the last 4.00:08
sdake_anyway infra can handle 4k jobs per hour00:08
bmace_should be a lot quieter on a weekend at least00:09
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for MariaDB for devenv  https://review.openstack.org/42929600:09
sbezverkwhat are you doing?? did you check that all these configmaps have host specific info?????00:09
sbezverklike the creation date uid etc00:09
sbezverknot mentioning host specific interface names00:10
kfox1111sbezverk: uploading prebuilt stuff for devenv. sinc ethe devenv is static, having static stuff in it shoudl be ok.00:10
kfox1111at the ptg, we can discuss what the proper way to do configmaps with everyone,00:11
kfox1111and once we decide what that should look like,00:11
sbezverkI am sorry kfox1111 but I am not going to ack these until they sanitized00:11
kfox1111we can divy up the work and transform them into something more generic/reusable outside the devenv.00:11
sbezverkat the very leat metadat needs to be cleaned up00:11
sdake_ok - well we can sanitize them in a new commit00:11
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Nova for devenv  https://review.openstack.org/42932100:11
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Horizon for devenv  https://review.openstack.org/42932200:11
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Iscsid for devenv  https://review.openstack.org/42932300:11
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked Configmap for tgtd for devenv  https://review.openstack.org/42932400:11
bmace_done and done00:11
*** david-lyle has quit IRC00:12
kfox1111bmace_: cool thanks for the help.00:12
kfox1111sbezverk: yeah. thats a good idea.00:12
kfox1111lets merge them, then make a ps that cuts those lines out?00:12
sdake_agreed00:12
sdake_we are already pressuring the gate with that set - and there is no wy to undo that00:13
sbezverkwell, two of you can merge, I do not support it00:13
sdake_i've seen patches merge in kolla with 5 -1's and 2 +2's before :)00:13
sdake_btw, where are all the releasenotes?00:15
kfox1111sbezverk: yeah. I get your concern. and I appreciate it. what can we do to aleviatge it?00:15
sdake_for features00:15
sdake_kfox1111 i think the solution you proposed is fine00:16
kfox1111we seemed to have some miscommunication about how we were going to do the work, so it ended up as a whole ps chain, so would be really difficult to fix without doing a huge amount of rebases too.00:16
kfox1111so if we can do it right after merging, would that eleviate the concern?00:16
sbezverkyou do not need me to move forward.. but I want my -1 to be noted as the way it is done is completely wrong..00:17
sdake_sbezverk - we can always change it00:18
kfox1111ah. a procedural objection?00:18
kfox1111yeah. I get that.00:18
sdake_sbezverk that is the beauty of source code :)00:18
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for RabbitMQ for devenv  https://review.openstack.org/42929700:18
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for memcached for devenv  https://review.openstack.org/42929800:18
sbezverksdake_: Oh really ?!?!?! -1 for ws is ok, but -1 for what you did it not ok give me a break00:19
sdake_for ws?00:19
sbezverkwhite space00:19
sdake_i didnt' do that patch00:19
sdake_bmace_ did00:19
sdake_anywy lets move on00:19
openstackgerritSteven Dake proposed openstack/kolla-kubernetes master: Remove minikkube guide  https://review.openstack.org/42932500:21
bmace_huh? what whitespace?  in the generated configmaps?  or my containerized devenv review?00:22
sdake_a -1 indicates more work is needed00:23
sdake_the -1 is noted sbezverk - someone will sanitize whatever bmace proposed as an additional patch00:23
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for keepalived for devenv  https://review.openstack.org/42929900:23
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Keystone for devenv  https://review.openstack.org/42930000:23
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for resolv-conf for devenv  https://review.openstack.org/42930100:23
sdake_in this csae, there wer a slew of patches submitted which rebasing is not reasonable00:24
sdake_this could use an ack: https://review.openstack.org/#/c/429325/00:24
*** Pavo has quit IRC00:24
sdake_wierd -t here are ducks flying south and its the middle of winter00:25
sdake_sbezverk the helm meta chart looks good00:32
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Openvswitch-db-server for devenv  https://review.openstack.org/42930200:33
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for Openvswitch-vswitchd for devenv  https://review.openstack.org/42930300:34
bmace_sbezverk: so my plan is to abandon the reviews that aren't merged, go through all of the ones that aren't committed and even the ones that are, clean up all the metadata to include only name, and push up reviews per-service rather than the finer granularity they were at originally.  sound ok?00:35
*** rhallisey has quit IRC00:35
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute-kit chart  https://review.openstack.org/42815000:36
