| sdake | sbezverk ready for webex | 00:00 |
|---|---|---|
| sdake | but if you need to jet totally understand | 00:00 |
| sdake | my wie is ready t ofire me too | 00:00 |
| *** haplo37__ has quit IRC | 00:02 | |
| sdake | sbezverk ack on foreground state | 00:03 |
| *** sdake_ has joined #openstack-kolla | 00:06 | |
| *** zhurong has quit IRC | 00:07 | |
| *** daneyon has joined #openstack-kolla | 00:08 | |
| kfox1111 | sbezverk: the idea is to run the screen command via exec to see what's wrong with the screen command. | 00:08 |
| kfox1111 | the while sleep thing just lets you keep the container running long enough to do the screen. | 00:08 |
| *** sdake has quit IRC | 00:09 | |
| *** daneyon has quit IRC | 00:12 | |
| sbezverk | sdake kfox1111 https://cisco.webex.com/join/sbezverk | 00:20 |
| sbezverk | but in 30 minutes I am gone :-) | 00:20 |
| kfox1111 | I gota head out in like 10 min. :/ | 00:21 |
| sbezverk | good :-) let's call it a day, we can do it tomorrow | 00:21 |
| sbezverk | I filed a bug with kuberenetes | 00:22 |
| kfox1111 | k. | 00:22 |
| sbezverk | maybe somebody get back to me with some ideas | 00:22 |
| kfox1111 | +1. :) | 00:22 |
| kfox1111 | it might be interesting to see if adding [ -f /configmap_file ] || echo configmap file not there; real command | 00:23 |
| kfox1111 | does anything. | 00:23 |
| sdake_ | kfox1111 you can't use exec - exec is a bash operation | 00:24 |
| sdake_ | we couldn't get bash -c to work | 00:25 |
| kfox1111 | no, I mean docker exec. | 00:25 |
| kfox1111 | you run a dummy command in the container so it will start and stay up, | 00:25 |
| sdake_ | oh thats not helpful because it only happens during startup | 00:25 |
| kfox1111 | then you docker exec in, and try the real command you want to test there. | 00:25 |
| sdake_ | we tried that | 00:25 |
| sdake_ | worked every time | 00:26 |
| kfox1111 | that makes sure you have the right command line figurd out to put in the docker command: | 00:26 |
| kfox1111 | but did you try that with the screen command line above? | 00:26 |
| sdake_ | not sure how that helps? | 00:26 |
| sdake_ | what we tried is sleep infinity | 00:26 |
| sdake_ | followed by a docker exec | 00:26 |
| kfox1111 | justo make sure screen isn't the thing crashing too. | 00:27 |
| sdake_ | screen was a new technique to try to replicate what we were doing with bash -c which didn't work because of json parsing errors | 00:27 |
| kfox1111 | and all the screen command line swithces are correct. | 00:27 |
| sdake_ | the issue is screen doesn't run in foreground mode | 00:27 |
| sdake_ | there is probably a way to make it do so | 00:28 |
| kfox1111 | ah. | 00:28 |
| sdake_ | that is what is needed :) | 00:28 |
| sdake_ | since it backgrounds, the container exits immediately | 00:28 |
| kfox1111 | wait $!? :) | 00:28 |
| sdake_ | if we could do two commands we wouldn't have a problem | 00:29 |
| sdake_ | the problem is we can't do two commands | 00:29 |
| kfox1111 | ah. | 00:29 |
| sdake_ | kill -9 1 doesn't work | 00:29 |
| sdake_ | (inside a container) | 00:29 |
| sdake_ | or outside kill -9 PIDOFCONTAINERSPID1 | 00:29 |
| kfox1111 | maybe put the commands in configmap and execute it from there as a bash script? | 00:29 |
| kfox1111 | or does that munge the json too? | 00:30 |
| sdake_ | a pretty easy solution is to put a bash script in that container that doe smultiple things | 00:30 |
| sdake_ | so pid 1 isn't openvswitch-db | 00:30 |
| sdake_ | and then run that bash script fro mthe daemonset | 00:31 |
| kfox1111 | yeah. | 00:31 |
| kfox1111 | that works too. | 00:31 |
| dcwangmit01 | sbezverk: Add this to the container spec, and it will pause the container on startup. " command: [ "/bin/bash", "-c", "while :; do echo 'Hit CTRL+C'; sleep 100; done" ]" . Then, you can "kubectl exec -it <pod> -c <container> bash" and then manually kick off the container startup scripts | 00:32 |
| dcwangmit01 | ^ I've only glanced at this conversation, so i may be off topic :) | 00:33 |
| sdake_ | dcwangmit01 that doesn't work as i've said 3 or 4 imes :) | 00:39 |
| *** banix has quit IRC | 00:43 | |
| *** rhallisey has joined #openstack-kolla | 00:43 | |
| *** diogogmt has quit IRC | 00:46 | |
| *** zhiwei has joined #openstack-kolla | 00:48 | |
| *** phuongnh has joined #openstack-kolla | 00:49 | |
| dcwangmit01 | Damn. It totally works for me, in the general sense. You may have a different container, without bash or something | 00:53 |
| dcwangmit01 | I've used this to debug mariadb, as well as mariadb-bootstrap | 00:54 |
| dcwangmit01 | spec: | 00:54 |
| dcwangmit01 | containers: | 00:54 |
| dcwangmit01 | - image: kollaglue/centos-binary-mariadb:2.0.0 | 00:54 |
| dcwangmit01 | name: mariadb | 00:54 |
| dcwangmit01 | command: [ "/bin/bash", "-c", "while :; do echo 'Hit CTRL+C'; sleep 100; done" ] | 00:54 |
| dcwangmit01 | ^ I got from wirehead_ , so I know he uses it too | 00:55 |
| dcwangmit01 | Debugging containers on kubernetes is hard :) | 00:56 |
| *** dcwangmit01 has quit IRC | 01:06 | |
| *** ayoung has joined #openstack-kolla | 01:08 | |
| sbezverk | dcwangmit01 the problem, this command is in json file | 01:10 |
| sbezverk | the one we stick into config map | 01:10 |
| sdake_ | sbezverk spend time with wife - dont want to get wifefired :) | 01:10 |
| sbezverk | :-) gone gone gone | 01:11 |
| *** zhurong has joined #openstack-kolla | 01:16 | |
| *** zhurong has quit IRC | 01:21 | |
| *** zhurong has joined #openstack-kolla | 01:22 | |
| *** daneyon has joined #openstack-kolla | 01:56 | |
| *** dcwangmit01 has joined #openstack-kolla | 01:59 | |
| *** daneyon has quit IRC | 02:01 | |
| *** sdake has joined #openstack-kolla | 02:03 | |
| *** sdake_ has quit IRC | 02:04 | |
| *** rhallisey has quit IRC | 02:05 | |
| *** fragatina has joined #openstack-kolla | 02:12 | |
| *** sdake_ has joined #openstack-kolla | 02:14 | |
| openstackgerrit | Matthew Montgomery proposed openstack/kolla: Customizations for keepalived https://review.openstack.org/347030 | 02:16 |
| *** sdake has quit IRC | 02:17 | |
| *** dave-mcc_ has joined #openstack-kolla | 02:17 | |
| *** dave-mccowan has quit IRC | 02:17 | |
| *** unicell has quit IRC | 02:25 | |
| *** fragatina has quit IRC | 02:27 | |
| *** zhurong_ has joined #openstack-kolla | 02:37 | |
| *** zhurong has quit IRC | 02:39 | |
| *** yuanying has quit IRC | 02:49 | |
| openstackgerrit | Merged openstack/kolla: Add customizations for keystone https://review.openstack.org/349736 | 02:50 |
| *** banix has joined #openstack-kolla | 02:56 | |
| *** dave-mcc_ has quit IRC | 02:59 | |
| britthouser | Looks like the automation for retrieving MAC addresses got the wrong address for 729534-comp-s3500-050.cloud7.osic.rackspace.com | 03:11 |
| *** unicell has joined #openstack-kolla | 03:13 | |
| openstackgerrit | Matthew Montgomery proposed openstack/kolla: Customizations for keepalived https://review.openstack.org/347030 | 03:16 |
| britthouser | Ditto for 729564-comp-s3700-042.cloud7.osic.rackspace.com | 03:20 |
| *** unicell has joined #openstack-kolla | 03:21 | |
| openstackgerrit | Merged openstack/kolla: Enable central logging without deploying elasticsearch/kibana https://review.openstack.org/320012 | 03:23 |
| openstackgerrit | Merged openstack/kolla: Customizations for mongodb https://review.openstack.org/347073 | 03:23 |
| openstackgerrit | Merged openstack/kolla: Updated from global requirements https://review.openstack.org/350892 | 03:24 |
| openstackgerrit | Merged openstack/kolla: Change workers to be static when using kubernates https://review.openstack.org/350356 | 03:24 |
| sdake_ | down to 8 pages of reviews | 03:25 |
| *** unicell has quit IRC | 03:26 | |
| *** ayoung is now known as ayoung_ZZZzzz | 03:26 | |
| britthouser | Go sdake_! | 03:27 |
| sdake_ | i'm wiped out | 03:27 |
| sdake_ | time for bed | 03:27 |
| sdake_ | night | 03:27 |
| britthouser | same here! | 03:27 |
| britthouser | gnight! | 03:27 |
| *** chip_ has joined #openstack-kolla | 03:29 | |
| *** chip_ has quit IRC | 03:30 | |
| *** unicell has joined #openstack-kolla | 03:31 | |
| openstackgerrit | Merged openstack/kolla: Cleanup images as part of kolla-ansible cleanup https://review.openstack.org/342860 | 03:34 |
| openstackgerrit | Merged openstack/kolla: Fix the kolla_docker issue with docker 1.12 https://review.openstack.org/349354 | 03:34 |
| *** unicell has quit IRC | 03:36 | |
| openstackgerrit | Merged openstack/kolla: Customizations for ceilometer https://review.openstack.org/347005 | 03:39 |
| *** sdake_ has quit IRC | 03:42 | |
| *** yuanying has joined #openstack-kolla | 03:48 | |
| *** dwalsh has joined #openstack-kolla | 03:54 | |
| *** sdake has joined #openstack-kolla | 03:57 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add aodh role https://review.openstack.org/351027 | 03:57 |
| *** zhubingbing has joined #openstack-kolla | 04:03 | |
| zhubingbing | hi | 04:03 |
| *** banix has quit IRC | 04:04 | |
| *** dwalsh has quit IRC | 04:12 | |
| *** sdake has quit IRC | 04:14 | |
| *** chip_ has joined #openstack-kolla | 04:16 | |
| *** dwalsh has joined #openstack-kolla | 04:17 | |
| *** chip_ has quit IRC | 04:22 | |
| *** chip_ has joined #openstack-kolla | 04:23 | |
| *** sdake has joined #openstack-kolla | 04:24 | |
| *** chip_ has quit IRC | 04:25 | |
| *** chip_ has joined #openstack-kolla | 04:25 | |
| *** fragatina has joined #openstack-kolla | 04:27 | |
| sdake | hey zhubingbing | 04:30 |
| zhubingbing | sdake， i have a probleme ，can you help me | 04:31 |
| sdake | shoot | 04:31 |
| *** fragatina has quit IRC | 04:31 | |
| zhubingbing | ADD openstack-base-archive /openstack-base-source | 04:31 |
| zhubingbing | I didn't see the openstack-package | 04:32 |
| sdake | for which? | 04:32 |
| sdake | it gets created by the build tools | 04:32 |
| sdake | for which contianer | 04:32 |
| sdake | openstack-base-archive is created by the mapping file for container versions | 04:32 |
| zhubingbing | o，build tools | 04:32 |
| zhubingbing | -i know | 04:32 |
| zhubingbing | I went to see | 04:33 |
| *** unicell has joined #openstack-kolla | 04:34 | |
| zhubingbing | i find it ，SOURCES = { | 04:36 |
| zhubingbing | 'openstack-base': | 04:36 |
| sdake | right | 04:36 |
| zhubingbing | in common/config.py | 04:37 |
| zhubingbing | thinks | 04:37 |
| *** dwalsh has quit IRC | 04:38 | |
| sdake | thinks or thanks? | 04:38 |
| *** unicell has quit IRC | 04:39 | |
| *** daneyon has joined #openstack-kolla | 04:39 | |
| zhubingbing | i'm fighting for tomorrow to finish Barbican and container | 04:39 |
| zhubingbing | sadake,i think monitoring is important, after all, I want to join zabbix in this, how do you feel | 04:40 |
| zhubingbing | thanks --- | 04:40 |
| zhubingbing | I often call the wrong word | 04:41 |
| sdake | atm we have a tig stack in the queue | 04:41 |
| sdake | are you familiar with tig? | 04:41 |
| zhubingbing | yes | 04:42 |
| zhubingbing | I'm quite familiar with zabbix | 04:42 |
| zhubingbing | I have done zabbix in this area of practice. | 04:43 |
| sdake | how is zabbix different from tig | 04:43 |
| zhubingbing | and magnum i think I can fix it. | 04:44 |
| *** daneyon has quit IRC | 04:44 | |
| zhubingbing | different? | 04:45 |
| zhubingbing | I think zabbix has a lot of enterprises in use | 04:46 |
| zhubingbing | sorry , I'm a little bit tig stack ,i don't know what it means. | 04:47 |
| zhubingbing | - - - | 04:47 |
| sdake | tig = telegraph/influxdb/grafana | 04:48 |
| zhubingbing | Oh, I see | 04:48 |
| zhubingbing | thanks | 04:49 |
| sdake | our current thinking is we are not limiting choice on monitoring tools | 04:51 |
| *** yee_ has joined #openstack-kolla | 04:52 | |
| zhubingbing | i also think so, i want kolla to be able to support more monitoring tools | 04:52 |
| *** yee_ has quit IRC | 04:55 | |
| *** dwalsh has joined #openstack-kolla | 04:57 | |
| sdake | what i'd like to understand is how zabbix integreates or odesn't integreate with tig | 04:59 |
| *** unicell has joined #openstack-kolla | 04:59 | |
| *** chip_ has quit IRC | 05:02 | |
| sdake | could you answer i fyou know or indicate you dont know? | 05:02 |
| zhubingbing | I don't know much about for tig | 05:02 |
| zhubingbing | I dont' know -- | 05:03 |
| sdake | cool thanks | 05:03 |
| zhubingbing | Grafana also support zabbix | 05:06 |
| sdake | support in what way | 05:06 |
| sdake | zabbix layer below or layer above? | 05:06 |
| zhubingbing | layer | 05:08 |
| zhubingbing | https://github.com/alexanderzobnin/grafana-zabbix | 05:08 |
| zhubingbing | link | 05:08 |
| *** dwalsh has quit IRC | 05:13 | |
| *** dwalsh has joined #openstack-kolla | 05:14 | |
| sdake | i'd prefer not to cross-integrate anything for newton | 05:20 |
| sdake | i think it just makes thigns more complicated and risks tig at minimum not working | 05:21 |
| sdake | or zabbix for that matter | 05:21 |
| zhubingbing | ok , i agree | 05:24 |
