*** blahRus has quit IRC | 00:06 | |
*** jtriley has quit IRC | 00:10 | |
SamYaple | sdake: gate doesnt like the pep8 | 00:13 |
---|---|---|
openstackgerrit | Sam Yaple proposed openstack/kolla: Remove keystone admin token https://review.openstack.org/285625 | 00:14 |
*** sdake has quit IRC | 00:14 | |
*** rhallisey has joined #kolla | 00:16 | |
dave-mccowan | i'm getting this error message on deploy: One or more undefined variables: 'nova_keystone_user' is undefined | 00:30 |
dave-mccowan | nova_keystone_user looks defined to me. do i need to rebuild my containers? | 00:30 |
dave-mccowan | sdake samyaple ^^ | 00:31 |
SamYaple | dave-mccowan: this was fixed recently | 00:38 |
SamYaple | check mastr | 00:38 |
*** jtriley has joined #kolla | 00:41 | |
openstackgerrit | Hui Kang proposed openstack/kolla: Fix missing libraries for ceilometer https://review.openstack.org/286274 | 00:57 |
*** skape has joined #kolla | 01:00 | |
skape | Hi | 01:01 |
skape | I'm having problems running ./tools/kolla-ansible deploy and when I tried ./tools/kolla-ansible prechecks it failed also | 01:03 |
skape | TASK [prechecks : Checking Docker version] fails but I have installed docker 1.8 with yum on centos 7 | 01:05 |
skape | what is the correct version for docker ? | 01:05 |
SamYaple | skape: for master the correct version is >=1.10.0 | 01:08 |
skape | with the command curl -sSL https://get.docker.io | bash will install this version ???? | 01:09 |
*** jasonsb has joined #kolla | 01:12 | |
skape | tkanks for the info! | 01:16 |
*** iceyao has joined #kolla | 01:17 | |
*** skape has left #kolla | 01:17 | |
*** alisonh has joined #kolla | 01:19 | |
*** prithiv1 has quit IRC | 01:29 | |
*** iceyao has quit IRC | 01:32 | |
*** iceyao_ has joined #kolla | 01:32 | |
*** Jeffrey4l has joined #kolla | 01:47 | |
*** tfukushima has joined #kolla | 01:51 | |
*** tfukushima has quit IRC | 01:54 | |
*** rhallisey has quit IRC | 02:04 | |
*** rtmie has quit IRC | 02:04 | |
*** dims has quit IRC | 02:04 | |
*** iceyao has joined #kolla | 02:04 | |
*** rtmie has joined #kolla | 02:04 | |
*** iceyao_ has quit IRC | 02:06 | |
*** dims has joined #kolla | 02:08 | |
*** dims has quit IRC | 02:13 | |
*** Marga__ has joined #kolla | 02:16 | |
*** dims has joined #kolla | 02:17 | |
*** jtriley has quit IRC | 02:19 | |
*** Marga_ has quit IRC | 02:19 | |
*** tfukushima has joined #kolla | 02:19 | |
*** tfukushima has quit IRC | 02:22 | |
*** tfukushima has joined #kolla | 02:24 | |
*** tfukushima has quit IRC | 02:26 | |
*** jtriley has joined #kolla | 02:28 | |
*** tfukushima has joined #kolla | 02:35 | |
*** pbourke_ has quit IRC | 02:46 | |
*** iceyao_ has joined #kolla | 02:47 | |
*** pbourke_ has joined #kolla | 02:47 | |
*** jtriley has quit IRC | 02:47 | |
*** iceyao has quit IRC | 02:50 | |
*** Allen_Gao has quit IRC | 02:50 | |
*** Allen_Gao has joined #kolla | 02:51 | |
*** jtriley has joined #kolla | 02:58 | |
*** jtriley has quit IRC | 03:18 | |
*** dims has quit IRC | 03:19 | |
*** yuanying has joined #kolla | 03:21 | |
*** yuanying_ has quit IRC | 03:23 | |
*** Marga__ has quit IRC | 03:26 | |
*** yuanying has quit IRC | 03:27 | |
*** Marga_ has joined #kolla | 03:30 | |
*** Marga_ has quit IRC | 03:35 | |
*** iceyao has joined #kolla | 03:35 | |
*** iceyao_ has quit IRC | 03:40 | |
*** Marga_ has joined #kolla | 03:43 | |
*** klint has joined #kolla | 03:44 | |
*** dave-mccowan has quit IRC | 03:45 | |
*** jtriley has joined #kolla | 03:45 | |
openstackgerrit | Hui Kang proposed openstack/kolla: Fix missing libraries for ceilometer https://review.openstack.org/286274 | 03:47 |
*** Marga_ has quit IRC | 03:47 | |
*** iceyao has quit IRC | 04:06 | |
*** yuanying has joined #kolla | 04:10 | |
*** Marga_ has joined #kolla | 04:16 | |
*** Marga_ has quit IRC | 04:16 | |
*** Marga_ has joined #kolla | 04:17 | |
*** daneyon_ has quit IRC | 04:26 | |
*** vhosakot has joined #kolla | 04:46 | |
*** phuongnh has joined #kolla | 04:56 | |
openstackgerrit | MD NADEEM proposed openstack/kolla: Manila ubuntu binary container https://review.openstack.org/286379 | 04:59 |
*** Allen_Gao has quit IRC | 05:08 | |
*** jtriley has quit IRC | 05:16 | |
*** salv-orl_ has joined #kolla | 05:19 | |
*** Marga_ has quit IRC | 05:20 | |
*** Allen_Gao has joined #kolla | 05:20 | |
*** salv-orlando has quit IRC | 05:22 | |
*** daneyon has joined #kolla | 05:25 | |
*** Marga_ has joined #kolla | 05:30 | |
*** Allen_Gao has quit IRC | 05:35 | |
*** Allen_Gao has joined #kolla | 05:40 | |
openstackgerrit | Merged openstack/kolla: Remove keystone admin token https://review.openstack.org/285625 | 06:11 |
*** sdake has joined #kolla | 06:11 | |
asalkeld | the gate is unhappy | 06:16 |
nihilifer | but only for kolla-mesos :/ | 06:16 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Restart daemons if their config changes https://review.openstack.org/284664 | 06:17 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Make write_file return if the contents changed https://review.openstack.org/286398 | 06:17 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Move generate_configs into the Command class https://review.openstack.org/286399 | 06:17 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Extend kolla-mesos-config to be able to set zookeeper variables https://review.openstack.org/286400 | 06:17 |
asalkeld | some package has a broken wheel | 06:18 |
*** daneyon has quit IRC | 06:23 | |
*** sdake has quit IRC | 06:27 | |
*** sdake has joined #kolla | 06:33 | |
sdake | evening | 06:35 |
sdake | is gate still bsuted? | 06:37 |
sdake | there were changes to requirements.txt recently | 06:37 |
sdake | one of them was setuptools | 06:38 |
sdake | asalkeld ^^ | 06:38 |
sdake | that might interact with wheel | 06:38 |
asalkeld | na, it's local too | 06:39 |
asalkeld | seems like someone uploaded a broken package | 06:39 |
nihilifer | umm... local too? i removed .tox/py27 dir, runned tests, and they work perfectly | 06:42 |
asalkeld | https://pypi.python.org/pypi/portalocker#downloads | 06:42 |
asalkeld | uploaded on 2016-02-29 | 06:42 |
nihilifer | do we know which of our deps depend on portalocker? | 06:44 |
asalkeld | dcos i think | 06:44 |
nihilifer | :/ | 06:44 |
nihilifer | yes, you're right | 06:45 |
asalkeld | i think 0.5.4 portalocker works | 06:46 |
asalkeld | maybe we can restrict it to that | 06:46 |
nihilifer | sure. but it's not in our openstack deps | 06:46 |
sdake | hard pins will be removed y global requirements bot | 06:47 |
nihilifer | so we can only pin in mesosphere's repo | 06:47 |
asalkeld | ~/work/requirements$ (master) git grep portalocker | 06:47 |
asalkeld | upper-constraints.txt:portalocker===0.5.5 | 06:47 |
sdake | oh ya upper constrInts | 06:47 |
nihilifer | ah, ok | 06:47 |
asalkeld | git pull and now ... | 06:47 |
nihilifer | asalkeld: you will prepare a patch to reqirements or should i do this? | 06:47 |
asalkeld | git grep portalocker | 06:48 |
asalkeld | upper-constraints.txt:portalocker===0.5.7 | 06:48 |
asalkeld | doh | 06:48 |
asalkeld | let's see who did that | 06:48 |
nihilifer | openstack proposal bot :P | 06:48 |
openstackgerrit | Steven Dake proposed openstack/kolla: DO NOT MERGE https://review.openstack.org/286409 | 06:49 |
asalkeld | nihilifer: worth asking infra what to do, as that job may just revert our change | 06:50 |
nihilifer | ok | 06:51 |
sdake | it wont remove it if we dont a pprove the changes | 06:51 |
sdake | if this is some short term pain | 06:51 |
asalkeld | nihilifer: getting to 5pm here - need to take a break (have a meetings later :( ) | 06:51 |
sdake | just revert it and send a note asking foks not to merge the proposal ot for the shor tterm | 06:51 |
asalkeld | nihilifer: maybe email the author Wolph@wol.ph | 06:52 |
asalkeld | if he uploads a new one, it might save some time | 06:53 |
nihilifer | asalkeld: ok, will do this today | 06:53 |
mdnadeem | nihilifer, Jeffrey4l , if get time, please review it : https://review.openstack.org/#/c/285138/ | 06:54 |
Jeffrey4l | mdnadeem, roger | 06:55 |
sdake | Jeffrey4l evening | 06:57 |
sdake | Jeffrey4l nice job on the fast turnaround on reconfigure | 06:57 |
sdake | tbh I didn't think it would be ready to go ;) | 06:57 |
*** Marga_ has quit IRC | 06:57 | |
sdake | did glance merge yet? | 06:57 |
Jeffrey4l | sam found some issue about the reconfigure. I should fix that first. | 06:58 |
*** achanda has quit IRC | 06:58 | |
Jeffrey4l | it impact keystone too. sdake | 06:59 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: [WIP] Add heka support https://review.openstack.org/286301 | 06:59 |
*** daneyon has joined #kolla | 06:59 | |
sdake | Jeffrey4l could you expand please | 06:59 |
sdake | yes please get keysstone fied, then make glance the reference implementation that people can copy | 07:00 |
*** tfukushima has quit IRC | 07:00 | |
Jeffrey4l | sdake, see the sam's comment https://review.openstack.org/#/c/285873/3/ansible/roles/glance/tasks/do_reconfigure.yml | 07:00 |
sdake | if eveyroen copies 2 or 3 in the core team, we should be able to hit the 2-4th with it | 07:00 |
*** tfukushima has joined #kolla | 07:00 | |
Jeffrey4l | the process should be ok. now need improve the 1) docker restart command to kolla_docker 2) docker exec <> printenv to kolla_docker | 07:01 |
sdake | Jeffrey4l right i just read the review | 07:01 |
sdake | Jeffrey4l if you coudl make that your top priority i'd super appreciate it :) | 07:02 |
Jeffrey4l | sdake, i am working on it. | 07:02 |
sdake | do you have other kolla work you think is higher priority? | 07:02 |
Jeffrey4l | about reconfigure? or some thing else? | 07:02 |
sdake | bugs -> RC | 07:02 |
sdake | anything else | 07:02 |
Jeffrey4l | no. | 07:02 |
sdake | pretty much jerked aroud with getting my kids computer back to the manufactuer today | 07:03 |
sdake | ALL DAY | 07:03 |
*** daneyon has quit IRC | 07:04 | |
sdake | tomorrow morning i am putting my lab back together, then tearing it apart on the 15th | 07:04 |
Jeffrey4l | should buy a new one. lol | 07:04 |
sdake | so i'll have 2 weeks of multinode availale | 07:04 |
sdake | it cost 10k | 07:04 |
sdake | not buying a new one ;) | 07:04 |
Jeffrey4l | :o | 07:05 |
sdake | https://drive.google.com/file/d/0B8q6xDPETSkHUlJjQWtFVzhPb2s/view?usp=sharing | 07:06 |
Allen_Gao | sdake: looks great! | 07:06 |
sdake | the ups guy complimented us on how well we packaged the palette ;) | 07:06 |
sdake | he said normally when they do pallette pickups at peoples homes its a disaster | 07:07 |
Jeffrey4l | water-cooled ? cool | 07:07 |
sdake | he was in and out in 5 mins | 07:07 |
sha_ | Hi guys... | 07:07 |
sdake | sha_ if you have a bug to report, please use passtebin :) | 07:07 |
sha_ | I have created local repository for docker.. | 07:07 |
sha_ | ha.. sure :) | 07:07 |
mdnadeem | sdake, Please let me know if https://blueprints.launchpad.net/kolla/+spec/ansible-unit-tests can be merge in Mitaka ? | 07:08 |
sdake | 1% power, gotta run | 07:08 |
sdake | mdnadeem the answer is yes | 07:08 |
sdake | get it in good progress state (the blueprint) | 07:08 |
mdnadeem | sdake, ok, i will try to cover all test case with in this week | 07:08 |
sdake | mdmit looks like your blueprint is in good progress to me | 07:08 |
mdnadeem | sdake, if possible please enroll my name in kolla team list on launch pad, my launchpad id is mail2nadeem92, Thanks :) | 07:10 |
*** sdake has quit IRC | 07:10 | |
*** daneyon has joined #kolla | 07:10 | |
sha_ | I have created local registry for docker images. But when I tried building the images using ./tools/build.py -b ubuntu -t source --registry 172.22.2.81:4000 --push | 07:10 |
sha_ | Iam getting this error: | 07:11 |
sha_ | 07:11 | |
*** iceyao has joined #kolla | 07:12 | |
*** sdake has joined #kolla | 07:12 | |
sha_ | failed: [localhost] => {"failed": true} msg: Unknown error message: Error while pulling image: Get http://172.22.2.81:4000/v1/repositories/kollaglue/ubuntu-source-kolla-toolbox/images: dial tcp 172.22.2.81:4000: | 07:13 |
sha_ | getsockopt: connection refused | 07:13 |
Jeffrey4l | sha_, paste, paste, paste | 07:13 |
Jeffrey4l | sha_, http://paste.openstack.org/ http://paste.openstack.org/ http://paste.openstack.org/ | 07:13 |
Jeffrey4l | pls | 07:13 |
nihilifer | yes, we told about this yesterday, didn't we? | 07:13 |
nihilifer | second thing | 07:13 |
nihilifer | 1) check whether docker registry is running | 07:14 |
Jeffrey4l | many times. | 07:14 |
nihilifer | 2) check where it's running, what's its IP address / hostname / hwatever | 07:14 |
nihilifer | 3) set it in /etc/kolla/globals.yml | 07:14 |
nihilifer | and once again, paste plz | 07:14 |
sdake | sha_ run docker ps | 07:14 |
sdake | into a paste plz | 07:14 |
sdake | if your on centos you can run docker ps | fpaste | 07:15 |
sha_ | http://paste.openstack.org/show/488702/ | 07:15 |
sdake | i dont know if ubuntu has a paste service built into the distro | 07:15 |
*** achanda has joined #kolla | 07:15 | |
*** mikelk has joined #kolla | 07:16 | |
sdake | sha_ I need to see docker ps - to see how you started your registry | 07:16 |
*** Allen_Gao has quit IRC | 07:17 | |
vhosakot | good night! | 07:17 |
sdake | vhosakot night :) | 07:17 |
sha_ | docker ps | 07:17 |
sha_ | http://paste.openstack.org/show/488704/ | 07:17 |
*** vhosakot has quit IRC | 07:18 | |
sdake | sha_ try telnet host_ip_of_registry 4000 | 07:18 |
sdake | sha_ what ip is your registry running on? | 07:19 |
sha_ | I have kept in /global.yml as docker_registry: "172.22.2.81:4000" | 07:19 |
Jeffrey4l | memcached container is up. the docker-registry should be ok. | 07:19 |
sdake | is 172.22.2.81 the node where your registry is running? | 07:19 |
sdake | my first guess would be a firewall problem | 07:20 |
