| *** eaguilar has quit IRC | 00:00 | |
| kfox1111 | inc0: what do you think of this: | 00:02 |
|---|---|---|
| kfox1111 | which pip 2>/dev/null && echo pip && pip freeze | sort; which dpkg 2>/dev/null && echo dpkg && dpkg -l | awk '{print $1, $2, $3, $4}' | sort -k 2; which rpm 2>/dev/null && echo rpm && rpm -qa | sort | 00:02 |
| * inc0 parsing shell | 00:02 | |
| kfox1111 | if container versions without revisions match, compare those two to see if containers are diffrent and revision needs incremented. | 00:03 |
| inc0 | so one thing | 00:03 |
| inc0 | for source builds - I think it's pointless | 00:03 |
| inc0 | commits go in every day | 00:03 |
| kfox1111 | to pip? | 00:03 |
| inc0 | we don't use pip during source build | 00:03 |
| inc0 | not for main servie anyway | 00:03 |
| kfox1111 | in ocata/newton/mitaka? | 00:04 |
| kfox1111 | Trunk, yeah for sure. | 00:04 |
| inc0 | mitaka is eol | 00:04 |
| jascott1 | kfox1111 thanks | 00:04 |
| jascott1 | oh that issue | 00:04 |
| inc0 | so we maybe just build last tag of mitaka and leave it there | 00:04 |
| kfox1111 | jascott1: yeah... a big long issue. :) | 00:04 |
| jascott1 | the 1.6 explosion catch all | 00:05 |
| kfox1111 | jascott1: yeah. :/ but a lot of good detail on what went wrong. | 00:05 |
| jascott1 | was this not more correct than our approach? https://github.com/kubernetes/kubernetes/issues/43815#issuecomment-292341902 | 00:05 |
| kfox1111 | inc0: do you think the stable releases really get updated that frequently? | 00:05 |
| inc0 | kfox1111: https://github.com/openstack/neutron/commits/stable/newton | 00:06 |
| kfox1111 | jascott1: for the permission thing? | 00:06 |
| inc0 | every day | 00:07 |
| jascott1 | ya | 00:07 |
| jascott1 | AllowsAllow over the unauthenticated | 00:07 |
| jascott1 | guess its minor issue for dev env | 00:07 |
| kfox1111 | inc0: interesting.... so really, source should be tracking stable release tags, not the tip of the branch? | 00:07 |
| inc0 | I think we should have option for both | 00:07 |
| inc0 | security patches.. | 00:08 |
| jascott1 | wouldnt want to propagate a less secure anything if there is a better solution within reason | 00:08 |
| kfox1111 | jascott1: yeah. we put the hack in just to get things going again in the gate. totally believe its not the right solution. :) | 00:08 |
| inc0 | tags are every few weeks, you don't want to wait that long | 00:08 |
| *** chas has joined #openstack-kolla | 00:08 | |
| kfox1111 | inc0: +1 to having both. a -source-stable, and a -source-tip or something? | 00:08 |
| kfox1111 | every few weeks would track with binary packages. | 00:09 |
| kfox1111 | agreed that we might not want to deal with revisions and stable-tip's. | 00:09 |
| *** eaguilar has joined #openstack-kolla | 00:10 | |
| inc0 | speaking of which, this need fixing https://github.com/openstack/kolla/blob/stable/ocata/kolla/common/config.py#L442 | 00:11 |
| inc0 | also I took a look at barbican, and 4.0.0 tag is newest from Feb | 00:11 |
| *** rwellum has quit IRC | 00:11 | |
| inc0 | damn our version management needs a lot of work | 00:12 |
| *** chas has quit IRC | 00:13 | |
| inc0 | so we have few things to handle here | 00:16 |
| inc0 | 1- tip of branch, pretty straighforward | 00:17 |
| kfox1111 | inc0: +1. | 00:17 |
| jascott1 | hey kfox1111 what about docker diff on the containers? | 00:17 |
| kfox1111 | the version stuff really hurts. :/ | 00:17 |
| inc0 | 2- latest released tag from repo | 00:17 |
| inc0 | harder stuff :( | 00:17 |
| kfox1111 | jascott1: the concern is, there could be a lot of random things that change that are not important. | 00:17 |
| jascott1 | but we could ignore | 00:18 |
| kfox1111 | yeah. | 00:18 |
| jascott1 | yea im saying add that to what you are doing | 00:18 |
| jascott1 | haha | 00:18 |
| jascott1 | do more work! twwfpt | 00:18 |
| kfox1111 | indeed. :) | 00:18 |
| kfox1111 | "more's always better" :) | 00:18 |
| kfox1111 | hmm.. docker diff does look like it only shows changes between layers? | 00:19 |
| kfox1111 | not sure if it even can be used to compare two differently rooted containers, which would be always the case. | 00:19 |
| jascott1 | ah | 00:19 |
| jascott1 | i forget the complex image building | 00:20 |
| kfox1111 | inc0: so, not considering the stable-tip case, do you think that fingerprinting command would gather what info we would need? | 00:20 |
| inc0 | what I'm concerned with this is that lots of things get changed | 00:21 |
| inc0 | each service has a lot of deps | 00:21 |
| kfox1111 | yeah. but if anyu of them change, it really should be a different container. | 00:21 |
| kfox1111 | tag | 00:21 |
| inc0 | we share base image so if something changes in openstack-base, which holds a lot of deps | 00:22 |
| inc0 | everything gets rebuilt | 00:22 |
| jascott1 | contents hash and date and some diff along the lines kfox1111 is doing might be workable | 00:22 |
| kfox1111 | inc0: ah. yeah, that could probably use a thining out. | 00:22 |
| inc0 | what I'm saying is, if we rebuild and nothing changes, nothing gets uploaded | 00:22 |
| inc0 | because layers in docker are checksums | 00:22 |
| kfox1111 | inc0: yeah. and thats what we need the fingerprinting for. | 00:22 |
| inc0 | so we have "fingerprinting" already | 00:23 |
| jascott1 | hash says its different. date helps figure out what world liked like wehn it was built. and then the change long to say which packages changed | 00:23 |
| kfox1111 | so that if a log file or install thingy or whatever drops a 'date > foo' then it doesn't trigger the container to be different. | 00:23 |
| kfox1111 | inc0: how so? | 00:23 |
| kfox1111 | no a checksum is not enough for a fingerprint. | 00:23 |
| kfox1111 | its way way too touchy. | 00:23 |
| kfox1111 | the container didn't really change if all it did was have date > /var/log/whatever. | 00:24 |
| kfox1111 | we only care if software changed in it. | 00:24 |
| inc0 | well still, base is rebuilt - we rebuild whole set | 00:24 |
| kfox1111 | the docker checksum changes if any byte on any file changes, including stuff we really don't care about. | 00:24 |
| kfox1111 | inc0: true. but how often do we think that will happen? | 00:25 |
| kfox1111 | Its less then ideal, for sure. that can be fixed seperately though. | 00:25 |
| inc0 | not sure, I'd bet on quite often | 00:25 |
| kfox1111 | I'd bet no more frequently then once a week. | 00:25 |
| inc0 | but that's researchable | 00:25 |
| kfox1111 | I'd guess its more like once a month. | 00:25 |
| inc0 | kfox1111: if you can crack out rudimentary manifesto generating shellscript | 00:26 |
| inc0 | we could put it into gates and log it after gate build | 00:26 |
| kfox1111 | inc0: yeah. thats the shell snippet from above. | 00:26 |
| inc0 | gate run | 00:26 |
| kfox1111 | ah. that would be interesting. yeah. | 00:26 |
| kfox1111 | does kolla have a periodic gate for binary container builds yet? | 00:27 |
| inc0 | let's add this to deploy gates run in kolla gates | 00:27 |
| inc0 | every commit also gates on binary | 00:27 |
| kfox1111 | I could stick image building in the kolla-kubernetes periodic gate. | 00:27 |
| inc0 | it's just not voting that's all | 00:27 |
| kfox1111 | well, we need to gather the fingerprints somehow too to mine it. | 00:28 |
| kfox1111 | if it was in a periodic gate, we could upload them to tarballs.o.o. | 00:28 |
| *** eanylin has joined #openstack-kolla | 00:28 | |
| inc0 | infra keeps logs from changes for quite a while | 00:28 |
| kfox1111 | inc0: true. just not sure how to mine it out of review's. | 00:28 |
| inc0 | we just need to look at them and download file output every now and then | 00:28 |
| *** chas has joined #openstack-kolla | 00:29 | |
| inc0 | same as we publish logs... | 00:29 |
| kfox1111 | hmm... yeah, that might work. | 00:29 |
| inc0 | or even dump it to stdout | 00:29 |
| kfox1111 | put it in $LOGS/container_prints or something. | 00:29 |
| *** duonghq has joined #openstack-kolla | 00:29 | |
| inc0 | exactly | 00:29 |
| kfox1111 | then can easily fetch them by filename. | 00:29 |
| duonghq | kfox1111, are you there? | 00:29 |
| inc0 | yup | 00:29 |
| kfox1111 | inc0: k. I'll post a ps for that. | 00:29 |
| inc0 | thanks | 00:29 |
| kfox1111 | duonghq: yeah. whats up? | 00:29 |
| duonghq | hi inc0 kfox1111 | 00:29 |
| inc0 | you are better at gates than myself | 00:30 |
| inc0 | good afternoon Duong | 00:30 |
| kfox1111 | I know nothing about the kolla gate. but can take a stab at it. | 00:30 |
| duonghq | kfox1111, I have a small question about my ps, let me know when you free | 00:30 |
| *** huzhengchuan has joined #openstack-kolla | 00:30 | |
| kfox1111 | duonghq: nows a good time. whats up? | 00:30 |
| inc0 | well I gtg now, but I'll be happy to help tomororrow | 00:30 |
| kfox1111 | inc0: uh... one issue... | 00:31 |
| inc0 | go ahead | 00:31 |
| duonghq | kfox1111, https://review.openstack.org/#/c/437252/9/tests/bin/basic_tests.sh@181 about the last line, what is it for? | 00:31 |
| kfox1111 | that would be against trunk builds. and we know those will change a lot? | 00:31 |
| kfox1111 | or is this just on base? | 00:31 |
| inc0 | kfox1111: good point, we'd need that change in stable branch I guess | 00:31 |
| inc0 | but I think we can do it for ocata at least | 00:32 |
| inc0 | will give us idea anyway | 00:32 |
| kfox1111 | duonghq: it belongs to the test right before your changes. what that test does, is | 00:32 |
| kfox1111 | create a cinder volume on vm1, write a file to it. moves the cinder volume to vm2, and verify the data in it is the same as was written. | 00:32 |
| kfox1111 | the diff does the actual compare of the data written to what was read back out. | 00:33 |
| inc0 | for glory of science, I'll allow this backport:P | 00:33 |
| duonghq | kfox1111, understood, thanks | 00:33 |
| kfox1111 | duonghq: np. :) | 00:34 |
| *** chas has quit IRC | 00:34 | |
| *** tervancovan has joined #openstack-kolla | 00:34 | |
| inc0 | ok, I'm going off | 00:34 |
| inc0 | talk to y'all tomorrow | 00:35 |
| duonghq | inc0, cya | 00:35 |
| *** inc0 has quit IRC | 00:35 | |
| *** huzhengchuan has quit IRC | 00:37 | |
| spsurya | kfox1111: small item for you https://review.openstack.org/#/c/457614/ | 00:37 |
| *** sayantan_ has joined #openstack-kolla | 00:38 | |
| spsurya | duonghq one more https://review.openstack.org/#/c/451654/ | 00:38 |
| kfox1111 | spsurya: looks good. :) | 00:39 |
| spsurya | kfox1111 one more https://review.openstack.org/#/c/451654/ | 00:39 |
| kfox1111 | done. thanks for fixing that. :) | 00:39 |
| kfox1111 | gotta head out for a while. bbl. have a good one. :) | 00:39 |
| spsurya | kfox1111 thanks for review | 00:40 |
| kfox1111 | np. thanks for the ps. :) | 00:41 |
| *** openstackgerrit has joined #openstack-kolla | 00:43 | |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Deployment guide updated with few more information https://review.openstack.org/457614 | 00:43 |
| openstackgerrit | Merged openstack/kolla-kubernetes master: README.rst character improvement https://review.openstack.org/451654 | 00:44 |
| *** david-lyle has joined #openstack-kolla | 00:44 | |
| *** huzhengchuan has joined #openstack-kolla | 00:49 | |
| *** sayantan_ has quit IRC | 00:50 | |
| *** chas has joined #openstack-kolla | 00:50 | |
| *** cuongnv has joined #openstack-kolla | 00:51 | |
| *** chas has quit IRC | 00:54 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes master: Heat service deployment https://review.openstack.org/437252 | 00:57 |
| tervancovan | Hi | 01:04 |
| *** StephenWang1991 has joined #openstack-kolla | 01:10 | |
| *** yangyapeng has joined #openstack-kolla | 01:15 | |
| openstackgerrit | Chen proposed openstack/kolla master: Use LABEL instead of MAINTAINER (deprecated) in all Dockerfile.j2 https://review.openstack.org/457513 | 01:18 |
| tervancovan | Could somebody tell me your own dev environment include kubernetes , in my case I am trying to openstack-helm's kubeadm-aio for kolla-kubernetes | 01:20 |
| *** caowei has joined #openstack-kolla | 01:21 | |
| openstackgerrit | Taejeoung Kim proposed openstack/kolla-kubernetes master: fix error on document https://review.openstack.org/457829 | 01:24 |
| *** hrw has quit IRC | 01:31 | |
| *** MarginHu2017 has joined #openstack-kolla | 01:31 | |
| *** iceyao has joined #openstack-kolla | 01:31 | |
| *** chas has joined #openstack-kolla | 01:31 | |
| *** hrw has joined #openstack-kolla | 01:32 | |
| *** zhubingbing has joined #openstack-kolla | 01:34 | |
| *** goldyfruit has joined #openstack-kolla | 01:36 | |
| *** chas has quit IRC | 01:36 | |
| *** japestinho has joined #openstack-kolla | 01:40 | |
| *** iceyao has quit IRC | 01:45 | |
| *** zhubingbing has quit IRC | 01:48 | |
| spsurya | sdake http://paste.openstack.org/show/607093/ | 01:49 |
| spsurya | helm installation | 01:49 |
| spsurya | step 3 | 01:49 |
| sdake | spsurya your running asa the root user | 01:49 |
| sdake | /usr/local/bin is notin the root user's path for security reasons | 01:50 |
| spsurya | ohhh | 01:50 |
| sdake | spsurya when i create a vagrant box using your vagrant scipt it logs me in not as teh root user | 01:50 |
| spsurya | you can do that | 01:51 |
| spsurya | sudo su | 01:51 |
| spsurya | one more step to be root | 01:51 |
| spsurya | just after logged in | 01:51 |
| *** chas has joined #openstack-kolla | 01:52 | |
| sdake | spsurya the guide doesn't say sudo -i :) | 01:53 |
| sdake | or suod su | 01:53 |
| sdake | the guide is meant to be run as a local user | 01:53 |
| sdake | with passwordless sudo | 01:53 |
| *** chas has quit IRC | 01:57 | |
| spsurya | sdake r0ger | 01:57 |
| spsurya | got thhis time | 01:57 |
| *** iceyao has joined #openstack-kolla | 02:01 | |
| *** zhubingbing_ has joined #openstack-kolla | 02:03 | |
| *** ssurana has quit IRC | 02:03 | |
| spsurya | but tiller heads up :( | 02:04 |
| spsurya | https://www.irccloud.com/pastebin/1eHLocWA/ | 02:04 |
| spsurya | sdake | 02:04 |
| spsurya | though helm is configured | 02:04 |
| sdake | hmm | 02:04 |
| sdake | i ahve actually never looksd at the get script | 02:04 |
| sdake | can you run kubeadm get pods -n kube-system | 02:05 |
| spsurya | yes ran | 02:05 |
| spsurya | running fine | 02:05 |
| sdake | -> paste | 02:05 |
| sdake | tiller is running? | 02:05 |
| sdake | how about helm version | 02:05 |
| spsurya | https://www.irccloud.com/pastebin/yas2ZX5i/ | 02:06 |
| sdake | no tiller running | 02:06 |
| sdake | run helm version -> paste | 02:07 |
| spsurya | Helm v2.3.1 | 02:08 |
| sdake | run helm init | 02:08 |
| spsurya | ran connection refused | 02:09 |
| spsurya | https://www.irccloud.com/pastebin/MsRZucwZ/ | 02:09 |
| sdake | you are no long errunning in the root creds | 02:09 |
| spsurya | yes | 02:10 |
| sdake | so you need to do the step which copies over the creds fiel in /etc/kubernetes/config | 02:10 |
| spsurya | no root | 02:10 |
| sdake | admin.conf that is | 02:10 |
| sdake | mkdir -p $HOME/.kube | 02:10 |
| sdake | sudo -H cp /etc/kubernetes/admin.conf $HOME/.kube/config | 02:10 |
| sdake | sudo -H chown $(id -u):$(id -g) $HOME/.kube/config | 02:11 |
| spsurya | already followed in Step 1 | 02:12 |
| sdake | you ran it as root | 02:12 |
| spsurya | yes | 02:12 |
| sdake | it copied it to /root/.kube.config | 02:12 |
| sdake | not /spurya/.kube/config | 02:12 |
| sdake | or wahtever your login is | 02:12 |
| spsurya | ohhh | 02:12 |
| sdake | could probably use a doc change that these instructions are meant to be run as the non-root user and explain why | 02:13 |
| sdake | because tiller cannot be found in /usr/local/bin | 02:13 |
| *** chas has joined #openstack-kolla | 02:13 | |
| spsurya | now worked when copied as non root as well | 02:14 |