| zhubingbing | sdake, i changed it https://review.openstack.org/#/c/351027/7 | 05:27 |
| patchbot | zhubingbing: patch 351027 - kolla - Add aodh role | 05:27 |
| *** egonzalez has joined #openstack-kolla | 05:29 | |
| *** egonzalez has quit IRC | 05:29 | |
| sdake | Jeffrey4l around? | 05:36 |
| Jeffrey4l | sdake, yep | 05:37 |
| sdake | any word on ubuntu gate? | 05:37 |
| sdake | Jeffrey4l - note the bare metal is deployed in the osic cluster | 05:37 |
| sdake | what isn't deployed is the dependencies | 05:38 |
| sdake | are the dependnecies that is | 05:38 |
| Jeffrey4l | sdake, no :( i can not even reproduce it locally. | 05:40 |
| Jeffrey4l | that's cool | 05:40 |
| Jeffrey4l | sdake, ubuntu gate issue? | 05:40 |
| Jeffrey4l | about the osic, how the usage time is assign to every one? | 05:40 |
| sdake | Jeffrey4l we haven't sorted that out yet | 05:40 |
| sdake | Jeffrey4l inc0 floated the idea of geo coordinators | 05:41 |
| sdake | so, apac, us, emea are geos | 05:41 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla: Customizations for magnum https://review.openstack.org/351492 | 05:41 |
| sdake | and someone from eachgeo would be responsible for coordinating activities for our plan for the next 3 weeks | 05:41 |
| Jeffrey4l | yes. that's will be high efficiency | 05:42 |
| *** yee_ has joined #openstack-kolla | 05:43 | |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla: Customizations for magnum https://review.openstack.org/351492 | 05:44 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla: Customizations for magnum https://review.openstack.org/351492 | 05:45 |
| sdake | Jeffrey4l can you review https://review.openstack.org/#/c/347030/3 | 05:47 |
| patchbot | sdake: patch 347030 - kolla - Customizations for keepalived | 05:47 |
| *** yee_ has quit IRC | 05:48 | |
| Jeffrey4l | done | 05:49 |
| sdake | Jeffrey4l also review that last customization patch if you wouldn't mind but don't workflow until gate passes | 05:49 |
| sdake | Jeffrey4l our review queue is 8 pages long | 05:49 |
| *** yee_ has joined #openstack-kolla | 05:49 | |
| sdake | we need to make progress on reducing its length | 05:49 |
| sdake | where i could really use your help the most is anything that is starred | 05:50 |
| Jeffrey4l | no problem | 05:50 |
| sdake | also customizations since those are easy reviews | 05:50 |
| sdake | the starred ones are highly complex | 05:51 |
| sdake | for thte most part mostof them look pretty good to me | 05:51 |
| Jeffrey4l | \i have PSs ready for review. but some of them haven't a +1 | 05:52 |
| sdake | i reviewed about 40 patches tonight | 05:52 |
| sdake | but review fatigue set in | 05:52 |
| sdake | Jeffrey4l if you need some reviews link em and ill tak a look | 05:53 |
| Jeffrey4l | complex PS consume energy.. | 05:53 |
| sdake | right | 05:53 |
| Jeffrey4l | sdake, https://review.openstack.org/349197 | 05:54 |
| openstackgerrit | Merged openstack/kolla: Customizations for keepalived https://review.openstack.org/347030 | 05:54 |
| Jeffrey4l | sdake, https://review.openstack.org/348382 | 05:54 |
| Jeffrey4l | sdake, https://review.openstack.org/348432 | 05:54 |
| Jeffrey4l | there three. not that complex :) | 05:55 |
| sdake | any more? | 06:02 |
| sdake | that last one should probably have a release note on second thought | 06:02 |
| sdake | so going to -1 it | 06:02 |
| *** yee_ has quit IRC | 06:04 | |
| sdake | Jeffrey4l this one would be nice to merge - if its correct - so we can easily set up our deps on the osic gear | 06:04 |
| Jeffrey4l | got. thanks. | 06:05 |
| *** yee_ has joined #openstack-kolla | 06:05 | |
| Jeffrey4l | others is more complicate. | 06:05 |
| Jeffrey4l | https://review.openstack.org/346400 | 06:05 |
| Jeffrey4l | https://review.openstack.org/347606 | 06:05 |
| Jeffrey4l | https://review.openstack.org/346444 | 06:06 |
| Jeffrey4l | https://review.openstack.org/345751 | 06:06 |
| Jeffrey4l | https://review.openstack.org/347457 | 06:06 |
| Jeffrey4l | https://review.openstack.org/349764 this one is simple, without it openstack can not support live-migration. but it may be danger to bind vnc to all interface | 06:08 |
| *** yee_ has quit IRC | 06:08 | |
| *** hieulq_ has joined #openstack-kolla | 06:08 | |
| *** yee_ has joined #openstack-kolla | 06:08 | |
| Jeffrey4l | sdake, btw, what tool are we used to install the osic machine? bifrost? or? | 06:11 |
| sdake | at present it was done via cobbler | 06:11 |
| sdake | we are reserving the last week of testing for bifrost scale testing - hope to have bifrost mrged by then | 06:12 |
| Jeffrey4l | roger | 06:12 |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the kolla_docker issue with docker 1.12 https://review.openstack.org/351498 | 06:12 |
| sdake | if you see a starred review with -1 on it already pleasae also review it again for correctness | 06:12 |
| sdake | we need to beat bifrost code into mergable shape in next couple of weeks | 06:13 |
| Jeffrey4l | OK. | 06:13 |
| sdake | more eyes = better | 06:13 |
| sdake | one reviewer has hard time picking out all problems | 06:13 |
| Jeffrey4l | yes. I will test it recently. | 06:14 |
| *** coolsvap has joined #openstack-kolla | 06:14 | |
| sdake | hey coolsvap | 06:15 |
| sdake | coolsvap review queue is long - need reviews if possible | 06:16 |
| sdake | especially starred reviews - if you see a starred review with a -1 on it, please also review it again and see if you can pick out any other problems | 06:16 |
| sdake | coolsvap also osic cluster is deployed | 06:18 |
| sdake | coolsvap in terms of the bare metal - but no dependencies and no kolla | 06:18 |
| coolsvap | sdake: sure will do reviews | 06:19 |
| coolsvap | my laptop was busted yesterday had to reconfigure everything | 06:19 |
| coolsvap | jsut getting started | 06:19 |
| sdake | ya both my laptops are damaged | 06:21 |
| sdake | one has a completely tanked keyboard | 06:21 |
| sdake | other has a screen breakage | 06:21 |
| sdake | my productivity is not high as a result :( | 06:21 |
| sdake | Jeffrey4l http://logs.openstack.org/27/351027/7/check/gate-kolla-dsvm-deploy-centos-binary/679bed3/console.html#_2016-08-05_05_46_49_140357 | 06:24 |
| *** yee_ has quit IRC | 06:24 | |
| sdake | Jeffrey4l i recently merged a patch which enabled external elasticsearch - it was passing the gate iirc | 06:24 |
| sdake | but it is possible it wasn't | 06:24 |
| Jeffrey4l | i am checking it. | 06:25 |
| sdake | if our gate is bust we need to fix it asap | 06:25 |
| sdake | easiest solution is a revert if that is the root cause | 06:25 |
| sdake | let me see if i can find the patch which may have caused it moment | 06:25 |
| sdake | https://review.openstack.org/#/c/320012/ | 06:26 |
| patchbot | sdake: patch 320012 - kolla - Enable central logging without deploying elasticse... (MERGED) | 06:26 |
| sdake | guess it wsn't passing the gate | 06:26 |
| sdake | guess i need to not be merging patches on fumes | 06:26 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add barbican dockerfile https://review.openstack.org/351506 | 06:26 |
| *** yee_ has joined #openstack-kolla | 06:26 | |
| Jeffrey4l | OK. let me take a look. | 06:27 |
| *** daneyon has joined #openstack-kolla | 06:27 | |
| openstackgerrit | Steven Dake proposed openstack/kolla: Revert "Enable central logging without deploying elasticsearch/kibana" https://review.openstack.org/351507 | 06:28 |
| *** phuongnh has quit IRC | 06:29 | |
| Jeffrey4l | sdake, i think "{{ 'yes' if enable_central_logging | bool else 'no' }}" may fix this. | 06:30 |
| *** zhubingbing_ has joined #openstack-kolla | 06:30 | |
| sdake | ok well its your idea - can you fix it and try - we can merge if that works - or revert if not | 06:30 |
| *** phuongnh has joined #openstack-kolla | 06:30 | |
| sdake | need to do this now | 06:30 |
| sdake | one or the other | 06:30 |
| sdake | gate breakage = everything grinds to halt | 06:30 |
| *** daneyon has quit IRC | 06:31 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add aodh role https://review.openstack.org/351027 | 06:32 |
| *** zhubingbing has quit IRC | 06:33 | |
| *** mewald has joined #openstack-kolla | 06:38 | |
| *** dmueller-evoila has joined #openstack-kolla | 06:41 | |
| *** yee_ has quit IRC | 06:43 | |
| *** yee_ has joined #openstack-kolla | 06:43 | |
| *** zhubingbing__ has joined #openstack-kolla | 06:47 | |
| *** zhubingbing_ has quit IRC | 06:50 | |
| *** duonghq has joined #openstack-kolla | 06:52 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix the evaluating conditional for enable es and kibana https://review.openstack.org/351518 | 06:54 |
| *** Mr_Broken has joined #openstack-kolla | 06:55 | |
| *** mnasiadka has joined #openstack-kolla | 06:59 | |
| *** dwalsh has quit IRC | 07:00 | |
| *** belmoreira has joined #openstack-kolla | 07:03 | |
| openstackgerrit | Merged openstack/kolla: Fix the kolla_docker issue with docker 1.12 https://review.openstack.org/351498 | 07:05 |
| *** phuongnh has quit IRC | 07:05 | |
| sdake | Jeffrey4l your patch is faliing the gate | 07:21 |
| *** b_bezak has joined #openstack-kolla | 07:22 | |
| *** zhubingbing__ has quit IRC | 07:22 | |
| Jeffrey4l | sdake, which one? | 07:22 |
| sdake | the one that just merged | 07:23 |
| sdake | http://logs.openstack.org/98/351498/1/check/gate-kolla-dsvm-deploy-centos-binary/98ce29d/console.html#_2016-08-05_06_44_32_858929 | 07:23 |
| openstackgerrit | Merged openstack/kolla: Correct mistral-executor image name https://review.openstack.org/350638 | 07:25 |
| sdake | that patch is definately the problem - see https://review.openstack.org/#/c/351507/ | 07:25 |
| patchbot | sdake: patch 351507 - kolla - Revert "Enable central logging without deploying e... | 07:25 |
| Jeffrey4l | hmm is it related? about the publicURL. | 07:26 |
| sdake | Jeffrey4l oh your patch to fix the ansible code hasn't gone through ci yet | 07:26 |
| sdake | my mistake | 07:26 |
| Jeffrey4l | this should come from a oslo.cache bug. | 07:27 |
| Jeffrey4l | but i remember i have a workaround fix. | 07:27 |
| Jeffrey4l | sdake, my fix is on the way. wait for minute. | 07:28 |
| sdake | https://review.openstack.org/351518 | 07:28 |
| sdake | this one? | 07:28 |
| Jeffrey4l | yes. | 07:29 |
| *** godleon has quit IRC | 07:29 | |
| *** gfidente has joined #openstack-kolla | 07:32 | |
| *** flaper87 has joined #openstack-kolla | 07:36 | |
| *** flaper87 is now known as Guest73548 | 07:36 | |
| sdake | Jeffrey4l that syntax on that ansible is funky | 07:40 |
| sdake | i guess it was only tested in teh "yes" mode | 07:40 |
| *** Mr_Broken has quit IRC | 07:44 | |
| Jeffrey4l | `else` should be used | 07:44 |
| *** Mr_Broken has joined #openstack-kolla | 07:44 | |
| Jeffrey4l | checked from zuul. seems my fix works too. | 07:45 |
| sdake | coolsvap I'll +2 it and head to bed - if it passes the gate feel free to workflow it | 07:46 |
| coolsvap | Sure | 07:48 |
| sdake | you have else in the code? | 07:48 |
| sdake | oh sorry ws reading wrong part of screen :) | 07:48 |
| *** shardy has joined #openstack-kolla | 07:55 | |
| *** openstackgerrit has quit IRC | 08:03 | |
| *** openstackgerrit has joined #openstack-kolla | 08:03 | |
| Jeffrey4l | coolsvap, it green again. https://review.openstack.org/351518 | 08:03 |
| coolsvap | Jeffrey4l: yeah sorr ywas away for lunch | 08:05 |
| Jeffrey4l | thanks :) | 08:08 |
| *** shardy has quit IRC | 08:14 | |
| *** shardy has joined #openstack-kolla | 08:14 | |
| *** yee_ has quit IRC | 08:15 | |
| *** daneyon has joined #openstack-kolla | 08:15 | |
| *** daneyon has quit IRC | 08:20 | |
| *** yee_ has joined #openstack-kolla | 08:20 | |
| *** yee_ has quit IRC | 08:22 | |
| *** yee_ has joined #openstack-kolla | 08:25 | |
| *** yee_ has quit IRC | 08:27 | |
| *** yee_ has joined #openstack-kolla | 08:30 | |
| *** yee_ has quit IRC | 08:34 | |
| *** yee_ has joined #openstack-kolla | 08:35 | |
| openstackgerrit | Merged openstack/kolla: Volume specification for ironic-api was invalid. https://review.openstack.org/351208 | 08:37 |
| *** sdake has quit IRC | 08:37 | |
| openstackgerrit | Merged openstack/kolla: Customizations for iscsid https://review.openstack.org/351319 | 08:38 |
| *** Guest73548 is now known as flaper87 | 08:39 | |
| *** flaper87 is now known as Guest78641 | 08:39 | |
| *** wrouesnel1 has joined #openstack-kolla | 08:41 | |
| *** yee_ has quit IRC | 08:42 | |
| *** yee_ has joined #openstack-kolla | 08:43 | |
| *** Guest78641 has quit IRC | 08:44 | |
| *** wrouesnel1 has quit IRC | 08:52 | |
| *** kbaegis has joined #openstack-kolla | 08:58 | |
| openstackgerrit | Merged openstack/kolla: Added influxdb container https://review.openstack.org/346446 | 08:59 |
| kbaegis | I keep having kolla-build fail | 09:01 |
| kbaegis | My host is telling me that I can't access a public url. For Centos and ubuntu both is percona, | 09:02 |
| kbaegis | Can't access: INFO:kolla.image.build.base:Cannot open: http://repo.percona.com/release/7/RPMS/x86_64/percona-release-0.1-3.noarch.rpm. Skipping | 09:02 |