sbezverkbmace_: sorry for all trouble but I think it is the right thing to do00:36
bmace_sbezverk: not a big deal really.  isn't that much work :)00:37
bmace_sbezverk: happy to have things done the right way.00:37
sbezverkbmace_: thanks, appreciate it00:37
kfox1111sbezverk: yeah, its good. thats why we have reviewers.00:37
*** rhallisey has joined #openstack-kolla00:38
sdake_sup rhallisey00:38
rhalliseyyo00:38
rhalliseywhat's happenin00:38
sdake_ready for dinner00:38
kfox1111sbezverk: you still have the missing lines in all_values.yaml00:38
sbezverkkfox1111: I approached it from different angle00:39
sdake_this time margaritas :)00:39
sdake_have big ass superbowl party tomrorow00:39
sbezverkI modified service's values.yaml00:39
sdake_100 people orso00:39
kfox1111sbezverk: it will fail microservices then.00:39
sbezverkbut whether it works or not, not sure00:39
sdake_thank god i'm not hosting it :)00:39
kfox1111being in all_values has it working for both service packages and microservice packages.00:40
sdake_at some point we will need a gate job for the meta package00:40
sbezverkkfox1111: I am lost a bit.. you asked to remove element_name from all.yaml00:40
kfox1111sdake_: yeah, we wer talking about that in the review.00:40
sdake_i know i saw it00:41
kfox1111sbezverk: sorry, I think we miscommunicated there.00:41
sdake_"what we have here, is a failure to communicate" :)00:41
kfox1111sbezverk: I intended to say, we should revert the changes in all_values.yaml.00:41
kfox1111so, no changes needed in either all_values.yaml or service's at all.00:42
kfox1111and just the element_name overrides in the compute-kit values.yaml.00:42
sbezverkkfox1111: but that option failed00:42
kfox1111I don't think it did.00:42
kfox1111you tested with removing the lines out of all_values.yaml.00:42
sbezverkit did, the previos PS helm failed to render mariadb00:42
kfox1111but not with having the unchanged file.00:42
sbezverkoh I must have then removed more. let me double check00:43
kfox1111so, if you thakne the current ps,00:43
kfox1111and git diff x | patch -p1 -R  on:00:44
kfox1111helm/all_values.yaml00:44
kfox1111helm/service/mariadb/values.yaml00:44
kfox1111helm/service/memcached/values.yaml00:44
kfox1111I think it will work.00:44
sbezverkI see, but I used to have original yaml, and that was when the problem started00:44
kfox1111I think that was before we firgured out the right changes for helm/compute-kits/compute-kit-0.4.0/values.yaml ?00:45
sbezverkthat was exectly the case, I have not changed any values.yaml when I got failer00:45
sbezverkok fine I will restore all.yaml and service's yaml and give it a run00:45
*** Pavo has joined #openstack-kolla00:45
kfox1111thx.00:46
*** Pavo has quit IRC00:49
*** salv-orlando has joined #openstack-kolla00:51
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute-kit chart  https://review.openstack.org/42815000:53
sbezverkkfox1111: ok so now there is no any changes to all.yaml and either of services..00:53
sdake_ok dinner time00:54
sdake_bbl00:54
sdake_or maybe never :)00:54
kfox1111sbezverk: ok. lets see what happens.00:54
*** salv-orlando has quit IRC00:55
sbezverkkfox1111: while we are waiting, I notices that helm_prebuild and helm_build are essentially the same for microservice/service/compute_kit, would you be ok to make them generic and select between three posibilities via command args?00:56
kfox1111sounds good to me. :)00:56
sbezverkkfox1111: great.. less code ;)00:58
kfox1111yeah. I'm all for removing redundant code. :)00:58
kfox1111so, as for the version thing.01:03
kfox1111can we talk about that a little more?01:03
kfox1111are you thinking there will be other ones, like with/without heat that kind of thing,01:04
kfox1111or trying to support the 0.4.0 release with trunk?01:04
kfox1111or something else?01:04
kfox1111just trying to understand your use cases.01:04
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Cinder in devenv  https://review.openstack.org/42933001:05
bmace_sbezverk: can you take a look at that?  more like how you want it?01:06
bmace_sbezverk: all configmap data other than name that should be auto-generated is now nuked01:06
bmace_and commit is per-service.01:06
sbezverkbmace_: man that looks much better :) kfox1111: there is one thing, how you are going to address hardcoded password in openstack config files? all dev installation will use the same passwords?01:08
kfox1111sbezverk: yeah. for now. we can work on templating that bit out asap.01:08
kfox1111or users can still use genconfig if they really want secure.01:09
bmace_sbezverk: thanks for the feedback.. i think i have a few more to fix up / push out.01:09
bmace_kfox1111: sbezverk: yeah, for devenv I don't expect the security side to be a big deal at all.01:09
sbezverkbmace_: this one looks good, can I ask you if you see hardcoded interface name like enp13s0, replace it with eth001:09
sbezverketh0 I think will be more common and will cover more possible cases01:10