sha_ | When i do telnet it is giving : http://paste.openstack.org/show/488706/ | 07:21 |
sdake | is 172.22.2.81 a machine in your network? | 07:21 |
*** Allen_Gao has joined #kolla | 07:24 | |
sha_ | sdake : able to ping that ip | 07:24 |
sdake | where is your registry running? | 07:24 |
sha_ | on 172.22.2.81:4000 | 07:25 |
sdake | on the machine 172.22.2.81 your running docker? | 07:25 |
sha_ | No docker is not running on that machine | 07:26 |
sdake | well how could the registry possibly run there? | 07:26 |
sdake | what machine is docker running on - ip address | 07:26 |
sha_ | on local machine | 07:26 |
sdake | ok well that is where your registry is running | 07:26 |
sdake | wht is the ip addresss of your local machine | 07:26 |
sdake | ip link show will show you | 07:27 |
sdake | eth0 is what your looking for | 07:28 |
sdake | paste ip link show if you need help parsing it | 07:28 |
sha_ | http://paste.openstack.org/show/488707/ | 07:28 |
sha_ | can I run docker registery on local system? | 07:29 |
sdake | is eth0 on a virtual machine? | 07:29 |
sha_ | No.. | 07:29 |
sdake | run ifconfig and paset it | 07:29 |
sdake | eth0 has no ip address assigned to it | 07:29 |
sdake | according to ip link show | 07:29 |
sdake | also you have to turn off libvirtd to run kolla | 07:30 |
sdake | i can tell yourrunning libvirt from virbr0 | 07:30 |
sha_ | http://paste.openstack.org/show/488708/ | 07:30 |
sdake | sha_ your registry is running at 172.16.73.114:4000 | 07:32 |
sdake | try telnet 172.6.73.114 4000 | 07:32 |
sdake | the registry doesn't run at sosme random address just because its in globals.yml | 07:32 |
sdake | the registry runs on the ip address of the machine docker is running on | 07:32 |
sdake | ./build.py --regisstry 172.6.73.114:4000 --type source --base ubuntu | 07:33 |
sha_ | ok..Then I will change the Ip address and build it | 07:33 |
sdake | ./build.py --registry 172.6.73.114:4000 --push | 07:34 |
sdake | run those two comamnds | 07:34 |
sha_ | Thank you soo much!! | 07:34 |
sdake | that will put your images in your registry | 07:34 |
sdake | roger your welcome | 07:34 |
*** Allen_Gao has quit IRC | 07:34 | |
sdake | once you get your images in your regisstry, you need to change globals.yml to pull form it | 07:35 |
sdake | it takes me 12 minutes to build and 5 minutes to push | 07:35 |
sdake | but I am on gige to the internet with PCIE NVMe SSD | 07:35 |
sdake | so look at 17 mins upwards of one hour if your net/gear is slow | 07:35 |
sdake | also i dont know if master is unbroken yet | 07:36 |
sdake | i have a gate test running now to find out | 07:36 |
sdake | but the gate has a 2-3 hour lag because its deadline time so eveyrone jamming on patch submissions | 07:36 |
*** akwasnie has joined #kolla | 07:41 | |
sdake | ok 1am time to ptfo | 07:41 |
sdake | night all | 07:41 |
sdake | hi akwasnie :) | 07:41 |
sdake | akwasnie elemoine can you both give me a quick update on the state of EHK | 07:42 |
sdake | i saw there is a patch to integrate them | 07:42 |
sdake | is that all that is needed? | 07:42 |
openstackgerrit | Martin André proposed openstack/kolla: Add a new tool to help find the latest versions https://review.openstack.org/248481 | 07:43 |
elemoine_ | sdake: yes | 07:43 |
akwasnie | hi sdake | 07:43 |
sdake | nice | 07:43 |
sdake | well i can sleep well knowing that is done ;) | 07:43 |
elemoine_ | that patch connects Heka, Elasticsearch and Kibana together | 07:44 |
akwasnie | sdake we tested it and it is ok:) | 07:44 |
sdake | are logs stored on the local nodes as well? | 07:44 |
sdake | that was an unexpressed requirement | 07:44 |
elemoine_ | I've tested it in both AIO and multinode modes | 07:44 |
elemoine_ | sdake: yes | 07:44 |
sdake | nice | 07:44 |
elemoine_ | in the kolla_logs volume | 07:44 |
sdake | sweet | 07:44 |
sdake | i will check it out tomorrow when I reassemble my lab | 07:45 |
sdake | working in vms just not viable | 07:45 |
akwasnie | good, thanks | 07:45 |
elemoine_ | out of curiosity, how machines do you use for kolla? do you test in multinode? | 07:45 |
sdake | 3 machines for deploy plus 1 to deploy from | 07:46 |
*** sdake has quit IRC | 07:48 | |
*** Allen_Gao has joined #kolla | 07:49 | |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Set graphics_ip and _port in Vagrantfile https://review.openstack.org/286427 | 07:54 |
*** fgimenez has joined #kolla | 08:00 | |
*** fgimenez has joined #kolla | 08:00 | |
*** achanda_ has joined #kolla | 08:02 | |
*** achanda has quit IRC | 08:04 | |
*** achanda_ has quit IRC | 08:04 | |
*** chandankumar has joined #kolla | 08:06 | |
*** slagle has quit IRC | 08:06 | |
*** achanda has joined #kolla | 08:08 | |
*** ccesario has quit IRC | 08:09 | |
*** akwasnie has quit IRC | 08:10 | |
*** daneyon has quit IRC | 08:17 | |
*** akwasnie has joined #kolla | 08:19 | |
*** achanda has quit IRC | 08:27 | |
*** britthouser has quit IRC | 08:29 | |
*** britthouser has joined #kolla | 08:30 | |
openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/kolla: Add required packages to ironic-conductor container https://review.openstack.org/286150 | 08:32 |
*** ccesario has joined #kolla | 08:38 | |
*** salv-orl_ has quit IRC | 08:47 | |
*** openstackgerrit has quit IRC | 08:48 | |
*** openstackgerrit has joined #kolla | 08:48 | |
*** salv-orlando has joined #kolla | 08:48 | |
openstackgerrit | MD NADEEM proposed openstack/kolla: Manila ubuntu binary container https://review.openstack.org/286379 | 08:56 |
openstackgerrit | Martin André proposed openstack/kolla: Add a new tool to help find the latest versions https://review.openstack.org/248481 | 08:57 |
*** gfidente has joined #kolla | 08:58 | |
*** athomas has joined #kolla | 09:01 | |
*** chandankumar has quit IRC | 09:02 | |
*** mbound has joined #kolla | 09:06 | |
*** chandankumar has joined #kolla | 09:09 | |
*** salv-orlando has quit IRC | 09:11 | |
*** iceyao has quit IRC | 09:15 | |
*** iceyao has joined #kolla | 09:16 | |
*** Serlex has joined #kolla | 09:22 | |
*** prithiv has joined #kolla | 09:34 | |
*** prithiv has quit IRC | 09:35 | |
*** Allen_Gao has quit IRC | 09:41 | |
*** amnk has joined #kolla | 09:48 | |
*** prithiv has joined #kolla | 09:53 | |
*** kproskurin has joined #kolla | 09:54 | |
*** tfukushima has quit IRC | 10:27 | |
-openstackstatus- NOTICE: Gerrit is going to be restarted due to poor performance | 10:34 | |
*** ChanServ changes topic to "Gerrit is going to be restarted due to poor performance" | 10:34 | |
*** prithiv has quit IRC | 10:36 | |
*** prithiv has joined #kolla | 10:38 | |
*** ChanServ changes topic to "Kolla IRC meetings on Wednesday 16:30 UTC (Day light savings doesn't apply) - Agenda @ https://wiki.openstack.org/wiki/Meetings/Kolla - IRC channel is *LOGGED* @ http://eavesdrop.openstack.org/irclogs/%23kolla/" | 10:40 | |
-openstackstatus- NOTICE: gerrit finished restartign | 10:40 | |
*** prithiv has quit IRC | 10:42 | |
*** kproskurin_ has joined #kolla | 10:43 | |
*** prithiv has joined #kolla | 10:43 | |
*** kproskurin has quit IRC | 10:43 | |
*** kproskurin_ is now known as kproskurin | 10:43 | |
*** phuongnh has quit IRC | 10:45 | |
openstackgerrit | Merged openstack/kolla-mesos: Add ANSIBLE_NOCOLOR to chronos job https://review.openstack.org/284648 | 10:46 |
openstackgerrit | Merged openstack/kolla-mesos: Add explicit calls to extend_start and skip the default https://review.openstack.org/283356 | 10:47 |
*** iceyao has quit IRC | 10:48 | |
openstackgerrit | Merged openstack/kolla-mesos: Fix the running of apache based services https://review.openstack.org/283972 | 10:48 |
*** dims has joined #kolla | 11:03 | |
openstackgerrit | MD NADEEM proposed openstack/kolla: WIP: Added unit test for kolla_docker https://review.openstack.org/285138 | 11:10 |
*** Sara_ has joined #kolla | 11:19 | |
Sara_ | Hi | 11:19 |
*** vhosakot has joined #kolla | 11:19 | |
*** vhosakot has quit IRC | 11:19 | |
*** vhosakot has joined #kolla | 11:20 | |
*** Allen_Gao has joined #kolla | 11:21 | |
openstackgerrit | MD NADEEM proposed openstack/kolla: WIP: Added unit test for kolla_docker https://review.openstack.org/285138 | 11:22 |
openstackgerrit | MD NADEEM proposed openstack/kolla: Added unit test for check container https://review.openstack.org/286513 | 11:28 |
sha_ | Hi.. | 11:33 |
*** akwasnie has quit IRC | 11:33 | |
sha_ | I am facing problem with keystone during Kolla deploy | 11:33 |
sha_ | Error : http://paste.openstack.org/show/488737/ | 11:34 |
sha_ | Can anyone help me? | 11:34 |
*** iceyao has joined #kolla | 11:34 | |
*** iceyao_ has joined #kolla | 11:38 | |
*** salv-orlando has joined #kolla | 11:41 | |
*** iceyao has quit IRC | 11:42 | |
*** akwasnie has joined #kolla | 11:46 | |
*** Jeffrey4l has quit IRC | 11:51 | |
*** prithiv has quit IRC | 11:54 | |
*** prithiv has joined #kolla | 11:57 | |
*** salv-orlando has quit IRC | 11:58 | |
openstackgerrit | Mauricio Lima proposed openstack/kolla: Update cleanup-containers to remove ceph containers https://review.openstack.org/282451 | 11:59 |
*** limamauricio has joined #kolla | 11:59 | |
limamauricio | morning guys | 12:00 |
*** prithiv has quit IRC | 12:05 | |
*** prithiv has joined #kolla | 12:06 | |
*** prithiv has quit IRC | 12:08 | |
*** Jeffrey4l has joined #kolla | 12:09 | |
*** ccesario has quit IRC | 12:24 | |
sha_ | Hi guys.. | 12:24 |
sha_ | I am facing problem with keystone during Kolla deploy | 12:24 |
sha_ | Error : http://paste.openstack.org/show/488737/ | 12:24 |
sha_ | Can anyone help me? | 12:24 |
*** dave-mccowan has joined #kolla | 12:27 | |
*** jtriley has joined #kolla | 12:28 | |
*** iceyao has joined #kolla | 12:32 | |
*** iceyao_ has quit IRC | 12:35 | |
*** jtriley has quit IRC | 12:36 | |
*** iceyao has quit IRC | 12:36 | |
*** iceyao_ has joined #kolla | 12:36 | |
limamauricio | Hey Jeffrey4l :) | 12:43 |
*** rhallisey has joined #kolla | 12:43 | |
Jeffrey4l | limamauricio, hi | 12:43 |
Jeffrey4l | sup | 12:43 |
limamauricio | I updated my patch, check if everything is ok :) | 12:43 |
limamauricio | https://review.openstack.org/#/c/269688/ | 12:43 |
nihilifer | sha_: i'd start with doing | 12:44 |
limamauricio | your review is important | 12:44 |
limamauricio | ^^ | 12:44 |
nihilifer | source openrc | 12:44 |
nihilifer | and trying to execute some basic cli commands to verify whether keystone is working at all | 12:44 |
nihilifer | in example | 12:44 |
nihilifer | openstack user list | 12:44 |
nihilifer | openstack project list | 12:44 |
nihilifer | etc. | 12:44 |
nihilifer | you may use --debug option for that | 12:45 |
Jeffrey4l | limamauricio, Cannot Merge? you need a base first. | 12:45 |
nihilifer | and after executing these commands, check the loks of keystone - "docker logs keystone" | 12:45 |
Jeffrey4l | limamauricio, the gate is OK now. So ensure all the jenkins check is green. | 12:45 |
nihilifer | sha_: and please don't repost the same thing in hour period... | 12:45 |
sha_ | ok | 12:46 |
*** shardy has quit IRC | 12:47 | |
*** macsz has joined #kolla | 12:51 | |
Jeffrey4l | limamauricio, you patch can not merge now. Please rebase to master. leave a recheck is not a good idea. | 12:52 |
elemoine_ | the services' initial logs can be viewed using "docker logs <container_name>", but keep in mind that the services logs now are in the kolla_logs volume (/var/lib/docker/volumes/kolla_logs/_data) | 12:52 |
limamauricio | i know | 12:52 |
limamauricio | just wait ;) | 12:52 |
Jeffrey4l | :D | 12:52 |
*** salv-orlando has joined #kolla | 12:55 | |
openstackgerrit | Merged openstack/kolla: Add required packages to ironic-conductor container https://review.openstack.org/286150 | 12:59 |
openstackgerrit | Merged openstack/kolla: Move ironic-discoverd to ironic-inspector https://review.openstack.org/280655 | 13:00 |
*** m3m0 has joined #kolla | 13:00 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add kolla_docker action for reconfigure https://review.openstack.org/286570 | 13:01 |
*** dims has quit IRC | 13:02 | |
*** jmccarthy has quit IRC | 13:05 | |
*** jmccarthy has joined #kolla | 13:06 | |
*** dims has joined #kolla | 13:08 | |
*** prithiv has joined #kolla | 13:13 | |
openstackgerrit | Mauricio Lima proposed openstack/kolla: Add ansible support for Manila https://review.openstack.org/269688 | 13:14 |
limamauricio | Jeffrey4l, Done :) | 13:14 |
Jeffrey4l | cool | 13:15 |
*** sdake has joined #kolla | 13:16 | |
*** ccesario has joined #kolla | 13:18 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Reconfigure glance service https://review.openstack.org/285873 | 13:18 |
*** sdake has quit IRC | 13:21 | |
*** akwasnie has quit IRC | 13:21 | |
*** sdake has joined #kolla | 13:21 | |
*** kproskurin has quit IRC | 13:24 | |
limamauricio | morning sdake, manila have many drivers, maybe can be interesting to think of a way to make the kolla support these drivers in the future | 13:25 |
limamauricio | SamYaple ^ | 13:25 |
sdake | limamauricio yes i'd like to sort out plugins in the future in a reasonable way | 13:25 |
sdake | atm we don't hve them sorted out | 13:25 |
sdake | we have an austin summit session on plugins | 13:25 |
*** alger has joined #kolla | 13:26 | |
limamauricio | sdake, interesting | 13:26 |
*** amnk has quit IRC | 13:28 | |
*** Serlex has quit IRC | 13:33 | |
*** slagle has joined #kolla | 13:33 | |
limamauricio | 2016-03-01 13:33:59.617 | ERROR:kolla.cmd.build:keystone:Error'd with the following message 2016-03-01 13:33:59.617 | ERROR:kolla.cmd.build:keystone:The command '/bin/sh -c yum -y install httpd mod_wsgi && yum clean all && sed -i -r 's,^(Listen 80),#\1,' /etc/httpd/conf/httpd.conf' returned a non-zero code: 1 | 13:35 |
limamauricio | gate-kolla-dsvm-deploy-centos-source ^ | 13:36 |
limamauricio | :/ | 13:36 |