| kbaegis | Anyone seen this issue before? I can curl the url from the docker host | 09:02 |
| kbaegis | Is this just some dumb networking setup I'm not familiar with w/ docker? | 09:02 |
| openstackgerrit | Hiroki Ito proposed openstack/kolla: Prechecks fails when using multinode deploy using a single node and haproxy disabled https://review.openstack.org/351588 | 09:05 |
| pbourke | kbaegis: are you building behind a proxy? | 09:11 |
| *** hieulq has quit IRC | 09:12 | |
| kbaegis | pbourke I'm building on a linux box one hop from my edge | 09:19 |
| *** anbehl has joined #openstack-kolla | 09:19 | |
| kbaegis | No proxy | 09:19 |
| kbaegis | (My 1st kolla build) :) | 09:19 |
| kbaegis | Most of them succeeded in my ubuntu vm | 09:20 |
| openstackgerrit | Merged openstack/kolla: Add doc8 test and improve rst syntax https://review.openstack.org/351173 | 09:20 |
| openstackgerrit | Merged openstack/kolla: Added neutron-metering-agent https://review.openstack.org/346629 | 09:20 |
| kbaegis | I can make whatever networking changes are necessary. The docker default subnet isn't routed by my edge. It's pulling 172.17.0.0/24 addresses on all hosts | 09:21 |
| kbaegis | oh | 09:21 |
| kbaegis | Taking a closer look that's 172.17/16 | 09:21 |
| *** Mr_Broken has quit IRC | 09:21 | |
| kbaegis | Still, that's taken on the edge | 09:21 |
| kbaegis | <— obviously not as familiar with docker networking | 09:22 |
| *** Mr_Broken has joined #openstack-kolla | 09:22 | |
| kbaegis | can I get kolla to use a different range of IPs? Setting a prefix on my -p (ex. 192.168.20.30:4000:5000) is what gets my registry network accessiblity | 09:23 |
| kbaegis | I'll keep reading :) | 09:28 |
| *** zhiwei has quit IRC | 09:31 | |
| *** flaper87 has joined #openstack-kolla | 09:40 | |
| *** flaper87 has quit IRC | 09:40 | |
| *** flaper87 has joined #openstack-kolla | 09:40 | |
| *** yee_ has quit IRC | 09:50 | |
| kbaegis | resolved. I needed —bip appended to the daemon on startup and a route out | 09:56 |
| *** daneyon has joined #openstack-kolla | 10:04 | |
| *** daneyon has quit IRC | 10:08 | |
| *** duonghq has left #openstack-kolla | 10:09 | |
| *** zhurong_ has quit IRC | 10:10 | |
| *** hieulq_ has quit IRC | 10:11 | |
| *** Mr_Broken has quit IRC | 10:25 | |
| *** kbaegis has quit IRC | 10:42 | |
| *** dasTor has quit IRC | 10:43 | |
| *** kbaegis has joined #openstack-kolla | 10:49 | |
| *** yee_ has joined #openstack-kolla | 10:50 | |
| *** yee_ has quit IRC | 10:54 | |
| *** rhallisey has joined #openstack-kolla | 11:07 | |
| *** Mr_Broken has joined #openstack-kolla | 11:09 | |
| *** ad_rien_ has joined #openstack-kolla | 11:19 | |
| ad_rien_ | HI all, | 11:19 |
| *** mliima has joined #openstack-kolla | 11:20 | |
| ad_rien_ | I'm wondering whether it is possible to deploy several compute containers per host without using the fake driver. In other words, I'm trying to understand why putting several times the same hostname in the multinodes inventory (in the compute group) does not enable ansible to create several instances. | 11:21 |
| kbaegis | ad_rien_ once nova is up, it should be able to create a large number of guests | 11:23 |
| kbaegis | ad_rien_ you're trying to artificially simulate having a large number of chassis or something? | 11:24 |
| ad_rien_ | yepp | 11:24 |
| kbaegis | Well, I don't work with the kolla team. I'm just an openstack guy | 11:24 |
| kbaegis | But I'd say the easiest way is baremetal->vm->docker | 11:24 |
| kbaegis | and then use vms to compartmentalize | 11:24 |
| kbaegis | docker just uses cgroups | 11:25 |
| kbaegis | much more lightweight, but no dedicated resource compartments | 11:25 |
| ad_rien_ | we thought about such an approach (using vagrant) but we would like to remove the virtualization layer now | 11:25 |
| kbaegis | Well, using kolla one libvirt driver should do | 11:26 |
| kbaegis | You only need to talk to the baremetal lxc | 11:26 |
| kbaegis | So you don't need "multiple libvirts" or anything | 11:26 |
| kbaegis | You should be able to spawn tons of instances | 11:26 |
| kbaegis | cpu_overcommit I think is still 16:1 | 11:26 |
| openstackgerrit | Merged openstack/kolla: Customizations for haproxy https://review.openstack.org/348390 | 11:27 |
| kbaegis | Oh, it's not 16:1 by default :) | 11:28 |
| *** zhurong has joined #openstack-kolla | 11:32 | |
| openstackgerrit | Merged openstack/kolla: Customizations for gnocchi https://review.openstack.org/346300 | 11:32 |
| openstackgerrit | Merged openstack/kolla: Customizations for magnum https://review.openstack.org/351492 | 11:32 |
| openstackgerrit | Merged openstack/kolla: Customizations for Horizon https://review.openstack.org/350476 | 11:32 |
| kbaegis | Yay! | 11:33 |
| *** kbaegis has quit IRC | 11:36 | |
| *** kbaegis has joined #openstack-kolla | 11:36 | |
| *** kbaegis has quit IRC | 11:41 | |
| *** kbaegis has joined #openstack-kolla | 11:42 | |
| *** kbaegis1 has joined #openstack-kolla | 11:43 | |
| *** kbaegis has quit IRC | 11:43 | |
| *** yee_ has joined #openstack-kolla | 11:51 | |
| *** daneyon has joined #openstack-kolla | 11:52 | |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Customizations for Nova https://review.openstack.org/351676 | 11:53 |
| *** daneyon has quit IRC | 11:56 | |
| sean-k-mooney | o/ | 11:57 |
| sean-k-mooney | hello everyone | 11:57 |
| *** dave-mccowan has joined #openstack-kolla | 12:08 | |
| *** mkoderer_ has joined #openstack-kolla | 12:15 | |
| *** lmiccini_ has joined #openstack-kolla | 12:17 | |
| *** lmiccini has quit IRC | 12:17 | |
| *** mkoderer has quit IRC | 12:18 | |
| openstackgerrit | zhubingbing proposed openstack/kolla: repair gnocchi image dockerfile https://review.openstack.org/348388 | 12:19 |
| *** yee_ has quit IRC | 12:22 | |
| *** JoseMello has joined #openstack-kolla | 12:35 | |
| *** egonzalez has joined #openstack-kolla | 12:37 | |
| *** lulei has quit IRC | 12:37 | |
| *** inc0 has joined #openstack-kolla | 12:42 | |
| inc0 | good morning | 12:42 |
| inc0 | pbourke, we have deployed bare metal:) | 12:43 |
| *** kbaegis1 has quit IRC | 12:45 | |
| *** kbaegis has joined #openstack-kolla | 12:46 | |
| pbourke | hey inc0 | 12:50 |
| pbourke | yeah been chatting with britthouser | 12:50 |
| pbourke | have deployed ssh keys to all, so ready for ansible | 12:51 |
| pbourke | inc0: currently want to set up bonds, could you check my config when you have a chance? (open on the tmux) | 12:51 |
| inc0 | omw | 12:51 |
| pbourke | basically want to try it on one node first before doing the rest | 12:51 |
| pbourke | also unsure if I need to change routes | 12:52 |
| pbourke | inc0: britthouser already spotted some missing stuff | 12:54 |
| pbourke | britthouser: can you update it with what you think? | 12:55 |
| britthouser | yup | 12:55 |
| inc0 | what was deployment node ip? | 12:56 |
| britthouser | 8.253 | 12:56 |
| inc0 | first you need bond- | 12:57 |
| inc0 | bond0 | 12:57 |
| inc0 | then you can add bond0.702 | 12:57 |
| britthouser | like that? | 12:57 |
| britthouser | ah tx | 12:59 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add gnocchi ansible role https://review.openstack.org/349351 | 12:59 |
| inc0 | ok it looks good | 12:59 |
| inc0 | to me | 12:59 |
| sean-k-mooney | inc0: speaking of bonds i need to add a patch to allow configuing networking interfaces as part of kolla-host to supprot ovs-dpdk | 13:00 |
| sean-k-mooney | inc0: if ye keep a list of manual steps ye need to do i can add support to the kolla-host playbook for them | 13:00 |
| inc0 | that's gonna be funky as network setup might differ | 13:00 |
| inc0 | I'd rather setup some requrements | 13:01 |
| inc0 | like "make a bond and put a name into globals" | 13:01 |
| inc0 | you need to restart networking britthouser pb | 13:01 |
| inc0 | pbourke, | 13:02 |
| pbourke | soo | 13:02 |
| pbourke | the node is no longer on the pxe vlan? | 13:02 |
| sean-k-mooney | ya well we support bonding in our devstack pluging but yes the code is not simple. | 13:02 |
| britthouser | yeah pbourke - we removed that interface altogether. | 13:02 |
| britthouser | kinda why I wanted todo this via cobbler....=/ | 13:02 |
| britthouser | if we add that interface back in, we'll have two default routes | 13:03 |
| britthouser | would like to add it in with DHCP but say, "ignore dhcp default route" | 13:03 |
| britthouser | lemme google and see if there is ubuntu network config to override DHCP default route | 13:03 |
| sean-k-mooney | inc0: we do it with two arrays of pairs. bond->bridge mappins and port>bond mappings | 13:04 |
| pbourke | nice | 13:04 |
| britthouser | that "Partner Mac Address:" is what lets you know its talking to the switch | 13:04 |
| inc0 | britthouser, we didnt | 13:05 |
| inc0 | we made bond on p1p2 and p4p2 | 13:05 |
| pbourke | inc0: britthouser: I can ansible the same onto the other nodes | 13:05 |
| inc0 | pxe goes over p1p1 | 13:05 |
| pbourke | if that's good with you | 13:05 |
| inc0 | that's perfect;) | 13:05 |
| inc0 | you can use ip from p1p1 to get last octet for 702 | 13:05 |
| britthouser | so inc0 you're saying dont' movet he default route? | 13:05 |
| inc0 | do we need to really? | 13:05 |
| britthouser | What traffic would use the mgmt network if we don;t? | 13:06 |
| inc0 | mariadb, rabbit, APIs | 13:06 |
| inc0 | that sort of stuff | 13:06 |
| inc0 | we'll use it as kolla_internal | 13:06 |
| britthouser | yeah lets discuss all the networks we'll need | 13:07 |
| britthouser | so you're saying kolla_internal=bond0.702 | 13:07 |
| sean-k-mooney | are ye going to share networks or keep them seperate per function ? | 13:07 |
| inc0 | and neutron_interface = bond0 | 13:07 |
| inc0 | sean-k-mooney, they'll be separated as 702 is vlan | 13:08 |
| inc0 | on that note | 13:08 |
| britthouser | The way I read the switch configs, all the bond0 IPs need to be in a VLAN | 13:08 |
| inc0 | I wonder if neutron will like tagged interface from same bond | 13:08 |
| britthouser | for VXLAN it should be fine. | 13:08 |
| sean-k-mooney | inc0: if you use vxlan it wont care | 13:09 |
| inc0 | maybe we should use p4p2.702 for mgmt? | 13:09 |
| inc0 | ok, we'll see soon enough;) | 13:09 |
| britthouser | So neutron_interface=bond0.730 ? | 13:09 |
| britthouser | kolla_internal=bond0.702 | 13:09 |
| inc0 | no, bond0 | 13:09 |
| inc0 | or... | 13:10 |
| inc0 | yeah 730 will be better | 13:10 |
| britthouser | Pretty sure you won't get any packets passed on just bond0 w/o a tag | 13:10 |
| inc0 | but all traffic will go throush 702 anyway | 13:10 |
| inc0 | yay for vxlan | 13:10 |
| inc0 | britthouser, but neutron knows how to vlan if we setup provisioner: vlan | 13:11 |
| britthouser | I'm adding notes from here to etherpad | 13:11 |
| britthouser | is that on our list inc0? I can't remember | 13:11 |
| sean-k-mooney | by the way would it be good to test using a different interface for each ot the interface types | 13:11 |
| sean-k-mooney | #kolla_external_vip_interface: "{{ network_interface }}" #api_interface: "{{ network_interface }}" #storage_interface: "{{ network_interface }}" #cluster_interface: "{{ network_interface }}" #tunnel_interface: "{{ network_interface }}" | 13:11 |
| britthouser | if we're gonna do VLAN tenant networks, we should use p4p1 dedicated for it. | 13:11 |
| *** sdake has joined #openstack-kolla | 13:11 | |
| britthouser | thx for that list sean-k-mooney | 13:12 |
| openstackgerrit | zhubingbing proposed openstack/kolla: repair aodh image dockerfile https://review.openstack.org/350715 | 13:12 |
| sdake | hey cats | 13:12 |
| *** jtriley has joined #openstack-kolla | 13:12 | |
| britthouser | hey sdake | 13:12 |
| inc0 | gmorning sdake | 13:12 |
| sdake | mornign britthouser inc0 | 13:13 |
| britthouser | We're discussing host networking layout | 13:13 |
| britthouser | Right now we have p1p1=pxe network and also our default route | 13:13 |
| sean-k-mooney | in the ceph case i think it makes sense to test a dedicated storage network. | 13:13 |
| sdake | inc0 need me to kick off a geo coordinators thread on ml? | 13:13 |
| sean-k-mooney | and hi sdake good morning | 13:13 |
| britthouser | bond0 of p1p2 and p4p2 being used for kolla_internal on Vlan702 | 13:13 |
| sdake | hey sean-k-mooney :) | 13:13 |
| britthouser | need to to decide what else should be broken out. | 13:14 |
| *** egonzalez90 has joined #openstack-kolla | 13:14 | |
| *** mkoderer has joined #openstack-kolla | 13:14 | |
| *** zhubingbing__ has joined #openstack-kolla | 13:14 | |
| britthouser | Do we need to test both VLAN and VXLAN? | 13:14 |
| britthouser | I see LB and OVS being called out as permutations | 13:15 |
| britthouser | but if we can keep to just VXLAN, then we can make p4p1 our storage cluster network | 13:15 |
| zhubingbing__ | can i help me | 13:15 |
| *** lmiccini has joined #openstack-kolla | 13:15 | |
| britthouser | and the do the rest on VLANs on the bond | 13:16 |