| spsurya | and ran | 02:14 |
| spsurya | helm init | 02:14 |
| spsurya | now tiller got the tiller too | 02:15 |
| spsurya | https://www.irccloud.com/pastebin/BWaXaTss/ | 02:15 |
| spsurya | noted to updated the doc for this non root running and the reason | 02:16 |
| *** unicell has quit IRC | 02:17 | |
| spsurya | I think i should run again | 02:17 |
| spsurya | destroy and up | 02:17 |
| sdake | nah your good to go | 02:18 |
| *** chas has quit IRC | 02:18 | |
| spsurya | k8s setup | 02:18 |
| sdake | just proceed from here | 02:18 |
| spsurya | ok | 02:18 |
| *** huzhengchuan has quit IRC | 02:18 | |
| *** yangyapeng has quit IRC | 02:21 | |
| *** jayahn has joined #openstack-kolla | 02:26 | |
| *** schwicht has quit IRC | 02:29 | |
| spsurya | sdake as i ran into error unable to execute gcc | 02:30 |
| spsurya | should i include it into guide | 02:30 |
| spsurya | so that user go | 02:31 |
| *** huzhengchuan has joined #openstack-kolla | 02:31 | |
| spsurya | uniterupted | 02:31 |
| sdake | spsurya yum groupinstall "Development Tools" | 02:32 |
| sdake | spsurya yum -y groupinstall "Development Tools" | 02:33 |
| spsurya | that needs to be in doc too | 02:34 |
| *** chas has joined #openstack-kolla | 02:34 | |
| spsurya | sdake right ? | 02:34 |
| portdirect | Hey tervancovan, i wrote the kubeadm-aio container in openstack-helm, there's nothing special about the k8s it installs: so you should be able to run kolla-k8s ontop of it no problem unless it has specific requirements above and beyond vanilla k8s. | 02:34 |
| sdake | spsurya indeed | 02:37 |
| sdake | spsurya probably separate patches | 02:37 |
| sdake | on efor that and one for root | 02:37 |
| sdake | portdirect you sure read far back in the scrollback :) | 02:38 |
| portdirect | Lol, only an hour:20. | 02:38 |
| *** chas has quit IRC | 02:38 | |
| portdirect | Though be good to discuss strategy for Dev/poc moving forward | 02:39 |
| portdirect | I'm pretty sure we can come up with something pretty awesome together :) | 02:40 |
| spsurya | ok | 02:40 |
| *** unicell has joined #openstack-kolla | 02:40 | |
| sdake | portdirect that may be - i'd like new contribs to come up with the approach rather then the core reviewers defining a strategy | 02:40 |
| sdake | strategy fails for many reasons | 02:41 |
| *** huzhengchuan has quit IRC | 02:41 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes master: Heat service deployment https://review.openstack.org/437252 | 02:42 |
| portdirect | Your telling me ;) organic growth wins every time | 02:42 |
| sdake | i guess it depends on your definition of organic :) | 02:43 |
| sdake | anyway busy helping others have no time for debate | 02:43 |
| portdirect | All good brother, if there's anything I can help with just holler :) | 02:44 |
| *** dave-mccowan has quit IRC | 02:47 | |
| sdake | spsurya i'm headed to bed soon | 02:52 |
| sdake | spsurya where ya at in the instructions :) | 02:52 |
| spsurya | ok | 02:52 |
| spsurya | updating the guide | 02:53 |
| sdake | did you get through the deploy guide sucessfullY? | 02:53 |
| spsurya | then proceed further | 02:54 |
| spsurya | just copied the kolla and k8s files toetc | 02:54 |
| spsurya | etc/kolla and etc/k8s | 02:54 |
| sdake | ok well get throug hthe guide | 02:54 |
| sdake | i'll wait up | 02:54 |
| sdake | and submit patches after :) | 02:55 |
| sdake | that way I can help ya out | 02:55 |
| spsurya | ok | 02:55 |
| *** chas has joined #openstack-kolla | 02:55 | |
| sdake | legimately help you out | 02:55 |
| spsurya | :) | 02:56 |
| *** huzhengchuan has joined #openstack-kolla | 02:56 | |
| *** lamt has joined #openstack-kolla | 02:59 | |
| *** chas has quit IRC | 02:59 | |
| *** unicell has quit IRC | 03:04 | |
| spsurya | sdake just want to be precaucious... As in my case vm has ip 192.168.121.163 at eth0. So i should put kolla internal ip 192.168.121.163, network interface as eth0 and neutron external interface to eth1 as per below $ip addr | 03:09 |
| spsurya | https://www.irccloud.com/pastebin/9va4X5zK/ | 03:09 |
| *** iceyao has quit IRC | 03:10 | |
| spsurya | in globals.yml | 03:10 |
| *** MasterOfBugs has quit IRC | 03:11 | |
| sdake | spsurya where is your eth1? :) | 03:12 |
| spsurya | no eth1 ;) | 03:12 |
| sdake | right | 03:12 |
| sdake | neutron needs an eth1 to operate | 03:12 |
| sdake | there is probably a way to handle that - I think the easiest is to change your vagrant file to export a second interface | 03:13 |
| sdake | spsurya as to your quesiton, network_iterface hsould be 192.168.121.63 | 03:13 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: Add fwaas extension in l3 agent when neutron fwaas is enabled https://review.openstack.org/457846 | 03:14 |
| *** shasha___ has joined #openstack-kolla | 03:15 | |
| spsurya | sdake: or i can create eth1 | 03:15 |
| *** chas has joined #openstack-kolla | 03:16 | |
| sdake | spsurya how ? | 03:16 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: Cleanup /run/netns host directory volume mount https://review.openstack.org/457848 | 03:17 |
| *** shasha___ has quit IRC | 03:18 | |
| *** kencjohnston_ has quit IRC | 03:19 | |
| spsurya | sdake: with system-config-network | 03:19 |
| sdake | ok if that works | 03:19 |
| sdake | haven't used that toolbefore | 03:19 |
| spsurya | via creating new device and assigning new address | 03:19 |
| *** kencjohnston has joined #openstack-kolla | 03:19 | |
| spsurya | i will check let you know | 03:20 |
| *** chas has quit IRC | 03:20 | |
| portdirect | spsurya: a dummy device should be ok for dev/testing | 03:24 |
| spsurya | portdirect yes, it should | 03:26 |
| portdirect | https://www.irccloud.com/pastebin/ATHFWoTq/ | 03:28 |
| portdirect | should be all you need i think? | 03:28 |
| spsurya | yes | 03:32 |
| spsurya | thanks | 03:32 |
| *** lamt has quit IRC | 03:34 | |
| spsurya | portdirect: i think it will goes away once get down | 03:35 |
| spsurya | i.e vm reboot | 03:35 |
| portdirect | yup | 03:35 |
| spsurya | need to keep it in file | 03:36 |
| spsurya | more reliable | 03:36 |
| spsurya | for now your one is good | 03:36 |
| *** chas has joined #openstack-kolla | 03:37 | |
| *** iceyao has joined #openstack-kolla | 03:37 | |
| portdirect | spsurya: makes sense, though I'm not sure if theres a PS for the vagrant env your working on up atm, so just spitballing options :) | 03:39 |
| spsurya | :) | 03:40 |
| spsurya | nice | 03:41 |
| *** yangyapeng has joined #openstack-kolla | 03:41 | |
| *** chas has quit IRC | 03:41 | |
| *** lamt has joined #openstack-kolla | 03:41 | |
| *** iceyao has quit IRC | 03:42 | |
| *** hieulq_ has joined #openstack-kolla | 03:42 | |
| *** gkadam has joined #openstack-kolla | 03:43 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla stable/ocata: Bump OpenStack Service Version https://review.openstack.org/457851 | 03:43 |
| *** gkadam is now known as Guest79394 | 03:43 | |
| *** Guest79394 has quit IRC | 03:44 | |
| *** gkadam_ has joined #openstack-kolla | 03:44 | |
| *** gkadam_ is now known as gkadam | 03:45 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla stable/ocata: Bump OpenStack service version https://review.openstack.org/457851 | 03:47 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla stable/ocata: Bump OpenStack service version https://review.openstack.org/457851 | 03:52 |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible stable/ocata: neutron_vpnaas_agent: kolla_docker parameter 'volumes' is wrong https://review.openstack.org/457856 | 03:54 |
| *** bmace has quit IRC | 03:54 | |
| openstackgerrit | Merged openstack/kolla-ansible master: neutron_vpnaas_agent: kolla_docker parameter 'volumes' is wrong https://review.openstack.org/457505 | 03:57 |
| *** chas has joined #openstack-kolla | 03:58 | |
| *** eaguilar has quit IRC | 03:58 | |
| *** chas has quit IRC | 04:02 | |
| *** goldyfruit has quit IRC | 04:03 | |
| sdake | spsurya i've hit the wall dude - headed to bed :) | 04:07 |
| spsurya | np | 04:07 |
| sdake | spsurya I think you should be set now - you have worked thorugh all the deployment guide issues you ahve found | 04:07 |
| spsurya | sdake | 04:07 |
| spsurya | i will update in guide | 04:08 |
| sdake | spsurya not sure how to handle the second interface - I am sure it is solveable in your environment - google probely holds the answer :) | 04:08 |
| sdake | thanks | 04:08 |
| sdake | night all | 04:08 |
| spsurya | got the eth1 with no ip that what i required | 04:08 |
| spsurya | thanks | 04:08 |
| spsurya | night | 04:09 |
| *** yangyapeng has quit IRC | 04:16 | |
| tervancovan | Thanks portdirect | 04:17 |
| *** yangyapeng has joined #openstack-kolla | 04:17 | |
| *** chas has joined #openstack-kolla | 04:18 | |
| tervancovan | But to use kolla-k8s We have to set CIDR in kubeadm-aio | 04:19 |
| *** iceyao has joined #openstack-kolla | 04:21 | |
| *** salv-orlando has joined #openstack-kolla | 04:23 | |
| *** chas has quit IRC | 04:23 | |
| openstackgerrit | Merged openstack/kolla-kubernetes master: Version 0.7.0 https://review.openstack.org/457657 | 04:25 |
| *** iceyao has quit IRC | 04:26 | |
| *** lamt has quit IRC | 04:26 | |
| *** salv-orlando has quit IRC | 04:28 | |
| *** sudipto has joined #openstack-kolla | 04:39 | |
| *** chas has joined #openstack-kolla | 04:39 | |
| *** hieulq_ has quit IRC | 04:40 | |
| *** chas has quit IRC | 04:44 | |
| *** skramaja has joined #openstack-kolla | 04:46 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes master: Heat service deployment https://review.openstack.org/437252 | 04:47 |
| *** shasha___ has joined #openstack-kolla | 04:54 | |
| *** salv-orlando has joined #openstack-kolla | 04:56 | |
| *** hieulq_ has joined #openstack-kolla | 04:57 | |
| *** shasha___ has quit IRC | 04:58 | |
| *** hieulq_ has quit IRC | 04:58 | |
| *** chas has joined #openstack-kolla | 05:00 | |
| *** lamt has joined #openstack-kolla | 05:01 | |
| *** iceyao has joined #openstack-kolla | 05:01 | |
| *** unicell has joined #openstack-kolla | 05:04 | |
| *** chas has quit IRC | 05:05 | |
| *** lamt has quit IRC | 05:05 | |
| *** iceyao has quit IRC | 05:06 | |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla-kubernetes master: Deployment guide updation for better followup https://review.openstack.org/457870 | 05:14 |
| *** tovin07_ has joined #openstack-kolla | 05:26 | |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla-kubernetes master: Development Tools package installation added in deployment guide https://review.openstack.org/457871 | 05:28 |
| portdirect | tervancovan: though undocumented you can set the default device/ip to bind to when launching the container using the KUBE_BIND_DEV env var. I'll try to add customisation of the CNI CIDR in the next few days. :) | 05:28 |
| tervancovan | thanks portdirect | 05:34 |
| tervancovan | I just check that env var | 05:35 |
| *** zhurong has joined #openstack-kolla | 05:42 | |
| *** chas has joined #openstack-kolla | 05:42 | |
| tervancovan | Can you share your CIDR configuration for running kolla-k8s in kubeadm-aio container? | 05:44 |
| tervancovan | portdirect? | 05:44 |
| *** iceyao has joined #openstack-kolla | 05:46 | |
| *** chas has quit IRC | 05:46 | |
| *** iceyao has quit IRC | 05:51 | |
| *** Serlex has joined #openstack-kolla | 06:03 | |
| *** unicell1 has joined #openstack-kolla | 06:03 | |
| *** unicell has quit IRC | 06:03 | |
| *** salv-orlando has quit IRC | 06:04 | |
| openstackgerrit | zhubingbing proposed openstack/kolla-ansible stable/ocata: neutron_vpnaas_agent: kolla_docker parameter 'volumes' is wrong https://review.openstack.org/457879 | 06:10 |
| *** yangyapeng has quit IRC | 06:15 | |
| *** dvx76 has quit IRC | 06:15 | |
| *** chas has joined #openstack-kolla | 06:24 | |
| tervancovan | portdirect After apply "canal CNI driver" in kubeadm-aio container -> Canal pods were in "ClashLoopBackoff" -> https://thepasteb.in/p/lOhOx4lpROJFB | 06:28 |
| *** iceyao has joined #openstack-kolla | 06:28 | |
| *** chas has quit IRC | 06:28 | |
| *** StephenWang1991 has quit IRC | 06:33 | |
| *** pcaruana has joined #openstack-kolla | 06:33 | |
| *** iceyao has quit IRC | 06:33 | |
| Jeffrey4l | duonghq, could u review https://review.openstack.org/457848 | 06:34 |
| duonghq | Jeffrey4l, mins | 06:37 |
| Jeffrey4l | thanks :) | 06:37 |
| duonghq | done ;) | 06:37 |
| duonghq | hope that some day we can get more stable gate | 06:38 |
| Jeffrey4l | aha, ;) gate is hard | 06:40 |
| duonghq | yes | 06:40 |
| openstackgerrit | Merged openstack/kolla-ansible stable/ocata: Cleanup /run/netns host directory volume mount https://review.openstack.org/457848 | 06:41 |
| *** lamt has joined #openstack-kolla | 06:44 | |
| *** lamt has quit IRC | 06:44 | |
| *** chas has joined #openstack-kolla | 06:44 | |
| Jeffrey4l | zhubingbing_, please review https://review.openstack.org/457851 | 06:47 |
| *** chas has quit IRC | 06:49 | |
| *** yangyapeng has joined #openstack-kolla | 06:49 | |
| *** jaosorior has joined #openstack-kolla | 06:53 | |
| *** chas has joined #openstack-kolla | 06:56 | |
| *** jayahn has quit IRC | 07:00 | |
| *** matrohon has joined #openstack-kolla | 07:02 | |
| *** MasterOfBugs has joined #openstack-kolla | 07:03 | |
| *** manheim has joined #openstack-kolla | 07:08 | |
| *** manheim has quit IRC | 07:08 | |
| *** manheim has joined #openstack-kolla | 07:08 | |
| *** matrohon has quit IRC | 07:08 | |
| *** athomas has joined #openstack-kolla | 07:09 | |
| *** StephenWang1991 has joined #openstack-kolla | 07:11 | |
| *** iceyao has joined #openstack-kolla | 07:12 | |
| *** tovin07 has joined #openstack-kolla | 07:12 | |
| *** shardy has joined #openstack-kolla | 07:14 | |
| zhubingbing_ | ok | 07:17 |
| *** salv-orlando has joined #openstack-kolla | 07:17 | |
| zhubingbing_ | done | 07:18 |
| *** iniazi has quit IRC | 07:29 | |
| *** egonzalez has joined #openstack-kolla | 07:38 | |
| *** mgoddard has joined #openstack-kolla | 07:43 | |
| *** blallau has joined #openstack-kolla | 07:46 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible stable/ocata: Set fluentd output to elasticsearch to use elasticsearch_address https://review.openstack.org/457907 | 07:52 |
| *** matrohon has joined #openstack-kolla | 07:53 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible stable/ocata: Fix panko mongo idempotency https://review.openstack.org/457908 | 07:54 |
| *** jascott1 has quit IRC | 07:56 | |
| *** iceyao has quit IRC | 07:56 | |
| *** jascott1 has joined #openstack-kolla | 07:57 | |
| sbezverk | Jeffrey4l: ping | 07:58 |
| *** magicboiz has joined #openstack-kolla | 07:59 | |
| Jeffrey4l | sbezverk, pong. | 08:01 |
| *** athomas has quit IRC | 08:02 | |
| *** jascott1 has quit IRC | 08:02 | |
| *** matrohon_ has joined #openstack-kolla | 08:03 | |
| sbezverk | Jeffrey4l: thanks but I found my misconfig :) sorry to bug you | 08:04 |
| Jeffrey4l | np ;) | 08:05 |
| *** matrohon has quit IRC | 08:05 | |
| *** mgoddard has quit IRC | 08:07 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: [TEST][OCATA][KOLLA-ANSIBLE] TEST HEAT RELATED SERVICES https://review.openstack.org/434636 | 08:08 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: [TEST][OCATA][KOLLA-ANSIBLE] TEST CEILOMETER RELATED SERVICE https://review.openstack.org/434637 | 08:08 |
| *** zhubingbing_ has quit IRC | 08:14 | |
| *** zhubingbing_ has joined #openstack-kolla | 08:23 | |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla-kubernetes master: By default neutron interface corrected in guide https://review.openstack.org/457922 | 08:26 |
| *** gfidente has joined #openstack-kolla | 08:29 | |
| openstackgerrit | Merged openstack/kolla stable/ocata: Bump OpenStack service version https://review.openstack.org/457851 | 08:34 |