*** iceyao_ has quit IRC | 13:38 | |
sdake | lima thanks i'll look at the source | 13:40 |
sdake | limamauricio ^ | 13:40 |
Jeffrey4l | limamauricio, done | 13:43 |
limamauricio | i think someone needs to define when and where to use two blank lines | 13:44 |
limamauricio | in some files are one in the other are two | 13:44 |
limamauricio | is one* | 13:45 |
sdake | limamauricio the code looks correct | 13:46 |
limamauricio | ok, sdake | 13:46 |
*** Marga_ has joined #kolla | 13:46 | |
sdake | but clearly something is wrong | 13:46 |
sdake | limamauricio let me connect to not my vpn and run a build there | 13:49 |
*** Marga_ has quit IRC | 13:50 | |
*** shardy has joined #kolla | 13:50 | |
*** sdake_ has joined #kolla | 13:50 | |
*** chandankumar has quit IRC | 13:52 | |
*** sdake has quit IRC | 13:54 | |
*** Serlex has joined #kolla | 13:55 | |
*** mbound has quit IRC | 13:56 | |
SamYaple | morning | 13:56 |
sdake_ | limamauricio centos source builds for me | 13:56 |
sdake_ | keystone atleast | 13:56 |
*** rmart04 has joined #kolla | 13:56 | |
sdake_ | limamauricio try a recheck | 13:56 |
sdake_ | limamauricio do you have a link tothe review in question? | 13:56 |
*** limamauricio_ has joined #kolla | 13:57 | |
sdake_ | hey SamYaple | 13:57 |
*** limamauricio has quit IRC | 13:57 | |
sdake_ | gate is green from last night https://review.openstack.org/#/c/286409/ yay ;) | 13:57 |
limamauricio_ | sdake_ gate-kolla-dsvm-deploy-centos-source was finished successfully | 13:57 |
sdake_ | which gate is failing for you lima? | 13:58 |
*** klint has quit IRC | 13:59 | |
limamauricio_ | It was this gate that the error occurred sdake_ | 13:59 |
sdake_ | oh so you mean the error is on and off? | 13:59 |
limamauricio_ | maybe, i don't know | 13:59 |
SamYaple | sdake_: gates not actually all good | 13:59 |
SamYaple | it mostly works but i need to finish out the v3 keystone or horizon will still have issues | 14:00 |
limamauricio_ | I do not know how this gate works | 14:00 |
sdake_ | limamauricio_ where did you see the error you pasted above | 14:00 |
SamYaple | limamauricio_: it just runs the playbooks and launches an instance | 14:00 |
SamYaple | if all of that is successfully it passes | 14:00 |
limamauricio_ | i understood :) | 14:01 |
limamauricio_ | I have to change a lot in pacth manila now. God, it has no end :D | 14:02 |
sdake_ | SamYaple well the v3 keystone missing from horizon doesn't appear to impact the gate | 14:02 |
*** vhosakot has quit IRC | 14:03 | |
*** chandankumar has joined #kolla | 14:03 | |
*** chandankumar has quit IRC | 14:05 | |
*** chandankumar has joined #kolla | 14:05 | |
SamYaple | it does not | 14:07 |
*** akwasnie has joined #kolla | 14:11 | |
*** mbound has joined #kolla | 14:15 | |
openstackgerrit | Merged openstack/kolla: Update cleanup-containers to remove ceph containers https://review.openstack.org/282451 | 14:16 |
*** XuChao918 has joined #kolla | 14:16 | |
SamYaple | sdake_: 1.1.0 | 14:18 |
SamYaple | keeping the gate stable while we are merging into that branch feels nigh impossible | 14:18 |
SamYaple | at least with the time available | 14:18 |
sdake_ | http://logs.openstack.org/73/285873/4/check/gate-kolla-pep8/c35e63a/console.html#_2016-03-01_13_30_29_166 | 14:18 |
*** XuChao918 has quit IRC | 14:19 | |
*** sdake_ is now known as sdake | 14:19 | |
sdake | SamYaple i hear what your saying | 14:19 |
sdake | if we can have a patch stream | 14:19 |
sdake | where at the end is a green gate | 14:19 |
SamYaple | so is a long patchset that gets it from where it is to stable is that ok? | 14:19 |
sdake | i think thatwould be good enough | 14:19 |
SamYaple | ok that works for me | 14:19 |
SamYaple | so block the first one, get all approved, then merge first | 14:19 |
sdake | yup | 14:19 |
SamYaple | ok i can make that work | 14:20 |
sdake | this is atypical, but since we don't have voting gates it doesn't matter ;) | 14:20 |
sdake | SamYaple please take a look at that pep8 failure | 14:20 |
sdake | its in kolla_docker.py looks like a simple resolution | 14:20 |
*** Xu_Chao has joined #kolla | 14:20 | |
sdake | i think some pep8 tools are running differently on different notes | 14:20 |
sdake | (the link above) | 14:20 |
SamYaple | sdake: why? | 14:20 |
sdake | i would have thought kolla_docker was in the exclude lsit | 14:21 |
SamYaple | no | 14:21 |
SamYaple | we havent exluded ansible modules for a long time | 14:21 |
sdake | because pep8 works in some cases and not in others | 14:21 |
SamYaple | that patch you were looking at added that funciton... | 14:21 |
SamYaple | its a valid failure | 14:21 |
sdake | oh duh | 14:21 |
sdake | so used to invalid failures :) | 14:21 |
sdake | and just woke up ;) | 14:22 |
Xu_Chao | openstack-ansible project support the deployment of it HA architecture? | 14:22 |
SamYaple | Xu_Chao: of course! | 14:23 |
SamYaple | i use it almost exlusively | 14:23 |
sdake | Xu_Chao you mean kolla or openstack-ansible? | 14:23 |
sdake | I don't know what openstack-ansible does | 14:23 |
SamYaple | Xu_Chao: yea sorry i missted that. i read Kolla. but yes both projects support HA | 14:24 |
SamYaple | I use Kolla almost exlusively :) | 14:24 |
Xu_Chao | I mean openstack-ansible | 14:25 |
Xu_Chao | thanks | 14:25 |
sdake | Xu_Chao openstack-ansible has a irc channel if you want to follow up | 14:25 |
sdake | #openstack-ansible | 14:25 |
Xu_Chao | thanks.:-) | 14:27 |
*** mbound has quit IRC | 14:30 | |
*** stvnoyes has quit IRC | 14:32 | |
*** shardy has quit IRC | 14:33 | |
sha_ | Hi.. Where to specify domain for keystone while creating admin, project and role during Kolla? | 14:33 |
sha_ | am getting error for that | 14:33 |
sdake | sha_ are you using msater? | 14:33 |
*** stvnoyes has joined #kolla | 14:33 | |
sdake | as in master from t-6 hours ago | 14:33 |
sdake | if not, git pull --rebase | 14:33 |
sdake | and copy over the new config files | 14:33 |
sdake | and rebuild | 14:34 |
*** mbound has joined #kolla | 14:34 | |
sha_ | ansible | 14:34 |
sha_ | yes.. master | 14:35 |
*** shardy has joined #kolla | 14:35 | |
*** akwasnie has quit IRC | 14:41 | |
sha_ | its using v2/.. as per my knowledge it doesnt require domain (for v2)..even then it is throwing error for "could not find domain" | 14:41 |
SamYaple | sha_: after rebuild did you reset your environment? | 14:42 |
SamYaple | master is a constantly changing thing you need to build new containre and deploy a _new_ environment | 14:42 |
sha_ | which environment? | 14:43 |
sha_ | Kolla? | 14:43 |
sha_ | I built thrice.. | 14:44 |
*** chandankumar has quit IRC | 14:44 | |
*** akwasnie has joined #kolla | 14:44 | |
*** jtriley has joined #kolla | 14:45 | |
SamYaple | sha_: it sounds like you didnt remove the existing environment then | 14:45 |
dave-mccowan | sdake what do you kolla folks do on Friday of summit? | 14:46 |
sha_ | ok.. I will check | 14:46 |
SamYaple | dave-mccowan: thats where we do most planning | 14:47 |
dave-mccowan | we don't know yet if kolla will have the AM or PM slot, right? /me is booking plane tickets | 14:47 |
*** sdake has quit IRC | 14:49 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Fix Keystone v3 and Horizon https://review.openstack.org/286637 | 14:49 |
Jeffrey4l | SamYaple, what do you want about the {{ container_name }} in the kolla_docker restart PS? | 14:50 |
SamYaple | Jeffrey4l: remove it and use with_items | 14:50 |
Jeffrey4l | use with_items with include ? | 14:51 |
SamYaple | no | 14:51 |
SamYaple | inside do_reconfigure | 14:51 |
openstackgerrit | Merged openstack/kolla-mesos: Add state info to commands https://review.openstack.org/271422 | 14:51 |
SamYaple | like we do everywhere else | 14:51 |
Jeffrey4l | But it will be only one item forever. | 14:51 |
SamYaple | you can include multiple items | 14:52 |
Jeffrey4l | I want to restart/reconfigure the service, one by one. | 14:52 |
SamYaple | i dont agree with that | 14:52 |
SamYaple | and i certainly do agree with this variable abstraction | 14:52 |
SamYaple | we have (mostly) gotten away from that | 14:52 |
SamYaple | it broke many times before we learned our lesson | 14:53 |
SamYaple | the only place we still use it is for ceph pools | 14:53 |
*** sdake has joined #kolla | 14:56 | |
sdake | dave-mccowan friday is the contributor meetup | 14:56 |
sdake | dave-mccowan my hotel I think is until friday morning iirc | 14:56 |
sdake | I haven' tbooked my flgiht yet | 14:57 |
sdake | I may change my hotel to leave saturday | 14:57 |
sdake | since it looks like we can get a full day contributormeetup friday | 14:57 |
sdake | at the contributor meetup last time we sorted out how to do upgrades in a general fashion | 14:57 |
*** dwalsh has joined #kolla | 14:57 | |
sdake | 4 hours of upgrade discussion :) | 14:57 |
*** shardy has quit IRC | 15:00 | |
Jeffrey4l | SamYaple, you mean , using the varilabe in module argument will be the same using variable in task name? | 15:00 |
dave-mccowan | i always seem to regret my travel plans. when i leave early on Friday, i'm regret I may have missed a good session. if i stay, i wonder why i stayed behind in a mostly empty room with (other project) folks who can't wait to get out and go touring. :-) | 15:01 |
sdake | dave-mccowan usually we have a 4 hour half day contributor meetup followed by profuse drinking | 15:02 |
sdake | the half day contributor meetup will be packed with action im sure | 15:02 |
sdake | it was actuallh our best session at summit last cycle ;) | 15:02 |
Jeffrey4l | the hardest part is this https://review.openstack.org/#/c/286050/1/ansible/roles/cinder/tasks/reconfigure.yml | 15:03 |
openstackgerrit | Hui Kang proposed openstack/kolla: Fix missing libraries for ceilometer https://review.openstack.org/286274 | 15:03 |
Jeffrey4l | no all serice running on one host. | 15:04 |
Jeffrey4l | config.yml do some wrong that, it will copy the cinder-backup configure file to the master even it is not used. | 15:04 |
Jeffrey4l | SamYaple, https://github.com/openstack/kolla/blob/master/ansible/roles/cinder/tasks/config.yml | 15:05 |
Jeffrey4l | reconfigure is not the same. | 15:06 |
SamYaple | oh yea i forgot to go back and add that filter Jeffrey4l | 15:13 |
SamYaple | you can do a when filter on the item | 15:13 |
SamYaple | so when: {{ item.in_group }} | 15:13 |
SamYaple | thats for pointing that back out | 15:13 |
Jeffrey4l | yea. | 15:13 |
SamYaple | im suprised no one has noticed it for this long | 15:14 |
SamYaple | i made a comment when i did the merge_configs thing about that and it jsut slipped my mind | 15:14 |
Jeffrey4l | In that way, there will be much of duplicated and long when: xxxx | 15:15 |
Jeffrey4l | it doesn't raise any issue. So no one notice it :D | 15:15 |
SamYaple | thats probably true | 15:15 |
SamYaple | so for the skipped conditionals it wont raise any (unless you arent running that service i suppose) | 15:15 |
SamYaple | since its inline | 15:16 |
Jeffrey4l | I still doubt, why we can not use variable like the way i used? will raise issue? | 15:16 |
SamYaple | you werent around before the merge_config stuff, but it always caused issues for people | 15:17 |
SamYaple | it never worked _quite_ right | 15:17 |
SamYaple | and it was also confusing to read | 15:17 |
SamYaple | with_items was far easier to read, even if it does generate some "skipped" tasks | 15:17 |
*** spisarski has joined #kolla | 15:18 | |
Jeffrey4l | just one variable, not that hard to read. with_items will increate the duplicated code and the number of the lines. | 15:19 |
Jeffrey4l | but anyway, i will try the with_item, and see the feedback. | 15:19 |
*** sdake_ has joined #kolla | 15:20 | |
Jeffrey4l | one last question: in my implemention, the reconfigure is run one service by one service, then one node by node. it is slow and will can cause bit down time. | 15:21 |
SamYaple | i would tend to agree with you Jeffrey4l, thats why the initial implementation used what you are doing all over the place | 15:21 |
SamYaple | but we changed it for a reason | 15:21 |
SamYaple | Jeffrey4l: is there a question in there? | 15:22 |
*** sdake has quit IRC | 15:23 | |
Jeffrey4l | but in the with_item, the reconfigure will run one node by node. it is fast but will make all service(for example, all the nova related service on the controller) on one node crashed( because, all the nova related container may be removed at the same time on one node) | 15:24 |
Jeffrey4l | (I do not like this.) | 15:24 |
Jeffrey4l | SamYaple, ^ | 15:24 |
*** haplo37 has joined #kolla | 15:27 | |
*** spisarski has quit IRC | 15:28 | |
*** akwasnie has quit IRC | 15:28 | |
sdake_ | mdnadeem ping re https://review.openstack.org/#/c/286379/ | 15:28 |
sdake_ | mdnadeem sam had a question about manilla-share, can you answer so we can either merge or you can submit a new patch? | 15:29 |
*** spisarski has joined #kolla | 15:30 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add kolla_docker action for reconfigure https://review.openstack.org/286570 | 15:30 |
SamYaple | Jeffrey4l: well it wont be at the same time | 15:31 |
SamYaple | it still goes service by serivice | 15:31 |
Jeffrey4l | SamYaple, we have two tasks now. one is remove the container, another is `include start.yml`. I do think it will remove all the container, then start all the container. | 15:32 |
Jeffrey4l | I will leave keystone implementation as it, and change the glance's implementation by using with_items. | 15:33 |
*** blahRus has joined #kolla | 15:33 | |
SamYaple | thats only if you use COPY_ONCE | 15:34 |
Jeffrey4l | SamYaple, yep. | 15:34 |
SamYaple | stupid COPY_ONCE it shouldnt be used | 15:34 |
SamYaple | no one likes it and it causes problems | 15:34 |
Jeffrey4l | lol | 15:34 |
SamYaple | operators hate it since it breaks the nor | 15:34 |
SamYaple | norm* | 15:34 |