| britthouser | thoughts? | 13:16 |
| inc0 | for ceph we'll need 2 networks really | 13:16 |
| sdake | zhubingbing__ of course you can hep yourself :) | 13:16 |
| sean-k-mooney | britthouser: i think that would work fine. i dont thing vlan tenent networks buy us alot | 13:16 |
| *** egonzalez has quit IRC | 13:16 | |
| inc0 | one for replication and one for storage | 13:16 |
| inc0 | gonna be interesting | 13:16 |
| sdake | zhubingbing__ what do you need :) | 13:16 |
| sean-k-mooney | britthouser: if we want to test tenant provider networks we could do it with a flat network | 13:17 |
| *** mkoderer_ has quit IRC | 13:17 | |
| mewald | britthouser: I dont really like pxe boot NIC to be the default route. That would mean, we need to put all nodes together into some internal api network. That means running multiple different host types in the same network within the same scalability boundaries. In our installations, we usually give each host group (control, network. compute) it's own api network and share only very few where absolutely necessary. | 13:17 |
| britthouser | So inc0 you're saying bond0.731 shoudl be storage (clients) and p4p1 should be storage (replications? | 13:17 |
| zhubingbing__ | {"failed": true, "msg": "The conditional check 'enable_elasticsearch | bool' failed. The error was: error while evaluating conditional (enable_elasticsearch | bool): {{ 'yes' if enable_central_logging | bool }}: the inline if-expression on line 1 evaluated to false and no else section was defined. i don't know | 13:18 |
| zhubingbing__ | http://logs.openstack.org/27/351027/8/check/gate-kolla-dsvm-deploy-ubuntu-source/9810828/console.html | 13:18 |
| zhubingbing__ | i | 13:18 |
| zhubingbing__ | I don't know why this is the wrong, I was good before | 13:18 |
| *** lmiccini_ has quit IRC | 13:18 | |
| *** egonzalez90 has quit IRC | 13:18 | |
| britthouser | yeah mewald I'd rather see the default route be on the bond0.702 management network. But the cobbler preseed given to us doesn't setup anything but the PXE interface. So we have todo the rest after the install. | 13:19 |
| sdake | zhubingbing__ you need to recheck | 13:19 |
| sdake | zhubingbing__ Jeffrey4l fixed that bug last night | 13:19 |
| britthouser | restarting networking from ansible gives me angina | 13:19 |
| *** yee_ has joined #openstack-kolla | 13:19 | |
| sean-k-mooney | britthouser: it might be easier to discuss the networkign with an etherpad | 13:19 |
| britthouser | https://etherpad.openstack.org/p/kolla-N-midcycle-osic | 13:19 |
| britthouser | scroll down to the "Working Notes" section | 13:20 |
| sdake | zhubingbing__ we merged a patch which was broken - it broke your ci results | 13:20 |
| sean-k-mooney | ah thanks britthouser | 13:20 |
| sean-k-mooney | so since p1p1 is our default route that should also be the interface we use for the external vip right? | 13:21 |
| britthouser | yeah | 13:21 |
| britthouser | unless there is a safe way to move the default route with ansible. | 13:22 |
| sbezverk | sdake: I got daemonset running by adding sleep 1 before ovs-db start. works 100% now.. | 13:22 |
| sean-k-mooney | i dont think we need to move the defualt route though it will only be used for internet faceing traffic | 13:23 |
| zhubingbing__ | This is my fault? what do I need to do? | 13:23 |
| inc0 | ok...so do we setup docker and stuff now? | 13:23 |
| pbourke | britthouser: our default route is not on the mgmt network | 13:23 |
| pbourke | is that a problem? | 13:23 |
| britthouser | I dno't think so | 13:23 |
| inc0 | me neither | 13:23 |
| britthouser | jsut require a little re-learning | 13:23 |
| britthouser | =) | 13:24 |
| pbourke | ok | 13:24 |
| sean-k-mooney | if you can reach the internet of the pxe network then all is good | 13:24 |
| pbourke | are we going to use sean-k-mooney's playbook or something else | 13:24 |
| inc0 | seans play is for ubuntu 16 | 13:24 |
| inc0 | but we can use most of it | 13:24 |
| inc0 | and maybe modify it a bit | 13:24 |
| britthouser | if you look a little higher in etherpad, lets lay out the networking | 13:24 |
| *** banix has joined #openstack-kolla | 13:24 | |
| sean-k-mooney | inc0: it might work on 14.04 also but i actully need to go test that now | 13:24 |
| britthouser | b/c I think we'll need to add some interfaces to your ansible pbourke | 13:25 |
| inc0 | well 14.04 doesnt use systemd | 13:25 |
| inc0 | so I doubt it will just fly | 13:25 |
| sdake | sbezverk yes i expected that outcome | 13:25 |
| britthouser | Any objections to making cluster_interface dedicated interface p4p1 ? | 13:25 |
| sean-k-mooney | ya i think im mising the insecure registry config for 14.04 and the storage driver change | 13:26 |
| sdake | sbezverk change sleep 1 to /bin/true pls | 13:26 |
| britthouser | and then api_interface can be same as mgmt | 13:26 |
| inc0 | yeah | 13:27 |
| inc0 | that sounds good | 13:27 |
| britthouser | storage_interface same as mgmt too? or its own VLAN? | 13:27 |
| inc0 | own vlan I'd say | 13:27 |
| britthouser | ok here is what I put in etherpad: | 13:27 |
| britthouser | kolla_internal: bond0.702 | 13:27 |
| britthouser | kolla_external_vip_interface: p1p1 (also pxe, default route) | 13:27 |
| britthouser | api_interface: bond0.702 | 13:27 |
| britthouser | storage_interface: bond0.731 | 13:27 |
| britthouser | cluster_interface: p4p1 | 13:27 |
| britthouser | tunnel_interface: bond0.730 | 13:27 |
| britthouser | +1 -1 ? | 13:28 |
| sean-k-mooney | what is the neutron_external_interface set too | 13:28 |
| *** Mr_Broken has quit IRC | 13:28 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Remove files from /var/lib/apt/lists when cleaning up on Ubuntu/Debian https://review.openstack.org/351738 | 13:29 |
| britthouser | taht wasn't in your previous list sean-k-mooney. =P | 13:29 |
| *** Mr_Broken has joined #openstack-kolla | 13:29 | |
| inc0 | seems legit | 13:29 |
| sean-k-mooney | :) correct neiter was neutron_interface | 13:29 |
| britthouser | lol | 13:29 |
| inc0 | I'll set up rest of vlans | 13:29 |
| zhubingbing__ | ok | 13:29 |
| sean-k-mooney | sorry not neutron_interface | 13:29 |
| sean-k-mooney | network_interface | 13:29 |
| zhubingbing__ | sdake,ok | 13:29 |
| britthouser | is that different than tunnel_interface? | 13:30 |
| britthouser | I thikn we're still missing external inc0 | 13:30 |
| britthouser | for FIP | 13:30 |
| sean-k-mooney | the network_interface i think is what you have called kolla_internal | 13:30 |
| inc0 | so, won't it create flat on bond0 | 13:31 |
| inc0 | ? | 13:31 |
| inc0 | or rather neutron interface? | 13:31 |
| sean-k-mooney | neutron_external_interface is for flat/vlan neutron networks | 13:31 |
| inc0 | britthouser, we need vlan at p4p1 too I think | 13:32 |
| sean-k-mooney | if we dont use vlan tenent networks then neutron_external_interface can be a vlan on bond0 | 13:32 |
| inc0 | as native would be pxe I'd guess | 13:32 |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Customizations for Cinder https://review.openstack.org/350587 | 13:32 |
| *** diogogmt has joined #openstack-kolla | 13:32 | |
| britthouser | yeah neutron_external_interface will be vlan on bond0 | 13:33 |
| britthouser | lets say 732? | 13:33 |
| *** Mr_Broken has quit IRC | 13:33 | |
| inc0 | btw how do you setup tunnel_network? | 13:33 |
| britthouser | Yeah we can make p4p1.733 for storage cluster? | 13:33 |
| inc0 | I'd go for bond0 as neutron external | 13:34 |
| inc0 | and setup fip network as a vlan via neutron | 13:34 |
| britthouser | I'm 95% sure untagged packets on bond0 will be dropped. | 13:34 |
| inc0 | we'll make it provisioner:vlan | 13:34 |
| inc0 | and let neutron tag it | 13:34 |
| britthouser | I thought we were doing vxlan? | 13:35 |
| inc0 | one doesn't rule out the other | 13:35 |
| britthouser | ...... | 13:35 |
| sdake | inc0 re geo coordinators - prefer i send a list to get that rolling - or will you send today? | 13:35 |
| sdake | send an email to list that is | 13:35 |
| britthouser | Lets stick with jsut vxlan here at the beginning | 13:35 |
| britthouser | if we want to get fancy with multiple varieties of neutron networks, then lets save that for later. | 13:36 |
| britthouser | yes/no? | 13:36 |
| inc0 | sdake, mail with geo-coordinators? | 13:36 |
| inc0 | ok | 13:36 |
| inc0 | britthouser, ok | 13:36 |
| inc0 | so I'll add bond0.732 | 13:36 |
| sdake | inc0 ya in last team meeting you had indicated you would send an email so we could establish who would be geo coordinators for our 3 geos (APAC, EMEA, and US) | 13:36 |
| sdake | that way we could use a follow-the-sun strategy for making use of osic cluster | 13:37 |
| pbourke | sdake: that's on the etherpad | 13:38 |
| sdake | geo coordinator would be responsible for knowing the current state of osic cluster | 13:38 |
| sean-k-mooney | britthouser: with vxlan for the teneant network you will still want the neutron external network to be a flat network that will use bond0.732 | 13:38 |
| sdake | oh - didn't know people had signed up htere - my bad | 13:38 |
| pbourke | or maybe we didn't designate one for each | 13:38 |
| sdake | i haven't looked at thte etherpad in last day | 13:38 |
| pbourke | no you're right | 13:38 |
| pbourke | we only have the people in each zone | 13:38 |
| inc0 | britthouser, check tmux and look if I miss something | 13:38 |
| sdake | inc0 if your busy i can tackle it - up to you :) | 13:39 |
| sdake | inc0 yes or no will work :) | 13:39 |
| inc0 | sdake, can you send it plz? | 13:39 |
| sdake | roger | 13:39 |
| inc0 | I'm ok with mail, I'm just trying to get this hardware rolling;) | 13:39 |
| sdake | pbourke mind linking etherpad? | 13:39 |
| sdake | inc0 right your busy | 13:39 |
| britthouser | sdake: https://etherpad.openstack.org/p/kolla-N-midcycle-osic | 13:40 |
| sdake | thanks folks | 13:40 |
| sdake | email inc | 13:40 |
| pbourke | inc0: makes sense to me but will wait for britthouser | 13:40 |
| britthouser | inc0 the cluster is not on the bond, right? | 13:40 |
| sean-k-mooney | inc0: britthouser is it possibly to make sure that vlan 732 is routed to the internet so that the tenent vms have internet access? | 13:40 |
| britthouser | It does sean-k-mooney | 13:41 |
| pbourke | sean-k-mooney: 732 has internet access | 13:41 |
| britthouser | at lest the docs say it does | 13:41 |
| inc0 | ah good catch - forgot to setup p4p1 | 13:41 |
| inc0 | sean-k-mooney, all of vlans have internet access | 13:41 |
| sean-k-mooney | ah cool that makes things easy | 13:42 |
| *** b_bezak has quit IRC | 13:42 | |
| *** b_bezak has joined #openstack-kolla | 13:42 | |
| *** b_bezak has quit IRC | 13:42 | |
| sean-k-mooney | thenant internet acess will be needed for the 1000 vms each running kolla test in week 3 | 13:42 |
| inc0 | yeah | 13:43 |
| inc0 | btw, that's going to be interesting | 13:43 |
| sean-k-mooney | interesting is one word to discribe it | 13:43 |
| inc0 | :D | 13:43 |
| inc0 | brb | 13:44 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Reducing disk footprint for Ubuntu/Debian images https://review.openstack.org/351744 | 13:44 |
| sean-k-mooney | yes actully ill be back in 20mins also | 13:44 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Reducing disk footprint for Ubuntu/Debian images https://review.openstack.org/351744 | 13:45 |
| britthouser | driving for a bit inc0 | 13:45 |
| *** b_bezak has joined #openstack-kolla | 13:47 | |
| sbezverk | sdake: it works with true as well, it looks like it needs an extra millisecond ;-) to finish container initialization | 13:47 |
| britthouser | ok pbourke and inc0 I htink this is good | 13:50 |
| pbourke | sweet | 13:50 |
| britthouser | how do I run this thing? | 13:50 |
| britthouser | thx! | 13:51 |
| britthouser | back in a few mins.... | 13:51 |
| inc0 | just had to make one small change | 13:51 |
| pbourke | look at it go | 13:52 |
| pbourke | beautiful | 13:52 |
| inc0 | :D | 13:52 |
| inc0 | sdake, so yesterday I watched full 1hr documentary on honey badgers | 13:52 |
| inc0 | it's on netflix | 13:52 |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Reducing disk footprint for Ubuntu/Debian images https://review.openstack.org/351744 | 13:53 |
| * inc0 approves | 13:53 | |
| sdake | inc0 sweet i'll definatley have to check that out tonight | 13:53 |
| sdake | recall title - or just search for honey badgers? | 13:53 |
| inc0 | yeah, just search honey badgers | 13:53 |
| pbourke | i tried watching that before | 13:53 |
| *** b_bezak has quit IRC | 13:55 | |
| *** b_bezak has joined #openstack-kolla | 13:55 | |
| *** yee_ has quit IRC | 13:56 | |
| inc0 | :( | 13:57 |
| inc0 | hmm | 13:58 |
| sdake | now on to emergency #4 for the day :) | 13:58 |
| inc0 | I might have fucked up our networking cluster-wide | 13:59 |
| inc0 | yay | 13:59 |
| sdake | sbezverk cool - goal is to get a backtrace - if that can't be done it can't be done | 13:59 |
| sbezverk | sdake I opened a bug with kube folks | 13:59 |