| *** MasterOfBugs has quit IRC | 08:37 | |
| openstackgerrit | Chen proposed openstack/kolla master: Use LABEL instead of MAINTAINER (deprecated) in all Dockerfile.j2 https://review.openstack.org/457513 | 08:39 |
| *** mgoddard has joined #openstack-kolla | 08:42 | |
| *** sambetts|afk is now known as sambetts | 08:52 | |
| *** zhubingbing_ has quit IRC | 08:54 | |
| *** zhubingbing has joined #openstack-kolla | 08:57 | |
| *** g3ek has quit IRC | 09:00 | |
| *** britthouser has quit IRC | 09:01 | |
| *** britthouser has joined #openstack-kolla | 09:01 | |
| hrw | zhubingbing: which distros are supported by grafana image? Ubuntu one fails terribly | 09:04 |
| *** g3ek has joined #openstack-kolla | 09:09 | |
| *** zhugaoxiao has joined #openstack-kolla | 09:13 | |
| *** rmart04 has joined #openstack-kolla | 09:16 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: grafana: disable for non-x86 https://review.openstack.org/457946 | 09:18 |
| hrw | sbezverk: can you update helm to 2.2.0? it would then be buildable for non-x86 targets too | 09:22 |
| sbezverk | hrw: we are currently using 2.3.0 and we cannot downgrade it to 2.2.0 | 09:23 |
| sbezverk | due to some features we use | 09:23 |
| hrw | sbezverk: hm have to update probably as Dockerfile.j2 here says 2.1.3 | 09:23 |
| sbezverk | hrw: helm does not come in docker, well at least we do not use one | 09:25 |
| hrw | sbezverk: ok, I meant helm-repository one. srry | 09:25 |
| sbezverk | hrw: we install helm as a part of deployment process onto kubernetes cluster | 09:25 |
| sbezverk | hrw: oh, that one for sure | 09:26 |
| sbezverk | hrw: please use version 2.3.0 to be more consistent | 09:26 |
| hrw | sbezverk: so far I only build so no idea how to test 2.1.3->2.3.0 update for being ok | 09:27 |
| hrw | sbezverk: I can send one patch which will bump to 2.3.0 and then on top of it one to make it buildable on non-x86. ok? | 09:29 |
| sbezverk | hrw: sounds good | 09:31 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: helm-repository: update to v2.3.0 https://review.openstack.org/457967 | 09:34 |
| hrw | and now let's try that on aarch64 ;D | 09:36 |
| openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla master: Reparent kolla-toolbox from openstack-base https://review.openstack.org/435023 | 09:42 |
| *** salv-orl_ has joined #openstack-kolla | 09:48 | |
| hrw | INFO:kolla.image.build.helm-repository:Happy Helming! | 09:48 |
| hrw | ;D | 09:49 |
| hrw | on aarch64 | 09:49 |
| *** salv-orlando has quit IRC | 09:51 | |
| japestinho | Hey folks, I am able to deploy kolla-k8s AIO on my centos-7 desktop machine. the problem is I can't ping or ssh VM address although I can assign floating IP to VM. My AIO host address 10.10.10.13/24, pod-network-cidr=10.11.0.0/16 service-cidr=10.13.3.0/24, public net openstack address is same with AIO host network 10.10.10.0/24, and private openstack address | 09:51 |
| japestinho | 10.10.11.0/24 | 09:51 |
| japestinho | is that correct I use same address for public net on openstack with AIO host network address? | 09:52 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla master: Fix cinder-api ubuntu binary https://review.openstack.org/457982 | 09:52 |
| sbezverk | hrw: so now you can helm on your phone ;) ?? | 09:53 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: helm-repository: make it work on non-x86 https://review.openstack.org/457983 | 09:53 |
| hrw | sbezverk: My arm64 box is normal server | 09:53 |
| hrw | sbezverk: 8 cores, few sata disks, pci express, 10GbE network ports | 09:53 |
| sbezverk | hrw: ok got it.. | 09:53 |
| hrw | sbezverk: there are servers with 48cores/cpu (up to 2 cpus per board), 1TB ram per cpu... | 09:54 |
| *** zhubingbing has quit IRC | 09:57 | |
| *** jascott1 has joined #openstack-kolla | 09:59 | |
| *** Seyong has quit IRC | 10:00 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix cinder-api ubuntu binary deploy https://review.openstack.org/457993 | 10:01 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix cinder-api ubuntu binary deploy https://review.openstack.org/457993 | 10:03 |
| hrw | sbezverk: you convinced me to finally write a post on my blog ;D | 10:04 |
| *** jascott1 has quit IRC | 10:05 | |
| sbezverk | hrw: sure, good idea, until you mentioned I was under impression arm processors were meant for something really small | 10:05 |
| gema | sbezverk: they are, but there are also server chips | 10:06 |
| hrw | sbezverk: 32bit arm had some desktops in past. There were servers as well. Some high density ones too (64 quad core systems in 2-4U). But it was never popular. | 10:06 |
| gema | sbezverk: so we are working on bringing the new AArch64 servers up there in terms of software supported with the x86_64 ones | 10:08 |
| gema | sbezverk: it really feels no different than an intel machine | 10:08 |
| hrw | and as a side effect I get kolla working on ppc64le as well | 10:08 |
| hrw | have access to some power8 machine with 128GB ram and 80 cpu threads | 10:09 |
| sbezverk | gema: curious is MS supports this arch as well? | 10:09 |
| hrw | sbezverk: yes | 10:09 |
| sbezverk | hrw: 80 cpu threads sounds nice.. | 10:10 |
| hrw | sbezverk: they presented MS Windows running Azure load | 10:10 |
| hrw | sbezverk: no idea will they make desktop version | 10:10 |
| hrw | sbezverk: x86-64 feels limited when you work with other architectures ;d | 10:10 |
| hrw | 8/16 systems (cpu/threads) compared to 10/80 or 8-48-240 cores | 10:11 |
| *** MarginHu2017 has quit IRC | 10:12 | |
| *** cuongnv has quit IRC | 10:13 | |
| sbezverk | hrw: I have quad server each with 36 threads so I would not say it is too limited ;) | 10:14 |
| hrw | ;D | 10:14 |
| *** yangyapeng has quit IRC | 10:16 | |
| hrw | sbezverk: I spoke with sdake recently about which aarch64 server to buy. The main problem was to find one with 1GbE port as it usually means extra cost because SoC does not support anything <10GbE so you need extra pcie controller | 10:16 |
| *** tovin07_ has quit IRC | 10:19 | |
| sbezverk | hrw: I see, always some gotchas ;) | 10:21 |
| *** zhurong has quit IRC | 10:34 | |
| openstackgerrit | Merged openstack/kolla-ansible stable/ocata: Add fwaas extension in l3 agent when neutron fwaas is enabled https://review.openstack.org/457846 | 10:35 |
| *** zhubingbing has joined #openstack-kolla | 10:39 | |
| openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla master: Reparent kolla-toolbox from openstack-base https://review.openstack.org/435023 | 10:39 |
| openstackgerrit | Paul Bourke (pbourke) proposed openstack/kolla master: Reparent kolla-toolbox from openstack-base https://review.openstack.org/435023 | 10:41 |
| *** jayahn has joined #openstack-kolla | 10:44 | |
| egonzalez | when is ocata 4.0.1 released? | 10:45 |
| *** StephenWang1991 has quit IRC | 10:47 | |
| *** salv-orl_ has quit IRC | 10:52 | |
| *** tervancovan has quit IRC | 10:54 | |
| openstackgerrit | caoyuan proposed openstack/kolla master: Install pankoclient in cloudkitty base container https://review.openstack.org/458024 | 10:59 |
| *** dmk0202 has joined #openstack-kolla | 11:02 | |
| *** sudipto has quit IRC | 11:03 | |
| *** sudipto has joined #openstack-kolla | 11:03 | |
| *** sudipto has quit IRC | 11:03 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix cinder-api ubuntu binary deploy https://review.openstack.org/457993 | 11:04 |
| *** magicboiz has quit IRC | 11:04 | |
| *** tervancovan has joined #openstack-kolla | 11:04 | |
| egonzalez | can I get reviews on https://review.openstack.org/#/c/457993/ and https://review.openstack.org/#/c/457982/ | 11:05 |
| egonzalez | atm cinder binary is broken in ocata | 11:05 |
| egonzalez | *ubuntu cinder binary | 11:05 |
| *** jascott1 has joined #openstack-kolla | 11:06 | |
| hrw | egonzalez: "USER cinder" line is not used then on debian/ubuntu? | 11:07 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix cinder-api ubuntu binary deploy https://review.openstack.org/457993 | 11:07 |
| *** StephenWang1991 has joined #openstack-kolla | 11:08 | |
| egonzalez | hrw, ubuntu removed support for cinder-api system service, only works with apache | 11:08 |
| hrw | ok | 11:08 |
| egonzalez | hrw, debian uses same packages as ubuntu right? | 11:08 |
| hrw | egonzalez: Debian lacks ocata packages | 11:09 |
| hrw | egonzalez: ubuntu has own packages | 11:09 |
| hrw | they are based on Debian ones but done only for ubuntu | 11:10 |
| *** jascott1 has quit IRC | 11:11 | |
| *** StephenWang1991 has quit IRC | 11:12 | |
| sdake | morning folks | 11:15 |
| sdake | tervancovan if you are in need of help let me know | 11:15 |
| *** rwallner has joined #openstack-kolla | 11:16 | |
| openstackgerrit | caoyuan proposed openstack/kolla master: Install pankoclient in cloudkitty base container https://review.openstack.org/458024 | 11:16 |
| hrw | hi sdake | 11:16 |
| *** caowei has quit IRC | 11:16 | |
| egonzalez | hrw, thanks for the explanation | 11:18 |
| sdake | hey hrw | 11:19 |
| sdake | hrw actually what I wa safter with an arm server is something that had 10gig ports that were copper :) | 11:19 |
| sdake | hrw I looked around quite a bit, and really couldn't find anything which means I'd need a 6k 10gig QSFP+ switch | 11:19 |
| *** rwallner has quit IRC | 11:20 | |
| *** dmk0202 has quit IRC | 11:21 | |
| *** dmk0202 has joined #openstack-kolla | 11:21 | |
| *** mgoddard has quit IRC | 11:27 | |
| *** rwallner has joined #openstack-kolla | 11:29 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Move Cloudkitty options into globals.yml https://review.openstack.org/458032 | 11:35 |
| portdirect | morning o/ | 11:38 |
| egonzalez | morning portdirect | 11:39 |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Move Cloudkitty options into globals.yml https://review.openstack.org/458032 | 11:40 |
| *** salv-orlando has joined #openstack-kolla | 11:42 | |
| *** salv-orlando has quit IRC | 11:48 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Allow use of any backup backend for any volume backend https://review.openstack.org/457554 | 11:51 |
| portdirect | tervancovan: the crash loop you were seeing is because calico is already running, you shouldnt need to install any cni or play with the cidr | 11:52 |
| portdirect | (unless it clashes with your network) | 11:52 |
| tervancovan | Yes I saw the code portdirect | 11:53 |
| tervancovan | but it is not same as kolla-k8s guide documents | 11:53 |
| *** yangyapeng has joined #openstack-kolla | 11:54 | |
| hrw | https://marcin.juszkiewicz.com.pl/2017/04/19/so-you-run-openstack-on-your-phone/ | 11:54 |
| tervancovan | this is kolla-k8s documents ---> https://thepasteb.in/p/Anhr4q4xAQoFv | 11:55 |
| tervancovan | and this code is from kubeadm-aio -> https://thepasteb.in/p/P1hvWVW8gXOul | 11:56 |
| tervancovan | hi sdake | 11:56 |
| portdirect | tervancovan: if you want to use canal you'll need to modify the container - just swapping out the bits you have in those pastes should do it | 11:57 |
| sdake | tervancovan kubeadm-aio is not cimpatible iwth the deploy documentation | 11:57 |
| sdake | portdirect also the deployment guide recomends setting the service and pod cidr | 11:57 |
| tervancovan | Thanks sdake , so I just fix it a little bit | 11:59 |
| tervancovan | So I applied canal for kolla-k8s in the container | 12:00 |
| sdake | tervancovan how did you run kubeadm init? | 12:01 |
| tervancovan | just add some code | 12:02 |
| hrw | sdake, sbezverk: I hope that you enjoy post ;D | 12:02 |
| portdirect | sdake: you dont need to do that, but its great to show people how they can (change the cidr) | 12:02 |
| sdake | portdirect i recognize it is not mandatory for many people | 12:02 |
| sdake | portdirect however to really understand how to work on kolla - i find it was necessary persoanlly to understand how kubernetes deploys (atleast with kubeadm) | 12:03 |
| tervancovan | this is snip code which i fixed it -> https://thepasteb.in/p/xGhmkNk087juM in kubeadm-aio | 12:04 |
| sdake | tervancovan there is more needed | 12:04 |
| sdake | tervancovan kubeadm doesn't do correct operations when the pod and service cidr are changed | 12:04 |
| *** srwilkers has joined #openstack-kolla | 12:05 | |
| *** satyar has joined #openstack-kolla | 12:06 | |
| satyar | Hi Kolla | 12:06 |
| sdake | tervancovan specifically: sudo sed -i 's/10.96.0.10/10.3.3.10/g' /etc/systemd/system/kubelet.service.d/10-kubeadm.conf | 12:06 |
| sdake | sup satyar | 12:06 |
| tervancovan | sdake yes right it was not working when cider was changed | 12:06 |
| satyar | hey sdake | 12:06 |
| satyar | how are you? | 12:06 |
| egonzalez | sup satyar | 12:06 |
| sdake | satyar good | 12:06 |
| satyar | it seems the upgrade is breaking the whole system | 12:06 |
| sdake | tervancovan i'd recommend just starting fresh without trying the kubeadm-aio container | 12:07 |
| sdake | tervancovan this will let you learn how kubeadm works | 12:07 |
| satyar | this time i removed the code for jumbo frame support | 12:07 |
| satyar | all MTU is setup to 1500 | 12:07 |
| satyar | created a VM | 12:07 |
| sdake | tervancovan understanding how kubeadm works is crucial to doinganything with kuberntes imo :) | 12:07 |
| satyar | the VM got 1450 MTU and able to ssh the VM | 12:07 |
| tervancovan | Yes i | 12:08 |
| satyar | upgraded the system to stable/ocata | 12:08 |
| tervancovan | I thinks so | 12:08 |
| satyar | then not able to access the VM any more | 12:08 |
| satyar | restarted the VM and saw VM is having 1500 MTU | 12:08 |
| sdake | tervancovan kubeadm-aio is really a nice peice of tech, however, it is pretty much a black box | 12:08 |
| satyar | i do have DVR and L3 HA | 12:08 |
| sdake | tervancovan much like any container it is declrative, eitherit works or it doesn't | 12:09 |
| *** sudipto has joined #openstack-kolla | 12:09 | |
| satyar | any help? | 12:09 |
| sdake | tervancovan if it does't work - it is very difficult to debug | 12:09 |
| sdake | just like the reest of the continer world :) | 12:10 |
| sdake | tervancovan tell me what kind of hardware you ahve | 12:10 |
| sdake | tervancovan linux machine? | 12:10 |
| egonzalez | hrw, interesting article | 12:10 |
| hrw | egonzalez: thx | 12:10 |
| *** Pavo has joined #openstack-kolla | 12:11 | |
| tervancovan | yes right and our team is going to try testing on baremetal for production | 12:14 |
| portdirect | sdake: I think its not much of a black-box (but here we differ) by being a cuut down environment it shows exactly how things like kubeadm actually work | 12:14 |
| portdirect | thats one of the great things i see about docker style containers | 12:15 |
| sdake | tervancovan if your after an eval of kolla-kubernetes, I'd recommend using a) baremental with two nics b) a vm with two nics | 12:16 |
| *** jayahn has quit IRC | 12:16 | |
| tervancovan | I think a is better | 12:17 |
| sdake | tervancovan yup - either will work for an AIO eval | 12:17 |
| sdake | tervancovan I use both :) | 12:17 |
| sdake | tervancovan if you ahve linux baremetal to work with, the guide works flawlessly | 12:17 |
| sdake | tervancovan once you learn how kubeadm works - you wlil find kubeadm prints out a big "this is alpha, do not use in production" warning | 12:18 |
| sdake | tervancovan there are other options for production deployments of kubernetes - if your end goal is to go to production, you will need in great detail to understand how kubernetes works | 12:18 |
| sdake | tervancovan also note kolla-kubernetes has not released a 1.0.0 - while good for eval, kolla-kubernetes is currently unsuitable for produciton | 12:19 |
| tervancovan | My boss should know that this issue .... | 12:19 |