SamYaple | sdake_: ^ | 15:34 |
openstackgerrit | Paul Bourke proposed openstack/kolla: Add new gates for oraclelinux https://review.openstack.org/246437 | 15:34 |
SamYaple | so yea Jeffrey4l if you use copy once that will happen | 15:34 |
SamYaple | suggest you dont use COPY_ONCE | 15:35 |
sdake_ | SamYaple read all the scrollback | 15:35 |
sdake_ | what specifically was the question | 15:35 |
sdake_ | pbourke_ thx ;-) | 15:35 |
SamYaple | just pointing out COPY_ONCE is causing problems yet again | 15:35 |
sdake_ | well we have two options (in newton) | 15:36 |
sdake_ | remove config-strategy | 15:36 |
sdake_ | or make copy_once work properly | 15:36 |
pbourke_ | sdake_: thanks for the reminder about it! is there anyway to test these before adding them to infra? | 15:36 |
sdake_ | in fact, copy once should work properly in mitaka ;) | 15:36 |
sdake_ | pbourke_ you can make an experimental job | 15:37 |
sdake_ | then run check experimental | 15:37 |
SamYaple | sdake_: it works properly, it just means more downtime | 15:37 |
sdake_ | once that is going well you can add it to always be in the check | 15:37 |
SamYaple | but i would vote remove it | 15:37 |
Jeffrey4l | copy_once is a little wired, do not see this kinda concept in other place. | 15:37 |
sdake_ | if it works properly with more downtime we can document that | 15:37 |
SamYaple | Jeffrey4l: agreed | 15:37 |
sdake_ | what is the downtime caused by, the container removal? | 15:37 |
SamYaple | two tasks | 15:37 |
SamYaple | removal then creation | 15:38 |
SamYaple | thats also more IO intensive | 15:38 |
*** chandankumar has joined #kolla | 15:38 | |
sdake_ | so that is like 1 second ? :) | 15:38 |
SamYaple | no its ansible | 15:38 |
Jeffrey4l | ( i will submit the code ) | 15:38 |
SamYaple | so like 10 or more seconds | 15:38 |
sdake_ | ok | 15:38 |
SamYaple | and it depends on your backend | 15:38 |
SamYaple | and your ssh connections | 15:38 |
SamYaple | its a bad default for sure | 15:38 |
sdake_ | well its way too late to remove it for newton, so lets shelve that conversation until summit | 15:38 |
sdake_ | open to changing the default prior to newton | 15:39 |
SamYaple | Jeffrey4l: what about that? | 15:39 |
Jeffrey4l | but SamYaple , I still want use my variable version :D. it will not cause this kinda issue. | 15:39 |
SamYaple | Jeffrey4l: but it will cause all the other issues | 15:39 |
sdake_ | rather too late to remove for mitaka | 15:39 |
SamYaple | sdake_: i agree with removal | 15:39 |
SamYaple | i mean too late for removal | 15:39 |
SamYaple | Jeffrey4l: really what we want is blocks, but thats ansible 2.0 | 15:39 |
*** mbound has quit IRC | 15:40 | |
SamYaple | how about we do it with with_items for mitaka, and ansible 2.0 we switch to blocks, that resolves all issues | 15:40 |
sdake_ | the problem with removing it completely is then the containers are not immutable | 15:40 |
SamYaple | i do understand your concerns about it | 15:40 |
Jeffrey4l | SamYaple, for example? i do not meet any issue. it just a normal variable, like those in the globals. | 15:40 |
*** mbound has joined #kolla | 15:40 | |
sdake_ | ok lets get out on the table what the issue with using the variabls in do_configure are? | 15:40 |
SamYaple | Jeffrey4l: as i said before, we encoutnered many issues doing this that didnt pop out at first | 15:40 |
SamYaple | Jeffrey4l: i also did what you were doing _at_ _first_, but had to change due to one issue after another | 15:41 |
SamYaple | and then the ansible bugs.... | 15:41 |
SamYaple | there are many edge cases here, but we moved away from this and im not keen to move back | 15:41 |
sdake_ | the issues we ran into with the variables were related to the docker module | 15:41 |
* Jeffrey4l go to rush the code.. two edition will be submited and compare. | 15:41 | |
SamYaple | sdake_: you are wrong | 15:41 |
SamYaple | sdake_: i am specifically refering to the merge_config predecessor | 15:41 |
SamYaple | it was a big variable include thing too | 15:41 |
sdake_ | i see | 15:42 |
SamYaple | and it always had some issue | 15:42 |
SamYaple | thats why it was replaced with merge_config and with_items | 15:42 |
sdake_ | i was referring to the fact that we had the inheritence thing with variables which didn't work with iirc host=pid mode | 15:42 |
Jeffrey4l | SamYaple, I only know, use the variable in name with the with_items will raise issue. | 15:42 |
SamYaple | Jeffrey4l: no it wont | 15:42 |
SamYaple | feel free to test it | 15:42 |
SamYaple | but it works fine | 15:42 |
sdake_ | ok so the question here is does jeffrey's do_reocnfgiure work when someone types kolla-ansible reconfigure | 15:43 |
sdake_ | yes/no? | 15:44 |
*** inc0 has joined #kolla | 15:44 | |
SamYaple | yea it should work, until you hit those edge cases that broke the config playbook :) | 15:44 |
SamYaple | you know, the bonafide ansible bugs | 15:44 |
inc0 | good mornign | 15:45 |
sdake_ | hey inc0 hows your fine morning doing | 15:45 |
sdake_ | Jeffrey4l SamYaple i'd like something merged *today* so people can start c&p the work | 15:45 |
inc0 | started out my day by 2 political meetings, so it's a breeze | 15:45 |
sdake_ | if it means more downtime, i dont really care | 15:45 |
sdake_ | for the copy-once case | 15:45 |
sdake_ | inc0 welcome to your third political meeting :) | 15:45 |
inc0 | duh | 15:46 |
inc0 | I guess you have something in your mind sdake_ ? | 15:46 |
sdake_ | copy-once gurantees immutability, if there is some pain involved with that guaranatee, we can docuemnt it | 15:46 |
Jeffrey4l | SamYaple, you are right, it works fine.. ( I really meet some similar issue) | 15:47 |
sdake_ | copy-alwlays does not gurantee immutability, if there is less pain with the lack of that guarantee, we can document it | 15:47 |
Jeffrey4l | sdake_, I will push the code soon. | 15:47 |
sbezverk | is python-magnumclient part of any magnum related containers? | 15:47 |
SamYaple | Jeffrey4l: im just trying to make sure we dont hit all those edge cases we just took out again | 15:47 |
inc0 | sdake_, copy once by no means guarantee immutabilitu | 15:47 |
SamYaple | Jeffrey4l: i know it seems arbitrary but the with_items we know will work, the include we know doesnt always work | 15:47 |
sdake_ | inc0 it gurantees immutability of the config | 15:48 |
inc0 | if image changes, containers gets rebuild just the same | 15:48 |
Jeffrey4l | cool | 15:48 |
sdake_ | inc0 yes that is an upgrade | 15:48 |
inc0 | ok | 15:48 |
SamYaple | i agree with you about it not being ideal Jeffrey4l, but i dont have another suggest at this time | 15:48 |
sdake_ | sbezverk i think its part of openstack-base but not certain | 15:48 |
Jeffrey4l | let's make it clear about the edge cases. it will help us to make choice. | 15:49 |
Jeffrey4l | SamYaple, ^ | 15:49 |
sdake_ | Jeffrey4l i think Sam was mentioning bugs in ansible | 15:49 |
sdake_ | which bugs they are, is unknown | 15:49 |
sdake_ | arguablity merge_configs was much more complex then do_reconfigure | 15:50 |
*** chandankumar has quit IRC | 15:50 | |
sdake_ | we shouldn't repeat past mistakes | 15:50 |
sdake_ | unless something in the environment has changed | 15:50 |
Jeffrey4l | sdake_, i known. I don't miunderstand. I am talking the ansible bug too. | 15:50 |
sdake_ | (for example moving to ansible 2.0) | 15:50 |
*** shardy has joined #kolla | 15:51 | |
sbezverk | sdake_ if it were, then I should have been able to run magnum cli commands on my host, but it is not the case.. | 15:51 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Change default CONFIG_STRATEGY https://review.openstack.org/286671 | 15:51 |
SamYaple | there sdake_ https://review.openstack.org/286671 | 15:51 |
sdake_ | we can trust SamYaple that if he says there are ansible bugs, there are ansible bugs | 15:51 |
inc0 | hey shardy :) since you're here | 15:51 |
inc0 | I have question for you | 15:51 |
sdake_ | sup shardzofglass | 15:51 |
inc0 | should we enable convergence as default in Mitaka kolla? | 15:51 |
sdake_ | inc0 try to make it harder plz ;) | 15:52 |
SamYaple | Jeffrey4l: i do not have time to run back through, but if you look through the bugs in kolla youll see 10-20 around this whole variable include stuff. one of those bugs contains links to ansible bugs, and some of them we just said "we dont know what the issue is, we are switching to with_items" | 15:52 |
Jeffrey4l | SamYaple, ok. i will check it. | 15:52 |
sdake_ | Jeffrey4l the bottom line with ansible is 1.9.5 iss never coming | 15:53 |
sbezverk | sdake_ http://paste.openstack.org/show/488779/ | 15:53 |
SamYaple | sdake_: i was promised! | 15:53 |
sdake_ | so if there are ansible bugs, we need to avoid them | 15:53 |
SamYaple | multiple times! | 15:53 |
sdake_ | SamYaple first i've heard of it, ill ask around | 15:53 |
SamYaple | lol its not ive told you about it several times | 15:53 |
sdake_ | sbezverk why don't you pip install python-magnumclient | 15:53 |
SamYaple | that said i am also doubtful about 1.9.5 even after th promise | 15:53 |
sdake_ | ansible is hiring more cats | 15:54 |
sdake_ | their budget is big and their burn is small | 15:54 |
sdake_ | so 1.9.5 could be possible - who knows | 15:55 |
sdake_ | but I bitch daily at them about abanodning 1.9.z series | 15:55 |
sbezverk | sdake_ as a workaround I will, but should it not be installed during kolla deployment? | 15:55 |
openstackgerrit | Mauricio Lima proposed openstack/kolla: Add ansible support for Manila https://review.openstack.org/269688 | 15:55 |
sdake_ | nothing during kolla deployment should instlal the clients | 15:57 |
sdake_ | the clients run on a different machine then the target deployment nodes | 15:57 |
sdake_ | let me correct | 15:57 |
sdake_ | a cieint may be instlaled in a container for use by the service | 15:57 |
sdake_ | for example nova, installs python-neutronclient | 15:57 |
sdake_ | and a bunch of others | 15:57 |
sdake_ | but these are not for end user usage | 15:57 |
sdake_ | they are for integration with nova | 15:58 |
*** vhosakot has joined #kolla | 15:58 | |
sbezverk | sdake_ ok, what then in containerized deployment can be use to run nova service-list for example? Or you exepect everything should be done via horizon?!?! | 16:01 |
*** chandankumar has joined #kolla | 16:01 | |
sdake_ | no here is how i operate | 16:02 |
sdake_ | i have a lpatop | 16:02 |
sdake_ | i install all my cleints on my laptop | 16:02 |
sdake_ | i have 4 machines | 16:02 |
sdake_ | 1 of my machines is a deployment host | 16:02 |
sdake_ | the other 3 are deployment targets | 16:02 |
sdake_ | i ansible deploy from my deployment host | 16:02 |
sdake_ | i connect to openstack from my laptop | 16:02 |
*** Xu_Chao has quit IRC | 16:02 | |
sdake_ | clients can go on the all-in-one machine | 16:02 |
sdake_ | or htey can g anywhere else | 16:02 |
SamYaple | sbezverk: clients go anywhere. you can install them whereever you want | 16:03 |
sdake_ | all that is sepcial about the above is i have n openrc file | 16:03 |
SamYaple | kolla doesnt determine where your clients run | 16:03 |
sdake_ | the openrc file specifies my keystone server | 16:03 |
sbezverk | sdake_ SamYaple, got it so clients are not part of deployment for user's usage, only for internal things.. | 16:04 |
sdake_ | roger | 16:04 |
*** haplo37 has quit IRC | 16:04 | |
sdake_ | for exxample, nova needs to communicate to neutron so it instals neutroncleint | 16:04 |
sdake_ | heat needs to communicate to everything so it installs all clients | 16:04 |
sdake_ | these clients are installed in the related containers | 16:05 |
sdake_ | but for the end user, there is no toolbox container with all of the clients | 16:05 |
sdake_ | good feature idea | 16:05 |
sdake_ | but hasn't been done yet | 16:05 |
sdake_ | has been suggested before | 16:05 |
sdake_ | Jeffrey4l are we all set then? what is the course of action you intend to take? | 16:07 |
Jeffrey4l | what's it? | 16:08 |
Jeffrey4l | sdake_, ^ | 16:09 |
sdake_ | Jeffrey4l what i mean is the course of action to just go with the more downtime, or what? | 16:09 |
sdake_ | Jeffrey4l tht isn't to say i don't like your implementation, i think the idea rocks | 16:10 |
sdake_ | but if ansible is busted - we can't use it | 16:10 |
*** macsz has quit IRC | 16:10 | |
Jeffrey4l | base on the ansible bug, this is the only option technically. | 16:11 |
Jeffrey4l | sdake_, i have no other ideat now. | 16:11 |
sdake_ | whcih is the only option, the more downtime? | 16:12 |
Jeffrey4l | yep | 16:12 |
sdake_ | ok well lets make it so | 16:12 |
sdake_ | and deal with optimizing downtime in newton | 16:12 |
*** macsz has joined #kolla | 16:12 | |
Jeffrey4l | yea | 16:12 |
sdake_ | put a docimpact in the commit log please | 16:12 |
sdake_ | so someone can document the imapct ;) | 16:12 |
Jeffrey4l | ok. | 16:13 |
sdake_ | i'll document the impact if your commit message is soid enough | 16:13 |
sdake_ | i suspect we will be going to nasible 2.0 in newton | 16:14 |
sdake_ | although unknown | 16:14 |
Jeffrey4l | roger that. | 16:14 |
Jeffrey4l | I'd like to move to ansible 2 ASAP in newton. | 16:14 |
Jeffrey4l | I should not so hard. | 16:14 |
*** haplo37 has joined #kolla | 16:16 | |
sdake_ | Jeffrey4l https://bugs.launchpad.net/kolla/+bug/1551314 please respodn quilcly | 16:18 |
openstack | sdake_: Error: malone bug 1551314 not found | 16:18 |
sdake_ | SamYaple that info from clarkb is pretty sweet huh, we dont have to rebases everything | 16:19 |
sdake_ | that is not how i thought it worked ;) | 16:19 |
sdake_ | may be a recent change | 16:20 |
SamYaple | sdake_: im of the opinion that needs to be tested | 16:20 |
sdake_ | agree | 16:20 |
sdake_ | i found an old patch and rechecked it, watching for the results | 16:20 |
* Jeffrey4l reading | 16:20 | |