bmace_yup, i agree.. looks like only keepalived and openvswitch refer to that..01:10
bmace_so that should be easy01:10
sbezverkgreat, thanks a lot!01:10
bmace_all my other devenv work has been based on a centos base01:10
bmace_which uses the ethX nomenclature01:11
kfox1111sbezverk: in the dev env its not eth0 though.01:11
*** furlongm has quit IRC01:11
kfox1111once we can get it up and running, then we can template that out so it can run under other envs.01:11
sbezverkkfox1111: which will it be then?01:11
kfox1111sbezverk: halcyon.01:12
kfox1111they are working on the docs to stand it all up in vm's using halcyon.01:12
bmace_well, halcyon using centos is ethX based vs. enps based01:12
sbezverkkfox1111: right but in halcyon I think they use eth0 eth101:12
bmace_halcyon on ubuntu is enpsX01:12
kfox1111fun.01:12
kfox1111so either we need to make the docs always use ubuntu,01:13
bmace_i sort of wish that the halcyon stuff only supported centos :)01:13
sbezverkbmace_: damn ;) no consistency01:13
kfox1111or we need to make that configurable.01:13
kfox1111there's no reason we couldn't specify only one in the first pass at thte docs.01:13
sbezverkkfox1111: I think using the same interface on centos and ubunutu is not too much to ask01:13
bmace_right now all my devenv container stuff is based on centos but other than the ethX stuff probably works ok in either env. but will need some testing01:13
kfox1111k. so,l lets default to centos then, hard code it,01:14
kfox1111and then plan on making it configurable to support ubuntu soon?01:14
sbezverkkfox1111: +101:14
bmace_kfox1111: that would be my pref also for the time being01:14
*** furlongm has joined #openstack-kolla01:15
kfox1111ok. cool.01:15
openstackgerritjimmygc proposed openstack/kolla master: Use install-pip macro in most source images  https://review.openstack.org/41386101:18
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Neutron in devenv  https://review.openstack.org/42933101:22
*** furlongm has quit IRC01:23
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Glance in devenv  https://review.openstack.org/42933201:25
sbezverkkfox1111: it fails with no changes to all.yaml.. I think it is helm bug.. Error: release compute-kit-0.4.0 failed: services "rabbitmq" already exists01:30
kfox1111looking...01:31
sbezverkas rabbitmq/mariadb and memcached have element_name at the same level01:31
sbezverkwhen they get merged only one value survive and is used for all three services01:31
sbezverkhence duplicate name error01:31
*** furlongm has joined #openstack-kolla01:32
sbezverksince it is a special chart all values gets computed into the same table01:33
kfox1111global.element_name shouldnt be used at all though. thats the thing confusing me.01:33
sbezverkso to separate them we need to keep them at lower levels01:33
kfox1111yeah. but the upper layers value shoudln't matter at all.01:34
kfox1111let me look...01:34
sbezverksure, maybe I am missing something01:34
kfox1111mariadb-statefulset has: {{- $elementSearchPath := ":global.kolla.mariadb.statefulset:global.kolla.mariadb.all:global.kolla.all:global" }}01:35
kfox1111so global.kolla.mariadb.all.element_name should override global.element_name....01:35
kfox1111maybe theres template with a different path...01:35
sbezverkkfox1111: right but we do not have it set01:35
kfox1111should be set in computekit/values though?01:36
sbezverkthe only element_name we have is at global. level01:36
kfox1111these: https://review.openstack.org/#/c/428150/25/helm/compute-kits/compute-kit-0.4.0/values.yaml01:36
sbezverkhmm, it is there01:36
kfox1111those should be searched in the path first.01:37
sbezverkwell, I can do --debug and dry-run01:37
sbezverkthen we can figure out where it comes from01:37
kfox1111k. sounds good.01:37
kfox1111I'll keep digging.01:37
sbezverkcause now even if I do --debug it will not generate anything useful01:37
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute-kit chart  https://review.openstack.org/42815001:39
kfox1111its always complaining about an svc too, right?01:40
sbezverkonly svc01:40
kfox1111hm...01:40
sbezverkbut maybe it is because it stops at first error found01:41
kfox1111yeah. could be.01:41
kfox1111svc's are pretty early.01:41
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Nova in devenv  https://review.openstack.org/42933401:41
*** zhurong has joined #openstack-kolla01:42
kfox1111I never really tested it, but theres another possibility. that global.X works the same way as X.01:43
sbezverkkfox1111: bmace_: we also must include the same secrets, otherwise it will not work.01:45
kfox1111yeah.01:45
kfox1111or capture the passwords.yaml and use the secret-generator.py.01:46
kfox1111and/or.01:46
bmace_kfox1111: sbezverk: sure.. if the plan is to have some pre-built stuff probably better to have the generated secrets versus requiring they generate them though?01:46
kfox1111yeah. true.01:47
sbezverkbmace_: +1 for prebuilt then01:47
kfox1111the search paths seem to be ok...01:48
kfox1111building the ps locally...01:49