| sbezverk | got initial communication | 14:00 |
| sdake | sbezverk openvswitch folks would appreciate a backtrace i'm sure | 14:00 |
| *** banix has quit IRC | 14:00 | |
| sdake | sbezverk mind providing link or tldr? | 14:00 |
| sbezverk | https://github.com/kubernetes/kubernetes/issues/30106 | 14:01 |
| *** b_bezak has quit IRC | 14:02 | |
| *** openstackgerrit has quit IRC | 14:03 | |
| *** openstackgerrit has joined #openstack-kolla | 14:03 | |
| *** mliima_ has joined #openstack-kolla | 14:03 | |
| *** banix has joined #openstack-kolla | 14:04 | |
| inc0 | ok networking looks good | 14:05 |
| inc0 | pbourke, wanna try Sean's playbook | 14:05 |
| inc0 | ? | 14:05 |
| pbourke | inc0: yup | 14:05 |
| inc0 | do the honors | 14:05 |
| pbourke | is it merged/ | 14:05 |
| pbourke | or in a review | 14:05 |
| inc0 | I'll have to go to appointment soon so will leave you to it | 14:06 |
| inc0 | still in review | 14:06 |
| pbourke | grand | 14:06 |
| inc0 | let me check if we need to change much in it | 14:06 |
| inc0 | as it was desitned for 16.04 | 14:06 |
| inc0 | pbourke, https://review.openstack.org/#/c/325631/ | 14:06 |
| patchbot | inc0: patch 325631 - kolla - add baremetal role to install kolla deps. | 14:06 |
| pbourke | got it | 14:06 |
| pbourke | i'll have a look | 14:07 |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Customizations for Nova https://review.openstack.org/351676 | 14:07 |
| *** mliima has quit IRC | 14:07 | |
| britthouser | pbourke - back now. | 14:07 |
| pbourke | britthouser: cool, just having a look now at sean's patch above | 14:08 |
| britthouser | which nodes did we designate control/compute/ceph? I just wanted to check their partitioning | 14:08 |
| inc0 | britthouser, we don't run ceph yet I think | 14:08 |
| *** mkoderer has quit IRC | 14:08 | |
| inc0 | and most of them are running cinder image | 14:08 |
| inc0 | let me double check | 14:08 |
| pbourke | britthouser: oh yeah, that would be useful if you could partition the inventory? | 14:08 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add sahara ansible roles https://review.openstack.org/351294 | 14:09 |
| *** duonghq has joined #openstack-kolla | 14:09 | |
| britthouser | Oh....ok you marked it done, so I thought it was done | 14:09 |
| *** lmiccini has quit IRC | 14:09 | |
| *** ad_rien_ has quit IRC | 14:09 | |
| inc0 | britthouser, I think all of nodes runs generic now | 14:09 |
| britthouser | Yeah...we want to move compute to the cinder profile and storage....well I guess to the cinder profile as well and then later to the swift profile when we move to ceph/ | 14:10 |
| britthouser | Yeah...we want to move compute to the cinder profile and storage....well I guess to the cinder profile as well and then later to the swift profile when we move to ceph? | 14:10 |
| inc0 | yeah | 14:10 |
| inc0 | let's just deploy kolla now | 14:10 |
| inc0 | without storage | 14:10 |
| inc0 | and once we have all the plays written | 14:11 |
| inc0 | repaving will be a breeze | 14:11 |
| *** coolsvap has quit IRC | 14:11 | |
| britthouser | Ok so I'll pick three nodes to be control nodes | 14:12 |
| britthouser | and the rest I'll re-kickstart with the cinder profile | 14:12 |
| britthouser | sound good? | 14:12 |
| inc0 | let's try to deploy now | 14:12 |
| inc0 | ok? | 14:12 |
| inc0 | I mean without rekicking | 14:13 |
| inc0 | just basic no-cinder installation | 14:13 |
| *** kbaegis has quit IRC | 14:13 | |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Customizations for Nova https://review.openstack.org/351676 | 14:13 |
| britthouser | So right now everyone is deployed as generic, they have a single RAID1 array | 14:14 |
| inc0 | we can rekick them eod so we won't waste time on waiting for them boot | 14:14 |
| inc0 | yeah which will be ok for super basic scenerio | 14:14 |
| britthouser | Ok. While you do that, I'll work on changing up the input files for the EOD scenario | 14:14 |
| britthouser | Basically you're doing #3, right inc0? | 14:15 |
| britthouser | or step3 will be after EOD re-kickstart? | 14:15 |
| inc0 | it sill be semi-3 as no real storage | 14:16 |
| britthouser | Ok. Roll with it. =) | 14:16 |
| *** mkoderer has joined #openstack-kolla | 14:17 | |
| *** narasimha_SV has joined #openstack-kolla | 14:17 | |
| narasimha_SV | when deploying ironic using kolla, there is no specification for drivers needed to be used ? | 14:18 |
| narasimha_SV | is there a way that we can specify these things | 14:18 |
| britthouser | We should think bout backing up all the files we generated in the osic-prep container..... | 14:18 |
| *** lmiccini has joined #openstack-kolla | 14:19 | |
| sdake | narasimha_SV ironic as a service is not functional | 14:19 |
| inc0 | 8.254 will be keepalived-floating ok? | 14:20 |
| sdake | narasimha_SV its a work in progress - needs love - feel free to help as you see fit - we need help there :) | 14:20 |
| sdake | narasimha_SV where we are focusing at present related to ironic is bare metal deployment so kolla isn't half a solution but a full solution | 14:20 |
| *** mnasiadka has quit IRC | 14:20 | |
| britthouser | no inc0 | 14:21 |
| britthouser | that is the lxc container | 14:21 |
| sdake | narasimha_SV if you want to follow that work - look at owner: sean k mooney in gerrit | 14:21 |
| inc0 | I changed it to 200 | 14:21 |
| sdake | narasimha_SV work expected to land in newton | 14:21 |
| narasimha_SV | ok | 14:22 |
| sdake | narasimha_SV notice the stars on the reviews - that indicates its important for crs to review | 14:22 |
| narasimha_SV | sdake: sure . I am ready to start working on this. | 14:22 |
| sean-k-mooney | narasimha_SV: the ironic support almost works in kolla | 14:23 |
| sdake | narasimha_SV sweet the best way to help is to review those starred patches from sean to make sure they are correct | 14:23 |
| britthouser | actually I was wrong inc0 | 14:23 |
| sean-k-mooney | the issue is the pxe boot | 14:23 |
| britthouser | 0.254 is the lxc | 14:23 |
| britthouser | 8.254 could be the VIP | 14:23 |
| *** diogogmt has quit IRC | 14:23 | |
| inc0 | yeah, but let's not mix last octets | 14:23 |
| sean-k-mooney | narasimha_SV: when ironic pxe boots the nodes the node send the chassis uuid | 14:23 |
| inc0 | I like to have last octet same across the vlans | 14:23 |
| narasimha_SV | sdake: sure I will check them | 14:23 |
| sdake | sean-k-mooney do you mean t he bifrost work or hte bare metal as a service work | 14:23 |
| sdake | he/the | 14:24 |
| sdake | (fast with lots of mistakes;) | 14:24 |
| sean-k-mooney | narasimha_SV: the dhcp server is configured to respond to the neutron port uuid. | 14:24 |
| sean-k-mooney | sdake: ironic not bifrost | 14:24 |
| sdake | sean-k-mooney thanks for clarification | 14:24 |
| openstackgerrit | zhubingbing proposed openstack/kolla: fix sahara dockerfile https://review.openstack.org/351320 | 14:24 |
| sdake | it would be fantastic to have that work land in newton too - not sure who cn take it up atm | 14:24 |
| sdake | we only have about 3 weeks left in the cycle | 14:25 |
| britthouser | so your docker registry is same IP as LXC inc0 | 14:25 |
| inc0 | yeah, we can run registry in it | 14:25 |
| inc0 | why not | 14:25 |
| sean-k-mooney | sdake: we have been looking at it on and off over the cycle. have not had time to fix the ironic supprot properly | 14:25 |
| britthouser | Run a docker container inside the LXC? | 14:25 |
| sean-k-mooney | sdake: the pxe boot/uuid issue was the last one we think need to be fixed to get it running | 14:26 |
| sdake | sean-k-mooney roger- imo your working on the right thing at the right time | 14:26 |
| sdake | narasimha_SV if your a dev or ops interested in that functionality we could really use the help on it :) | 14:26 |
| sdake | sean-k-mooney is closest to the work - he has a thourough understanding of what is needed | 14:27 |
| inc0 | sean-k-mooney, soo...how do you add kolla_ssh_key to your playbook? | 14:27 |
| narasimha_SV | sdake: we are planning to implement ironic in our environment | 14:27 |
| *** zhurong has quit IRC | 14:27 | |
| narasimha_SV | we are using kolla for this to setup our env | 14:27 |
| sean-k-mooney | inc0: oh ya soory i manually copied the bifrost ones and renamed them locally in my passwords.yml | 14:28 |
| inc0 | gonna put that into review | 14:28 |
| sean-k-mooney | inc0: ill do that now | 14:28 |
| pbourke | inc0: already done ;) | 14:28 |
| inc0 | yeah saw it | 14:28 |
| inc0 | good catch | 14:28 |
| narasimha_SV | if that is not there then I will start looking at it so that I can push to the master | 14:29 |
| narasimha_SV | I just raised bug in kolla | 14:29 |
| sean-k-mooney | speaking of that i need to figure out a good way to test this also in the ci | 14:29 |
| *** Jeffrey4l has quit IRC | 14:29 | |
| *** Jeffrey4l has joined #openstack-kolla | 14:29 | |
| inc0 | sean-k-mooney, we're complementing you for tremendous job | 14:30 |
| inc0 | just ran this play, cool stuff! | 14:30 |
| narasimha_SV | https://bugs.launchpad.net/kolla/+bug/1610272 | 14:30 |
| openstack | Launchpad bug 1610272 in kolla "ironic.conf file missing enabled_drivers info" [Undecided,New] | 14:30 |
| sean-k-mooney | inc0: did it work other then the ssh key? | 14:31 |
| sdake | sean-k-mooney ci is for later - we hae a whole slew of ci work to do | 14:31 |
| *** michauds has joined #openstack-kolla | 14:31 | |
| inc0 | yeah | 14:32 |
| sean-k-mooney | sdake: yep. i think we can test it in the ci the same way bifrost does via vms. but i have not looked into it | 14:32 |
| sdake | at next summit i'd like to have another repeat of our contributor day on how to work with project-config and the gate | 14:32 |
| sdake | sean-k-mooney will you be in barcelona? | 14:33 |
| sean-k-mooney | sdake: it looking that way yes | 14:33 |
| sdake | cool - if we get a speaking spot i'd really like to highlight your &teams work on bifrost integration | 14:33 |
| sean-k-mooney | still no confirmation if travel is aproved but im on the list to travel. | 14:34 |
| sdake | sean-k-mooney well whether you are there or not, we will be highlighting that work if we get a speaking slot | 14:34 |
| *** daneyon has joined #openstack-kolla | 14:34 | |
| sdake | sean-k-mooney would be nice to get you on stage with the rest of the cores that have travel approval to acknowledge this extremely difficult objective | 14:35 |
| *** senk has joined #openstack-kolla | 14:35 | |
| sdake | since from my pov its about done :) | 14:35 |
| senk | can someone please add +1 worklow or readd the workflow on https://review.openstack.org/#/c/350222/2 | 14:35 |
| patchbot | senk: patch 350222 - kolla - Fix inconsistencies in git url | 14:35 |
| inc0 | +1 to that | 14:35 |
| *** diogogmt has joined #openstack-kolla | 14:35 | |
| sdake | senk looking | 14:35 |
| senk | were some issues with jenkins in the morning | 14:36 |
| senk | sdake: thanks | 14:36 |
| sdake | senk ya that was my fault for merging a patch that didn't pass the gate | 14:36 |
| sdake | moment looking at patch to see if that issue is your issue | 14:36 |
| sean-k-mooney | sdake well that more insentive to get it completed i guess. im not much of a public speaking kind of guy but i would be honored. | 14:36 |
| sdake | senk not sure what is going on there -sometimes ci flakes out - what I'd recommend is rebasing your patch on master (git pull --rebase; git review) | 14:37 |
| sdake | and it will be reacked | 14:37 |
| sean-k-mooney | oh i just realise i probly should add a release note for the kolla-host and bifrost changes... | 14:37 |
| sdake | sean-k-mooney yes please | 14:37 |
| narasimha_SV | sean-k-mooney: not that one I dont see the supported driver info in the ironinc.conf file | 14:38 |
| *** banix has quit IRC | 14:38 | |
| senk | sdake: na zuul simply skipped docs jobs. infra fixed it at https://review.openstack.org/#/c/351626/ | 14:38 |
| patchbot | senk: patch 351626 - openstack-infra/project-config - Fix kolla docs skip (MERGED) | 14:38 |
| inc0 | gtg for an hour or so | 14:38 |
| inc0 | cyas | 14:38 |
| narasimha_SV | I have even check ansible code there also there is no specification metioned either in templates or tasks | 14:38 |
| *** inc0 has quit IRC | 14:38 | |
| senk | ah my workflowchange restarted it | 14:39 |
| *** daneyon has quit IRC | 14:39 | |
| *** mliima_ has quit IRC | 14:40 | |
| sean-k-mooney | narasimha_SV: am in the context of kolla ironic support? sorry im context switching so lost context of what we were talking about | 14:40 |
| sdake | one thing that would help is defining the work as either "bifrost" or "ironic as a service" | 14:41 |
| sdake | since the term ironic is overloaded | 14:41 |
| *** ayoung_ZZZzzz is now known as ayoung | 14:42 | |
| sean-k-mooney | sdake: i think narasimha_SV was refering to ironic as a service | 14:43 |
| sdake | sean-k-mooney agree | 14:43 |
| duonghq | hello sdake | 14:44 |