inc0 | pbourke_, around | 16:21 |
inc0 | ? | 16:21 |
pbourke_ | inc0: hey | 16:21 |
inc0 | hey Paul, do you want mariadb upgrades? | 16:21 |
inc0 | ;) | 16:22 |
inc0 | we have fire going and I won't be able to handle it | 16:22 |
*** daneyon has joined #kolla | 16:25 | |
pbourke_ | not really! but if there's noone else I'll take it | 16:25 |
inc0 | soo...my dear kolla community, show of hands | 16:26 |
*** daneyon_ has joined #kolla | 16:26 | |
inc0 | we need someone to take on mariadb | 16:26 |
Jeffrey4l | sdake_, done | 16:26 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: Autodetect number of instances https://review.openstack.org/284643 | 16:26 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: Support dependencies inside one host https://review.openstack.org/284737 | 16:26 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: [WIP] Add heka support https://review.openstack.org/286301 | 16:26 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: Dependency between nova-libvirt and nova-compute inside host https://review.openstack.org/284815 | 16:26 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: Use constraints only on multinode deployment https://review.openstack.org/285970 | 16:26 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: Allow to override resources in globals.yml https://review.openstack.org/284836 | 16:26 |
pbourke_ | inc0: have you looked at it much, how hairy is it | 16:27 |
inc0 | SamYaple, will have most of info | 16:27 |
inc0 | it's not super trivial, but not too hard as well | 16:27 |
inc0 | will require a bit more than just include start.yml | 16:28 |
inc0 | and logic will be similar to what I'm working on - rabbitmq | 16:28 |
pbourke_ | go on then | 16:28 |
pbourke_ | I'll have a look | 16:28 |
inc0 | ok, I'll let you know as soon as I push anytihn up | 16:29 |
sbezverk | SamYaple after getting organized client I am trying to get magnum service-list and I get Not Authorized (admin-openrc.sh is loaded). In keystone log I see this error every time I try to execute magnum commands: "Expecting to find domain in project - the server could not comply with the request since it is either malformed or otherwise incorrect. The client is assumed to be in error." Do you | 16:30 |
sbezverk | think something might be missing in magnum configutation? | 16:30 |
*** daneyon has quit IRC | 16:30 | |
*** chandankumar has quit IRC | 16:32 | |
*** absubram has joined #kolla | 16:32 | |
*** absubram_ has joined #kolla | 16:33 | |
*** absubram has quit IRC | 16:37 | |
*** absubram_ is now known as absubram | 16:37 | |
Jeffrey4l | SamYaple, sdake_ see this implementation for cinder reconfigure. https://gist.github.com/jeffrey4l/d18ff4c6b841f7d523bd | 16:37 |
Jeffrey4l | 1. code become more complicated 2. code can not run, with follow error ERROR: [DEPRECATED]: include + with_items is a removed deprecated feature | 16:38 |
Jeffrey4l | SamYaple, sdake_ any idea? | 16:38 |
*** iceyao has joined #kolla | 16:38 | |
sdake_ | Jeffrey4l busy can't context switch at the moment, perhaps SamYaple can help you out with the include nd with_items deprecation problem | 16:39 |
Jeffrey4l | SamYaple, around? | 16:41 |
*** iceyao has quit IRC | 16:43 | |
SamYaple | Jeffrey4l: yea include + with_items is deprecated | 16:46 |
sdake_ | Jeffrey4l interested in joining the kolla-coresec team (VMT) | 16:47 |
SamYaple | but you can variablize the start | 16:47 |
Jeffrey4l | SamYaple, found a solution. | 16:47 |
Jeffrey4l | will push later. | 16:47 |
SamYaple | pbourke_: inc0: i have mariadb | 16:47 |
SamYaple | ill post a WIP moment | 16:47 |
sdake_ | i'll be looking for volunteers for kolla-coresec, sending out mail in a bit | 16:47 |
pbourke_ | SamYaple: cool thanks | 16:48 |
inc0 | what's kolla-coresec? | 16:48 |
SamYaple | sdake_: ill volunteer, that should be no suprise though | 16:48 |
sdake_ | sounds good sam | 16:48 |
SamYaple | inc0: sdake is getting a security team together | 16:48 |
sdake_ | i'll add you | 16:48 |
inc0 | add me too plz | 16:48 |
SamYaple | sbezverk: that sounds like the partial v3 implementation we have | 16:49 |
SamYaple | i have a WIP patch to fix that | 16:49 |
SamYaple | horizon alo suffers | 16:49 |
Jeffrey4l | I think i can. but do not know what's kolla-coresec? | 16:49 |
rhallisey | sdake_, I'll join also | 16:50 |
sdake_ | ok so i need to limit this tema to about 4 people | 16:50 |
sdake_ | let me send out email to mailing list describing what the responsiblities are | 16:51 |
sdake_ | which i am in process of adding | 16:51 |
sdake_ | its just that Jeffrey4l had found a VMT bug potentially | 16:51 |
*** chandankumar has joined #kolla | 16:51 | |
sdake_ | process of writing | 16:52 |
rhallisey | I don't need to be on the team, but plan on writing selinux policy for kolla | 16:52 |
elemoine_ | question about docker image documentation: would it make sense to have each Dockerfile directory include a README file documenting on the image should be used? Has this been discussed already? I am currently writing some documentation for the Heka image and I was wondering if that documentation could go to Kolla. | 16:52 |
Jeffrey4l | btw, VMT means? | 16:53 |
sbezverk | SamYaple, it looks like I was using old globals.yml where new variables like kolla_internal_fqdn were not defined in globals.yml but were used in magnum/defaults/main.yml | 16:53 |
inc0 | elemoine_, more docs will not hurt | 16:53 |
pbourke_ | elemoine_: I think you should just add an entry for heka the same as the ones at the bottom of http://docs.openstack.org/developer/kolla/ | 16:54 |
pbourke_ | elemoine_: or actually maybe an over all "logging in kolla" under the main section would be better, depends on the angle you're going for | 16:54 |
elemoine_ | I agree that a "logging in kolla" page would be useful | 16:55 |
vhosakot | Jeffrey4l: sdake_, I have started neutron reconfig work... shall I update the bp with steps to test reconfiguring services in https://blueprints.launchpad.net/kolla/+spec/kolla-reconfig ? | 16:55 |
elemoine_ | but this is not what I was referring to in my initial question | 16:55 |
elemoine_ | my initial question relates to documenting the kolla images (what they provide, their usage, what volumes they expect to run, etc.) | 16:56 |
sdake_ | steps to test would be great - in the whiteboard or in an etherpad linked from the blueprint | 16:56 |
sdake_ | etherpad would be preferrable | 16:56 |
Jeffrey4l | vhosakot, please, that will be helpful. | 16:56 |
vhosakot | Jeffrey4l: sdake_ ok | 16:56 |
*** mikelk has quit IRC | 16:56 | |
vhosakot | Jeffrey4l: I want you to review the steps to test, please :) | 16:56 |
Jeffrey4l | vhosakot, np | 16:56 |
sdake_ | rhallisey please respond to that email | 16:57 |
sdake_ | vhosakot moment letme make you an etherpad to start with | 16:57 |
vhosakot | Jeffrey4l: sdake_, cool, I think all BPs should have well documented/updated steps to test that work :) | 16:57 |
elemoine_ | that docker image discussion is also related to making our images more visible and useful to others (who do not necessarily deploy using kolla-ansible or kolla-mesos) | 16:57 |
pbourke_ | elemoine_: that sounds good to me. Though I think something under the doc/ directory would be best | 16:58 |
sdake_ | vhosakot https://etherpad.openstack.org/p/kolla-mitaka-testing-reconfigure | 16:58 |
vhosakot | sdake_: I was going to add under the "reconfigure services on a redeploy" section in the Mitaka midcycle roadmap etherpad at https://etherpad.openstack.org/p/kolla-mitaka-midcycle-roadmap | 16:58 |
vhosakot | sdake_: cool, thanks! | 16:58 |
pbourke_ | elemoine_: just my opinion :) | 16:58 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Reconfigure glance service https://review.openstack.org/285873 | 16:59 |
elemoine_ | pbourke_: yeah, not scattering the docs across the repo makes sense to me as well | 16:59 |
vhosakot | sdake_: Jeffrey4l, so, let us add for each service like "how to test reconfiguring keystone", "how to test reconfiguring glance", "how to test reconfiguring neutron" blah blah | 16:59 |
sdake_ | vhosakot don't put stuff in the midcycle docs plz | 16:59 |
vhosakot | sdake_: ok, sure.. won't :) | 16:59 |
*** jtriley has quit IRC | 16:59 | |
sdake_ | there is a link at the blueprint now to use | 16:59 |
sdake_ | on etherpad | 16:59 |
vhosakot | awesome! | 17:00 |
SamYaple | sdake_: thats not a VMT bug i told you that in the review | 17:00 |
vhosakot | the deadline for implemeting reconfig is March 4th ? | 17:00 |
sdake_ | SamYaple are you subscribed to the bug? VMT dicssion shoudl happen in bugs, not in gerrit | 17:00 |
SamYaple | sdake_: auth is not a security hole | 17:01 |
sdake_ | in fact, from whati understand gerrit should not be used at all - the vmt handles merging the changes | 17:01 |
SamYaple | the bug is superfolus | 17:01 |
elemoine_ | pbourke_: but having a doc/images directory where only one or two images are documented is a bit ridiculous. By using README files in the Dockerfile directories we make it less official… | 17:01 |
*** mbound has quit IRC | 17:01 | |
pbourke_ | elemoine_: it can be a work in progress | 17:01 |
sdake_ | vhosakot more or less | 17:01 |
elemoine_ | pbourke_: so this would be my preference until we have a decent number of documented images | 17:02 |
sdake_ | kolla has a bunch of outstanding work | 17:02 |
sdake_ | upgrades is #! | 17:02 |
elemoine_ | pbourke_: just my opinion :) | 17:02 |
sdake_ | diagnostics is #@ | 17:02 |
pbourke_ | elemoine_: there can also be a doc describing overall approach | 17:02 |
pbourke_ | elemoine_: including the jinja mechanisms etc. | 17:02 |
sdake_ | reconfigure is #3 | 17:02 |
sdake_ | security is #4 | 17:03 |
sdake_ | I think :) | 17:03 |
sdake_ | I think we should have a design document in the main docs, not in readmes | 17:03 |
elemoine_ | pbourke_: true | 17:03 |
*** jtriley has joined #kolla | 17:03 | |
sdake_ | lets not clutter up the repos with a whole slew of readmes | 17:03 |
sdake_ | because nobody will read them or keep them up to date | 17:03 |
vhosakot | sdake_: you mean that VMT... :) I was the one who added the spec for it the the security-doc repo :) https://github.com/openstack/security-doc/blob/master/README.rst | 17:03 |
sdake_ | we can't even keep the vagrant docs up to date | 17:03 |
vhosakot | sdake_: do you think kolla team-members need to join VMT ? | 17:04 |
elemoine_ | sdake_: would you accept doc/images/ with just heka.rst in a Change Request? | 17:04 |
sdake_ | vhosakot its limited to core reviewers by the tag | 17:05 |
sdake_ | or I'd ask dave-mccowan to join | 17:05 |
vhosakot | ah ok... got it :) | 17:05 |
elemoine_ | sdake_: if I do that I will also add a general doc file explaining the overall approach that pbourke_ mentioned | 17:05 |
sdake_ | I might ask dave-mccowan to execute a security audit in any regard ;-) | 17:05 |
*** bmace has quit IRC | 17:05 | |
vhosakot | I did some work for vulnerability testing and incorporating security notes for neutron | 17:06 |
sdake_ | doc/diagnostics.rst | 17:06 |
sdake_ | doc/design.rst | 17:06 |
vhosakot | yep | 17:06 |
*** bmace has joined #kolla | 17:06 | |
elemoine_ | so I'd create a CRÂ that adds two files : doc/images/heka.rst and doc/images/general.rst | 17:06 |
sdake_ | no | 17:06 |
sdake_ | its not heka to document, its diagnostics to document | 17:06 |
sdake_ | doc/diagnostics.rst | 17:06 |
elemoine_ | sdake_: as I said my main goal here is to describe the kolla images | 17:06 |
elemoine_ | doc/diagnostics.rst makes a lot of sense, I totally agree | 17:07 |
sdake_ | if you want to describe the kolla images, we might as ewll get started ona kolla/design.rst | 17:07 |
elemoine_ | but it's not what I am referring to here | 17:07 |
sdake_ | or kolla/doc/implementation.rst | 17:07 |
elemoine_ | sdake_: I want to describe each image individually as well | 17:07 |
sdake_ | i htink that becomes difficult to maintain | 17:08 |
limamauricio_ | sdake_, SamYaple -> https://review.openstack.org/#/c/269688/ | 17:08 |
pbourke_ | " If there are more then 4 individuals interested in joining this team, I will form the team from the most active members based upon liberty + mitaka commits, reviews, and PDE spent. | 17:08 |
pbourke_ | " | 17:08 |
sdake_ | perhaps it makes more sense to describe our implementation in terms of best practices and go fro mthere | 17:08 |
pbourke_ | SamYaple you're screwed :p | 17:08 |
dave-mccowan | vhosakot why did you add a security-doc repo? | 17:08 |
sdake_ | pbourke lol | 17:08 |
elemoine_ | my primary goal is to describe the heka image, which other folks are going to use outside kolla-ansible | 17:08 |
openstackgerrit | Merged openstack/kolla: Change default CONFIG_STRATEGY https://review.openstack.org/286671 | 17:08 |
SamYaple | pbourke_: why what? | 17:08 |
sdake_ | SamYaple he was obviously being sarcastic ;) | 17:09 |
vhosakot | dave-mccowan: I did not add the repo.. I added the spec doc describing VMT and security practices in OpenStack when I worked on testing disabling neutron port-secutiry for multicast in neutron | 17:09 |
SamYaple | ohh i didnt see the quote | 17:09 |
SamYaple | lol | 17:09 |
elemoine_ | sdake_ it's a "private" goal, but I was wondering it it could benefit kolla | 17:09 |
SamYaple | wow. that was the quickest merge in history https://review.openstack.org/286671 | 17:09 |
sdake_ | elemoine_ i think it is a slipperly slope - we dont want to document how all 115 of our containers work | 17:10 |
sdake_ | we want to tlak in general terms | 17:10 |
vhosakot | dave-mccowan: did you have any specific Qs about the security-doc repo ? | 17:10 |
sdake_ | i dont want to set a precedent where we are encouraged to write those docs for every container | 17:10 |