*** salv-orlando has joined #openstack-kolla01:52
sbezverkkfox1111: bmace_: family calls, will check back in about an hour will ack new ps then if sdake will not do it earlier01:53
kfox1111k. thanks. :)01:54
bmace_sbezverk: thanks much.  most are in other than 2 or 3 at this point.  i need to go through all the ones that were merged already and fix them up too.01:54
*** salv-orlando has quit IRC01:56
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Fixed Pre-Baked k8s Configmaps to remove metadata  https://review.openstack.org/42933501:58
kfox1111bmace_: I guess the other possibility, is if there isn't too much difference between centos/ubuntu in the configs,02:08
kfox1111we could template just that little bit out for now.02:08
bmace_kfox1111: expecting most of it will be network related.. though i never did the ceph config on ubuntu.  not sure if there are differences there also.02:11
kfox1111yeah. I'm sure there will be some differences there. :/02:12
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Cinder in devenv  https://review.openstack.org/42933002:15
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Neutron in devenv  https://review.openstack.org/42933102:15
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Glance in devenv  https://review.openstack.org/42933202:15
openstackgerritMerged openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for Nova in devenv  https://review.openstack.org/42933402:15
*** jmontleon has quit IRC02:17
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Optimize reconfigure action for horizon  https://review.openstack.org/42271902:24
openstackgerritMerged openstack/kolla master: Remove all kolla-ansible related docs  https://review.openstack.org/42574902:27
openstackgerritMerged openstack/kolla-ansible master: Integrate designate with neutron  https://review.openstack.org/42467902:36
openstackgerritJeffrey Zhang proposed openstack/kolla master: Remove include_header and include_footer in all Dockerfiles  https://review.openstack.org/42801702:45
kfox1111sbezverk: uh, seeing something very strange.02:46
*** zhubingbing_ has joined #openstack-kolla02:46
kfox1111in the generated values, I see a global.kolla.all.element_name: rabbit02:47
kfox1111no where in the source base, can I find anyone setting an elmeent_name at that level.02:48
kfox1111oh. there's one....02:50
kfox1111rabbitmq has a global.kolla.all.element_name....02:50
zhubingbing_ping  kfox111102:51
kfox1111zhubingbing_: hey02:51
zhubingbing_ kolla-K8S denpend kolla-ansible ?02:51
kfox1111zhubingbing_: currently.02:52
kfox1111for configuration generation.02:52
zhubingbing_so we should install kolla-ansible ?02:52
kfox1111yeah.02:52
zhubingbing_ok02:52
zhubingbing_understand02:52
zhubingbing_thanks02:52
kfox1111we're working on removing it as a dep.02:52
zhubingbing_i think we should remove it02:52
kfox1111but it was much easier to bootstrap using kolla-ansible.02:52
*** salv-orlando has joined #openstack-kolla02:53
kfox1111so we didn't have to do everyting from the begining.02:53
kfox1111now that the rest of the stuff is looking pretty good, we can look at that part.02:53
zhubingbing_now how to do ?02:54
kfox1111how to do what?02:54
zhubingbing_bootstrap02:55
kfox1111ah.02:55
kfox1111bmace_: and sdake_ are working on some install docs.02:56
zhubingbing_ok。02:56
kfox1111the ones in doc/ are quite out of date. :/02:56
*** salv-orlando has quit IRC02:57
zhubingbing_i install kolla-k8s yesterday, doc is very good02:57
zhubingbing_;)02:57
Jeffrey4lfyi: infra team have increased the post job timeout from 10 min to 30 min. then centos+binary image should be uploaded successfully02:59
kfox1111zhubingbing_: which doc did you use?02:59
zhubingbing_http://docs.openstack.org/developer/kolla-kubernetes/03:00
kfox1111which branch of those docs did you use?03:01
kfox1111the Dev environment?03:01
zhubingbing_no03:01
zhubingbing_not dev environment03:01
zhubingbing_http://docs.openstack.org/developer/kolla-kubernetes/host-setup.html03:02
zhubingbing_Start with this document03:03
kfox1111ah.03:03
kfox1111but haven't tried to launch kolla-kubernetes yet?03:03
kfox1111just the prep stuff?03:04
zhubingbing_i began contribute kolla-K8S03:04
zhubingbing_began working03:04
zhubingbing_so i began  study kolla-k8s03:05
*** qiliang27 has joined #openstack-kolla03:05
sbezverkkfox1111: I think it one comes from all.yaml03:06
kfox1111zhubingbing_: the docs are pretty old unfortunately past that point.03:07
kfox1111 the gate is the most complete instructions to install. but is source code. :/03:07
kfox1111currently.03:07
kfox1111tools/setup_gate.sh.03:07
zhubingbing_ok03:07
zhubingbing_1, the current kolla-k8s completed a number of projects (keystone nova)03:07
zhubingbing_2, how to do03:07
zhubingbing_thanks03:07
kfox1111but bmace_ and sdake_ are working on getting the docs update.03:07
zhubingbing_i should read gate03:07
kfox1111updated.03:07
kfox1111yeah. we have keystone,glance,cinder,nova,neutron, and horizon working now.03:08