| sdake | narasimha_SV the kolla team hasn't committed to supporting ironic as a service in newton - unless someone comes along nd makes it work | 14:44 |
| sdake | sup duonghq wht can i do for ya | 14:44 |
| sean-k-mooney | sdake: narasimha_SV im not actully that familar with seting up ironic itself as i let bifrost do that form me so not sure if we need to specify the supproted dirvers | 14:44 |
| duonghq | As I see nobody take the bp nova-kubernetes up to now | 14:44 |
| duonghq | Can I try on this bp? | 14:45 |
| sdake | duonghq pretty sure wirehead_ is working on that | 14:45 |
| sdake | duonghq can you coordinate with him? | 14:45 |
| openstackgerrit | Robin Naundorf proposed openstack/kolla: Fix inconsistencies in git url https://review.openstack.org/350222 | 14:45 |
| sdake | senk ^^ | 14:45 |
| sean-k-mooney | narasimha_SV: if you need to tweak anyting in the ironic.conf you should be able to drop a file into /etc/kolla/config/ironic/ironic.conf | 14:46 |
| sdake | i always thought it was a shame batman didn't follow up with a & robyn after "dark knight rises" | 14:46 |
| senk | sdake: still learning to work with gerrit..^^ | 14:46 |
| *** fragatina has joined #openstack-kolla | 14:46 | |
| sdake | senk learning experience efor all of us - thanks for the contribution :) | 14:46 |
| duonghq | I don't know his timezone, but I can, | 14:46 |
| sdake | duonghq he is PST | 14:47 |
| sdake | its 7:47 PST atm | 14:47 |
| sdake | he is usually alive by 9am or 10am | 14:47 |
| *** dmueller-evoila has quit IRC | 14:48 | |
| sdake | duonghq another person to coordinate with is dcwangmit01 | 14:48 |
| sdake | he is also working on this iiuc | 14:48 |
| *** fragatin_ has joined #openstack-kolla | 14:48 | |
| duonghq | I'm on UTC+7 | 14:48 |
| *** srwilkers has joined #openstack-kolla | 14:48 | |
| *** diogogmt has quit IRC | 14:48 | |
| sdake | so 8pm there? | 14:48 |
| duonghq | near 10pm here | 14:48 |
| sdake | yikes | 14:48 |
| sdake | apac and us hard to coordinate because of timezones | 14:48 |
| sdake | one option is to leave a note in the blueprint | 14:49 |
| sdake | (if you want to head to bed) | 14:49 |
| sdake | rhallisey leading up kolla-kubernetes work - woudl be another contact (less close to the actual nova blueprint in question) | 14:50 |
| duonghq | I can work late but the main problem is I cannot reach my dev PC at this time, but I can try to grab one alternative, | 14:50 |
| *** mkoderer_ has joined #openstack-kolla | 14:51 | |
| duonghq | but as I see, it seem that we do not communicate much on mailing list? | 14:51 |
| *** lmiccini_ has joined #openstack-kolla | 14:51 | |
| sdake | mailing list works fine | 14:51 |
| *** fragatina has quit IRC | 14:51 | |
| sdake | irc is faster | 14:51 |
| duonghq | sure | 14:52 |
| *** diogogmt has joined #openstack-kolla | 14:52 | |
| narasimha_SV | sean-k-mooney: this directory is not default right we need to add this so that we can change | 14:52 |
| sdake | wirehead_ dcwangmit01 can give you a faster update on what is happening on that blueprint | 14:52 |
| *** mliima has joined #openstack-kolla | 14:52 | |
| sdake | via irc - mailing list ha coule days lag time | 14:52 |
| narasimha_SV | sean-k-mooney: what values do I need to keep here all the values in template form or new values which i want to add | 14:52 |
| *** mewald has quit IRC | 14:52 | |
| duonghq | yeah, I can discuss with them on IRC and work on my dev machine later (in the worst case) | 14:53 |
| sean-k-mooney | narasimha_SV: kolla will merge the overdies from the config director with the template so you only need to add your custom values | 14:53 |
| *** mkoderer has quit IRC | 14:54 | |
| *** lmiccini has quit IRC | 14:54 | |
| narasimha_SV | sean-k-mooney: can I add this config value as default value in ironic.conf by sepcifying all default drivers ? | 14:55 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla: Introduce a script to launch ovsdb-server process https://review.openstack.org/351784 | 14:55 |
| duonghq | sdake: I'm fine with wirehead_ working hours | 15:00 |
| sdake | duonghq that said given its friday - not sure if he is on pto today or not, so reach out to dcwangmit01 as well | 15:01 |
| duonghq | okay, dcwangmit01 is on PST, isn't him? | 15:02 |
| sdake | duonghq believe so | 15:02 |
| sdake | probably same hours as wirehead_ | 15:02 |
| duonghq | okay | 15:03 |
| sdake | one of em can get you integrated in that effort | 15:03 |
| *** banix has joined #openstack-kolla | 15:03 | |
| *** mliima has quit IRC | 15:04 | |
| duonghq | nice, I'll ask them when somebody online | 15:04 |
| sdake | rhallisey is typically around at this time - but may be afk atm | 15:05 |
| sdake | rhallisey has a deeper understanding of what wirehead_ and dcwangmit01 are up to | 15:05 |
| rhallisey | hey | 15:05 |
| rhallisey | had my sound off | 15:05 |
| sdake | duonghq we had planned at midcycle to put work items in the blueprints | 15:05 |
| * rhallisey reads back | 15:05 | |
| sdake | rhallisey ya my wife complains constantly about my buzzer interrupting herwork :) | 15:06 |
| sean-k-mooney | narasimha_SV: can you post a link to where you want to add it in the code. and yes you could | 15:07 |
| rhallisey | duonghq, wirehead_ has a patch for nova up now, | 15:07 |
| rhallisey | https://review.openstack.org/#/c/347572/ | 15:07 |
| patchbot | rhallisey: patch 347572 - kolla-kubernetes - Remove hostnetworking for Nova control plane and c... | 15:07 |
| duonghq | sorry sdake if it's my fault | 15:07 |
| *** shardy is now known as shardy_afk | 15:07 | |
| *** zhubingbing__ has quit IRC | 15:09 | |
| sdake | rhallisey can you star that review plelase | 15:09 |
| rhallisey | duonghq, the nova bp hasn't had any work items filled out yet | 15:10 |
| *** fragatin_ has quit IRC | 15:10 | |
| rhallisey | if you wan, apply that patch and see if you can get a vm up | 15:11 |
| * rhallisey will be doing so tmr | 15:11 | |
| sdake | duonghq nah its mine for not being omnipotent ;) | 15:11 |
| rhallisey | duonghq, there's also https://blueprints.launchpad.net/kolla-kubernetes/+spec/cinder-kubernetes | 15:11 |
| rhallisey | actully not sure if swift is fully fucntional | 15:12 |
| sdake | duonghq that is why we have a team - so folks dont have to know everything :) | 15:12 |
| rhallisey | swift would also be awwesome | 15:12 |
| *** zhubingbing__ has joined #openstack-kolla | 15:12 | |
| *** yee_ has joined #openstack-kolla | 15:13 | |
| duonghq | sdake absolutely :) | 15:13 |
| sdake | oh glorious rain | 15:13 |
| sdake | its not 115 today :) | 15:13 |
| duonghq | rhallisey: I am taking a look at those | 15:13 |
| rhallisey | duonghq, ok awesome. Just added https://blueprints.launchpad.net/kolla-kubernetes/+spec/swift-kubernetes if you're interested in that | 15:14 |
| rhallisey | swift already has pods/service, but needs a little more work | 15:15 |
| duonghq | rhallisey: do we have any bp work yet, so I can learn something from these | 15:16 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla: Introduce a script to launch ovsdb-server process https://review.openstack.org/351784 | 15:17 |
| rhallisey | duonghq, what do you mean by bp work? Workitems? | 15:17 |
| duonghq | I mean some pods already up and running | 15:18 |
| rhallisey | duonghq, yes, horizon, keystone, memcached, mariadb | 15:18 |
| *** mliima has joined #openstack-kolla | 15:18 | |
| rhallisey | glance, nova, and neutron are coming soon | 15:18 |
| rhallisey | duonghq, right now it's all CLI driven | 15:19 |
| rhallisey | I'll be adding ansible in front of that this weekend | 15:19 |
| * irtermite yawns | 15:19 | |
| irtermite | morning... Looks like things are rolling along | 15:19 |
| duonghq | awesome | 15:20 |
| sdake | irtermite appears so - thanks for the help :) | 15:20 |
| irtermite | sdake: of course | 15:20 |
| sdake | inc0 pbourke britthouser any blockers needing irtermite 's help? | 15:21 |
| irtermite | interesting... could you all do me a flavor << see what I did there? you will | 15:21 |
| pbourke | sdake: nope looking good | 15:21 |
| irtermite | flavor list? is it empty? | 15:21 |
| irtermite | take a look at the #openstack chan | 15:21 |
| irtermite | no default flavors when installing with kolla? | 15:22 |
| sdake | irtermite ya nova took out default flavors in master | 15:22 |
| sdake | stable branches have default flavors - that change wasnt backported | 15:23 |
| irtermite | as of Mitaka? | 15:23 |
| sdake | irtermite not certain | 15:23 |
| sdake | i thought it was a master only change | 15:23 |
| sdake | and didn't affect mitaka | 15:23 |
| sdake | but coud be wrong | 15:23 |
| *** inc0 has joined #openstack-kolla | 15:24 | |
| duonghq | rhallisey: I wonder why there is no patchset has bp/nova-kubernetes tag? | 15:24 |
| inc0 | I'm back | 15:25 |
| rhallisey | duonghq, he posted that before I made that bo | 15:25 |
| rhallisey | bp* | 15:25 |
| pbourke | sdake: britthouser inc0 prechecks passed! | 15:25 |
| pbourke | kolla is a go | 15:26 |
| duonghq | now I understand I cann't keep up with new bp | 15:26 |
| inc0 | we need to setup registry and push some images | 15:26 |
| Daviey | Chaps... anyone else seeing sudoer issues in kolla_toolbox? Ie, asking for a password? | 15:28 |
| britthouser | w00t pbourke! | 15:28 |
| duonghq | rhallisey: I need more time to find what people do for new bp, | 15:28 |
| *** Satya__ has joined #openstack-kolla | 15:29 | |
| sdake | duonghq the irst step is writing down the work items | 15:29 |
| rhallisey | inc0, ^^^^ kolla_toolbox issue | 15:29 |
| sdake | duonghq in launchpad - click yellow button | 15:29 |
| Satya__ | Hi sdake | 15:29 |
| sdake | hey Satya__ | 15:29 |
| rhallisey | inc0, did you get around that issue that Daviey is also having? | 15:30 |
| sdake | Satya__ mind reasking your question for folks to answer | 15:30 |
| Satya__ | i am not getting the flavors in nova flavor-list | 15:30 |
| Satya__ | after the deployment | 15:30 |
| sdake | Satya__ moment - let me look at codebase | 15:30 |
| Satya__ | sure | 15:30 |
| sean-k-mooney | go to run for a while be back in a few | 15:30 |
| duonghq | sdake: thank you | 15:30 |
| Daviey | rhallisey / inc0: Looking up disks to bootstrap for ceph OSD.... kolla_toolbox runnig find_disks.py, asking for sudoer password? | 15:31 |
| inc0 | hmm, yeah, Daviey did you rebuild lately? | 15:31 |
| Daviey | current Master | 15:31 |
| duonghq | I wish that other bp has description like the neutron one :P | 15:31 |
| inc0 | I had same issue and it turned out to be container missing ansible sudoers file | 15:31 |
| Daviey | inc0: rebuilt today, on Master | 15:31 |
| inc0 | for whatever reason | 15:31 |
| inc0 | I rebuilt and it was fixed | 15:32 |
| Daviey | huh | 15:32 |
| inc0 | but since you're getting the same issue...it's strange | 15:32 |
| Daviey | inc0: rebuilt world or just the toolbox? | 15:32 |
| inc0 | try just toolbox | 15:32 |
| inc0 | without cache | 15:32 |
| inc0 | pbourke, ssh 172.21.8.250 - no route to host... | 15:33 |
| duonghq | sdake, rhallisey: it is Friday night here, so I try on your suggestion next week | 15:33 |
| pbourke | uh oh | 15:34 |
| inc0 | but I can't access any node | 15:34 |
| inc0 | maybe my vpn is bronek | 15:34 |
| pbourke | inc0: no I cant either | 15:34 |
| duonghq | I hope that I can find some free time on weaken | 15:34 |
| pbourke | im logged in though | 15:34 |
| pbourke | but cant get a new connection | 15:34 |
| inc0 | eee? | 15:34 |
| rhallisey | duonghq, ok great! Talk to you later :) | 15:34 |
| britthouser | 8.253 | 15:35 |
| britthouser | inc0 | 15:35 |
| britthouser | ^^ | 15:35 |
| inc0 | ok | 15:35 |
| inc0 | gonna change etherpad | 15:35 |
| * britthouser thought he had done that....sorry | 15:35 | |
| sdake | Satya__ http://osdir.com/ml/openstack-dev/2016-04/msg00465.html | 15:36 |
| sdake | Satya__ are you using master or a stable branch? | 15:36 |
| pbourke | inc0: britthouser: trying to run the bootstrap role on deploy node also to get docker for registry etc. | 15:36 |
| Satya__ | stable mitaka | 15:36 |
| sdake | can you run pip show into a paste for me | 15:37 |
| sdake | rather pip show kolla | 15:37 |
| rhallisey | sdake, bunch of doc reworks at the top of the q that I just rebased | 15:37 |
| britthouser | Are you running from inside the LXC pbourke? | 15:37 |
| pbourke | :/ | 15:37 |
| inc0 | pbourke, we don't need this mount thing in registry tho | 15:37 |
| Daviey | inc0: did you see my mail to offer to help with OSIC? | 15:38 |
| pbourke | right but | 15:38 |
| pbourke | one momnt | 15:38 |
| inc0 | Daviey, must've missed it, can you plz give me your email? | 15:38 |
| Satya__ | from build node? | 15:39 |
| sdake | Satya__ right from build node please | 15:39 |
| Daviey | inc0: email@daviey.com subject: OSIC cluster for Kolla | 15:40 |
| *** belmoreira has quit IRC | 15:40 | |
| Satya__ | http://pastebin.com/NR11jQ5R | 15:41 |