sdake_ | just look at what has happened with the guide discussion on the nova review | 17:10 |
*** achanda has joined #kolla | 17:10 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Add kolla_docker action for reconfigure https://review.openstack.org/286570 | 17:11 |
sdake_ | we have 15-20 guides to write, and there is pushback on that, i dont want "115 container docs to write" | 17:11 |
elemoine_ | sdake_: yeah, I fully understand this | 17:11 |
sdake_ | how will people use the heka container outside kolla? | 17:11 |
sdake_ | you mean for other purposes? | 17:11 |
elemoine_ | sdake_: I just find it strange/stupid to document my heka image outside kolla | 17:12 |
sdake_ | well if you think it matters to the design, document it in the implementation.rst file | 17:12 |
elemoine_ | sdake_: I mean outside kolla-ansible | 17:12 |
SamYaple | elemoine_: i think it should be documented in kolla if it relates to kolla | 17:13 |
Jeffrey4l | glance reconfigure by using with_items https://review.openstack.org/285873 keystone reconfigure by using variable https://review.openstack.org/286570 two edition feel free to make comments. | 17:13 |
Jeffrey4l | SamYaple, sdake_ ^ | 17:13 |
elemoine_ | SamYaple: it's documentation about an image whose Dockerfile is in Kolla | 17:13 |
elemoine_ | this is how it related to kolla | 17:13 |
sdake_ | elemoine_ I have thought about the separate guides documentation, and prefer separate guides, but we can roll them into one file | 17:15 |
sdake_ | why not with the images in the implementation.rst | 17:15 |
sdake_ | elemoine_ we are after scalability for new develoeprs here, not for getting eople to reuse our containers | 17:15 |
sdake_ | elemoine_ what you ask is to allow people to reuse the heka container for their own purpose, which is great | 17:16 |
sdake_ | but it goes against our scalability goals for ramping up new devlopers | 17:16 |
sdake_ | because if there are 115 READMEs to read (and for us to write) that will be constalty dated | 17:16 |
elemoine_ | sdake_: yes exactly, I want to document how to use that image. period | 17:16 |
sdake_ | we have been down this road before with kubernetes and documetning the environment variabless, and that experiment imo was a failure | 17:16 |
elemoine_ | sdake_: my idea was to create docker/heka/readme.txt without setting any precedent | 17:17 |
*** dims_ has joined #kolla | 17:17 | |
sdake_ | if its inconsistent it sets precedent! | 17:17 |
sdake_ | documenation belongs in the docs directory | 17:17 |
Jeffrey4l | SamYaple, are you going to fix the config.yml issue we talked just now? | 17:17 |
sdake_ | in the doc directory | 17:18 |
elemoine_ | sdake_: it will be inconsistent, as I am not going to document the 100+ image we have :) | 17:18 |
elemoine_ | sdake_: ok for the doc directory, I understand | 17:18 |
sdake_ | ya but someone will see that and go "yay 115 commits i can make" | 17:18 |
sdake_ | and then we have to deal with why we made an *exception* for heka | 17:18 |
sdake_ | which to date, I haven't heard a good argument for | 17:18 |
elemoine_ | sdake_: exactly | 17:18 |
sdake_ | why not just document it on your blog post or somehting | 17:18 |
*** dims has quit IRC | 17:19 | |
sdake_ | we dont make exceptions in kolla ;-) | 17:19 |
*** salv-orl_ has joined #kolla | 17:19 | |
elemoine_ | sdake_: no problem for me | 17:19 |
sdake_ | because it leads exactly to the disaster I outlined above | 17:19 |
elemoine_ | sdake_: but it's also a shame | 17:19 |
sdake_ | do it once, have to do it again | 17:19 |
sdake_ | and again | 17:19 |
sdake_ | we make compromises, this is a normal part of democratic process | 17:19 |
sdake_ | but exceptions are not | 17:20 |
elemoine_ | sdake_: have you thought about publishing the kolla images to the official docker hub? | 17:20 |
SamYaple | Jeffrey4l: i doubt ill have time | 17:20 |
sdake_ | elemoine_ yes if you check the thrads, I've been tryin without much success to get infra to host a registry server | 17:20 |
SamYaple | Jeffrey4l: its not breaking anything, and i wont block somoen else fixing it | 17:20 |
SamYaple | going to lunch | 17:20 |
SamYaple | bbl | 17:20 |
Jeffrey4l | ok | 17:20 |
elemoine_ | sdake_: I know, so you never considered pushing the images to the official Docker hub, right? | 17:21 |
*** athomas has quit IRC | 17:21 | |
sdake_ | elemoine_ we have considered it | 17:21 |
elemoine_ | sdake_: I was asking because the standard is to have docs for the images you push | 17:21 |
sdake_ | the problem we have atm is we can't get credentials in the vm securely | 17:21 |
sdake_ | we dont hvae to follow that "standard" ;-) | 17:22 |
*** salv-orlando has quit IRC | 17:22 | |
elemoine_ | sdake_: I would find it strange to publish images without instructions on how to use those images | 17:22 |
sdake_ | we have instructions right on docs.openstack.org | 17:22 |
elemoine_ | no | 17:22 |
sdake_ | teh containers are to be used as a whole | 17:22 |
elemoine_ | yes | 17:22 |
sdake_ | greater then sum of parts ;) | 17:22 |
elemoine_ | not what docker hub users expect | 17:23 |
sdake_ | well we dont publis hto docker hub atm, so its a moot discussion at this point | 17:23 |
sdake_ | but even if we did, i think one link to our upstream docs would do the trick | 17:23 |
sdake_ | the only readme in the repo I want personally is in the TLD | 17:24 |
sdake_ | other then that,I'd prefer no readmes, and all doc/*rst files | 17:24 |
elemoine_ | I disagree, because that link would point to our Ansible/how to deploy kolla documentation, not to the image's documentation | 17:24 |
elemoine_ | not what people expect | 17:24 |
sdake_ | elemoine_ i am just one person | 17:24 |
sdake_ | start a mailing list thread plz | 17:24 |
sdake_ | get the community to weigh in | 17:24 |
sdake_ | i dont speak for the community i speak form yself in this case | 17:25 |
elemoine_ | ok let's close the discussion for now, thanks a lot | 17:25 |
sdake_ | the only time I represent the community is when I'm on stage or in an interview | 17:25 |
sdake_ | and I try not to touch on topics where there is not wide agreement | 17:25 |
sdake_ | try/dont | 17:25 |
elemoine_ | I may write to the mailing list about that, and see if people are interested in that topic | 17:25 |
*** athomas has joined #kolla | 17:26 | |
sdake_ | we should probably have a midcycle session on pushing to the docker hub or a registry hostd by openstck.org | 17:26 |
sdake_ | rahter not midcycle but austin ODS summit | 17:26 |
elemoine_ | sure | 17:26 |
sdake_ | we have a meeting wednesday at 23:00 UTC | 17:27 |
elemoine_ | again if we push kolla images to docker hub users will rightly expect documentation for theses images (and not a link to the kolla deploy documentation) | 17:27 |
elemoine_ | pushing to a registry hosted by openstack.org is another story | 17:28 |
limamauricio_ | Jeffrey4l, can you review it again https://review.openstack.org/#/c/269688/ | 17:28 |
Jeffrey4l | limamauricio_, sorry, i really need to go need. it is 1:30 AM ;D. I will review it tomorrow. | 17:30 |
Jeffrey4l | go bed* | 17:30 |
limamauricio_ | ok | 17:30 |
*** limamauricio_ is now known as mlima | 17:31 | |
*** fthiagogv_ has quit IRC | 17:31 | |
*** ftgomvier has joined #kolla | 17:32 | |
*** ftgomvier is now known as fthiagogv | 17:34 | |
sdake_ | elemoine_ because we are near the end of mitaka-3, we can't take up this topic during the meeting, but we will be adding austin summit discussions there | 17:34 |
sdake_ | we can't take up the topic because I think we will be short on time just taking care of normal business | 17:34 |
*** Serlex has left #kolla | 17:36 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Reconfigure glance service https://review.openstack.org/285873 | 17:37 |
*** Jeffrey4l has quit IRC | 17:43 | |
*** prithiv has quit IRC | 17:44 | |
*** absubram has quit IRC | 17:48 | |
*** sdake_ has quit IRC | 17:51 | |
*** fgimenez has quit IRC | 17:57 | |
*** kproskurin has joined #kolla | 18:01 | |
*** kproskurin has quit IRC | 18:02 | |
*** dwalsh has quit IRC | 18:09 | |
*** jmccarthy1 has joined #kolla | 18:09 | |
openstackgerrit | Merged openstack/kolla: Fix missing libraries for ceilometer https://review.openstack.org/286274 | 18:10 |
*** macsz has quit IRC | 18:14 | |
*** rmart04 has quit IRC | 18:15 | |
vhosakot | I've added the steps to test reconfigure bp in https://etherpad.openstack.org/p/kolla-mitaka-testing-reconfigure Could folks review it when there is a chance. Thanks! | 18:17 |
*** openstackgerrit has quit IRC | 18:18 | |
vhosakot | I have added an example how to reconfigure neutron-server and a generalized example for any service that needs to be reonfigured | 18:18 |
*** openstackgerrit has joined #kolla | 18:18 | |
vhosakot | The directory /etc/kolla/config does not exist on the host. Is this expected ? I just see /etc/kolla/ and not /etc/kolla/config/ | 18:18 |
mlima | sdake, gate-kolla-dsvm-build-centos-binary -> http://pastebin.com/UR49gPXy | 18:18 |
*** dwalsh has joined #kolla | 18:19 | |
*** JoseMello has joined #kolla | 18:23 | |
*** akwasnie has joined #kolla | 18:26 | |
*** serverascode has quit IRC | 18:26 | |
*** serverascode has joined #kolla | 18:28 | |
*** d_code has joined #kolla | 18:31 | |
d_code | if I want to expand compute resources in my OpenStack deployment using Kolla…do I just add it to the [compute] group and run kolla-deploy? | 18:31 |
*** chandankumar has quit IRC | 18:32 | |
d_code | and can I scale out a compute node by the same measure? | 18:34 |
d_code | i.e. reduce my compute resources | 18:34 |
*** dLobatog has joined #kolla | 18:35 | |
*** dLobatog has left #kolla | 18:35 | |
SamYaple | d_code: ye | 18:39 |
SamYaple | yes* | 18:39 |
d_code | sweet. thanks! | 18:39 |
*** d_code has quit IRC | 18:41 | |
*** d_code has joined #kolla | 18:41 | |
*** dims has joined #kolla | 18:45 | |
*** haplo37 has quit IRC | 18:46 | |
*** jmccarthy1 has quit IRC | 18:52 | |
*** dims_ has quit IRC | 18:52 | |
*** jmccarthy1 has joined #kolla | 18:52 | |
*** alger has quit IRC | 18:54 | |
*** openstack has joined #kolla | 19:14 | |
*** inc0 has quit IRC | 19:15 | |
*** openstackstatus has joined #kolla | 19:16 | |
*** ChanServ sets mode: +v openstackstatus | 19:16 | |
*** inc0 has joined #kolla | 19:18 | |
*** openstackgerrit has joined #kolla | 19:19 | |
ccesario | SamYaple, is there any estimated date for kolla documentation!? Scenarios, best practices, HA .... | 19:19 |
inc0 | ccesario, probably we'll do some after M3 | 19:20 |
inc0 | but is there any specific thing you want to ask about? | 19:20 |
inc0 | in quickstart guide we basically described what in our opinion is best deployment | 19:21 |
ccesario | inc0, yes I know about quickstart ....yes yes... I already did some tests using all in one deployment...... now I 'm looking for by advanced docs, as HA by example.... and the best way to usage the hardware and do not spent resources as physical deployments.... if you understand me... | 19:25 |
*** harmw has quit IRC | 19:25 | |
inc0 | so, kolla will deploy HA by default | 19:25 |
inc0 | what you want to do is to specify multiple controllers | 19:26 |
inc0 | best - odd number, but that's described in openstack ha guide itself | 19:26 |
ccesario | yes, in theory I know.... therefore I 'm asking about documents, like how to desgin a good scenario using kolla/containers ... and do not waste resources :) | 19:29 |
inc0 | ccesario, well...as I said, use multiple controllers and kolla will solve the rest | 19:29 |
inc0 | basically deploying ha in kolla is just deploying multinode with more than 1 controller | 19:30 |
inc0 | it's all default | 19:30 |
*** harmw has joined #kolla | 19:31 | |
ccesario | yes yes | 19:31 |
*** elmiko has joined #kolla | 19:32 | |
elmiko | hi, i'm curious if there is a recommended host distro to install kolla on? (i've been trying fedora21-23 and i keep getting errors) | 19:33 |
openstackgerrit | Thiago Gomes proposed openstack/kolla: Cleanup-images with images with duplicate id https://review.openstack.org/286791 | 19:34 |
inc0 | elmiko, so generally don't mix distros | 19:34 |
inc0 | centos on fedora should work | 19:34 |
inc0 | but ubuntu on fedora/centos had all sorts of issues | 19:34 |
inc0 | selinux and systemd usually | 19:34 |
elmiko | ok, so if i run f23 as the host i should be ok for centos images? | 19:35 |
elmiko | i had it go through the kolla-build, but then it kept failing to make the keystone image (it was missing a file i didn't recognize) | 19:35 |
inc0 | well....logs plz | 19:36 |
elmiko | yea, i'm gonna run it again and see what happens. this was all over the weekend, and i know i should have kept the logs... but, well, weekend ;) | 19:37 |
elmiko | inc0: thanks for the advice, i'll come back with some logs =D | 19:37 |
openstackgerrit | Andrey Pavlov proposed openstack/kolla-mesos: Initial commit for python library and CLI https://review.openstack.org/286799 | 19:42 |
openstackgerrit | Blake Smith proposed openstack/kolla: Fixed tool file path in image-building.rst Updated tools/build.py to tools/kolla-build.py Signed-off by: Blake Smith <blaksmit@cisco.com> https://review.openstack.org/286800 | 19:43 |
*** vhosakot has quit IRC | 19:46 | |
*** vhosakot has joined #kolla | 19:46 | |
openstackgerrit | Blake Smith proposed openstack/kolla: Fixed tool file path in image-building.rst https://review.openstack.org/286800 | 19:49 |
*** akwasnie has quit IRC | 19:51 | |
inc0 | SamYaple, around? | 19:54 |
SamYaple | inc0: momentarily | 19:54 |
inc0 | so, my setup looks like that | 19:54 |
inc0 | I have 2 nets, one for mgmt and other one for openstack stuff | 19:55 |
*** akwasnie has joined #kolla | 19:55 | |
inc0 | hostname of host bounds myhostname to mgmt | 19:55 |
inc0 | and my rabbit bounds to openstack stuff | 19:55 |