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Updated pre-built k8s configmaps for vagrant-centos  https://review.openstack.org/42933803:09
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmaps for ceph storage in devenv  https://review.openstack.org/42933903:09
openstackgerritBorne Mace proposed openstack/kolla-kubernetes master: Pre-Baked k8s Configmap for horizon in devenv  https://review.openstack.org/42934003:09
zhubingbing_hmm03:09
kfox1111heat's really close to working too.03:09
zhubingbing_good working03:09
zhubingbing_;)03:09
kfox1111been a team effort. :)03:09
bmace_kfox1111: sdake_: that should be it other than working out how we want to handle secrets / passwords.  all those are based on a working centos halcyon-vagrant deployment.03:09
sdake_zhubingbing_ reading gate will take you 30-40 hours - that is what it took me03:10
sdake_zhubingbing_ i have some notes to get you going quickly - moment03:10
kfox1111sdake_: it would be nice to finish cleaning up the gate too. a lot of forked files.03:10
kfox1111would be more understandable if we merged them back together.03:10
zhubingbing_thanks sdake03:11
sdake_zhubingbing_ http://paste.openstack.org/show/597726/03:11
*** crushil has joined #openstack-kolla03:12
sdake_zhubingbing_ i am going to rework the quickstart and turn thoseinto some basic docs03:12
zhubingbing_roger03:12
sdake_kfox1111 i dont follow03:12
sdake_kfox1111 the gate isworking and is complex to understand - I want to get people started on dev, not on learning the mysteries of the gate ;)03:13
zhubingbing_Kolla-K8S log  have  done ?03:13
sdake_zhubingbing_ do you mean fluentd?03:13
zhubingbing_yes03:13
sdake_zhubingbing_ that is tbd, would like it sooner rathe then later :)03:14
zhubingbing_ok03:14
kfox1111sdake_: I agree. just saying, the gate would e more understandable if it didn't fork to multiple copies of esentially the same code in places.03:14
kfox1111sbezverk: I think they do something odd with values sideways between subcharts.03:15
kfox1111sbezverk: can you remove the element_name line out of helm/service/rabbitmq/values.yaml and see what happens?03:16
bmace_kfox1111: it does feel like there is a lot of duplication in the tests/bin code with just minor differences between some of them.. like iscsi vs. ceph in some of them with most of the rest being the same.. hate to have a lot of logic in bash though :/03:16
sbezverkkfox1111: ok03:17
kfox1111bmace_: yeah. that duplication is a pain. :/03:17
sdake_on my quest for multinode by the 15th setup...03:18
sdake_sbezverk had indicated that there must be a frontend outside kubernetes that figures out where nova-api, cndier-api, etc run03:18
sdake_however, kubernetes has a built in proxy mechanism03:18
openstackgerritMerged openstack/kolla-kubernetes master: Fixed Pre-Baked k8s Configmaps to remove metadata  https://review.openstack.org/42933503:18
*** dave-mccowan has joined #openstack-kolla03:18
sdake_what is more accurate?03:19
openstackgerritSerguei Bezverkhi proposed openstack/kolla-kubernetes master: Adding compute-kit chart  https://review.openstack.org/42815003:19
*** dtturner has joined #openstack-kolla03:19
sdake_sup dave-mccowan03:19
kfox1111sdake_: the proxy really isn't intended for that.03:19
kfox1111making sure one of your nodes has an external_ip associated with it is the easiet way I've found.03:19
sbezverkkfox1111: pushed ps, will check back later to see the result.03:20
sdake_kfox1111 mind expanding - need more details03:20
kfox1111sbezverk: cool. thanks.03:20
sdake_assuming we have 3 nodes labeled as controllers, how on earth are we supposed to know which one to tlak to?03:20
kfox1111sdake_: each svc can be exported externally.03:20
sdake_right - so now you have 3 external ip addresses03:21
kfox1111k8s leaves how the ip gets external as an excersize to the reader.03:21
sdake_then what?03:21
kfox1111you add a 4th. :)03:21
kfox1111vip.03:21
sdake_and that vip is managed how?03:21
kfox1111then put like keepalived on the nodes so the vip can float between them on failure.03:21
sdake_why not put keepalived inside kubernetes itself?03:22
kfox1111there is one we provide, but not well tested yet.03:22
kfox1111yeah, thats how we do it.03:22
sdake_sbezverk had indicated keepalived should be external03:22
kfox1111managed by k8s. runs on the nodes you select for it.03:22
kfox1111runs with net=host.03:23
sdake_so dedicated keepalived nodes?03:23
kfox1111not strictly nessisary.03:23
sdake_can keepalived nodes run alongside the control nodes?03:24
kfox1111yeah.03:24
sdake_rather keepalived service03:24
sdake_sbezverk thoughts on the above?03:24
kfox1111k8s doesnt' caare where the vip lands, so long as its on a node.03:25
sdake_what does the vip connect to?03:25
kfox1111sbezverk: on my local box, that line removal seemed to help.03:25
kfox1111still investigating.03:26
kfox1111still a little puzzled why it had that effect though.03:26
kfox1111sdake_: k8s has something called kube-proxy03:28
kfox1111it inspects k8s svc's, and writes iptables rules for the host its running on.03:28