| pbourke | inc0: britthouser: sdake: probably a good idea to drop a summary to the ML after we complete each major step, for those who don't have time to keep up with tons of IRC logs. Will write something up before finishing for the day | 15:41 |
| *** Guest52040 has joined #openstack-kolla | 15:41 | |
| inc0 | good idea pbourke | 15:41 |
| sdake | Satya__ ok - your running a milestone version | 15:41 |
| sdake | satya_ not sure if that is root cause | 15:41 |
| sdake | but lets fix that first | 15:41 |
| britthouser | I was putting that in the etherpad, but ML is better | 15:41 |
| inc0 | pbourke, you installed it on host not in containrr | 15:41 |
| Satya__ | sure | 15:41 |
| pbourke | ah, keep forgetting | 15:41 |
| sdake | fix one problem at a time until none remain :) | 15:41 |
| sdake | pip remove kolla | 15:42 |
| sdake | or pip uninstall kolla | 15:42 |
| sdake | followed by pip show kolla | 15:42 |
| Satya__ | sonr | 15:42 |
| sdake | sonr? | 15:42 |
| Satya__ | done | 15:42 |
| sdake | ok | 15:42 |
| Satya__ | sorry fat fingure | 15:42 |
| Satya__ | fingre | 15:43 |
| sdake | can you paste pip show kolla again pls | 15:43 |
| *** mdnadeem has quit IRC | 15:43 | |
| Satya__ | root@test /]# pip show kolla [root@test /]# | 15:43 |
| sdake | nice | 15:43 |
| sdake | ok - so you ahve two options to run mitaka | 15:43 |
| Satya__ | i am using source code | 15:43 |
| sdake | you can run stable/mitaka or 2.0.2 | 15:44 |
| Satya__ | stable/mitaka | 15:44 |
| sdake | stable/mitaka has a few fixes on top of 2.0.2 atm | 15:44 |
| sdake | ok go to where kolla git repo is and run git status -> paste | 15:44 |
| pbourke | inc0: I want to move ansible to the host | 15:44 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla: Introduce a script to launch ovsdb-server process https://review.openstack.org/351784 | 15:45 |
| pbourke | whos typing atm | 15:45 |
| inc0 | pbourke, me | 15:45 |
| inc0 | you think so? | 15:45 |
| pbourke | inc0: what you doing | 15:45 |
| inc0 | installing docker inside container | 15:45 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add aodh role https://review.openstack.org/351027 | 15:45 |
| pbourke | I think we should move out of the container | 15:45 |
| inc0 | fair enoug | 15:46 |
| Satya__ | we have forked the stable mitaka to our local repo | 15:46 |
| Satya__ | and makin changes to it | 15:46 |
| Satya__ | so will it help if i do a git status? | 15:46 |
| sdake | Satya__ thought you were deploying upstream version - so that wont help your right | 15:46 |
| *** yee_ has quit IRC | 15:47 | |
| Satya__ | nope | 15:47 |
| sdake | satya_ ok run pip install -e . in kolla directory | 15:47 |
| sdake | Satya__ rather^ | 15:47 |
| sdake | followed by pip show kolla | 15:47 |
| Daviey | inc0: yeah, weird... i looked at the container, and it was indeed missing the sudoers file. Did a rebuild of that, and watched it go in. | 15:47 |
| Satya__ | [root@test nova]# pip install -e Usage: pip install [options] <requirement specifier> [package-index-options] ... pip install [options] -r <requirements file> [package-index-options] ... pip install [options] [-e] <vcs project url> ... pip install [options] [-e] <local project path> ... pip install [options] <archive url/path> ... -e option requires an argument | 15:48 |
| sdake | Daviey odd :( | 15:48 |
| openstackgerrit | Mauricio Lima proposed openstack/kolla: Customizations for Manila https://review.openstack.org/350601 | 15:48 |
| sdake | Satya__ from the root of the kolla dir | 15:48 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla: Introduce a script to launch ovsdb-server process https://review.openstack.org/351784 | 15:48 |
| Daviey | sdake: weird that 2 people have seen it in one day :/ | 15:48 |
| inc0 | Daviey, I had exactly same experience | 15:48 |
| inc0 | something is seriously wrong with that | 15:48 |
| inc0 | pbourke, did we set insecure registry? | 15:49 |
| inc0 | across the board? | 15:49 |
| Daviey | i can't understand how it isn't deterministic | 15:49 |
| Daviey | oh well | 15:49 |
| inc0 | Daviey, so since 2 of us had the same issue between few days | 15:49 |
| Satya__ | root@test openstack-kolla]# pip install -e Usage: pip install [options] <requirement specifier> [package-index-options] ... pip install [options] -r <requirements file> [package-index-options] ... pip install [options] [-e] <vcs project url> ... pip install [options] [-e] <local project path> ... pip install [options] <archive url/path> ... | 15:49 |
| inc0 | which driver are you using? aufs? | 15:49 |
| sdake | can a core reviewer review https://review.openstack.org/#/c/351784/5 pls | 15:49 |
| patchbot | sdake: patch 351784 - kolla - Introduce a script to launch ovsdb-server process | 15:49 |
| sdake | and if lgtu - just +2, I'll +w when ci passes | 15:50 |
| sdake | (this is blocking sbezverk) | 15:50 |
| sdake | Satya__ after you wrap up with the pip install -e . from ~/kolla (or whever your kolla repo is) | 15:51 |
| sdake | run pip show kolla again for me pls | 15:51 |
| *** egonzalez90 has joined #openstack-kolla | 15:52 | |
| duonghq | I filed a bug on launchpad some days ago: https://bugs.launchpad.net/kolla/+bug/1609502 I'm sorry but can somebody confirm this bug? | 15:52 |
| openstack | Launchpad bug 1609502 in kolla "Container bootstrap_mariadb exits due to Galera failed" [Undecided,New] | 15:52 |
| Satya__ | i just confirmed we are running 2 weeks behind the trunk | 15:52 |
| sdake | Satya__ - ok so your not running stable/mitaka :) | 15:53 |
| Satya__ | nope | 15:53 |
| sdake | the fact the flavors are not created in master is expected | 15:53 |
| sdake | nova took out flavor creation in newton-1 | 15:53 |
| sdake | the activity that took place there was pingint the operator list | 15:53 |
| Daviey | inc0: btrfs | 15:53 |
| sdake | operators responded "please remove flavor creation" | 15:54 |
| britthouser | so pbourke looks like you're moving from doing deployment from inside lxc to on the host? | 15:54 |
| pbourke | inc0: the baremetal play can do this for us | 15:54 |
| inc0 | ok...so it's not storage | 15:54 |
| inc0 | pbourke, ok | 15:54 |
| inc0 | pbourke, go ahead | 15:54 |
| sdake | Satya__ i would highly reocmmend notrunning on master - because your deploying openstack master as well | 15:54 |
| sdake | not just kolla master | 15:54 |
| sdake | Satya__ may i recommend rebasing your changes on stable/mitaka | 15:54 |
| inc0 | I had it with aufs | 15:54 |
| sdake | that will solve two problems - first flavor creation will happen - second you will have a stable openstack to run against | 15:55 |
| Satya__ | we are actually changing the code to use NDB rather than innodb | 15:55 |
| pbourke | britthouser: yeah I think it may avoid problems down the road | 15:55 |
| britthouser | agreed | 15:55 |
| pbourke | maybe not necessary | 15:55 |
| pbourke | but that ssh thing etc. was tripping me up | 15:55 |
| Satya__ | so we are working on upstream | 15:55 |
| sdake | Satya__ no idea what ndb is - but wfm :) | 15:55 |
| britthouser | keep teh lxc for jsut cobbler | 15:55 |
| Satya__ | Hi Britt | 15:56 |
| Satya__ | how are you? | 15:56 |
| Satya__ | after long time | 15:56 |
| sdake | Satya__ I still recommend against deploying master in a production deployment | 15:56 |
| britthouser | Hey Satya__ | 15:56 |
| Satya__ | remember me? | 15:56 |
| *** sdake has quit IRC | 15:56 | |
| *** sdake has joined #openstack-kolla | 15:56 | |
| duonghq | Satya_ I'm sorry but what does "trunk" mean? | 15:56 |
| britthouser | of course! I wasn't sure if it was your or not | 15:57 |
| Satya__ | 2 week behind the latest code | 15:57 |
| sdake | trunk means master | 15:57 |
| Satya__ | :P | 15:57 |
| Satya__ | now you git it britt | 15:57 |
| duonghq | Is it from svn term? | 15:57 |
| sdake | duonghq yes | 15:58 |
| duonghq | sdake: thanks | 15:58 |
| Satya__ | sdake what ever the latest code in master we just pull it in 2 weeks | 15:58 |
| sdake | Satya__ roger i hear ya - so the latest code of kolla deploys the latesst code of openstack | 15:58 |
| sdake | is that what you want? | 15:59 |
| sdake | if your doing a production cloud Ithink not | 15:59 |
| sdake | if your doing development, I htink so | 15:59 |
| Satya__ | nope | 15:59 |
| Satya__ | i am not working on production cloud | 15:59 |
| Satya__ | its a deployment to develop :) | 15:59 |
| sdake | ok - if doing development and using kolla as an upstream then your activites make sense to me | 16:00 |
| sdake | but you wont get flavor creation | 16:00 |
| sdake | becuase nova removed it | 16:00 |
| Satya__ | hmmm | 16:00 |
| sdake | you have to create flavors on your own | 16:00 |
| Satya__ | ok got it :) | 16:00 |
| sdake | sweet | 16:00 |
| sdake | happy to help | 16:00 |
| Satya__ | Thanks sdake | 16:00 |
| Satya__ | britt there? | 16:01 |
| Satya__ | please ping me your mail ID | 16:01 |
| Satya__ | need to drop a mail to you | 16:01 |
| Satya__ | :) | 16:01 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add Barbibcan container dockerfile https://review.openstack.org/351822 | 16:02 |
| sdake | zhubingbing__ on fire! | 16:03 |
| zhubingbing__ | - - | 16:03 |
| *** signed8bit has joined #openstack-kolla | 16:03 | |
| zhubingbing__ | thanks | 16:03 |
| sdake | zhubingbing__ (thats a us compliment to mean your doing good work) | 16:03 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla: Customizations for Ceph https://review.openstack.org/351823 | 16:04 |
| zhubingbing__ | thank you, i will work harder. | 16:05 |
| sdake | zhubingbing__ your working hard enough imo :) | 16:05 |
| zhubingbing__ | I have some questions about the Barbican barbican.conf, and role needs to be submitted tomorrow. | 16:06 |
| openstackgerrit | Eduardo Gonzalez proposed openstack/kolla: Customizations for Ceph https://review.openstack.org/351823 | 16:07 |
| sdake | zhubingbing__ dave-mccowan is a barbican core and expert in the code | 16:08 |
| sdake | zhubingbing__ recommend hitting him up for q&a | 16:08 |
| *** fragatina has joined #openstack-kolla | 16:08 | |
| *** senk has quit IRC | 16:09 | |
| *** fragatina has quit IRC | 16:09 | |
| zhubingbing__ | ok | 16:09 |
| dave-mccowan | zhubingbing__ hello. what questions to you have? | 16:10 |
| zhubingbing__ | If I want to run Barbican what I should do with the configuration for barbican.conf, use the default on the line well? | 16:11 |
| sean-k-mooney | ok im back meeting over. how is the osic cluster going | 16:11 |
| *** shardy_afk is now known as shardy | 16:12 | |
| *** yee_ has joined #openstack-kolla | 16:13 | |
| inc0 | pbourke, sean-k-mooney seems like we're good to kolla-go? | 16:14 |
| pbourke | inc0: assuming the insecure registry is configured correctly | 16:14 |
| pbourke | inc0: next step I think is to generate a build conf and start building | 16:14 |
| sean-k-mooney | inc0: cool | 16:15 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add Barbibcan container dockerfile https://review.openstack.org/351822 | 16:15 |
| sean-k-mooney | pbourke: if you used my kolla-host playbook on 14.04 i will not have updated the insecur registry. that support is currently missing for non systemd OSs but ill be adding it in the next version | 16:15 |
| pbourke | :) | 16:16 |
| sean-k-mooney | pbourke: eta 30-60 mins | 16:16 |
| pbourke | just noticing that now | 16:16 |
| inc0 | well...no insecure registry in /etc/default/docker tho | 16:16 |
| pbourke | no worries | 16:16 |
| pbourke | inc0: sorry should just have used your method | 16:16 |
| inc0 | :) | 16:16 |
| pbourke | actually | 16:16 |
| sdake | pbourke can we can timings via the time command on the dependnecy deploy operation | 16:17 |
| pbourke | we could just start building and pick up sean's change when it's ready | 16:17 |
| pbourke | sdake: yeah will start doing that | 16:17 |
| *** yee_ has quit IRC | 16:18 | |
| *** kbaegis has joined #openstack-kolla | 16:19 | |
| inc0 | crap | 16:22 |
| *** daneyon has joined #openstack-kolla | 16:23 | |
| inc0 | ehh... | 16:24 |
| *** fragatina has joined #openstack-kolla | 16:24 | |
| duonghq | inc0: I'm sorry but can you or somebody confirm my bug report: https://bugs.launchpad.net/kolla/+bug/1609502 | 16:24 |
| openstack | Launchpad bug 1609502 in kolla "Container bootstrap_mariadb exits due to Galera failed" [Undecided,New] | 16:24 |
| duonghq | I try and try many time | 16:24 |
| duonghq | but I still can not figure out any wrong with me | 16:24 |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Customizations for Nova https://review.openstack.org/351676 | 16:26 |
| sean-k-mooney | inc0: broke something? | 16:26 |
| inc0 | already fixed | 16:26 |
| inc0 | ok insecure registry is setup | 16:27 |
| inc0 | pbourke, wanna do the build honors? | 16:27 |
| *** daneyon has quit IRC | 16:27 | |
| pbourke | sure just gonna check the reg one last time | 16:28 |
| inc0 | duonghq, docker ps please | 16:29 |
| *** coolsvap has joined #openstack-kolla | 16:29 | |
| sdake | coolsvap do you have any bandiwdith to triage https://bugs.launchpad.net/kolla/+bug/1609502 | 16:30 |