SamYaple | i started working on this for ceph | 19:55 |
inc0 | which renders clusterer broken | 19:55 |
SamYaple | i know its broken | 19:56 |
SamYaple | look at ceph storage_hostname thing | 19:56 |
dave-mccowan | sdake samyaple any thoughts on this bug i opened related to two VIP support? https://bugs.launchpad.net/kolla/+bug/1550455 | 19:56 |
openstack | Launchpad bug 1550455 in kolla "Keepalive configuration can cause black hole with two VIPs" [Undecided,New] - Assigned to Dave McCowan (dave-mccowan) | 19:56 |
inc0 | maybe we should have hostnames inside dockers? | 19:56 |
SamYaple | inc0: thats not why its broken | 19:56 |
SamYaple | it has nothing to do with docker | 19:56 |
SamYaple | it has to do with your hostname resolution | 19:56 |
inc0 | I know | 19:56 |
inc0 | but inside docker we can have our own /etc/hosts | 19:56 |
SamYaple | it copies the host /etc/hosts | 19:56 |
inc0 | and it's not working | 19:57 |
inc0 | we can try to generate hostnames, but that will be hacky | 19:57 |
SamYaple | its not working because we havent implemneted it for rabbit | 19:57 |
SamYaple | check ceph out its implemented tehre | 19:57 |
SamYaple | dave-mccowan: thats not possible since they both use the same check script | 19:57 |
inc0 | SamYaple, so storage_hostname should be defined per node> | 19:58 |
inc0 | in inventory | 19:58 |
inc0 | ? | 19:58 |
dave-mccowan | samyaple ok. i'll setup a test it to prove to myself one way or the other. | 19:59 |
SamYaple | inc0: no its looked up on the host itself | 19:59 |
SamYaple | dave-mccowan: if the same script fails for one vip but not the other we have a bigger problem _however_ it would still flip to a new host | 19:59 |
vhosakot | Can somebody confirm is enable_haproxy is removed and not needed anymore in globals.yml | 19:59 |
SamYaple | vhosakot: the default is haproxy is on | 20:00 |
SamYaple | so you dont have t oenable it | 20:00 |
SamYaple | if you want to explicitly _disable_ it you must | 20:00 |
inc0 | ok I see it | 20:00 |
inc0 | hmm | 20:00 |
vhosakot | but, the field enable_haprosy is missing in globals.yml | 20:00 |
inc0 | ok I'll implement it for rabbit | 20:00 |
SamYaple | inc0: no it doesnt work | 20:00 |
SamYaple | inc0: i already tried it | 20:00 |
SamYaple | erlang does its own hostname fetching | 20:00 |
inc0 | damn | 20:00 |
SamYaple | you have to do other hacky stuff | 20:00 |
SamYaple | i havent worked it all out yet | 20:01 |
vhosakot | SamYaple: shall I add the enable_haproxy field back into globals,.yml ? | 20:01 |
SamYaple | for now, the only abstraction that works is storage and tunnel i think | 20:01 |
vhosakot | and submit for review ? | 20:01 |
SamYaple | vhosakot: no it was never there | 20:01 |
SamYaple | vhosakot: it doesnt belong there | 20:01 |
SamYaple | you can set it in your globals.yml to false if you want | 20:01 |
SamYaple | but it defaults to true | 20:01 |
vhosakot | ah, ok.. I thgout it was always there and commented out, per the wording in quickstart http://docs.openstack.org/developer/kolla/quickstart.html | 20:02 |
inc0 | SamYaple, and let me guess, clusterer doesn't work with IPs right? | 20:02 |
SamYaple | inc0: it *might* but its pretty late in the game to be changing that around | 20:03 |
*** alger has quit IRC | 20:04 | |
*** sdake has quit IRC | 20:05 | |
inc0 | yeah, just thinking of options | 20:06 |
inc0 | I'll just make hostnames | 20:06 |
*** akwasnie has quit IRC | 20:11 | |
*** akwasnie has joined #kolla | 20:12 | |
*** ssurana has joined #kolla | 20:16 | |
*** sdake has joined #kolla | 20:16 | |
*** alger has joined #kolla | 20:17 | |
*** jmccarthy1 has quit IRC | 20:17 | |
*** jasonsb has quit IRC | 20:18 | |
*** The_Ball has joined #kolla | 20:24 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Bump to latest stable ansible and deps https://review.openstack.org/286820 | 20:24 |
openstackgerrit | Thiago Gomes proposed openstack/kolla: Deleting only images from kolla https://review.openstack.org/286823 | 20:31 |
*** sdake has quit IRC | 20:31 | |
*** akwasnie has quit IRC | 20:36 | |
*** sdake has joined #kolla | 20:43 | |
vhosakot | SamYaple: can you comment on the script to automate kolla when you have a chance | 20:49 |
*** sdake has quit IRC | 20:56 | |
*** sdake has joined #kolla | 20:59 | |
*** esimone has joined #kolla | 21:02 | |
The_Ball | Is it easy to deploy just cinder and required services (keystone, mariadb) with kolla? | 21:03 |
SamYaple | vhosakot: i think the review you are talking about was blocked with -2 | 21:04 |
SamYaple | The_Ball: yea its fairly simple. you need to look at ansible/group_vars/all.yml and disable all the services you dont want by setting "enable_<service>: 'no'" in your globals.yml before deploy | 21:05 |
vhosakot | SamYaple: I have not even submitted for review... I needed to know if the script I am _planning_ to submit to automate steps in quickstart to install, build, and deploy kolla from scratch has not duplicated script in kolla :) | 21:05 |
vhosakot | SamYaple: I see many shell scripts for gate and vagrant (like tools/deploy_aio.sh, toos/setup_gate.sh and dev/vagrant/ubuntu-bootstrap.sh), but none of them install, build and deploy kolla from scratch... | 21:05 |
The_Ball | SamYaple, ah, that does sound simple :) I'll give it a try | 21:05 |
SamYaple | vhosakot: there was one like this already that got blocked | 21:05 |
vhosakot | ah, why ? | 21:06 |
SamYaple | let me see if i can find it | 21:06 |
vhosakot | the script I have written will install kolla in just one command from scratch on a new server | 21:06 |
vhosakot | cool, thanks! | 21:06 |
sdake | vhosakot there is already a review up for this | 21:09 |
sdake | vhosakot last i saw the -2 was removed with the idea of making it ansible down the road | 21:09 |
sdake | with possibly a kolla-host repo or something similar to be discussed at ODS | 21:09 |
*** akwasnie has joined #kolla | 21:11 | |
vhosakot | ah ok.. cool... | 21:11 |
sdake | a script in the repo that does the setup job isn't all that onerous | 21:16 |
sdake | i dont really want to automate the building step though | 21:16 |
sdake | i want people to understnad how that works | 21:16 |
sdake | otherwise it looks like a bunch of black magic "CONTAINERS" | 21:17 |
vhosakot | yes, agreed.. running each command by hand, we learn a lot.. true | 21:17 |
vhosakot | yep | 21:17 |
vhosakot | :) | 21:17 |
sdake | yu dont learn alot by turnning off seinux or that kind of stuff | 21:17 |
sdake | or turning off libvirt | 21:17 |
sdake | or configuring docker properly | 21:17 |
sdake | etc | 21:17 |
sdake | these are all steps that get in the way of an evaluation | 21:17 |
vhosakot | right | 21:17 |
sdake | and we need to automate those, shell works for now but not down the road | 21:18 |
vhosakot | so, are you saying a script is needed to _setup_ the env until kolla-build ? | 21:18 |
sdake | the purpose of that script in my mind is to get peopel to build.py faster | 21:18 |
sdake | ya and there is a patch in the repo for centos for jut that | 21:18 |
sdake | its not intended forproduction | 21:18 |
sdake | its intended for evaluation | 21:18 |
vhosakot | for centOS, shall I submit for Ubuntu... wanted to make sure I don't submit something the community does not want :) | 21:19 |
sdake | i like the idea | 21:19 |
sdake | not sure where others stnd on it | 21:19 |
vhosakot | yep, all for eval and dev | 21:19 |
sdake | inc0 wants it as ansible | 21:19 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Bump to latest stable ansible and deps https://review.openstack.org/286820 | 21:19 |
sdake | but i thnk he will accept a scriptfrist approach | 21:19 |
SamYaple | there you go sdake ^^ ansible bump | 21:19 |
SamYaple | no git cloning | 21:19 |
*** macsz has quit IRC | 21:20 | |
sdake | SamYaple what is the privs:USAGE thing | 21:22 |
sdake | never seen that before | 21:22 |
sdake | ya that git clone of a unstable beta no bueno :) | 21:23 |
SamYaple | sdake: thats the mysql "no privleges" option | 21:26 |
sdake | SamYaple even though i already added you to the coresec team can you respond on thread and say "count me in"or something else that you like | 21:26 |
SamYaple | basically it just lets haproxy sign into mysql and do nothing else | 21:26 |
sdake | soundsgood | 21:26 |
SamYaple | (its what weve been doing, this is just explictly granting priv because module changed) | 21:27 |
*** dims_ has joined #kolla | 21:31 | |
*** dims has quit IRC | 21:33 | |
SamYaple | so sdake... | 21:39 |
SamYaple | the shade domain module only exists in latest ansible-dev | 21:39 |
sdake | i have a patch that doesn't use shade in progress | 21:39 |
SamYaple | right. but we want to use shade | 21:40 |
sdake | somewhre on one of these damn vms on my laptop ;) | 21:40 |
SamYaple | its not _can_ we do it without shade, its we dont want to | 21:40 |
sdake | well if you get ansible-dev in there, that wfm, as long as its pinned to a specific tag | 21:40 |
SamYaple | it can be pinned to a commit | 21:40 |
sdake | i its not pinned to a specific tag, we can always rework it down the road | 21:40 |
sdake | pin to commit works too | 21:40 |
SamYaple | ok. let me do that quick like then | 21:41 |
sdake | as long as its reproducible | 21:41 |
SamYaple | we can pin to tag before mitaka most likely | 21:41 |
SamYaple | but even if we cant, if it works it works | 21:41 |
SamYaple | so thats fine | 21:41 |
*** akwasnie has quit IRC | 21:49 | |
elmiko | inc0: still around? | 21:51 |
elmiko | hey all, i am following the quickstart guide, Hhttp://docs.openstack.org/developer/kolla/quickstart.html | 21:53 |
elmiko | i have a fedora 23 host setup and have followed those instruction for a all-in-one install | 21:53 |
elmiko | but, i get this right out of the gate when i run kolla-build Hhttp://paste.openstack.org/show/488856/ | 21:53 |
elmiko | any ideas? | 21:53 |
*** rhallisey has quit IRC | 21:55 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Bump to latest ansible and deps https://review.openstack.org/286820 | 21:57 |
SamYaple | alright sdake that pins to a commit the best way we can (unfortantely that removes teh depth option) | 21:57 |
SamYaple | we can remove this once 2.1 lands | 21:57 |
SamYaple | and with that we have all the shade modules we need to switch entirely over to shade | 21:58 |
SamYaple | m/win 48 | 21:58 |
*** rhallisey has joined #kolla | 21:59 | |
sdake | rhallisey re you wanting to joi nthe security team | 21:59 |
sdake | rhallisey did you read he VMT document in the governance repository? | 21:59 |
rhallisey | ya I did | 21:59 |
sdake | basically the team attaches patches to bugs | 21:59 |
sdake | and then I think the security VMT team in openstack merges them in one go once the OSSA is ready | 22:00 |
sdake | but i'm not really sure | 22:00 |
sdake | its not about securitzing kolla, its about fixing security bugs we fuond and publishign them in a controlled coordinated fashion | 22:00 |
sdake | we/community/anyone | 22:01 |
sdake | I think it will take us a couple months to get up to speed | 22:01 |
rhallisey | right. I figured I'd add that because it was my goal | 22:01 |
sdake | i think the next step is a security audit of the code base by a third party | 22:02 |
sdake | red hat has a great security audit team - would be nice if we could get them to audit the codebase | 22:02 |
rhallisey | I figured I'd volunteer either way | 22:02 |
sdake | only 25k lines of code | 22:02 |
elmiko | might be worth looking at some of the work that the OSSP is doing with regards to threat analysis too | 22:02 |
sdake | should be pretty straightforward to tak them into it :) | 22:02 |
elmiko | (sorry to bud in) | 22:02 |
*** dims_ has quit IRC | 22:02 | |
sdake | elmiko welcome aboard :) | 22:02 |
elmiko | =) | 22:02 |
elmiko | i'm still climbing on, kolla-build is fighting me | 22:03 |
sdake | elmiko where areyou stuck | 22:03 |
sdake | i think we need some docs that indicate that you should have separate build and push steps and both require the --registry option | 22:03 |
sdake | i think that trips people up | 22:03 |
elmiko | well, i'm on round 3 of doing a clean quickstart install. but now kolla-build fails with http://paste.openstack.org/show/488856/ | 22:04 |
sdake | elmiko fedora may not be all that functioanl unfortunately | 22:04 |
elmiko | ahh, ok | 22:04 |
sdake | i'd go with centos or ubuntu or oracle linux at this time | 22:05 |
sdake | debian/fedora coming down the road i would hope | 22:05 |
elmiko | ok, i'll give this a try with centos. is centos7 a good candidate? | 22:05 |
elmiko | cool | 22:05 |
sdake | the big prbolem with fedora is the yum to dnf change | 22:05 |
sdake | yup centos7 | 22:05 |
sdake | build.py --base centos -t source --registry reginfo | 22:05 |
sdake | or if you want binary jsut | 22:06 |
sdake | build..py --base centos --registry reginfo | 22:06 |
elmiko | is fed23 ok for the host though? | 22:06 |
sdake | the nbuild.py --registry reginfo --push | 22:06 |
sdake | elmiko at one point fedora22 as a host was incompatible because the kernel mdoulse were compressed and something couldn't read them in the centos container | 22:06 |
sdake | this was on centos 7.0 | 22:06 |
elmiko | interesting | 22:07 |
elmiko | so, centos for host is recommended? | 22:07 |
sdake | i run centos on host | 22:07 |
elmiko | cool, i'll give that a try | 22:07 |
sdake | it would be cool to try fedora23 as a host | 22:07 |
sdake | so we can update the instructions | 22:07 |
sdake | its possible centos has updated their builds of guestfs or whatever was reading the modules to work prpoerly | 22:07 |
sdake | i dont recall the exact issue | 22:07 |
elmiko | i almost had it working this weekend, but the keystone container kept failing to build | 22:07 |