| sdake | tervancovan which one - I listed out two :) | 12:20 |
| sdake | tervancovan by the time you get done evaluating, I think you will have an idea how to actually use kolla-kubernetes in production | 12:20 |
| sdake | tervancovan first step there is learning how to use kubernetes in production | 12:21 |
| tervancovan | thanks for information sdake | 12:22 |
| sdake | tervancovan when is your deployment target month? | 12:22 |
| tervancovan | Maybe end of this year? | 12:23 |
| sdake | ya - thats no problem for kolla-kubernetes | 12:23 |
| sdake | it wll take you that long to learn how all this stuff works :) | 12:23 |
| sdake | by then i expect kubeadm to be out of alpha | 12:23 |
| sdake | or many of the other kubernetes deployment options | 12:23 |
| sdake | kolla doesn't really care how you deploy kubernetes - the motive for the guide is to help you understand how to get to a kolla deployment | 12:24 |
| sdake | if the kolla community could take out steps 1 and step 2 from the guide, we would do so - however the upstream docs on deployment of kubernetes are lacking | 12:24 |
| tervancovan | sdake thanks for informations | 12:26 |
| *** mgoddard has joined #openstack-kolla | 12:29 | |
| sdake | sup mgoddard | 12:29 |
| manheim | does anyone have problems building centos images? | 12:31 |
| *** zhurong has joined #openstack-kolla | 12:31 | |
| manheim | openstack-base fails to build as it fails to download some packages | 12:31 |
| *** iceyao has joined #openstack-kolla | 12:31 | |
| sdake | manheim master or a specific tag or branch? | 12:32 |
| manheim | permission denied on some rpm | 12:32 |
| manheim | stable/ocata | 12:32 |
| sdake | manheim moment let me give it a spin | 12:32 |
| manheim | hmmm, or master, let me check | 12:33 |
| sdake | manheim which distro/type are you building? | 12:34 |
| manheim | sudo kolla-build neutron-server -n kolla --base centos --base-image centos --base-tag 7 | 12:34 |
| manheim | the exact command | 12:34 |
| manheim | 4.0.0 | 12:35 |
| manheim | i'm on master, yes | 12:36 |
| sdake | manheim you just said you were on stable/ocata :) | 12:42 |
| *** Pavo has quit IRC | 12:42 | |
| manheim | sorry, my mistake | 12:42 |
| sdake | manheim stable/ocata wfm | 12:44 |
| sdake | manheim checking out master - moment | 12:44 |
| manheim | ok | 12:44 |
| sdake | manheim please run pip show kolla -> paste | 12:45 |
| *** zhurong has quit IRC | 12:45 | |
| manheim | http://paste.ubuntu.com/24413607/ | 12:46 |
| *** kbaegis has joined #openstack-kolla | 12:46 | |
| manheim | i was behind on my fork a few days, rebased now | 12:46 |
| *** lamt has joined #openstack-kolla | 12:47 | |
| sdake | manheim - master wfm | 12:51 |
| manheim | nice | 12:51 |
| sdake | manheim i'd recommend trying master - I believe a recent change fixed a yum repo location - grafana iirc | 12:52 |
| manheim | ok, i'll reinstall kolla after rebase and see, thanks | 12:52 |
| *** Pavo has joined #openstack-kolla | 12:53 | |
| sdake | pbourke ++ | 12:54 |
| *** goldyfruit has joined #openstack-kolla | 12:56 | |
| *** lamt has quit IRC | 12:58 | |
| *** rwellum has joined #openstack-kolla | 13:00 | |
| manheim | sdake it built with latest commits, thanks again | 13:01 |
| *** lamt has joined #openstack-kolla | 13:01 | |
| sdake | manheim enjoy | 13:01 |
| *** huzhengchuan has quit IRC | 13:03 | |
| *** srwilkers has quit IRC | 13:04 | |
| mgoddard | hey sdake, sorry for the slow reply :) | 13:04 |
| sdake | mgoddard no sweat - just a greeting | 13:04 |
| sdake | morning rwellum | 13:05 |
| rwellum | morning sdake | 13:05 |
| sdake | rwellum have you had an opportunity to run through the latest deployment guide? | 13:06 |
| rwellum | No - but definitely this week. | 13:06 |
| sdake | rwellum it has undergone a few changes - may need to refresh your browser | 13:06 |
| rwellum | Ok will do - looking at it now | 13:07 |
| sdake | rwellum cool - i am starting up a twice weekly "Demystifying the gate" webex session for 30 minutes | 13:07 |
| *** jascott1 has joined #openstack-kolla | 13:07 | |
| sdake | rwellum I think the gate is difficult for people to comprehend as jascott1 pointed out - and having such an effort will be beneficial for many | 13:08 |
| rwellum | yesterday I was previewing the Upstream University latest slides for Boston summit with one of the authors. I bet he'd be interested to see if there was any overlap they could include. | 13:09 |
| *** sean-k-mooney has joined #openstack-kolla | 13:09 | |
| *** calbers has quit IRC | 13:11 | |
| *** jascott1 has quit IRC | 13:12 | |
| rwellum | sdake: you'll send out invites? | 13:12 |
| sdake | rwellum first will vote on a time | 13:12 |
| rwellum | sgtm | 13:13 |
| *** lamt has quit IRC | 13:14 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix heat ec2 keystone auth https://review.openstack.org/444953 | 13:15 |
| *** lamt has joined #openstack-kolla | 13:16 | |
| *** gkadam has quit IRC | 13:17 | |
| *** schwicht has joined #openstack-kolla | 13:18 | |
| *** rwallner has quit IRC | 13:19 | |
| *** calbers has joined #openstack-kolla | 13:20 | |
| *** rwallner has joined #openstack-kolla | 13:20 | |
| openstackgerrit | caoyuan proposed openstack/kolla-ansible master: Move Cloudkitty options into globals.yml https://review.openstack.org/458032 | 13:21 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Comment designate options at globals.yml https://review.openstack.org/458074 | 13:21 |
| *** magicboiz has joined #openstack-kolla | 13:22 | |
| *** magicboiz has quit IRC | 13:26 | |
| *** srwilkers has joined #openstack-kolla | 13:27 | |
| *** super has quit IRC | 13:28 | |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla-ansible master: Fix neutron agents restarted on ml2 config change https://review.openstack.org/447992 | 13:34 |
| *** magicboiz has joined #openstack-kolla | 13:37 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: kubetoolbox: mark as x86-64 only https://review.openstack.org/458078 | 13:45 |
| *** eanylin has quit IRC | 13:45 | |
| *** huzhengchuan has joined #openstack-kolla | 13:46 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: [TEST][OCATA][KOLLA-ANSIBLE] TEST CEILOMETER RELATED SERVICE https://review.openstack.org/434637 | 13:46 |
| sdake | rwellum can you make today's team meeting? | 13:46 |
| sdake | rwellum 16:00 UTC | 13:47 |
| hrw | ~curse current development methods | 13:50 |
| hrw | there are few images which fetch kubectl binary from network. x86-64 binary. for Debian it can be installed from package. but not for ubuntu. | 13:51 |
| hrw | centos has package too | 13:51 |
| rwellum | sdake: I should be able to make that - have a meeting I run that finishes when that starts | 13:52 |
| *** lamt__ has joined #openstack-kolla | 13:52 | |
| *** lamt has quit IRC | 13:53 | |
| sdake | rwellum good to hear | 13:55 |
| sdake | spsurya hopefully you can make our team meeting as well | 13:55 |
| openstackgerrit | hosup choi proposed openstack/kolla-kubernetes master: Update deployment-guide.rst document for ubuntu https://review.openstack.org/458083 | 13:55 |
| hrw | anyway. another image to disable | 13:56 |
| sdake | query hrw | 13:56 |
| hrw | query sdake | 13:56 |
| *** al498u has joined #openstack-kolla | 13:57 | |
| sdake | winning | 13:57 |
| *** eanylin has joined #openstack-kolla | 13:57 | |
| *** ksumit has joined #openstack-kolla | 13:57 | |
| *** al498u_ has joined #openstack-kolla | 13:58 | |
| spsurya | sdake: hi.. | 13:59 |
| spsurya | meeting ?? | 13:59 |
| spsurya | didn't get you | 13:59 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: magnum-conductor: mark as x86-64 only https://review.openstack.org/458085 | 14:00 |
| hrw | spsurya: #openstack-meeting-4 in 2h from now - Kolla meeting | 14:01 |
| *** al498u has quit IRC | 14:02 | |
| *** jtriley has joined #openstack-kolla | 14:02 | |
| *** david-lyle has quit IRC | 14:04 | |
| *** tervancovan has quit IRC | 14:07 | |
| *** Pavo has quit IRC | 14:08 | |
| *** skramaja has quit IRC | 14:12 | |
| *** iceyao has quit IRC | 14:13 | |
| *** david-lyle has joined #openstack-kolla | 14:14 | |
| *** satyar has quit IRC | 14:19 | |
| *** eaguilar has joined #openstack-kolla | 14:19 | |
| *** satyar has joined #openstack-kolla | 14:28 | |
| *** satyar has quit IRC | 14:29 | |
| *** MarginHu2017 has joined #openstack-kolla | 14:29 | |
| *** satyar has joined #openstack-kolla | 14:29 | |
| sdake | spiette meeting at 16:00 UTC we have every wednesday on irc | 14:29 |
| sdake | spsurya ^^ | 14:29 |
| sdake | spsurya contributors generally attend the meeting | 14:29 |
| *** Pavo has joined #openstack-kolla | 14:29 | |
| *** zhubingbing has quit IRC | 14:29 | |
| spsurya | sdake i always attend | 14:30 |
| spsurya | got that | 14:30 |
| sdake | spsurya cool :) | 14:30 |
| spsurya | sdake on doodle now | 14:31 |
| *** iceyao has joined #openstack-kolla | 14:31 | |
| *** satyar has quit IRC | 14:31 | |
| spsurya | choosing the slot | 14:31 |
| *** satyar has joined #openstack-kolla | 14:31 | |
| *** StephenWang1991 has joined #openstack-kolla | 14:31 | |
| *** tomatopeel has left #openstack-kolla | 14:34 | |
| *** lamt__ has quit IRC | 14:35 | |
| *** iceyao has quit IRC | 14:35 | |
| openstackgerrit | Vladislav Belogrudov proposed openstack/kolla-ansible master: Allow different configuration of iSCSI helper https://review.openstack.org/458113 | 14:36 |
| *** MarginHu2017 has quit IRC | 14:38 | |
| sdake | rwellum do you have an email address you can share | 14:38 |
| *** MarginHu2017 has joined #openstack-kolla | 14:39 | |
| *** eaguilar has quit IRC | 14:44 | |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla-ansible master: Fix troubleshooting guide debugging section https://review.openstack.org/458115 | 14:50 |
| *** eaguilar has joined #openstack-kolla | 14:50 | |
| *** magicboiz has quit IRC | 14:50 | |
| *** MarginHu2017 has quit IRC | 14:51 | |
| *** magicboiz has joined #openstack-kolla | 14:53 | |
| *** duonghq_ has joined #openstack-kolla | 15:00 | |
| *** rmart04 has quit IRC | 15:00 | |
| *** zhubingbing has joined #openstack-kolla | 15:04 | |
| rwellum | sdake: sent it to you private message | 15:06 |
| duonghq_ | evening guys | 15:06 |
| sdake | rwellum ack getting to it :) | 15:06 |
| sdake | sup duonghq_ | 15:06 |
| *** duonghq has quit IRC | 15:06 | |
| *** duonghq_ is now known as duonghq | 15:06 | |
| *** duonghq_ has joined #openstack-kolla | 15:07 | |
| *** lamt has joined #openstack-kolla | 15:07 | |
| *** jascott1 has joined #openstack-kolla | 15:08 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Add panko.conf file in ceilometer container https://review.openstack.org/444719 | 15:08 |
| *** iceyao has joined #openstack-kolla | 15:10 | |
| *** jascott1 has quit IRC | 15:13 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Allow different configuration of iSCSI helper https://review.openstack.org/458113 | 15:14 |
| spsurya | duonghq: sup | 15:14 |
| duonghq | hi spsurya | 15:14 |
| *** iceyao has quit IRC | 15:15 | |
| *** manheim has quit IRC | 15:16 | |
| *** manheim has joined #openstack-kolla | 15:17 | |
| *** satyar has quit IRC | 15:17 | |
| *** satyar has joined #openstack-kolla | 15:18 | |
| *** manheim has quit IRC | 15:21 | |
| egonzalez | can I get reviews on https://review.openstack.org/#/c/457993/ and https://review.openstack.org/#/c/457982/ are critical bugs blocking cinder in ubuntu binary | 15:22 |
| *** pcaruana has quit IRC | 15:28 | |
| *** iceyao has joined #openstack-kolla | 15:30 | |
| pbourke | egonzalez: is the if/else at the end of the dockerfile necessary? | 15:31 |
| egonzalez | pbourke, yes for the user | 15:31 |
| egonzalez | pbourke, with cinder user cannot start apache process | 15:32 |
| pbourke | will we not be fixing cinder for source in ocata? | 15:32 |
| egonzalez | source works | 15:32 |
| egonzalez | is just ubuntu package who changed to apache only | 15:32 |
| pbourke | ah I see | 15:33 |
| egonzalez | is just to no backport features, only the fix necessary | 15:33 |
| *** iceyao has quit IRC | 15:35 | |
| *** inc0 has joined #openstack-kolla | 15:36 | |
| inc0 | good morning | 15:36 |
| zhubingbing | HI GUYS | 15:37 |
| zhubingbing | hi guys | 15:37 |
| srwilkers | o/ | 15:38 |
| duonghq | morning inc0 | 15:39 |
| duonghq | hi zhubingbing | 15:39 |
| duonghq | hi srwilkers | 15:39 |
| duonghq | nice to see you guys | 15:39 |
| zhubingbing | hi | 15:39 |
| zhubingbing | ;) | 15:39 |
| srwilkers | hey duonghq nice to see you too | 15:40 |
| duonghq | srwilkers, will you go to Boston? | 15:40 |
| *** david-lyle has quit IRC | 15:41 | |
| kfox1111 | morning. | 15:44 |
| srwilkers | duonghq, yeah ill be there. you? | 15:45 |
| srwilkers | kfox1111, morning man | 15:45 |
| sbezverk | good morning | 15:45 |
| duonghq | srwilkers, me too | 15:45 |
| srwilkers | duonghq, awesome :) | 15:45 |
| *** yee37916 has joined #openstack-kolla | 15:46 | |
| *** schwicht has quit IRC | 15:47 | |
| *** yee379 has quit IRC | 15:47 | |
| *** shardy has quit IRC | 15:48 | |
| sbezverk | anybody willing to get fluentd/eleasticsearch/kibana going in kube? ;) my experience with installing 5 node kolla-kube cluster strongly suggests we really need it :) | 15:49 |
| *** shardy has joined #openstack-kolla | 15:49 | |
| kfox1111 | sbezverk: been working on it a bit in a way. | 15:50 |
| kfox1111 | but kind of in a different way then we discussed. | 15:50 |
| kfox1111 | I found fluent-bit and have been playing around with it. | 15:50 |
| kfox1111 | then submitted this: | 15:50 |
| kfox1111 | https://github.com/kubernetes/charts/pull/895 | 15:50 |
| sbezverk | kfox1111: ok will you be committing it to kolla? | 15:51 |
| kfox1111 | was waiting till all the ansible config's merged (done I think) and then was going to be submitting patches to | 15:51 |
| srwilkers | sbezverk, kfox1111 ive been working on that as well | 15:51 |
| kfox1111 | support configing to log out to stdout instead of log files. | 15:51 |
| kfox1111 | with that in place, it will properly support log rolling/shipping with whatever k8s has installed. | 15:51 |
| kfox1111 | I tested log rolling with fluent-bit and k8s 1.6 and it seemed to work ok. | 15:52 |
| *** matrohon_ has quit IRC | 15:52 | |
| sbezverk | kfox1111: so fluent-bit runs as a side car ? | 15:52 |
| kfox1111 | sbezverk: daemonset. we could run it as a side car instead too, but would have to add a log roller too. :/ | 15:53 |
| kfox1111 | docker has a log roller built in apparently. | 15:53 |
| *** shardy has quit IRC | 15:53 | |
| rwellum | This is a bit out there - but does anyone have any advice/experience with configuring RHEL as a base OS for a server? Or point me to the correct place to ask? | 15:53 |
| kfox1111 | for dealing with stdout in its logs. | 15:53 |
| sbezverk | kfox1111: another thing do we have kolla container for it? | 15:53 |
| sdake | rwellum you might try #centos | 15:54 |
| kfox1111 | rwellum: what do you mean? | 15:54 |
| kfox1111 | rwellum: like, you want to do file serving, or whatever using RHEL? | 15:54 |
| kfox1111 | sbezverk: yeah, either we put it in kolla, or we use upstreams container/chart. | 15:55 |
| sbezverk | kfox1111: if it does not run asa side car, how do you plan to access openstack generated log files? | 15:55 |
| kfox1111 | I think there's other upstream charts we may want to use. | 15:55 |
| kfox1111 | prometheus/grafana too. | 15:55 |
| kfox1111 | been usign that for a while. | 15:55 |
| kfox1111 | sbezverk: the plan is to disable openstack log files. just push them directly to stdout. | 15:55 |
| rwellum | kfox1111: I mean a new server, no OS installed, I've used both Ubuntu and Centos in the past - wanting to try RHEL - had some issues with basic stuff like how it partitions, licensing post bringup etc. | 15:56 |