kfox1111any svc's marked external get rules that map from the external ip to the pool.03:29
kfox1111the rules don't have any effect unless the ip referenced though actually exists on the host.03:29
*** qiliang27 has quit IRC03:30
*** hfu has joined #openstack-kolla03:32
sdake_kfox1111 which pol?03:33
sdake_pool03:33
*** hfu has quit IRC03:35
kfox1111the pods attached to that svc03:35
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Optimize reconfigure action for neutron  https://review.openstack.org/42450403:36
kfox1111well, attached might not be the right word. the pods that the selector on the svc matches.03:36
*** hfu has joined #openstack-kolla03:36
*** dave-mccowan has quit IRC03:36
dtturnerHey folks.  Running kolla-ansible trunk and having issues adding compute nodes using --limit described here: https://review.openstack.org/#/c/398313/.    We already have this patch, but alas still hitting issue.03:37
dtturnerIdeas?03:37
kfox1111sdake one of the things I was hoping to do at some point was to document a few reference archetectures.03:41
*** hfu has quit IRC03:45
*** salv-orlando has joined #openstack-kolla03:53
sdake_dtturner i'd suggest mailing the mailing list as its the weekend - and the superbowl weekend at that03:53
sdake_dtturner i don't knwo the answer to your Q, and doubt anyone will be around to answer it until monday03:54
sdake_dtturner another option is to use ask.openstack.org03:54
*** sdake_ is now known as sdake03:54
dtturnerGood point.  Been so busy I just realized it was SB weekend. Didn't even know who was playing.03:55
dtturnerThanks sdake03:55
*** hfu has joined #openstack-kolla03:55
*** salv-orlando has quit IRC03:58
sdakedtturner me either - i dont follow sports-ball although i am going to a party tomrorow03:58
sdakedtturner my generation has wasted years of their lives individually watching sports03:59
kfox1111sbezverk: thats a weird result.03:59
kfox1111sbezverk: looks like it did solve the issue.03:59
sdakedtturner my chidren's generation has wasted years of their lives already watching videos of other people playing video games03:59
kfox1111but it faild in cinder-api. but cinder-api is in the endpoint list.03:59
kfox1111I think that really may be a k8s bug.04:00
kfox1111a pod shouldn't be in the svc if its initing.04:00
sdakedtturner sometimes I just want to shake my head at the whole state of things ;(04:00
sdakethe rebels are alerted to our presence04:02
sdakei am pretty sure george lucas didn't like digital04:06
sdakesince he killed off DAC in empire strikes back04:06
*** huzhengchuan has joined #openstack-kolla04:10
dtturnersdake Preaching to the choir.  My kids have never known cable television.04:10
dtturner;-)04:10
*** hfu has quit IRC04:11
sbezverkkfox1111: it looks better, so I guess that line was screwing things up04:19
*** rhallisey has quit IRC04:21
*** zhurong has quit IRC04:38
*** sdake_ has joined #openstack-kolla04:40
*** sdake has quit IRC04:41
*** Jeffrey4l has quit IRC04:46
*** salv-orlando has joined #openstack-kolla04:54
*** david-lyle has joined #openstack-kolla04:56
*** salv-orlando has quit IRC04:59
*** david-lyle has quit IRC05:01
*** crushil has quit IRC05:04
*** sdake_ has quit IRC05:05
*** hfu has joined #openstack-kolla05:05
*** Jeffrey4l has joined #openstack-kolla05:12
*** Jeffrey4l has quit IRC05:17
*** Jeffrey4l has joined #openstack-kolla05:18
*** Jeffrey4l has quit IRC05:55
*** david-lyle has joined #openstack-kolla05:57
*** Jeffrey4l has joined #openstack-kolla05:58
*** david-lyle has quit IRC06:01
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Optimize reconfigure action for neutron  https://review.openstack.org/42450406:24
*** jmontleon has joined #openstack-kolla06:46
*** david-lyle has joined #openstack-kolla06:58
*** david-lyle has quit IRC07:02
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: [wip] test scenario2 in gate  https://review.openstack.org/41227907:09
*** l4yerffeJ has joined #openstack-kolla07:09
*** bmace_ has quit IRC07:22
*** bmace_ has joined #openstack-kolla07:22
*** mdnadeem has joined #openstack-kolla07:28
*** huzhengchuan has quit IRC07:34
*** jascott1 has quit IRC07:38
*** jascott1 has joined #openstack-kolla07:38
*** jascott1 has quit IRC07:43
*** huzhengchuan has joined #openstack-kolla07:48
*** salv-orlando has joined #openstack-kolla07:56
*** david-lyle has joined #openstack-kolla07:59
*** salv-orlando has quit IRC08:01
*** david-lyle has quit IRC08:03
*** Jeffrey4l has quit IRC08:22
*** sdake has joined #openstack-kolla08:38
*** sdake_ has joined #openstack-kolla08:54
*** sdake has quit IRC08:57
*** salv-orlando has joined #openstack-kolla08:58
*** david-lyle has joined #openstack-kolla09:00
*** salv-orlando has quit IRC09:02
*** yangzb09 has joined #openstack-kolla09:03
*** david-lyle has quit IRC09:04
*** yangzb09 has quit IRC09:07
*** huzhengchuan has quit IRC09:11
*** zhurong has joined #openstack-kolla09:13
*** laki has joined #openstack-kolla09:16
lakiHello09:17
lakiWhere I fill find qemu.conf stored by kolla09:17
lakiI found few copies09:17