sdake | ya keystone is fixed now | 22:08 |
elmiko | any idea what the image dir error is about? | 22:08 |
sdake | try centos on fedora 23 - should work if centos is updated for the one binary | 22:08 |
sdake | no clue, but i'd recommend the following | 22:08 |
sdake | pip uninstall kolla | 22:08 |
sdake | go to the kolla directory | 22:08 |
sdake | sudo pip unisntall kolla i mean | 22:08 |
sdake | and run tools/build.py directly | 22:08 |
sdake | I think the pip installing kolla causes people all sorts of confusion | 22:09 |
sdake | especially since it looks like you have run into a bug of some kind | 22:09 |
elmiko | ok, gonna try the build.py directly | 22:09 |
sdake | make sure to uninstall kolla first | 22:09 |
sdake | because the buid tools are made to look in /usr/share/kolla/docker first then the local cwd second | 22:10 |
elmiko | got it | 22:10 |
sdake | same with kolla-ansible | 22:10 |
*** dims has joined #kolla | 22:10 | |
sdake | (iirc, this stuff changes all the time, have hard time keeping up ;) | 22:10 |
elmiko | hehe, it has that feeling to it ;) | 22:10 |
elmiko | gotta love the cutting edge! | 22:10 |
elmiko | ok, well it's building now =) | 22:12 |
elmiko | sdake++ | 22:12 |
sdake | 1000 commits since liberty | 22:12 |
sdake | only 25k lines of code | 22:12 |
elmiko | whoa | 22:12 |
sdake | including documentation | 22:12 |
sdake | and specs | 22:12 |
sdake | which alone are 5k of the codebase | 22:12 |
sdake | you can run ./loc to see for yourself | 22:12 |
elmiko | i applaud the team's effort to keep it tidy ;) | 22:13 |
sdake | ya we are moving past gold plating into solid gold ;) | 22:14 |
elmiko | haha | 22:14 |
*** shardy has quit IRC | 22:15 | |
sdake | SamYaple re https://blueprints.launchpad.net/kolla/+spec/mariadb-lights-out | 22:20 |
sdake | what is your thinking on this? | 22:20 |
sdake | I was thinking a "recovery" action in the playbooks which does whatever recovery is needed | 22:20 |
sdake | for mariadb this could recover by running the upgrade playbook | 22:21 |
sdake | neutron needs recovery of some sort as well | 22:21 |
sdake | after a reboot | 22:21 |
sdake | elemoine ping | 22:23 |
sdake | re https://blueprints.launchpad.net/kolla/+spec/heka | 22:23 |
sdake | shouldn't that be in needs review since you have finished the last patch needed? | 22:23 |
*** achanda has quit IRC | 22:23 | |
vhosakot | neutron coming out of lights out, I have seen that none of the agents come up until neutorn-server is up... neutron-server is up, virtual bridge agent is up next (neutron-linux-bridge agent or OVS agent), then, l3-agent and dhcp-agent... this order worked for me, although the order is not well documented anywhere about neutron recovering after lights out | 22:25 |
sdake | asalkeld what is status of https://blueprints.launchpad.net/kolla/+spec/zookeeper - good progress or compelte or what? | 22:26 |
sdake | vhosakot we may have to put recovery in a rc or punt it entirely | 22:26 |
sdake | vhosakot but reconfigure is ready to go - recommend focusing any development there ;) | 22:27 |
vhosakot | yep, I'm rebuilding with latest code... | 22:27 |
*** jtriley has quit IRC | 22:27 | |
sdake | rhallisey is this implemented: https://blueprints.launchpad.net/kolla/+spec/nfs-support-in-cinder | 22:27 |
vhosakot | sdake: https://etherpad.openstack.org/p/kolla-mitaka-testing-reconfigure | 22:28 |
sdake | vhosakot rockin :) | 22:29 |
vhosakot | cool :) | 22:29 |
vhosakot | Jeffrey will review it tomorrow | 22:29 |
sdake | can you do more services? | 22:29 |
sdake | you mentioned you would do nova as well iirc | 22:30 |
vhosakot | yep, neutron, nova and horizon | 22:33 |
vhosakot | sdake: we need all those services in bp implemented with reconfig before march 4th ? | 22:35 |
vhosakot | is marth 4th the right date (deadline) ? | 22:36 |
vhosakot | march* | 22:36 |
*** achanda has joined #kolla | 22:36 | |
*** JoseMello has quit IRC | 22:38 | |
*** inc0 has quit IRC | 22:40 | |
rhallisey | sdake, yes | 22:42 |
rhallisey | just changed it | 22:42 |
rhallisey | brb | 22:43 |
*** rhallisey has quit IRC | 22:43 | |
sbezverk | sdake I still get Not authorized, when I run any magnum commands. Something is not right in keystone as it generates these messages for every attempt: http://paste.openstack.org/show/488862/ | 22:43 |
sbezverk | sdake anybody managed to get magnum working with kolla deployment? | 22:44 |
sdake | sbezverk uncler if magnum is working at this time | 22:44 |
sdake | try a different service plz | 22:45 |
sdake | if magnum is busted which is likely file a bug | 22:45 |
sdake | there are some keystone changes which i didn't see make their way into any of the non-compute-kit containers like murano and heat etc | 22:45 |
sbezverk | sdake other services including horizon seems to be working fine.. I will file a bug then.. | 22:47 |
SamYaple | sdake: neutron is fine after reboot | 22:47 |
SamYaple | sdake: for mariadb thats a _full_ lights out which cannot be automatically recovered | 22:47 |
SamYaple | aio is fine for lights out right now | 22:47 |
*** dwalsh has quit IRC | 22:47 | |
SamYaple | the best I can give is a helper playbook and some docs | 22:48 |
SamYaple | sbezverk: dont file a bug | 22:49 |
SamYaple | its a known issue from the keystone stuff | 22:49 |
SamYaple | sbezverk: _when_ the keystone patch upstream merges your issue will be fixed, but also it should be fixed after https://review.openstack.org/#/c/286637/ | 22:49 |
*** vhosakot has quit IRC | 22:51 | |
sdake | SamYaple re neutron, I got impression from vhosakot that neutron needs its services restarted in a specific horder | 22:51 |
sdake | horder/order | 22:51 |
SamYaple | sdake: for ubuntu 14.04 you have to reapply the /run shared fix | 22:51 |
SamYaple | my suggestion would be an init script upgrade, but thats just what i do to solve the issue | 22:51 |
sdake | SamYaple is he running ubuntu 14.04? | 22:52 |
SamYaple | i believe so | 22:52 |
sdake | too bad he just dropped off | 22:52 |
sdake | ok so the answer is we dont at this time need a recover action I think from what you said | 22:52 |
sdake | but we need some documentation and a special lights out playbook | 22:52 |
SamYaple | not for kolla-ansible no | 22:52 |
SamYaple | some specific playbook for mariadb to assist in lights out, yes | 22:53 |
SamYaple | but its mostly going to be docs and the operator | 22:53 |
sdake | ok i'll add that to he blueprint | 22:53 |
SamYaple | basically with like a datacenter outage what will happen is the operator will need to essentially pick a random mariadb server to serve as new master | 22:53 |
SamYaple | we have a playbook that _might_ help there, but there is no gaurantee with a power outage it will have the needed info | 22:54 |
sbezverk | sdake I have filed a new bug: 1551992 | 22:54 |
openstack | bug 1551992 in kolla "After successful deployment any magnum commands return ERROR: Not Authorized" [Undecided,New] https://launchpad.net/bugs/1551992 | 22:54 |
sdake | sbezverk i think sam said don't file a bug | 22:54 |
sdake | but ok ;) | 22:54 |
sdake | i doubt magnum works out of the box as is anyway | 22:55 |
SamYaple | its ok sbezverk its just not a kolla bug | 22:55 |
sdake | it probably needs a rescrub | 22:55 |
sdake | it worked in kilo, notsure about liberty tbh | 22:55 |
SamYaple | oh. wait no thats magnum | 22:55 |
sdake | i think it worked in liberty | 22:55 |
SamYaple | thats using keystone v2 | 22:55 |
SamYaple | magnum doesnt support v3 yet | 22:55 |
sdake | magnum hasn't got v3 implemented yet? | 22:55 |
*** dims has quit IRC | 22:55 | |
SamYaple | nope | 22:56 |
SamYaple | https://review.openstack.org/#/c/274239/ | 22:56 |
sdake | groan | 22:57 |
sdake | it was almost a year ago that work was slatd to be done | 22:57 |
SamYaple | yup | 22:57 |
SamYaple | i found this funny https://blueprints.launchpad.net/magnum/+spec/magnum-keystone-v3 | 22:57 |
SamYaple | read the comments | 22:57 |
sdake | SamYaple join #openstack-containers | 22:58 |
sbezverk | SamYaple, sorry I missed your comment about bug :-( | 22:59 |
SamYaple | no problem sbezverk | 22:59 |
SamYaple | its not a big deal | 22:59 |
sbezverk | SamYaple, in theory if I change OS variable to V2 it should fix magnum issue then? | 23:00 |
sdake | something needs to be done about magnum, we will just repurpose that bug for doing so | 23:00 |
SamYaple | no we arent setting up v2 keystone at all | 23:00 |
SamYaple | so its likely not to work | 23:00 |
*** dims has joined #kolla | 23:00 | |
asalkeld | sdake: yip it's done (the zk bp) | 23:02 |
sdake | asalkeld can you mark it as so | 23:02 |
asalkeld | done | 23:02 |
sdake | https://bugs.launchpad.net/kolla/+bug/1551992 | 23:03 |
openstack | Launchpad bug 1551992 in kolla "Make Magnum use Keystone v3" [Undecided,New] | 23:03 |
sdake | there, bug repurposed | 23:03 |
SamYaple | thats still wrong but ok | 23:03 |
sdake | SamYaple you said your working on that one? | 23:03 |
SamYaple | we arent setting up v3 properly at _all_ | 23:03 |
SamYaple | its only partially working | 23:04 |
sdake | you mean with magnum | 23:04 |
SamYaple | no i mean period | 23:04 |
sdake | or in general | 23:04 |
sdake | got it | 23:04 |
elmiko | sdake: quick question, is kolla using, or planning to use, kubernetes? | 23:04 |
sdake | elmiko some folks have tlked about it | 23:04 |
sdake | i think for that to happen - it would take a team of 3-4 cats working full time on it uninterrupted | 23:04 |
sdake | asa was done with kolla-mesos | 23:04 |
elmiko | currently it is just docker+ansible though? | 23:05 |
sdake | when asalkeld proposed it, he said they would be brining 6 developers to the party to implement it | 23:05 |
elmiko | oh, kolla-mesos, that sounds interesting... ;) | 23:05 |
sdake | elmiko we also have a work in progress kolla-mesos implementation | 23:05 |
elmiko | neat! | 23:05 |
sdake | i'd like to see that working well before kubernetes is tackled | 23:05 |
elmiko | makes sense | 23:05 |
sdake | but if it was done in parallel that would wfm ;) | 23:05 |
sdake | I have alot of kubernetes experience | 23:06 |
sdake | and think the job is impossible | 23:06 |
openstackgerrit | Mauricio Lima proposed openstack/kolla: [WIP] Make Heka collect Manila logs https://review.openstack.org/286876 | 23:06 |
sdake | but things have changed | 23:06 |
sdake | which may make it possible | 23:06 |
elmiko | i'll take your word for it ;) | 23:06 |
sdake | and asalkeld has thought up ideas i didn't during our initial go of making kolla worko n kubernetes | 23:06 |
sdake | (when the project started, it was kolla on kubernetes on fedora atomic ;) | 23:06 |
* elmiko nods | 23:06 | |
sdake | talk about a whisky tango charlie foxtrot | 23:06 |
elmiko | haha | 23:06 |
elmiko | i know atomic has been making a lot of progress, but i haven't used it outside of openshift for some time now | 23:07 |
sdake | their upstream seemed pretty non-engaged | 23:07 |
sdake | when i worked on magnum quite a bit | 23:07 |
sdake | don't know if that has changed or not | 23:07 |
elmiko | that's sad =( | 23:07 |
*** intr1nsic has quit IRC | 23:08 | |
sdake | this is why coreos is clobbering atomic | 23:08 |
elmiko | interesting... | 23:08 |
sdake | (IMNSHO:) | 23:08 |
SamYaple | pbourke_: yo | 23:08 |
elmiko | totally fair, i'm honestly just curious about this stuff. i am doing an internal tech talk about containers in openstack, and i've found kolla to be a very exciting project | 23:08 |
elmiko | magnum too | 23:09 |
elmiko | although, i would love to switch from using devstack to using kolla. and, i'd really like to get see if we can get a sahara playbook submitted to the project. | 23:10 |
sdake | elmiko we will absolutely take any big tent project playbooks + containers | 23:10 |
elmiko | \o/ | 23:11 |
elmiko | i need to talk with our team, but i'm gonna bring it up at summit | 23:11 |
sdake | elmiko which place is your gig located | 23:11 |
sbezverk | vhosakot quick question, if I need to reconfigure neutron's ml2 I just create ml2_conf.ini with new configuration and follow the same steps reconfiguring netron-server container? | 23:12 |
elmiko | sdake: not sure i follow, my day job? | 23:12 |
sdake | ya where do you work | 23:12 |
elmiko | red hat =) | 23:12 |
sdake | cool | 23:12 |
sdake | we have alot of red hat core reviewers | 23:12 |
elmiko | nice, i should be bugging them! | 23:13 |
elmiko | ;) | 23:13 |
sdake | i think kolla is generally well known internally at RHT, but another tech talk wouldn't hurt :) | 23:13 |
elmiko | it's new for our group (emergin tech) | 23:13 |
*** haplo37 has quit IRC | 23:14 | |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Extend kolla-mesos-config to be able to set zookeeper variables https://review.openstack.org/286400 | 23:18 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Move generate_configs into the Command class https://review.openstack.org/286399 | 23:18 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Restart daemons if their config changes https://review.openstack.org/284664 | 23:18 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Make write_file return if the contents changed https://review.openstack.org/286398 | 23:18 |
*** salv-orl_ has joined #kolla | 23:19 | |
*** Jeffrey4l has joined #kolla | 23:21 | |
*** salv-orlando has quit IRC | 23:22 | |
*** blahRus has quit IRC | 23:26 | |
*** Jeffrey4l has quit IRC | 23:31 | |
*** intr1nsic has joined #kolla | 23:32 | |
*** spisarski has quit IRC | 23:43 | |
*** sdake has quit IRC | 23:47 | |
*** salv-orlando has joined #kolla | 23:50 | |
*** salv-orlando has quit IRC | 23:50 | |
*** salv-orl_ has quit IRC | 23:50 | |
*** salv-orlando has joined #kolla | 23:50 | |
*** unicell has joined #kolla | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!