| openstackgerrit | Surya Prakash (spsurya) proposed openstack/kolla-kubernetes master: Deployment guide updation for better followup https://review.openstack.org/457870 | 15:56 |
| rwellum | sdake: Centos I can do :) | 15:56 |
| *** jascott1 has joined #openstack-kolla | 15:56 | |
| kfox1111 | rwellum: its very similar to centos. really only just the extra licence registration step. | 15:57 |
| zhubingbing | meeting | 15:57 |
| kfox1111 | we've got a fair amount of rhel servers. | 15:57 |
| sbezverk | kfox1111: that would be a major change in behavior | 15:57 |
| zhubingbing | kfox1111 why user fluent-bit ? | 15:58 |
| kfox1111 | sbezverk: yeah. thats why I haven't talked about it too much yet. still experimenting. | 15:58 |
| inc0 | meeting peeps | 15:58 |
| kfox1111 | zhubingbing: its much lighter weight then fluentd, since it needs to be on all nodes. | 15:58 |
| kfox1111 | fluentd and logstash are basically equivilent. | 15:58 |
| rwellum | kfox1111: I stupidly chose default partition, and it put everything on /home. Pretty sure Centos doesn't do that. At least not the version I've been using... | 15:58 |
| kfox1111 | fluent-bit and filebeat are basically equivilent. | 15:58 |
| zhubingbing | hmm i konw | 15:59 |
| kfox1111 | rwellum: are you using the workstation edition or server edition? | 15:59 |
| kfox1111 | I think the defaults are a bit different. but they are otherwise essentially the same. | 16:00 |
| sdake | kfox1111 we have blueprints on the agenda today | 16:00 |
| sdake | would be nice to have you there #openstack-meeting-4 | 16:00 |
| kfox1111 | sdake: you started early. ;) | 16:01 |
| rwellum | kfox1111: I'll continue after the meeting! | 16:01 |
| *** unicell has joined #openstack-kolla | 16:01 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix cinder-api ubuntu binary deploy https://review.openstack.org/457993 | 16:02 |
| *** unicell1 has quit IRC | 16:04 | |
| *** Pavo has quit IRC | 16:04 | |
| *** schwicht has joined #openstack-kolla | 16:06 | |
| *** dave-mccowan has joined #openstack-kolla | 16:06 | |
| *** dmk0202 has quit IRC | 16:09 | |
| openstackgerrit | Merged openstack/kolla master: Fix cinder-api ubuntu binary https://review.openstack.org/457982 | 16:12 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla stable/ocata: Fix cinder-api ubuntu binary https://review.openstack.org/458152 | 16:12 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible master: Fix cinder-api ubuntu binary deploy https://review.openstack.org/457993 | 16:15 |
| *** huzhengchuan has quit IRC | 16:16 | |
| *** david-lyle has joined #openstack-kolla | 16:19 | |
| *** manheim has joined #openstack-kolla | 16:21 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible stable/ocata: Add panko.conf file in ceilometer container https://review.openstack.org/444720 | 16:21 |
| *** manheim has quit IRC | 16:25 | |
| openstackgerrit | Bertrand Lallau proposed openstack/kolla-ansible master: WIP: set kibana default index pattern https://review.openstack.org/458156 | 16:27 |
| * hrw drops from meeting - will leave session running | 16:27 | |
| *** blallau has quit IRC | 16:29 | |
| *** rwallner has quit IRC | 16:34 | |
| *** rwallner has joined #openstack-kolla | 16:35 | |
| *** eaguilar has quit IRC | 16:35 | |
| *** jheroux has joined #openstack-kolla | 16:38 | |
| *** daidv has joined #openstack-kolla | 16:41 | |
| *** unicell has quit IRC | 16:43 | |
| *** schwicht has quit IRC | 16:45 | |
| *** daidv has quit IRC | 16:45 | |
| *** zhubingbing has quit IRC | 16:49 | |
| *** daidv has joined #openstack-kolla | 16:49 | |
| *** daidv_ has joined #openstack-kolla | 16:50 | |
| *** daidv_ has quit IRC | 16:50 | |
| *** daidv_ has joined #openstack-kolla | 16:51 | |
| *** david-lyle has quit IRC | 16:51 | |
| *** Pavo has joined #openstack-kolla | 16:53 | |
| *** daidv_ has quit IRC | 16:53 | |
| *** daidv has quit IRC | 16:54 | |
| openstackgerrit | Duong Ha-Quang proposed openstack/kolla-kubernetes master: Heat service deployment https://review.openstack.org/437252 | 16:54 |
| *** salv-orlando has joined #openstack-kolla | 16:54 | |
| *** ksumit has quit IRC | 16:54 | |
| *** newmember has joined #openstack-kolla | 16:59 | |
| *** egonzalez has quit IRC | 17:00 | |
| *** inc0 has quit IRC | 17:00 | |
| *** inc0 has joined #openstack-kolla | 17:00 | |
| *** sambetts is now known as sambetts|afk | 17:00 | |
| duonghq | kfox1111, can you take a look at my ps: https://review.openstack.org/#/c/437252/ can you give me advice on error at http://logs.openstack.org/52/437252/12/check/gate-kolla-kubernetes-deploy-centos-binary-2-ceph-nv/923733d/console.html#_2017-04-19_07_02_14_825574 | 17:00 |
| *** rwallner has quit IRC | 17:00 | |
| jascott1 | relocatin' bbiab | 17:00 |
| kfox1111 | duonghq: sure. sec. | 17:01 |
| *** kbaegis has quit IRC | 17:01 | |
| inc0 | duonghq: we'll take your topics next week | 17:01 |
| duonghq | inc0, thank you | 17:01 |
| kfox1111 | dosaboy: oh. I know what that is. | 17:02 |
| *** rwallner has joined #openstack-kolla | 17:02 | |
| kfox1111 | duonghq: oh. I know what that is. | 17:02 |
| *** mgoddard has quit IRC | 17:02 | |
| kfox1111 | duonghq: trunk just got reversioned to 0.7.0. | 17:02 |
| *** rwallner has quit IRC | 17:02 | |
| *** magicboiz has quit IRC | 17:03 | |
| *** rwallner has joined #openstack-kolla | 17:03 | |
| kfox1111 | duonghq: but your ps is new, so didn't get the version update. please change all 0.6.0 refreences to 0.7.0. | 17:03 |
| duonghq | kfox1111, ah, so I should bump all heat-related microservice and service to 0.7.0? | 17:03 |
| kfox1111 | yeah. that should fix it. | 17:03 |
| duonghq | kfox1111, thanks | 17:03 |
| *** jascott1 has quit IRC | 17:03 | |
| kfox1111 | np. | 17:03 |
| duonghq | kfox1111, will you go to Boston? | 17:04 |
| *** jascott1 has joined #openstack-kolla | 17:04 | |
| kfox1111 | yup. will be there. you? | 17:04 |
| duonghq | me too | 17:04 |
| *** rwallner has quit IRC | 17:04 | |
| inc0 | duonghq: bring something tasty from your area:P | 17:04 |
| sdake | inc0 any chance you can form a signup list of who will actually be at summit since over half our contribs wont be because the US govt is crazy? | 17:04 |
| kfox1111 | duonghq: nice. looking forward to seeing you there. :) | 17:05 |
| rwellum | kfox1111 - when you get a chance look at this for RHEL. Is it normal to pack out /home like this? https://www.irccloud.com/pastebin/HqTwvVVY/ | 17:05 |
| sdake | inc0 lots of contribs/cores don't have visa clearance | 17:05 |
| duonghq | inc0, :P | 17:05 |
| inc0 | sdake: I already bubbled up visa issues to foundation | 17:05 |
| *** kbaegis has joined #openstack-kolla | 17:05 | |
| duonghq | I need to sleep, cya guys | 17:05 |
| sdake | inc0 sounds good | 17:05 |
| sdake | night duonghq | 17:05 |
| kfox1111 | rwellum: for a workstation edition of rhel, yeah, I think thats right. for a server edition, I think they have different, more centos like defaults. | 17:06 |
| kfox1111 | duonghq: have a good one. | 17:06 |
| *** duonghq has quit IRC | 17:06 | |
| rwellum | kfox1111: ok off to recheck wth I installed.... | 17:06 |
| sbezverk | rwellum: wow that is really wierd to have /home stretched accross two file systems | 17:07 |
| sbezverk | I mean two disks | 17:07 |
| *** newmember has quit IRC | 17:07 | |
| kfox1111 | sbezverk: yeah. for a work station, most disk space needs to be fore the user. | 17:07 |
| spsurya | night all | 17:07 |
| kfox1111 | spsurya: nite | 17:07 |
| *** newmember has joined #openstack-kolla | 17:07 | |
| sbezverk | kfox1111: for centos rhel I use the simplest and which I think most relable approch /boot - 500mb /swap 4096mb / - all the rest and it is using lvm | 17:08 |
| *** jascott1 has quit IRC | 17:09 | |
| kfox1111 | sbezverk: yeah. centos I think is based off of the server edition of rhel. so has all the server defaults. | 17:09 |
| sbezverk | need more space, add disk to lvm and then extend | 17:09 |
| *** unicell has joined #openstack-kolla | 17:09 | |
| *** hieulq_ has joined #openstack-kolla | 17:10 | |
| *** harlowja_ has quit IRC | 17:13 | |
| *** rwallner has joined #openstack-kolla | 17:13 | |
| *** rwallner has quit IRC | 17:13 | |
| *** rwallner has joined #openstack-kolla | 17:14 | |
| *** harlowja has joined #openstack-kolla | 17:15 | |
| *** StephenWang1991 has quit IRC | 17:15 | |
| *** harlowja has quit IRC | 17:17 | |
| *** harlowja has joined #openstack-kolla | 17:17 | |
| *** tonanhngo has joined #openstack-kolla | 17:20 | |
| sdake | for those of you that have not voted for the tc election, pleae do vote | 17:24 |
| *** tonanhngo has quit IRC | 17:24 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla master: Fingerprint containers. https://review.openstack.org/458174 | 17:24 |
| kfox1111 | inc0: -----^ | 17:24 |
| rwellum | sbezverk: I know right? Makes sense I guess for desktop - but still not convinced. | 17:24 |
| inc0 | kfox1111: have you seen ML discussion this morning? | 17:25 |
| kfox1111 | inc0: not yet. | 17:25 |
| inc0 | I think pbourke has great idea | 17:25 |
| rwellum | sbezverk: so for centos - you set that up manually correct? ( /boot - 500mb /swap 4096mb / - all the rest and it is using lvm) | 17:25 |
| sbezverk | rwellum: absolutely | 17:26 |
| sbezverk | only manual setup | 17:26 |
| sbezverk | and software packages (minimal server) | 17:26 |
| kfox1111 | inc0: not sure you can use a label for image pulling though. | 17:26 |
| kfox1111 | at least not in kubernetes. | 17:26 |
| rwellum | sbezverk: thanks - going to try that then :) | 17:26 |
| kfox1111 | though for k8s I don't think we want that anyway. | 17:26 |
| kfox1111 | so maybe its ok. | 17:27 |
| inc0 | kfox1111: you can't but what you can do is to use it to easily compare images | 17:27 |
| kfox1111 | yeah. we could store the fingerprint there. | 17:27 |
| sbezverk | rwellum: sure, I am confident it will work well as I use this approach for years | 17:27 |
| kfox1111 | then it gets shipped with the container. | 17:27 |
| rwellum | sbezverk: yes I've run into problems not using minimal server - especially something seems to mess up docker | 17:27 |
| kfox1111 | I do really like that idea. | 17:27 |
| inc0 | you can pull image to local | 17:27 |
| kfox1111 | but pbourke's intention was to avoid needing revisions in the tags. which I think we still need. | 17:28 |
| inc0 | and compare it with existing | 17:28 |
| kfox1111 | I think thats the place where it could shine though. | 17:28 |
| *** tonanhngo has joined #openstack-kolla | 17:28 | |
| kfox1111 | if yuou can get the labels without pulling the image, then a big win. | 17:28 |
| sbezverk | rwellum: right you let setup install some packages by default it will turn into a nightmare | 17:28 |
| kfox1111 | if you ahve to pull the image, computing the fingerprint is relatively light weight to do from scratch then. | 17:28 |
| sbezverk | due to versions collisions with packages you will need to install later | 17:29 |
| inc0 | not sure if docker supports that, but hmm | 17:29 |
| sbezverk | kfox1111: are you talking about adding md5 to the image name? | 17:30 |
| rwellum | sbezverk: yup - learned this the hard way | 17:30 |
| rwellum | But never understood exactly why. | 17:30 |
| kfox1111 | sbezverk: no. a revision number. | 17:30 |
| sbezverk | rwellum: different repos keeps different versions | 17:30 |
| kfox1111 | centos-binary-nova-api:4.0.0-1, centos-binary-nova-api:4.0.0-2, etc. | 17:31 |
| kfox1111 | so if its the same exact container, but has gotten rpm upgrades, its different. | 17:31 |
| sbezverk | example rhel repo might keep one version of rpm but epel another | 17:31 |
| sbezverk | so here is the collision | 17:31 |
| sbezverk | kfox1111: ah ok it will be used just at the tagging time | 17:32 |
| sbezverk | to see if new version needs to be published | 17:32 |
| kfox1111 | sbezverk: yeah. | 17:32 |
| kfox1111 | exactly. | 17:33 |
| kfox1111 | I want to be able to: | 17:33 |
| kfox1111 | helm upgrade nova-api kolla/nova-api | 17:33 |
| kfox1111 | done. :) | 17:33 |
| *** schwicht has joined #openstack-kolla | 17:33 | |
| *** gianpietro has joined #openstack-kolla | 17:33 | |
| *** gianpietro has quit IRC | 17:35 | |
| rwellum | kfox1111: This is what I installed: rhel-server-7.3-x86_64-dvd.iso | 17:36 |
| *** iceyao has joined #openstack-kolla | 17:37 | |
| kfox1111 | rwellum: hmm... weird. not sure then why it would default to that layout. | 17:37 |
| kfox1111 | lets see... | 17:37 |
| kfox1111 | oh... one other thought. | 17:37 |
| kfox1111 | I think rhel registration codes can be workstation/server related? | 17:38 |
| kfox1111 | its been a while since I played with them tough, so don't remember if that happens before/after partitioning. I would guess after. | 17:38 |
| kfox1111 | though if it was provided the code before partitioning, it might be able to figure it out. | 17:39 |
| sbezverk | kfox1111: yes, but rhel has WS flavor of image | 17:39 |
| rwellum | It happens well after partitioning :) | 17:39 |
| sbezverk | I used exactly the same image for my installation | 17:39 |
| sbezverk | rwellum: my rule of thumb, do not use defaults unless you established them ;) | 17:40 |
| rwellum | Well I am going to re-image and do exactly that :) | 17:40 |
| *** mgoddard has joined #openstack-kolla | 17:41 | |
| portdirect | sbezverk: i like your styles | 17:41 |
| rwellum | I'll be careful on the subscription as well - just in case kfox1111 has a point | 17:41 |
| portdirect | have you played with centos atomic host yet, if not I'm pretty sure you'd like it | 17:41 |
| *** iceyao has quit IRC | 17:41 | |
| portdirect | (and/or rhel atomic) | 17:42 |
| kfox1111 | I generally use centos and cobbler with a kickstart so always specify my partions manually anyway. | 17:42 |
| kfox1111 | well, manually but scripted. | 17:42 |
| kfox1111 | more magicical. :) | 17:42 |
| rwellum | Is Atomic stable portdirect ? | 17:42 |
| portdirect | rock solid | 17:43 |
| *** sudipto has quit IRC | 17:43 | |
| *** sudipto has joined #openstack-kolla | 17:43 | |
| *** sudipto has quit IRC | 17:44 | |
| *** hieulq_ has quit IRC | 17:44 | |
| *** gianpietro has joined #openstack-kolla | 17:44 | |
| *** gfidente is now known as gfidente|afk | 17:45 | |
| rwellum | portdirect: interesting. So does the default partitioning work? :) | 17:45 |
| rwellum | For containerization etc. | 17:45 |
| portdirect | 100% though I'd still reommend setting it up for your own needs | 17:46 |
| kfox1111 | inc0: responded to a bunch of the thread. | 17:46 |
| kfox1111 | inc0: I dont think the label idea is fully thought through. | 17:47 |
| portdirect | I run a 4gb root and put logs on their own part, set an appropriate swap up and then use the rest for docker. | 17:47 |
| kfox1111 | inc0: reguardless, what do you think of the ps I posted? we will need that either way. | 17:47 |
| *** eaguilar has joined #openstack-kolla | 17:48 | |
| rwellum | Thanks portdirect | 17:49 |
| inc0 | change with fingerprint in gate is good | 17:50 |
| openstackgerrit | Merged openstack/kolla master: Add OpenDaylight Container to kolla https://review.openstack.org/416369 | 17:50 |
| inc0 | regardless | 17:50 |
| inc0 | mgkwill ^ yay | 17:50 |
| openstackgerrit | Kevin Fox proposed openstack/kolla master: Fingerprint containers. https://review.openstack.org/458174 | 17:51 |
| kolla-slack | <jascott1> woah helm install tgz is missing all of a sudden https://kubernetes-helm.storage.googleapis.com/helm--linux-amd64.tar.gz | 17:51 |
| mgkwill | inc0: awesome! | 17:53 |
| *** jaosorior is now known as jaosorior_away | 17:54 | |
| kolla-slack | <jascott1> false alarm its just the IP im coming out of is rate limited at github | 17:59 |