lakiand changes done at /etc/libvirt/qemu.conf doesn't seem to make any change09:17
lakiand that one in /etc/kolla/nova-libvirt/ have only one line09:19
*** huzhengchuan has joined #openstack-kolla09:23
*** sdake_ has quit IRC09:28
openstackgerritJeffrey Zhang proposed openstack/kolla master: Fix panko log folder owner  https://review.openstack.org/42935809:33
openstackgerritJeffrey Zhang proposed openstack/kolla master: add errexit to all extend_start.sh  https://review.openstack.org/42936009:40
*** hrw has quit IRC09:41
*** hrw has joined #openstack-kolla09:43
*** salv-orlando has joined #openstack-kolla09:59
*** mannamne has joined #openstack-kolla10:00
*** david-lyle has joined #openstack-kolla10:00
*** huzhengchuan has quit IRC10:02
*** salv-orlando has quit IRC10:03
*** david-lyle has quit IRC10:05
*** hfu has quit IRC10:09
*** hfu has joined #openstack-kolla10:11
*** hfu has quit IRC10:11
*** hfu has joined #openstack-kolla10:11
*** hfu has quit IRC10:12
*** hfu has joined #openstack-kolla10:12
*** hfu has quit IRC10:13
*** hfu has joined #openstack-kolla10:13
*** hfu has quit IRC10:13
*** huzhengchuan has joined #openstack-kolla10:15
*** qiliang27 has joined #openstack-kolla10:18
*** l4yerffeJ has quit IRC10:19
*** l4yerffeJ has joined #openstack-kolla10:20
*** hfu has joined #openstack-kolla10:27
*** l4yerffeJ has quit IRC10:51
*** l4yerffeJ has joined #openstack-kolla10:51
*** l4yerffeJ has quit IRC10:52
*** l4yerffeJ has joined #openstack-kolla10:54
*** qiliang27 has quit IRC10:55
*** qiliang27 has joined #openstack-kolla10:58
*** zhurong has quit IRC10:58
*** salv-orlando has joined #openstack-kolla11:00
openstackgerritMerged openstack/kolla-ansible master: Ignore qemu-img process in cleanup-containers  https://review.openstack.org/42806111:01
*** david-lyle has joined #openstack-kolla11:01
*** salv-orlando has quit IRC11:04
*** david-lyle has quit IRC11:05
lakianyone out there?11:06
openstackgerritWei Cao proposed openstack/kolla master: Add dragonflow container  https://review.openstack.org/41819311:25
*** qiliang27 has quit IRC11:25
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Use new bootstrap and register method  https://review.openstack.org/42936611:30
openstackgerritWei Cao proposed openstack/kolla-ansible master: Add Freezer doc link in README.rst  https://review.openstack.org/42936711:37
*** david-lyle has joined #openstack-kolla12:02
*** pbourke has quit IRC12:04
*** pbourke has joined #openstack-kolla12:04
*** david-lyle has quit IRC12:06
*** mannamne has quit IRC12:08
*** qiliang27 has joined #openstack-kolla12:21
*** laki has quit IRC12:43
*** rhallisey has joined #openstack-kolla12:44
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Use new bootstrap and register method  https://review.openstack.org/42936612:46
*** qiliang27 has quit IRC12:47
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Fix the WARNING in copy keystone domain task  https://review.openstack.org/42937412:49
openstackgerritMerged openstack/kolla-ansible master: Fix grafana database creation  https://review.openstack.org/42910912:51
*** caowei has joined #openstack-kolla12:51
openstackgerritWei Cao proposed openstack/kolla master: Add dragonflow container  https://review.openstack.org/41819312:52
*** Jeffrey4l has joined #openstack-kolla12:52
*** mannamne has joined #openstack-kolla12:57
openstackgerritMerged openstack/kolla-ansible master: Enable tacker_horizon when enable_tacker is yes  https://review.openstack.org/40782212:57
*** salv-orlando has joined #openstack-kolla13:01
*** david-lyle has joined #openstack-kolla13:03
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Optimize reconfigure action for horizon  https://review.openstack.org/42271913:04
*** salv-orlando has quit IRC13:05
*** david-lyle has quit IRC13:07
openstackgerritzhubingbing proposed openstack/kolla master: test  https://review.openstack.org/42937713:07
openstackgerritzhubingbing proposed openstack/kolla master: Delete the extra blank lines  https://review.openstack.org/42937813:15
openstackgerritMerged openstack/kolla master: Install osprofiler in magnum base container  https://review.openstack.org/42920513:17
*** hfu has quit IRC13:18
openstackgerritWei Cao proposed openstack/kolla master: Install freezer dashboard into horizon image  https://review.openstack.org/42938013:31
openstackgerritWei Cao proposed openstack/kolla-ansible master: Load freezer dashboard  https://review.openstack.org/42938213:31
*** mdnadeem has quit IRC13:41
*** imcsk8 has quit IRC13:54
*** qiliang27 has joined #openstack-kolla13:55
*** imcsk8 has joined #openstack-kolla13:56
*** david-lyle has joined #openstack-kolla14:03
*** pcaruana has quit IRC14:07
*** david-lyle has quit IRC14:08
*** qiliang27 has quit IRC14:14
openstackgerritMerged openstack/kolla-kubernetes master: Unpin kolla-ansible  https://review.openstack.org/42736514:22