| kfox1111 | ah. you hit it too hard. :) | 18:00 |
| kolla-slack | <jascott1> lol | 18:00 |
| kolla-slack | <jascott1> so much for the non-proxy network that makes life livable and work doable | 18:00 |
| kfox1111 | ;) | 18:01 |
| kolla-slack | <jascott1> why do I come here? its a series of work speed bumps :S | 18:01 |
| kolla-slack | <jascott1> (here being the office) | 18:01 |
| kfox1111 | just need to tunnel back home. ;) | 18:01 |
| kfox1111 | and uncheck the yearly review "do I have everything nessisary to do my job" checkbox. :) | 18:02 |
| inc0 | I set up my nuc for dev that sits nicely on my desk at home | 18:02 |
| kolla-slack | <jascott1> hm i live 2 minutes away I could probably construct a tunnel in a few years ;) | 18:02 |
| kfox1111 | heh. thats one way to do it. ;) | 18:02 |
| kolla-slack | <jascott1> inc0 cool got it all dialed in? | 18:02 |
| openstackgerrit | David Moreau Simard proposed openstack/kolla-ansible master: Do not merge: Test ARA 0.12.5rc1 https://review.openstack.org/439963 | 18:03 |
| openstackgerrit | David Moreau Simard proposed openstack/kolla-ansible master: Do not merge: Test ARA 0.12.5rc1 https://review.openstack.org/439963 | 18:03 |
| inc0 | pretty much yeah | 18:03 |
| kfox1111 | ssh socks proxies make life sooo much easier. :) | 18:05 |
| *** jascott1 has joined #openstack-kolla | 18:10 | |
| *** kbaegis has quit IRC | 18:10 | |
| *** kbaegis has joined #openstack-kolla | 18:15 | |
| inc0 | kfox1111: tell me this, how would you define whether to pull from dockerhub or not? | 18:16 |
| *** gianpietro has quit IRC | 18:21 | |
| *** eaguilar has quit IRC | 18:23 | |
| *** unicell has quit IRC | 18:23 | |
| *** rwallner has quit IRC | 18:24 | |
| *** rwallner has joined #openstack-kolla | 18:24 | |
| *** satyar has quit IRC | 18:25 | |
| *** unicell has joined #openstack-kolla | 18:27 | |
| inc0 | kfox1111: I think for devenv we should have setup_gate_emptydir:) | 18:32 |
| *** rwallner has quit IRC | 18:32 | |
| *** Pavo has quit IRC | 18:44 | |
| *** lucasxu has joined #openstack-kolla | 18:45 | |
| *** rwallner has joined #openstack-kolla | 18:53 | |
| *** eaguilar has joined #openstack-kolla | 18:55 | |
| *** jmguzmanc has joined #openstack-kolla | 19:02 | |
| *** MasterOfBugs has joined #openstack-kolla | 19:03 | |
| *** david-lyle has joined #openstack-kolla | 19:06 | |
| *** lucasxu has quit IRC | 19:08 | |
| *** lucasxu has joined #openstack-kolla | 19:09 | |
| kfox1111 | inc0: say we have centos-binary-nova-api:4.0.0-1. | 19:10 |
| kfox1111 | user: | 19:10 |
| kfox1111 | helm install --name nova-api kolla/nova-api-deployment | 19:10 |
| kfox1111 | it pulls kolla/nova-api-deployment-4.0.0-1.tgz and installs it. | 19:11 |
| kfox1111 | it points to dockerhub/kolla/centos-binary-nova-api:4.0.0-1. now, redhat releases openssl that fixes a security update. | 19:11 |
| kfox1111 | the build system detects openssl changed, and pushes dockerhub/kolla/centos-binary-nova-api:4.0.0-2, | 19:12 |
| kfox1111 | and a helm package updated to point to it. | 19:12 |
| inc0 | are we going to publish new helm chart every time we publish image? | 19:12 |
| kfox1111 | for the updated package then to land on the users system, the user simply has to: | 19:13 |
| kfox1111 | helm upgrade nova-api kolla/nova-api-deployment | 19:13 |
| kfox1111 | I think so. | 19:13 |
| kfox1111 | we maintain a kolla-kubernetes helm repo with packages people can pull from, the same way we manage containers they can pull rom. | 19:14 |
| kfox1111 | same kind of use case. | 19:14 |
| inc0 | if you run helm upgrade nova-api while image and tag stays the same, but image has different sha1, will it upgrade? | 19:14 |
| kfox1111 | there are two modes of configuration. | 19:14 |
| kfox1111 | in the default, it will skip the upgrade, as we have imagePullPolicy=IfNotPresent | 19:14 |
| kfox1111 | it could be set to Always. | 19:15 |
| inc0 | so way I'd run it locally | 19:15 |
| kfox1111 | but actualy in that mode, nothing will change in the final k8s objects and k8s will fail to do the upgrade. | 19:15 |
| inc0 | (regardless of scale) | 19:15 |
| inc0 | ehh | 19:15 |
| kfox1111 | you would have to set it to Always, and skip the helm upgrade, and delete all the active pods. | 19:16 |
| inc0 | ok, so :latest tag will never work properly with helm+k8s? | 19:16 |
| kfox1111 | right. | 19:16 |
| inc0 | how fun | 19:16 |
| kfox1111 | it will work the first time. | 19:16 |
| kfox1111 | it won't support ugprades. | 19:16 |
| inc0 | -.- | 19:16 |
| kfox1111 | thats why I've been talking about tags so much. | 19:16 |
| kfox1111 | k8s is all about atomic images. | 19:16 |
| *** david-lyle_ has joined #openstack-kolla | 19:17 | |
| kfox1111 | that only works if you can distinguish containers, and thats done with tags. | 19:17 |
| *** david-lyle has quit IRC | 19:17 | |
| inc0 | so what I don't see happening then is constantly-upgrading use case | 19:18 |
| inc0 | I mean publishing helm charts alongside of images every day | 19:18 |
| inc0 | especially if we add caveat of publish only if something changes | 19:18 |
| inc0 | we create chain of co-dependency that's going to be hard to maintain | 19:19 |
| kfox1111 | I don't think its htat hard to maintain. | 19:20 |
| kfox1111 | we look at containers. and previously published helm packages. | 19:20 |
| kfox1111 | if they don't point to the newest, we update the values file and push to tarballs.o.o. | 19:20 |
| inc0 | but tags are in code right? | 19:21 |
| kfox1111 | values.yaml. | 19:21 |
| inc0 | which everyone will have on their own | 19:21 |
| kfox1111 | should be simple to poke in the values and helm package. | 19:21 |
| kfox1111 | so, that may need further integrations, yeah. | 19:22 |
| inc0 | I assume people will run their own values.yaml, no 2 openstack are exactlu the same | 19:22 |
| kfox1111 | but, this is in a catagory of, you want to use kolla-kubernetes and build your own helm packages, | 19:22 |
| kfox1111 | but don't want to build your own containers. | 19:22 |
| kfox1111 | think that will be common? | 19:23 |
| kfox1111 | (other then maybe for dev) | 19:23 |
| kfox1111 | there are multiple sets of values.yaml. | 19:23 |
| inc0 | for dev ironically I'd assume local registry | 19:23 |
| kfox1111 | there are the ones inside the packages, and there are users values.yaml | 19:23 |
| kfox1111 | I'm talking about the defaults in the packages. | 19:23 |
| inc0 | but values.yaml are for deploy phase not chart build right? | 19:23 |
| kfox1111 | yes/no. | 19:24 |
| inc0 | as in, if I want to do enable_neutron_lbaas, I still use kolla-prebuilt charts? | 19:24 |
| kfox1111 | the package has a values.yaml that has defaults. | 19:24 |
| kfox1111 | the user can provide a vaulues.yaml as well. | 19:24 |
| kfox1111 | at install/upgrade time, the values.yaml from the package is taken, then anything in the users one overrides that. | 19:25 |
| inc0 | yeah and I assume that will be the case most of the time | 19:25 |
| *** eaguilar has quit IRC | 19:25 | |
| inc0 | and defaults are in repo so I don't want to edit these automagically every time we update containers | 19:25 |
| kfox1111 | so if the user specified version 4.0.0-75 and the package 4.0.0-5, the users version would be taken. | 19:25 |
| kfox1111 | so if they are vbuilding their own stuff, and want their own revision numbers they are free to do so. | 19:25 |
| kfox1111 | in the case of prebuilt packages, | 19:26 |
| kfox1111 | they would point to 4.0.0-5, and if the user didn't specify, they woudl get the "most up to date version" | 19:26 |
| kfox1111 | and on upgrade, it would point to a newer version, and then do the right thing. | 19:26 |
| inc0 | how about defaults? | 19:27 |
| kfox1111 | defaults for what? | 19:27 |
| *** eaguilar has joined #openstack-kolla | 19:28 | |
| inc0 | https://github.com/openstack/kolla-kubernetes/blob/master/helm/microservice/cinder-api-deployment/templates/cinder-api.yaml#L5 | 19:28 |
| kfox1111 | right now, that comes from helm/all_values.yaml. | 19:28 |
| inc0 | or rather that | 19:28 |
| inc0 | https://github.com/openstack/kolla-kubernetes/blob/master/helm/all_values.yaml | 19:28 |
| inc0 | yeah | 19:28 |
| kfox1111 | yeah. | 19:28 |
| inc0 | so we publish image daily | 19:28 |
| inc0 | do we edit this file daily? | 19:29 |
| inc0 | to up a revision on image | 19:29 |
| kfox1111 | I guess the question is, do we push it back to git, or just do it in the prebuilt packages we ship? | 19:29 |
| kfox1111 | we up the revision only when it chagnes. | 19:29 |
| kfox1111 | no reason to make it look like updates are available when nothing changes. | 19:30 |
| inc0 | so one thing I'd like to have is full and easy repeatibility of helm build locally | 19:30 |
| inc0 | so no-no for automagical values.yaml | 19:30 |
| kfox1111 | we have that now. | 19:30 |
| inc0 | not officially really | 19:30 |
| kfox1111 | but are you talking pointing at local built containers, or for hub ones? | 19:30 |
| inc0 | mechanism will be the same reall | 19:30 |
| inc0 | y | 19:30 |
| kfox1111 | I think some folks will want to run this kind of process at their own sites with their own ci/cd. | 19:31 |
| inc0 | ok, I think your use case will be waaaay too complex to have a quick win:/ | 19:31 |
| kfox1111 | so the plumbing should be in tree. | 19:31 |
| inc0 | ofc | 19:31 |
| kfox1111 | I don't think it will take too long to implement really. | 19:31 |
| inc0 | so how about scenerio I described? | 19:31 |
| inc0 | you retag your own images | 19:31 |
| inc0 | and we'll provide tools to make it easier | 19:31 |
| inc0 | kolla-build pull-changed | 19:32 |
| inc0 | or sth like that | 19:32 |
| kfox1111 | I think maybe we have a tool that takes a yaml of container -> version's, and the tool tweaks the values.yaml as appropriate to build? | 19:32 |
| kfox1111 | you don't need to run it if you dont want. | 19:32 |
| kfox1111 | if you update and build revisions for things, then your tooling can call that and update the things, then build the packages? | 19:33 |
| kfox1111 | ideally, the file even would be spit out of the kolla build process. | 19:33 |
| inc0 | uhh | 19:34 |
| inc0 | we're talking about really complex system:/ | 19:34 |
| inc0 | in my mind at leasy | 19:34 |
| kfox1111 | maybe we make a kolla build mode that incorperates the fingerprinting. | 19:34 |
| kfox1111 | it does what it does today, but with a :latest tag. | 19:34 |
| inc0 | yeah, and LABELs | 19:34 |
| kfox1111 | then compares the containers it just built with a repo specified. | 19:35 |
| inc0 | or do something like | 19:35 |
| kfox1111 | it then tags the containers that changed with a rev+1, | 19:35 |
| inc0 | pull changed, retag locally, push to local registry | 19:35 |
| kfox1111 | and writes that to a yaml manifiest file. | 19:35 |
| inc0 | and we'll host tip of branch on dockerhub | 19:35 |
| kfox1111 | or even I guess the whole list of freshest revisions. | 19:35 |
| inc0 | so everyone will install from local registry | 19:35 |
| inc0 | but we'll provide tools to mainain it | 19:35 |
| kfox1111 | then kolla-kubernetes can take that file, update values.yaml with it. | 19:36 |
| kfox1111 | yeah. | 19:36 |
| inc0 | well if we keep all local, we can render values.yaml for kolla-k8s from kolla | 19:36 |
| *** pbourke has quit IRC | 19:36 | |
| inc0 | kolla-build pull-changes -o values.yaml | 19:36 |
| inc0 | or sth like that | 19:36 |
| *** pbourke has joined #openstack-kolla | 19:37 | |
| kfox1111 | yeah. | 19:37 |
| kfox1111 | then that values.yaml goes through something like kolla-kubernetes-update-values.py values.yaml | 19:37 |
| inc0 | kolla-build pull-changes -o values.yaml --repush my-registry:5000 --increment-revision | 19:37 |
| *** StephenWang1991 has joined #openstack-kolla | 19:37 | |
| kfox1111 | yeah. | 19:37 |
| inc0 | that thing could be run in cron | 19:38 |
| kfox1111 | yeah. :) | 19:38 |
| inc0 | values.yaml it outputs can be formated for helm | 19:38 |
| inc0 | so we won't need 2 separate tools | 19:39 |
| kfox1111 | or just kept very generic. key/value. | 19:39 |
| kfox1111 | then it could be used some day by kolla-ansible too if it wanted this kind of upgrade support. | 19:39 |
| inc0 | we'll figure this one out | 19:39 |
| inc0 | kolla-ansible is easier, you don't need to change tags;) | 19:39 |
| inc0 | but yeah | 19:39 |
| kfox1111 | the tool from generic yaml to helm specific is very small. | 19:39 |
| inc0 | anyway, that's detail | 19:40 |
| inc0 | so to summarize | 19:40 |
| inc0 | 1. we provide automatic pusher for dockerhub for tip of branch | 19:40 |
| inc0 | 2. we create tool to fingerprint versions, ideally in LABEL | 19:40 |
| inc0 | 3. we create tool to repush/retag changed containers to local registry | 19:41 |
| inc0 | it will output freshest list of image:tag | 19:41 |
| kfox1111 | the tags gotta get pushed to the hub too. | 19:41 |
| kfox1111 | local/remote registry. | 19:41 |
| inc0 | so that's where I don't really agree | 19:42 |
| inc0 | I don't want to maintain all that using dockerhub, not yet anyway | 19:42 |
| *** StephenWang1991 has quit IRC | 19:42 | |
| kfox1111 | whats to maintain? | 19:42 |
| kfox1111 | we keep the most recent tag. | 19:42 |
| inc0 | for one, kolla-ansible defaults | 19:42 |
| kfox1111 | we delete all older ones. | 19:42 |
| kfox1111 | keep 4.0.0, which is a link to the 4.0.0-N. | 19:42 |
| kfox1111 | when a new one comes in, 4.0.0 points to 4.0.0-N+1, and 4.0.0-N is deleted. | 19:43 |
| inc0 | if we keep just freshest, thats better | 19:43 |
| inc0 | we'll effectively just mirror on dockerhub some local registry somewhere | 19:43 |
| kfox1111 | it costs no extra space then we store now. | 19:43 |
| kfox1111 | I'm not arguing we should keep around 100's of old versions. | 19:43 |
| inc0 | we'll need to prep kolla-ansible for that tho | 19:43 |
| kfox1111 | just that the way we tag things allows for change detection. | 19:43 |
| inc0 | as today it doesn't support per-image tagging | 19:43 |
| kfox1111 | it would just point to the generically tagged 4.0.0 as it does today. | 19:44 |
| *** iceyao has joined #openstack-kolla | 19:44 | |
| kfox1111 | it would still work without changes. | 19:44 |
| kfox1111 | 4.0.0 is "the freshest", (latest). | 19:44 |
| kfox1111 | 4.0.0-1 is the specific build. | 19:44 |
| inc0 | no, it will be ocata-1 | 19:44 |
| inc0 | 4.0.0 is 4.0.0 tag in github | 19:44 |
| kfox1111 | pike-1 maybe. | 19:45 |
| inc0 | yeah or pike | 19:45 |
| kfox1111 | 4.0.0's already out there. | 19:45 |
| inc0 | yup | 19:45 |
| kfox1111 | well... lets think this through... | 19:45 |
| inc0 | that's what we're doing:P | 19:45 |
| kfox1111 | how would you handle changes to kolla code? would you just bump the revision again? | 19:45 |
| kfox1111 | so, using the current convention + revision, | 19:46 |
| kfox1111 | we'd have, | 19:46 |
| inc0 | yeah, I assume so | 19:46 |
| *** tonanhngo has quit IRC | 19:46 | |
| kfox1111 | 4.0.0-1, 4.0.0-2, 4.0.1-1, 4.0.1-2, etc. | 19:46 |
| inc0 | if we fix bug in kolla container | 19:46 |
| *** eaguilar has quit IRC | 19:46 | |
| kfox1111 | with pike, | 19:46 |
| inc0 | you want this bug to be fixed too | 19:46 |
| kfox1111 | those would be pike-1, pike-2, pike-3, pike-4? | 19:46 |
| inc0 | so I'm nto sure if we'll ever have 4.0.0-1 | 19:47 |
| inc0 | we'll have 4.0.0 and that's it | 19:47 |
| kfox1111 | Its our current "stable". we should have revisions of our current stable. | 19:47 |
| *** dmk0202 has joined #openstack-kolla | 19:47 | |
| inc0 | we'll have tip of the stable branch | 19:47 |
| kfox1111 | assuming this build system is completed while ocata is still stable. | 19:47 |
| inc0 | and we'll have tip-revision too | 19:47 |
| kfox1111 | if it takes till pile to implement, then ok. | 19:48 |
| *** eaguilar has joined #openstack-kolla | 19:48 | |
| inc0 | (I still want :latest like tag) | 19:48 |
| kfox1111 | 4.0.0 is the equiv of :latest today. | 19:48 |
| *** iceyao has quit IRC | 19:48 | |
| kfox1111 | agreed, we do want some kind of equiv of :latest for stable branches. | 19:48 |
| inc0 | well only because we don't have daily pushes | 19:48 |
| inc0 | so we'll have daily push ocata | 19:48 |
| inc0 | that will also be tagged ocata-1 | 19:48 |
| inc0 | same image, 2 tags | 19:49 |
| kfox1111 | k. | 19:49 |
| inc0 | then next one will be :ocata and :ocata-2 | 19:49 |
| kfox1111 | that would cover the use case. | 19:49 |
| kfox1111 | right. | 19:49 |
| kfox1111 | and we ignore kolla's versioning entirely? | 19:49 |
| inc0 | we'll push tags but that's going to be one time thing I'd assume | 19:49 |
| sbezverk | inc0: what -2 means in the image label? | 19:49 |
| inc0 | at least for now | 19:49 |
| inc0 | -2 means get away from this change;) | 19:50 |
| kfox1111 | sbezverk: container revision number. | 19:50 |
| inc0 | -2 is revision sbezverk | 19:50 |
| kfox1111 | container ocata-2 is a newer build then ocata-1 | 19:50 |
| kfox1111 | so its an "upgrade" | 19:50 |
| sbezverk | interesting, got it.. | 19:50 |
| *** tonanhngo has joined #openstack-kolla | 19:50 | |
| sbezverk | which will be the latest then? | 19:51 |
| inc0 | btw it's still going to be bitch to implement with docker becasue afaik docker registry currently doesn't allow you to quert all the tags available for image | 19:51 |
| inc0 | sbezverk: latest would be :ocata | 19:51 |
| inc0 | and :master | 19:51 |
| openstackgerrit | Kevin Fox proposed openstack/kolla master: Fingerprint containers. https://review.openstack.org/458174 | 19:51 |
| *** tonanhngo has quit IRC | 19:51 | |
| inc0 | and I don't think we'll revision :master | 19:51 |
| kfox1111 | sbezverk: just ocata | 19:51 |
| inc0 | as it will always change | 19:51 |
| sbezverk | kfox1111: cool push down stack kind of things | 19:52 |
| *** david-lyle_ has quit IRC | 19:52 | |
| kfox1111 | sbezverk: yeah. | 19:52 |
| kfox1111 | sbezverk: if you see you have ocata-2 and there's a ocata-4 out on the hub, you know you need to do some updateing. :) | 19:52 |
| *** tonanhngo has joined #openstack-kolla | 19:52 | |
| inc0 | I'll sumarize this on mailing list | 19:53 |
| inc0 | and you guys check if we understood each other | 19:53 |
| *** spsurya has quit IRC | 19:53 | |
| kfox1111 | k. | 19:53 |
| *** mgoddard has quit IRC | 19:56 | |
| *** manheim has joined #openstack-kolla | 19:58 | |
| *** tonanhngo has quit IRC | 19:59 | |
| jascott1 | kfox1111 PR for simple --net-host was just merged. yay | 19:59 |
| *** eaguilar has quit IRC | 20:01 | |
| kfox1111 | jascott1: cool. | 20:02 |
| *** manheim has quit IRC | 20:02 | |
| *** eaguilar has joined #openstack-kolla | 20:02 | |
| kfox1111 | so I think the next couple things for that is the DaemonSet support and nodeSelector. | 20:02 |
| *** eaguilar has quit IRC | 20:02 | |
| kfox1111 | with those two, I think the flannel helm chart will work. | 20:03 |
| kfox1111 | with 1.6. | 20:03 |
| *** iceyao has joined #openstack-kolla | 20:03 | |
| kfox1111 | the patch that just merged should work fine with k8s 1.5. | 20:03 |
| jascott1 | i might just ask technosophos about daemonset before starting a thread | 20:03 |
| kfox1111 | yeah. | 20:03 |
| *** tonanhngo has joined #openstack-kolla | 20:05 | |
| jascott1 | PSA: #helm is now #helm-users and a new channel #helm-dev has been created | 20:07 |
| kfox1111 | in slack too, or just irc? | 20:07 |
| *** iceyao has quit IRC | 20:08 | |
| jascott1 | slack afaik | 20:08 |
| kfox1111 | ok. thanks for the info. :) | 20:09 |
| jascott1 | we are talking about it now if you have a min | 20:12 |
| inc0 | kfox1111: mail sent, comment plz:) I'm going off for a few, need to clear my mind | 20:13 |
| inc0 | bbiaf | 20:13 |
| kfox1111 | k. | 20:13 |
| inc0 | I like this solution, seems flexible and elegant | 20:13 |
| *** lucasxu has quit IRC | 20:13 | |
| *** lucasxu has joined #openstack-kolla | 20:14 | |
| inc0 | in fact lately I started to think that we create so much awesome tools around container builds | 20:16 |
| inc0 | that it would be great to allow other people to use it outside openstack too | 20:16 |
| kfox1111 | +1 | 20:16 |
| inc0 | something to think about:) | 20:16 |
| kfox1111 | keeping fresh containers is a very unsolved problem in docker community outside of those containers that are purely ci/cd and are daily recieving commits. | 20:16 |
| inc0 | keeping up to date containers that aren't equivalent to curl | sh is even harder | 20:17 |
| kfox1111 | as the latter just punts on the problem and assumes frequency is good enough. | 20:17 |
| inc0 | I mean you know what's inside | 20:17 |
| kfox1111 | eyah. | 20:17 |
| kfox1111 | yeah. | 20:17 |
| inc0 | ok bbiaf | 20:18 |
| jascott1 | kfox1111 well the initial idea of daemonset was not shot down so that is something. I will get something going with it | 20:20 |
| kfox1111 | jascott1: cool. :) | 20:20 |
| kfox1111 | when I tried it in gate, I only had to do two things. kind=Deployment to kind=DaemonSet and removing .spec.replicas. | 20:21 |
| kfox1111 | for multinode we need --node-selector helm=true thingy too. | 20:21 |
| kfox1111 | so we can tag only one node with helm=true. | 20:21 |
| jascott1 | dont they run in all pods by default? so the flip is to make sure it is restricted right? | 20:22 |
| kfox1111 | with a daemonset, it will run on all selected nodes. | 20:22 |
| kfox1111 | I think tiller will malfunction if there is more then one though. | 20:22 |
| jascott1 | right just thinking about safe by default (not launching a million tillers) | 20:23 |
| kfox1111 | so being able to set a node selector and pin it to just one node ensures only one will ever run at a time. | 20:23 |
| jascott1 | so require a label or use default label | 20:23 |
| kfox1111 | safest thing is probably to automatically add --node-selector helm=true. | 20:23 |
| kfox1111 | yeah. | 20:23 |
| jascott1 | cool th | 20:24 |
| jascott1 | +x | 20:24 |
| *** iceyao has joined #openstack-kolla | 20:24 | |
| *** srwilkers has quit IRC | 20:25 | |
| *** sayantan_ has joined #openstack-kolla | 20:26 | |
| kfox1111 | jascott1: no, thank you. :) | 20:26 |
| *** iceyao has quit IRC | 20:29 | |
| *** rwallner has quit IRC | 20:29 | |
| *** srwilkers has joined #openstack-kolla | 20:30 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes master: WIP: Test out flannel helm package. https://review.openstack.org/456964 | 20:35 |
| *** david-lyle_ has joined #openstack-kolla | 20:38 | |
| *** Serlex has quit IRC | 20:41 | |
| *** mgoddard has joined #openstack-kolla | 20:44 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla master: Fingerprint containers. https://review.openstack.org/458174 | 20:50 |
| *** david-lyle_ has quit IRC | 20:53 | |
| *** jascott1 has quit IRC | 20:57 | |
| *** jascott1 has joined #openstack-kolla | 20:57 | |
| inc0 | back | 20:58 |
| kfox1111 | wb | 20:58 |
| inc0 | kfox1111: so, shall we start making kolla-registry toolset? | 20:58 |
| kfox1111 | +1 | 20:59 |
| kfox1111 | I just updated the fingerprinting code. one minor issue on ubuntu. | 20:59 |
| inc0 | good! python coding | 20:59 |
| kfox1111 | I think it will work though. | 20:59 |
| kfox1111 | +1 | 20:59 |
| inc0 | well, at the end of the day we need to figure out LABEL thingy | 20:59 |
| inc0 | or some other way to get it out of container and put it into metadata | 21:00 |
| kfox1111 | was wondering if there was a second option there... | 21:00 |
| kfox1111 | we're making a yaml file with teh manifest of the most recent stuff to pass to kolla-kubernetes. | 21:00 |
| kfox1111 | should we just post that to tarballs.o.o? | 21:00 |
| inc0 | file inside container is good idea, but hard to get out from inside container | 21:00 |
| *** jrobinson has joined #openstack-kolla | 21:00 | |
| *** jascott1 has quit IRC | 21:01 | |
| inc0 | so I want to make thing that's fully universal and based on 2 registries, no matter what they are | 21:01 |
| inc0 | so you can have cicd-fed registry in your infra | 21:01 |
| kfox1111 | yeah. its about as hard to extract from inside the container then it is to just finger print the container in place. | 21:01 |
| inc0 | and prod registry that hosts containers deployed | 21:01 |
| kfox1111 | fully universal good. | 21:01 |
| kfox1111 | whould work with either hub or local repo. | 21:01 |
| inc0 | that's the idea | 21:01 |
| inc0 | I kinda have idea how to do the labeling | 21:02 |
| *** kbaegis has quit IRC | 21:02 | |
| inc0 | we can output some well structured fingerprint to stdout | 21:02 |
| inc0 | during build | 21:02 |
| kfox1111 | for kolla + kolla-kubernetes, helm's the only tool the user shoudl have to interact with if they want. | 21:02 |
| inc0 | and kolla-build will screen-scrape it and put on image | 21:02 |
| kfox1111 | that would work. | 21:02 |
| kfox1111 | but if you wanted to make the tooling more generic, like we talked about, | 21:02 |
| kfox1111 | the aproach I took in the ps woudl work better. | 21:03 |
| inc0 | well labels are generic tho | 21:03 |
| inc0 | they're really good because you can always docker inspect running contanier | 21:03 |
| kfox1111 | no, the idea that you run the fingerprint script on a built container, then place the fingerprint in a label. | 21:03 |
| kfox1111 | rather then have the fingerprint spit out during build. | 21:03 |
| kfox1111 | the former works with all containers. the latter, only ones you built. | 21:04 |
| kfox1111 | and possibly modified the build of. | 21:04 |
| *** gianpietro has joined #openstack-kolla | 21:04 | |
| inc0 | hmm, well, let's see | 21:04 |
| *** jrobinson has quit IRC | 21:04 | |
| *** jrobinson has joined #openstack-kolla | 21:04 | |
| inc0 | that's implementation detail, I don't care that much for contianers not built by us | 21:04 |
| inc0 | I mean it's nice if we can make it this generic | 21:04 |
| kfox1111 | inc0: have a look at: https://review.openstack.org/#/c/458174/3/tools/gate_run.sh | 21:05 |
| kfox1111 | it should work on any container. | 21:05 |
| inc0 | kolla-build manifesto image-name:tag | 21:05 |
| kfox1111 | we just change the > logs/fingerprints.txt to docker label $line $fingerprint in the loop. | 21:05 |
| inc0 | yeah I guess | 21:05 |
| inc0 | ok, we can do both really | 21:05 |
| inc0 | I'll keep that in mind | 21:06 |
| kfox1111 | probably really should split that fingerprint thingy out to its own script even. | 21:06 |
| kfox1111 | then move the forloop outside. | 21:06 |
| sdake | kfox1111 - hey I know your in a conversation atm | 21:06 |
| inc0 | it's own cli call | 21:06 |
| kfox1111 | yeah. | 21:06 |
| sdake | kfox1111 but I want to intrude on a different topic | 21:06 |
| sdake | kfox1111 perhapas you can answere with a yes or no | 21:06 |
| kfox1111 | sdake: sure. whats up? | 21:06 |
| sdake | to do multinode kolla-kubernetes, you need keepalived only? | 21:06 |
| sdake | or keepalivd and haproxy? | 21:07 |
| kfox1111 | sdake: heh. well... | 21:07 |
| kfox1111 | keepalived to route traffic into the system. | 21:07 |
| kfox1111 | haproxy is currently being used in all the api pods to do connection tracking for safe rolling upgrades and some day ssl termination. | 21:07 |
| kfox1111 | but its not used in the way you would think of using haproxy as a load balancer. | 21:08 |
| sdake | haproxy is implemented already? | 21:08 |
| *** goldyfruit has quit IRC | 21:08 | |
| kfox1111 | sdake: sort of... see helm/kolla-common/templates/_common_api_python_deployment.yaml | 21:08 |
| kfox1111 | its not used as an l7 lb. | 21:08 |
| sdake | kfox1111 thanks - i'll take a look | 21:08 |
| inc0 | guys...am I understanding this correctly? docker is being renamed to moby? oO | 21:09 |
| kfox1111 | keepalived provides a vip. kube-proxy provides an l4 lb to the pods. | 21:09 |
| sdake | inc0 who woulda guessed :) | 21:09 |
| kfox1111 | inc0: docker's being renamed essentially to containerd I thought. | 21:09 |
| inc0 | scrach that, it uses containerd | 21:09 |
| sdake | bbiaf - need to collect child | 21:09 |
| inc0 | omfg | 21:09 |
| kfox1111 | sdake: then once the traffic is at the pod, it goes through haproxy to do connection tracking/ssl termination, then through 127.0.0.1 to the python service. | 21:10 |
| inc0 | they made openstack for containers | 21:10 |
| inc0 | each piece is swappable and such | 21:11 |
| *** jascott1 has joined #openstack-kolla | 21:11 | |
| inc0 | which is not a bad idea tbh | 21:11 |
| * kfox1111 looks up moby.... | 21:11 | |
| jascott1 | have we looked for 3rd party container comparison/introspection tools? | 21:12 |
| * inc0 grabs popcorn watching politics around containers | 21:12 | |
| jascott1 | man i would figure there was something `comprehensive` out there already | 21:12 |
| inc0 | we had some notes from summit | 21:12 |
| jascott1 | iirc there was security tool but dont recall comparison tool for 2 containers | 21:12 |
| *** chas has quit IRC | 21:14 | |
| *** chas has joined #openstack-kolla | 21:14 | |
| kfox1111 | this sounds like its more of a modular swarm. | 21:14 |
| kfox1111 | this is an attempt to un-upset all the parners they alientated with swarm I think. | 21:15 |
| kfox1111 | further atemts to try and get market share away from kubernetes. | 21:16 |
| jascott1 | know your role Docker! | 21:16 |
| jascott1 | j/k | 21:16 |
| *** jtriley has quit IRC | 21:16 | |
| kfox1111 | they as a company know what their role is. thye just don't wanto play with the emerging community. | 21:16 |
| kfox1111 | they want to try and own a community. | 21:17 |
| jascott1 | they do have that Canonical partnership | 21:17 |
| kfox1111 | last I looked they were alieneting redhat/ubuntu/coreos/google/etc. | 21:18 |
| kfox1111 | the 'partnerships' when from, | 21:18 |
| kfox1111 | company X is working on docker, | 21:18 |
| kfox1111 | to company X isn't blocking docker inc's stuff from working on their supported distro. | 21:18 |
| *** lucasxu has quit IRC | 21:19 | |
| jascott1 | i was refering to https://www.docker.com/docker-news-and-press/docker-and-canonical-partner-cs-docker-engine-millions-ubuntu-users | 21:19 |
| *** chas has quit IRC | 21:19 | |
| kfox1111 | and docker inc is doing the legwork to making docker inc;s version supported on distro X | 21:19 |
| kfox1111 | jascott1: exactly. | 21:19 |
| kfox1111 | all that means, is, canonical won't block docker inc's version from running well on ubuntu. | 21:20 |
| kfox1111 | and canonical tech support has an extra page in the playbook for answering very simple questions, then forwarding more complicated stuff back to docker inc. | 21:20 |
| sbezverk | inc0: naming docker moby is dangerous, green peace will be after them ;) | 21:21 |
| kfox1111 | thats very different I think then actively having sales people going out and marketing docker swarm like they do kubernetes. | 21:21 |
| jascott1 | and that pop music guy ive never listened to ;) | 21:21 |
| sbezverk | like some people forced kubernetes to rename petsets :) | 21:21 |
| kfox1111 | it looks like moby is to docker what fedora is to rhel potentially. | 21:21 |
| jascott1 | ugh | 21:22 |
| kfox1111 | but k8s doesn't really care about either moby or docker. | 21:22 |
| kfox1111 | it only cares about containerd now. | 21:22 |
| inc0 | what dick thought of renaming docker to moby... | 21:22 |
| inc0 | :P | 21:22 |
| kfox1111 | cause both "moby" and "docker" really are about coe functionality. | 21:23 |
| kfox1111 | its not a rename. its kind of more like a distro. | 21:23 |