*** mannamne has quit IRC14:28
*** Pavo has joined #openstack-kolla14:30
*** hachi has quit IRC14:35
*** sdake has joined #openstack-kolla14:43
*** hachi has joined #openstack-kolla14:44
openstackgerritcaoyuan proposed openstack/kolla-ansible master: Fix the creating haproxy mysql failed  https://review.openstack.org/42939015:01
*** david-lyle has joined #openstack-kolla15:04
openstackgerritMerged openstack/kolla-ansible master: Fix parameter issue in cinder handlers  https://review.openstack.org/42726815:09
*** david-lyle has quit IRC15:09
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Optimize reconfigure action for horizon  https://review.openstack.org/42271915:12
openstackgerritMerged openstack/kolla-ansible master: Add release note regarding nova upgrade  https://review.openstack.org/42913815:18
openstackgerritJeffrey Zhang proposed openstack/kolla-ansible master: Optimize reconfiguration for collectd  https://review.openstack.org/42413315:19
*** eaguilar has joined #openstack-kolla15:34
*** huzhengchuan has quit IRC15:36
*** klindgren__ has joined #openstack-kolla15:37
*** eaguilar has quit IRC15:37
*** klindgren_ has quit IRC15:39
*** hieulq has joined #openstack-kolla15:54
*** Pavo has quit IRC15:56
*** Pavo has joined #openstack-kolla15:57
*** sdake_ has joined #openstack-kolla16:03
*** Pavo has quit IRC16:04
*** david-lyle has joined #openstack-kolla16:05
*** sdake has quit IRC16:05
*** david-lyle has quit IRC16:10
*** saneax-_-|AFK is now known as saneax16:11
*** Pavo has joined #openstack-kolla16:16
*** Pavo has quit IRC16:23
*** skuda has joined #openstack-kolla16:31
*** pcaruana has joined #openstack-kolla16:39
*** salv-orlando has joined #openstack-kolla16:39
*** mgoddard_ has joined #openstack-kolla16:42
*** hieulq has quit IRC16:53
*** adrian_otto has joined #openstack-kolla16:53
*** saneax is now known as saneax-_-|AFK16:59
*** hieulq has joined #openstack-kolla17:00
*** david-lyle has joined #openstack-kolla17:06
*** zhubingbing_ has quit IRC17:10
*** david-lyle has quit IRC17:10
*** mannamne has joined #openstack-kolla17:15
*** mgoddard_ has quit IRC17:26
*** hieulq_ has joined #openstack-kolla17:37
*** hieulq has quit IRC17:38
*** Jeffrey4l has quit IRC17:50
*** Jeffrey4l has joined #openstack-kolla17:50
*** hieulq_ has quit IRC17:57
*** adrian_otto has quit IRC18:02
*** g3ek has quit IRC18:04
*** haplo37 has quit IRC18:05
*** adrian_otto has joined #openstack-kolla18:08
*** adrian_otto has quit IRC18:13
*** g3ek has joined #openstack-kolla18:16
*** haplo37 has joined #openstack-kolla18:17
*** salv-orlando has quit IRC18:19
*** salv-orlando has joined #openstack-kolla18:25
*** hachi has quit IRC18:25
*** salv-orlando has quit IRC18:28
*** hachi has joined #openstack-kolla18:30
*** v1k0d3n has quit IRC18:30
*** mgoddard_ has joined #openstack-kolla18:32
*** v1k0d3n has joined #openstack-kolla18:38
*** sdake_ has quit IRC18:38
*** jascott1 has joined #openstack-kolla18:39
*** jascott1 has quit IRC18:44
*** adrian_otto has joined #openstack-kolla18:45
*** sdake has joined #openstack-kolla18:48
*** sdake has quit IRC18:52
*** sdake has joined #openstack-kolla18:52
*** sdake has quit IRC18:57
*** adrian_otto has quit IRC19:12
*** adrian_otto has joined #openstack-kolla19:14
*** adrian_otto has quit IRC19:16
*** mgoddard_ has quit IRC19:22
*** dtturner has quit IRC19:56
*** dtturner has joined #openstack-kolla19:56
*** schwicht has joined #openstack-kolla20:15
*** pcaruana has quit IRC20:20
*** salv-orlando has joined #openstack-kolla20:24
*** salv-orlando has quit IRC20:28
*** salv-orlando has joined #openstack-kolla20:28
*** skuda has quit IRC20:33
*** rhallisey has quit IRC20:52
*** mgoddard_ has joined #openstack-kolla21:01
*** mgoddard_ has quit IRC21:20
*** Jeffrey4l has quit IRC21:35
*** l4yerffeJ has quit IRC21:35
*** laki has joined #openstack-kolla21:41
lakiHello21:41
lakiShort one21:41
lakiif I will make change - and disable ( for example cinder ) in globals21:41
lakiand then execute kolla-ansible deploy21:42
lakiwill it remove cinder from endpoints ?21:42
*** lamt has joined #openstack-kolla21:43
SamYaplelaki: no21:53
SamYapleno remove operations are implemented, and its unlikely they will be in kolla-ansible21:54
lakithx22:05
*** laki has quit IRC22:05
*** crushil has joined #openstack-kolla22:16
*** eaguilar has joined #openstack-kolla22:25
*** breitz has quit IRC22:28
*** breitz has joined #openstack-kolla22:28
*** adrian_otto has joined #openstack-kolla22:31
*** adrian_otto has quit IRC22:39
*** berendt has quit IRC22:40
*** berendt has joined #openstack-kolla22:40
*** suka_ has quit IRC22:44
*** suka_ has joined #openstack-kolla22:44
*** DuncanT has quit IRC22:44
*** mannamne has quit IRC22:46
*** DuncanT has joined #openstack-kolla22:46
*** eaguilar has quit IRC22:50
*** jrobinson has joined #openstack-kolla23:19
*** lamt has quit IRC23:19
*** zhurong has joined #openstack-kolla23:30
*** zhurong has quit IRC23:35
*** jrobinson has quit IRC23:37
*** crushil has quit IRC23:38
*** jrobinson has joined #openstack-kolla23:38
*** mannamne has joined #openstack-kolla23:39
*** salv-orlando has quit IRC23:42

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!