| *** jheroux has quit IRC | 21:23 | |
| jascott1 | moby... dick? | 21:23 |
| kfox1111 | look at http://mobyproject.org/ and see Audience. | 21:23 |
| kfox1111 | its a place to prototype changes in their coe. | 21:23 |
| jascott1 | always thought they shoulda named the whale "Lexi" as an homage to LXC roots | 21:24 |
| sbezverk | kfox1111: alternative name :) MOBER | 21:24 |
| kfox1111 | :) | 21:24 |
| inc0 | I like to moby moby \\o o// | 21:25 |
| inc0 | sorry I had to.. | 21:25 |
| kfox1111 | sbezverk: we probably should go back through all the k8s templates and move all the non node specific config changy bits to the templates now that we manage them. | 21:25 |
| *** lucasxu has joined #openstack-kolla | 21:26 | |
| *** kbaegis has joined #openstack-kolla | 21:26 | |
| *** kbaegis has quit IRC | 21:26 | |
| sdake | kfox1111 one thing just occured to me - does kolla-kubernetes work at all on multinode kubernetes? | 21:27 |
| sbezverk | I have running kolla-kube on 5 node kube nodes, works like a charm | 21:29 |
| sdake | sbezverk cool - are you using keepalived to handle the API endpoints? | 21:30 |
| sbezverk | btw sdake I will not be needing your hardware for demo | 21:30 |
| sdake | sbezverk roger | 21:30 |
| sbezverk | keepalived | 21:30 |
| sdake | sbezverk thanks - only keepalived is needed? | 21:30 |
| sbezverk | yes | 21:31 |
| sdake | sbezverk ++ | 21:31 |
| *** david-lyle_ has joined #openstack-kolla | 21:33 | |
| inc0 | kfox1111: so, any chances you can take over segregation of gate code <> non-gate code from me? I'll get to image mgmt tool | 21:34 |
| inc0 | will be more efficient this way imho | 21:34 |
| inc0 | basically lets turn our gate code to series of shellscripts which, when run in correct order, will deploy kolla-k8s | 21:35 |
| *** chas has joined #openstack-kolla | 21:35 | |
| inc0 | sdake: sbezverk ^ | 21:35 |
| jmguzmanc | Guys, sorry to interrupt... but we have been looking for some support with Kolla. Specifically, trying 4.0.0 and 4.0.1, we cannot see any styles in Horizon... I imagine it is something trivial to fix, but have not succeeded.... Any suggestions? | 21:36 |
| inc0 | jmguzmanc: fresh deploy of 4.0.0? | 21:37 |
| jmguzmanc | yes, and from stable/ocata for 4.0.1 | 21:37 |
| inc0 | dockerhub containers or built by you? | 21:37 |
| jmguzmanc | in 4.0.0 fetched containers. 4.0.1 built by me | 21:38 |
| jmguzmanc | I tried 4.0.1 expecting to have different results, but got stuck at the same point | 21:38 |
| inc0 | hmm | 21:38 |
| inc0 | ok, when you try to do http request for css stylesheet, what error do you get? | 21:39 |
| *** david-lyle_ has quit IRC | 21:39 | |
| inc0 | just stylesheet | 21:39 |
| *** chas has quit IRC | 21:40 | |
| jmguzmanc | weird thing, is that I don´t see any css code in the html retrieved.. I already fixed an issue with "Invalid Time Zone", that prevented to serve images, so static/ is fine.. but not css are being refered from html | 21:41 |
| *** sayantan_ has quit IRC | 21:41 | |
| inc0 | uhh | 21:41 |
| inc0 | strange | 21:41 |
| inc0 | you can find logs in /var/lib/docker/volumes/kolla_logs | 21:41 |
| inc0 | on host | 21:41 |
| inc0 | see if there is anything strange out there | 21:42 |
| sbezverk | inc0: I was under impression that we actually do exactly what you said, we runs a series of scripts to bring up the environment, I guess for me to understand what you want would require a specific example.. | 21:44 |
| inc0 | sbezverk: ok, so I ahve ubuntu machine, which scripts I run in which order to setup kolla-k8s? | 21:44 |
| inc0 | or centos machine | 21:44 |
| jmguzmanc | Found the logs, horizon.log with errors like Not Found: /manager/html | 21:44 |
| inc0 | jmguzmanc: never seen anything like this... | 21:45 |
| sdake | @inc0 I am hopeful that segration will occur as a result of the workshop i am holding upstream | 21:45 |
| inc0 | sdake: it won't occur on it's own | 21:46 |
| jmguzmanc | inc0: Ok, i will revert to 4.0.0.. Would that be easier to troubleshoot? | 21:46 |
| inc0 | we still need to have correct howto from shellscripts | 21:46 |
| *** tonanhngo has quit IRC | 21:46 | |
| inc0 | jmguzmanc: unlikely | 21:46 |
| inc0 | hmm | 21:46 |
| sdake | inc0 the workshop ends on may 4th | 21:46 |
| inc0 | Force will be with us that day | 21:46 |
| sdake | roger | 21:46 |
| sbezverk | inc0: you need to check the main gate setup to select which actual job you are interested in and run that scripts | 21:46 |
| sbezverk | we do have bunch of hard coded things, since it is after all gate environment | 21:47 |
| jmguzmanc | inc0: if you or somebody else wants to take a look by yourself, I can provide access to the EC2 instance where is running.. if it helps to fix something for the rest | 21:47 |
| inc0 | that's my point | 21:47 |
| sbezverk | but with minor modification you can use then without gate | 21:47 |
| inc0 | remove hardcoes to separate files | 21:47 |
| inc0 | so we can just run non-hardcoded pieces alone | 21:47 |
| inc0 | without any modifications | 21:48 |
| inc0 | jmguzmanc: if you allow me to log into it, I can take a look | 21:48 |
| inc0 | but I'm not horizon expert | 21:48 |
| jmguzmanc | Sure, send me your pubkey, and will allow you | 21:48 |
| *** salv-orl_ has joined #openstack-kolla | 21:49 | |
| *** tonanhngo has joined #openstack-kolla | 21:49 | |
| *** lucasxu has quit IRC | 21:49 | |
| jmguzmanc | inc0: jmguzman@whitestack.com | 21:49 |
| inc0 | jmguzmanc: http://paste.ubuntu.com/24416276/ | 21:50 |
| *** rwallner has joined #openstack-kolla | 21:50 | |
| kfox1111 | inc0: I can put it on the list, but unfortuately, time is scarce and can't be top priority I think. :/ but maybe I can focus on it during the summit. | 21:51 |
| *** tonanhngo has quit IRC | 21:51 | |
| *** salv-orlando has quit IRC | 21:52 | |
| jmguzmanc | inc: Try ubuntu@34.198.102.126, then sudo | 21:52 |
| inc0 | I'll start tmux session | 21:53 |
| *** lucasxu has joined #openstack-kolla | 21:54 | |
| inc0 | ok so these not-found things are from google crawler | 21:55 |
| *** rwallner has quit IRC | 21:55 | |
| jmguzmanc | inc0: yeah, unrelated to the main issue. | 21:56 |
| inc0 | jmguzmanc: fun fact...I can't see any notion of css in page source | 21:56 |
| jmguzmanc | inc0: Exactly.. it is weird. | 21:56 |
| jmguzmanc | inc0: http://34.198.102.126/ | 21:56 |
| inc0 | no css after login page too:/ | 21:57 |
| *** tonanhngo has joined #openstack-kolla | 21:58 | |
| jmguzmanc | inc0: yeap... It is my first kolla deploy, so no idea how to proceed | 21:58 |
| kfox1111 | sdake: cool. | 21:59 |
| inc0 | jmguzmanc: attach to tmux, you'll see me moving around;) | 21:59 |
| inc0 | but this is strange, I can't see any notion of style in html code | 21:59 |
| inc0 | looks like horizon issue... | 22:00 |
| sdake | anyone ever read moby dick? | 22:00 |
| inc0 | never got to | 22:01 |
| jascott1 | nope | 22:01 |
| *** mgoddard has quit IRC | 22:01 | |
| *** tonanhngo_ has joined #openstack-kolla | 22:01 | |
| sdake | the cliff notes version: https://en.wikipedia.org/wiki/Moby-Dick | 22:02 |
| jmguzmanc | inc0: FYI, I had to edit the static/ alias, to serve images | 22:03 |
| jmguzmanc | Alias /static /var/lib/kolla/venv/lib/python2.7/site-packages/openstack_dashboard/static | 22:03 |
| *** tonanhngo has quit IRC | 22:03 | |
| inc0 | hmm | 22:04 |
| jascott1 | bug? | 22:04 |
| inc0 | so issue I'm seeing is not static images | 22:04 |
| inc0 | I can't see any notion of styles alltogether in horizon | 22:05 |
| *** tonanhngo_ has quit IRC | 22:05 | |
| jascott1 | sdake all my free reading time goes to playing music... to my detriment im sure | 22:05 |
| jascott1 | oh and cat videos | 22:06 |
| sdake | jascott1 all my free reading time goes to openstack - definately to my detriment :) | 22:06 |
| jascott1 | maybe its after reading all day/night just cant really get into it | 22:06 |
| jascott1 | usually read on plane | 22:06 |
| jascott1 | i have started some really good books :) | 22:07 |
| *** lucasxu has quit IRC | 22:08 | |
| inc0 | lol when I turned on debug it worked | 22:09 |
| inc0 | jmguzmanc: | 22:09 |
| jascott1 | heisen bug! | 22:10 |
| jmguzmanc | inc0: :) | 22:10 |
| inc0 | jmguzmanc: well enjoy working css | 22:11 |
| *** tonanhngo has joined #openstack-kolla | 22:11 | |
| *** newmember has quit IRC | 22:11 | |
| inc0 | I guess by changing config I forced django to refresh cache | 22:11 |
| inc0 | or whatever happened | 22:11 |
| *** newmember has joined #openstack-kolla | 22:11 | |
| jmguzmanc | I will try to disable debug... and see what happens | 22:12 |
| inc0 | already done | 22:12 |
| jmguzmanc | I REALLY appreciate your help | 22:12 |
| inc0 | np | 22:12 |
| jmguzmanc | I owe you a beer...:) if you go to boston!! | 22:12 |
| inc0 | indeed I be there | 22:12 |
| inc0 | and beers are welcome | 22:13 |
| inc0 | jmguzmanc: are you coming to our hands on workshops for kolla-ansible? | 22:13 |
| jmguzmanc | Great! I hope to see you there.. traveling from Chile (south america)... | 22:13 |
| inc0 | interesting:) I've never been to south america, but Chile is high on my list | 22:14 |
| jmguzmanc | I expect to join as much as possible the Kolla activities.. Will check the sched for the workshop | 22:14 |
| inc0 | see you there then, make sure to say hello:) | 22:15 |
| jmguzmanc | Well, someday we will host a Openstack summit here... I hope:) | 22:15 |
| jmguzmanc | Thanks again for your magic touch... I saved me a lot of time.. see ya | 22:15 |
| inc0 | I don't know how much community OpenStack has in south america | 22:15 |
| inc0 | one of our cores, Mauricio is from Brazil | 22:16 |
| inc0 | other than that, I'm not sure if I know anyone | 22:16 |
| gianpietro | inc0: there are only a few of us :) (live from Lima, Peru) | 22:16 |
| inc0 | ha! | 22:16 |
| *** chas has joined #openstack-kolla | 22:16 | |
| inc0 | definetly need to visit then;) | 22:17 |
| inc0 | maybe openstack days at least? | 22:17 |
| *** dmk0202 has quit IRC | 22:17 | |
| inc0 | awesome, makes me happy to see how diverse and global our community is! | 22:17 |
| jmguzmanc | Let´s plan for these LATAM Openstack Days!! | 22:19 |
| inc0 | if you do, I'll try my best to visit and preach about Kolla;) | 22:20 |
| *** chas has quit IRC | 22:21 | |
| gianpietro | jmguzmanc: yours may be the first kolla-powered Ocata deployment "down here" :') | 22:21 |
| kfox1111 | so, will zuul3 support pipelines? | 22:35 |
| jascott1 | weird when I join nodes it breaks tiller even if its scheduled on master | 22:36 |
| kfox1111 | jascott1: in daemonset or deployment? | 22:36 |
| jascott1 | network teardown, some conntainer not found | 22:36 |
| kfox1111 | I'd expect daemonset to break. deployment to work. | 22:36 |
| jascott1 | still deployment, ive got my own ansible/ubuntu dev playbook going | 22:37 |
| kfox1111 | weird. | 22:37 |
| jascott1 | if I dont join the nodes everything works | 22:37 |
| kfox1111 | maybe the taint on master likes to push pods away if at all possible? | 22:37 |
| kfox1111 | maybe pull the taint before adding another node and see what happens? | 22:37 |
| jascott1 | good ideas | 22:37 |
| jascott1 | alright going to cut up that proposal ;) | 22:38 |
| *** chas has joined #openstack-kolla | 22:38 | |
| *** MarginHu2017 has joined #openstack-kolla | 22:41 | |
| *** manheim has joined #openstack-kolla | 22:42 | |
| *** chas has quit IRC | 22:43 | |
| *** salv-orl_ has quit IRC | 22:44 | |
| *** david-lyle has joined #openstack-kolla | 22:45 | |
| *** srwilkers has quit IRC | 22:50 | |
| *** chas has joined #openstack-kolla | 22:59 | |
| jascott1 | kfox1111 we still need tolerations tho right? | 23:00 |
| *** lamt has quit IRC | 23:01 | |
| kfox1111 | k8s 1.6.2. :) | 23:01 |
| kfox1111 | jascott1: ah. for single node, it would help. | 23:01 |
| kfox1111 | well.... actually for single node, you really gotta drop the taint to get anything to work. | 23:01 |
| jascott1 | but can you tolerate away NoSchedule? | 23:02 |
| jascott1 | yeah ok i was going to say | 23:02 |
| kfox1111 | no. :/ | 23:02 |
| jascott1 | so its nice but required? | 23:02 |
| kfox1111 | they want to make NoSchedule a proper taint, but its a funny inbetween state tat the moment. | 23:02 |
| kfox1111 | yeah. I think its a nice to have but not required for the use case. | 23:02 |
| jascott1 | prob easier to go ahead and do it. ok | 23:02 |
| kfox1111 | yeah, basically the same code as nodeSelector I think. | 23:03 |
| *** chas has quit IRC | 23:03 | |
| *** goldyfruit has joined #openstack-kolla | 23:03 | |
| *** schwicht has quit IRC | 23:04 | |
| jascott1 | hmmm --tolerations or --tolerate | 23:04 |
| jascott1 | guess im for the former although it is longer | 23:04 |
| *** gfidente|afk has quit IRC | 23:04 | |
| kfox1111 | yeah, consistency probably best. | 23:05 |
| *** MarginHu2017 has quit IRC | 23:06 | |
| *** MarginHu2017 has joined #openstack-kolla | 23:06 | |
| *** newmember has quit IRC | 23:11 | |
| jascott1 | https://github.com/kubernetes/helm/issues/2299 | 23:15 |
| jascott1 | kfox1111 ^ | 23:15 |
| jascott1 | oops forgot to add about default label | 23:16 |
| kfox1111 | jascott1: tolerations/node-selectors are key/value in k8s. | 23:16 |
| jascott1 | its key or key and value right? | 23:17 |
| jascott1 | hmm | 23:17 |
| kfox1111 | so would be like helm init --node-selector type=helm or something. | 23:17 |
| jascott1 | ooh right | 23:17 |
| jascott1 | i was thnking taints | 23:18 |
| jascott1 | you can do key:- or key:whatever- | 23:18 |
| kfox1111 | taints are also key/value? | 23:18 |
| jascott1 | key- | 23:18 |
| jascott1 | ya but weird | 23:18 |
| kfox1111 | yeah, matching whatever kubectl does would be good. | 23:18 |
| jascott1 | right! | 23:18 |
| jascott1 | ;) | 23:18 |
| kfox1111 | :) | 23:18 |
| jascott1 | might need to be key:value | 23:19 |
| kfox1111 | how kubectl run does node selectors (if it even supports them. | 23:19 |
| jascott1 | id | 23:19 |
| jascott1 | +k | 23:19 |
| *** chas has joined #openstack-kolla | 23:19 | |
| kfox1111 | [--env="key=value"] | 23:20 |
| jascott1 | yeah I will follow k8s | 23:20 |
| *** chas has quit IRC | 23:24 | |
| *** manheim has quit IRC | 23:25 | |
| *** Margin has joined #openstack-kolla | 23:32 | |
| *** MarginHu2017 has quit IRC | 23:32 | |
| *** Margin has quit IRC | 23:32 | |
| openstackgerrit | Kevin Fox proposed openstack/kolla master: Fingerprint containers. https://review.openstack.org/458174 | 23:33 |
| kfox1111 | inc0: the fingerprinting job I think is working. | 23:35 |
| inc0 | kfox1111: I just got to python to get it done | 23:37 |
| inc0 | still figuring out how to do it exactly | 23:37 |
| *** eaguilar has joined #openstack-kolla | 23:39 | |
| kfox1111 | inc0: its basically 2 lines of shell, and a simple shell script for runing in the container? maybe it doesn't need python? | 23:39 |
| kfox1111 | we make a fingerprint.sh? | 23:39 |
| inc0 | kfox1111: so 1. no for running in container | 23:39 |
| inc0 | it has to be done on image | 23:40 |
| inc0 | without running | 23:40 |
| kfox1111 | why? | 23:40 |
| inc0 | because we wont be running containers on push | 23:40 |
| kfox1111 | that makes it very kolla specific. | 23:40 |
| inc0 | too much of a hassle | 23:40 |
| inc0 | well, just let me figure it out | 23:40 |
| kfox1111 | docker run --rm -v /tmp/fingerprint.sh.$$:/fingerprint.sh $line /bin/bash /fingerprint.sh is a hastle? | 23:40 |
| kfox1111 | k. | 23:40 |
| *** chas has joined #openstack-kolla | 23:41 | |
| kfox1111 | just thinking this could be very useful outside of kolla. like, fingerprinting all sorts of containers in ci/cd. | 23:41 |
| *** chas has quit IRC | 23:45 | |
| *** gianpietro has quit IRC | 23:47 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!