| openstack | Launchpad bug 1609502 in kolla "Container bootstrap_mariadb exits due to Galera failed" [Undecided,New] | 16:30 |
| sdake | Satya__ sorry had to step awa yfor a moment - are you all set then? | 16:30 |
| Satya__ | yes Thanks sdake | 16:31 |
| Satya__ | just one query | 16:31 |
| coolsvap | sdake: will check later tonight | 16:31 |
| Satya__ | why nova thought of removing the flavors? | 16:31 |
| duonghq | irc0: I'm sorry but I do not have access to my developing computer right now, many sorry | 16:31 |
| sdake | Satya__ the discussion i saw about it was they were picked arbitrarily to copy aws, and they are not really useful for anyone except developers, so they were removed since operators by default just delete them anyway | 16:32 |
| duonghq | inc0: I'll retry and provide more information later | 16:32 |
| sdake | Satya__ yo uwould relaly have to ask on #nova for more details though | 16:32 |
| Satya__ | ok sure | 16:32 |
| sdake | duonghq work with coolsvap later tonight - he is on your timezone iiuc | 16:33 |
| sdake | duonghq when you may have access to your gear | 16:33 |
| *** kbaegis has quit IRC | 16:33 | |
| duonghq | sdake: okay, thank you | 16:33 |
| *** kbaegis has joined #openstack-kolla | 16:35 | |
| inc0 | we should time it | 16:35 |
| inc0 | pbourke, | 16:35 |
| inc0 | and yeah, after we moved to full threads ^C is broken, anyone want to fix? | 16:36 |
| duonghq | sdake: who has responsible for kolla namespace in docker hub? | 16:36 |
| inc0 | pbourke, I'll submit a bug for it | 16:37 |
| pbourke | thanks inc0 | 16:37 |
| inc0 | duonghq, core team | 16:37 |
| sdake | duonghq core reiew team | 16:37 |
| pbourke | was surprised no one was complaining so thought maybe it was just me | 16:37 |
| sdake | are ou talking aobut build not accepting ^c? | 16:37 |
| sdake | i noticed that - ANNOYING | 16:37 |
| pbourke | yeah | 16:37 |
| inc0 | nah, not just you | 16:37 |
| inc0 | yeah sdake | 16:37 |
| sdake | hae to ctrl-z kill -9 %1 | 16:37 |
| pbourke | kill % is shorted ;) | 16:37 |
| pbourke | *shorter | 16:38 |
| sdake | wfm | 16:38 |
| sdake | that needs fixing | 16:38 |
| inc0 | harlowja, ^ ;) | 16:38 |
| *** egonzalez90 has quit IRC | 16:38 | |
| openstackgerrit | Mauricio Lima proposed openstack/kolla: Customizations for Manila https://review.openstack.org/350601 | 16:38 |
| inc0 | https://bugs.launchpad.net/kolla/+bug/1610324 | 16:39 |
| openstack | Launchpad bug 1610324 in kolla "build doesn't accept ^C" [High,Confirmed] | 16:39 |
| zhubingbing__ | duonghq | 16:39 |
| inc0 | that being said, build is moving quickly | 16:39 |
| zhubingbing__ | i havre a registry for kolla | 16:40 |
| zhubingbing__ | i have | 16:40 |
| *** JoseMello has quit IRC | 16:40 | |
| duonghq | I think somebody should get images' tag update to master branch source code? | 16:41 |
| inc0 | duonghq, unless we'll have automatic pushing, it's always going to be outdated | 16:41 |
| inc0 | automatic pushing is something we want to have...one day | 16:41 |
| inc0 | but it's not trivial | 16:41 |
| sdake | duonghq we are not pushing master for some time | 16:41 |
| zhubingbing__ | INFO:__main__:Removing existing destination: /etc//my.cnf | 16:42 |
| zhubingbing__ | INFO:__main__:Copying /var/lib/kolla/config_files/galera.cnf to /etc//my.cnf | 16:42 |
| sdake | duonghq if you want master you have to build on your own | 16:42 |
| pbourke | inc0: looking good. Think I'll call it a day here | 16:42 |
| pbourke | inc0: will drop a mail to the list so people know where we're at | 16:42 |
| zhubingbing__ | duonghq i think you are wrong configuration file path | 16:42 |
| inc0 | 5min 30s for --profile default | 16:43 |
| * inc0 likes it | 16:43 | |
| duonghq | sdake, inc0: I think it outdate for some day is ok but for more than a week is quite long, especially for major tag like 3.0.0 | 16:43 |
| inc0 | duonghq, so master is meant only for developers really | 16:43 |
| inc0 | and developers should build their own images | 16:43 |
| inc0 | that's my take on it | 16:43 |
| *** egonzalez has joined #openstack-kolla | 16:44 | |
| *** senk has joined #openstack-kolla | 16:44 | |
| sdake | 3.0.0 is not a thing yet duonghq :) | 16:44 |
| duonghq | inc0: should we update the getting start page to notice this aspect? | 16:44 |
| *** egonzalez is now known as egonzalez90 | 16:44 | |
| sdake | this is why Daviey 's approach of using "current" is more appealing to me then calling our 3..0.0 tag 3.0.0 when its nto really 3.0.0 | 16:44 |
| sdake | duonghq docs here would help | 16:45 |
| duonghq | sdake, inc0: it is absolute ok to do source deployment, but the point it the documentation should point out that | 16:45 |
| duonghq | sdake, inc0: can I update the docs to reflect this aspect? | 16:46 |
| sdake | duonghq dont ask for permission, ask or forgiveness :) | 16:47 |
| duonghq | okay :) | 16:47 |
| inc0 | duonghq, of course you can | 16:48 |
| inc0 | you can publish whatever changes you want, worst case scenerio we won't merge it | 16:48 |
| duonghq | inc0: sure :) | 16:49 |
| inc0 | pbourke, make sure to include our build time in the mail | 16:49 |
| inc0 | #bragforkolla | 16:49 |
| inc0 | #buildporn | 16:49 |
| pbourke | inc0: I did :) it's brief enough as pretty tired, hopefully does the job anyway | 16:50 |
| *** senk has quit IRC | 16:50 | |
| inc0 | soo...who wants to see kolla deploying 130 nodes? | 16:50 |
| inc0 | sdake? | 16:50 |
| inc0 | britthouser? | 16:51 |
| sdake | i'd love to - have meeting in 9 minutes for 1 hour | 16:51 |
| sdake | dont wait for me | 16:51 |
| inc0 | kk\ | 16:51 |
| sdake | timings would be nice | 16:51 |
| inc0 | ofc | 16:51 |
| sean-k-mooney | inc0: are you going to make it a google hangouts | 16:52 |
| pbourke | inc0: keep general notes on the pad, dont want to miss all the fun :) | 16:52 |
| sean-k-mooney | inc0: by the way have you ran the precheks to see if they pass first on the 130+ nodes | 16:52 |
| duonghq | inc0: how can I see? | 16:52 |
| pbourke | heading off, chat later | 16:53 |
| *** fragatin_ has joined #openstack-kolla | 16:53 | |
| inc0 | hmmm | 16:53 |
| inc0 | interesting | 16:53 |
| inc0 | fouind first bug in kolla | 16:54 |
| inc0 | in osic | 16:54 |
| sdake | wow cold in phoenix for first time in 4 months | 16:55 |
| britthouser | in a meeting right now inc0....gimme a couple mins | 16:55 |
| *** fragatina has quit IRC | 16:55 | |
| *** kbaegis has quit IRC | 17:01 | |
| *** srwilkers has quit IRC | 17:01 | |
| *** zhubingbing__ has quit IRC | 17:02 | |
| *** duonghq has left #openstack-kolla | 17:02 | |
| britthouser | Ok I'm free inc0 | 17:03 |
| britthouser | On your mark....get set.... | 17:03 |
| inc0 | soo | 17:05 |
| britthouser | yes! | 17:05 |
| inc0 | it seems our little ansible_{{ api_interface }} can't grok a dot | 17:05 |
| britthouser | ew | 17:06 |
| *** kbaegis has joined #openstack-kolla | 17:06 | |
| britthouser | so no VLAN subinterfaces? | 17:06 |
| inc0 | I consider this a critical bug | 17:06 |
| britthouser | Yeah | 17:06 |
| inc0 | I'm trying to figure out how to get ansible to behave and publish a patchset today | 17:07 |
| kbaegis | You want sharded interfaces rather than ovs? | 17:07 |
| britthouser | this isn't for tenant network kbaegis, rather for the services networks | 17:07 |
| britthouser | like how you ssh into the baremetal | 17:08 |
| inc0 | nah. false alarm | 17:11 |
| inc0 | it's all good | 17:11 |
| *** diogogmt has quit IRC | 17:11 | |
| inc0 | we messed sth with globals | 17:11 |
| harlowja | inc0 well it does accept ctrl-c | 17:11 |
| harlowja | it just that it doesn't try to crap out randomly and not finish up cleanly | 17:11 |
| inc0 | maybe add something like "2*ctrl+c == die, just die" | 17:12 |
| inc0 | funny it works now | 17:13 |
| *** yee_ has joined #openstack-kolla | 17:13 | |
| *** diogogmt has joined #openstack-kolla | 17:14 | |
| * britthouser scratches my ehad | 17:14 | |
| *** mewald has joined #openstack-kolla | 17:14 | |
| sdake | inc0 groan | 17:14 |
| sdake | inc0 criticial | 17:14 |
| inc0 | sdake, false alarm | 17:14 |
| sdake | ok | 17:15 |
| inc0 | I have no idea what was wrong but it seems ok now | 17:15 |
| inc0 | doing prechecks | 17:15 |
| inc0 | and will see if I can record hangouts | 17:15 |
| *** signed8bit is now known as signed8bit_Zzz | 17:15 | |
| rhallisey | inc0, hangouts link? | 17:16 |
| inc0 | trying to get | 17:17 |
| inc0 | https://hangouts.google.com/call/c7fascv5bzgnzjpcaopqk2snjee | 17:17 |
| sean-k-mooney | mind if i join | 17:18 |
| rhallisey | sean-k-mooney, go for it | 17:18 |
| sdake | sean-k-mooney all our work is oepn to all | 17:18 |
| rhallisey | inc0, go through google plus | 17:20 |
| sdake | pbourke wow 5min build? | 17:20 |
| *** yee_ has quit IRC | 17:20 | |
| sdake | osic network on steroids :) | 17:20 |
| rhallisey | when you start a hangout there's a box that says enable hangounts on air | 17:21 |
| sean-k-mooney | sdake: 5 min build nice. didnt you say it took you 8min at home with your 1G fiber line and nvme storage | 17:21 |
| sdake | 15min for all | 17:22 |
| sdake | ut i ont build with the same profile as paul | 17:22 |
| *** signed8bit_Zzz is now known as signed8bit | 17:22 | |
| sean-k-mooney | in that case 5mins is awsome | 17:22 |
| sdake | or perhap i do | 17:22 |
| sdake | not sure | 17:22 |
| inc0 | https://plus.google.com/events/cib8giqjovmtfmesbe7h2foug1k | 17:22 |
| sdake | i build without profile option- i guess that would be default | 17:22 |
| inc0 | http://youtu.be/InU0Up_5H24 | 17:22 |
| rhallisey | lol | 17:22 |
| rhallisey | it has a youtube like nice | 17:23 |
| inc0 | yeah | 17:23 |
| inc0 | it's gonna be cool | 17:23 |
| *** diogogmt has quit IRC | 17:23 | |
| inc0 | although if we fail our deploy it's going to be recorded forever;) | 17:23 |
| inc0 | anyway, lets try shall we? | 17:23 |
| sean-k-mooney | inc0: your youtube carrer starts :) next you will be on twitch | 17:24 |
| sean-k-mooney | and yes | 17:24 |
| sean-k-mooney | we can hear you on youtube but we cant join the hangeout and talk as far as i know | 17:25 |
| rhallisey | inc0, timer? | 17:26 |
| rhallisey | 1:26 start | 17:26 |
| inc0 | I think I timed it | 17:26 |
| sean-k-mooney | inc0: you did | 17:26 |
| inc0 | we need to tweak ansible confs tho | 17:27 |
| Mech422 | I am soo sick of networking | 17:27 |
| inc0 | move to mountains and start herding sheeps Mech422 | 17:27 |
| Mech422 | I _think_ I almost have my dev lab mimicking the screwed up prod configs though | 17:27 |
| sean-k-mooney | so how long do you think the deploy will take out of interest? | 17:27 |
| Mech422 | inc0: actually, the VP of Olliance (SFO Open Source consulting company) actually moved to wisconsin(?) and started raising goats when we folded... | 17:28 |
| inc0 | sean-k-mooney, 64 nodes took ~25min | 17:28 |
| inc0 | so I'd guess circa 35? | 17:29 |
| sdake | goats | 17:29 |
| sdake | what an odd profession to choose | 17:29 |
| sdake | my cousin who has 20 mil became a rancher after he sold his business off | 17:29 |
| sean-k-mooney | inc0: thats pretty good scaling | 17:29 |
| inc0 | well compute nodes are faster than controller nodes | 17:29 |
| sdake | sean-k-mooney cmopute nodes take less time then control nodes to deploy | 17:30 |
| Mech422 | sdake: apperently, he makes world-renowned goat's milk cheese now... | 17:30 |
| inc0 | we'll see soon enough | 17:30 |
| *** diogogmt has joined #openstack-kolla | 17:30 | |
| sean-k-mooney | inc0: im guessing one of the bottlenecks will is coping the 5 or 6 imanges for the compute nodes | 17:30 |
| inc0 | yeah might be | 17:30 |
| inc0 | we'll see | 17:30 |
| Mech422 | sdake: but he made $$$$ on webmd, so .... | 17:30 |
| inc0 | also we should tweak ansible forks | 17:31 |
| inc0 | however this is as simple as it gets, we'll make series of videos, one of which will be "how fast it can get" | 17:31 |
| sean-k-mooney | inc0: it might be interesting to do a cleanup and redeploy the same config with the prechached images from the first deploy and compare the difference | 17:32 |
| inc0 | so with all the plays already done | 17:32 |
| inc0 | we can redeploy from bare metal at any point | 17:32 |
| inc0 | which is cool | 17:32 |
| inc0 | see? definetly we need more forks | 17:33 |
| sean-k-mooney | forks of kolla? | 17:34 |
| inc0 | no, of ansible | 17:35 |
| inc0 | so if we tweak ansible a bit it can be WAY faster | 17:35 |
| inc0 | after we're done with this one, I'll do the tweaks | 17:35 |
| sean-k-mooney | ah, got ya | 17:35 |
| sean-k-mooney | one thing i never looked into what does teh kolla-toolbox container actully do? | 17:39 |
| inc0 | sean-k-mooney, so historically we had to have ansible 2.0 for shade - ansible module for openstack | 17:42 |
| inc0 | so all keystone user creation and stuff could be done | 17:42 |
| sean-k-mooney | oh ok so now that we use ansibel 2 anyway is it still needed? | 17:43 |
| *** inc0_ has joined #openstack-kolla | 17:43 | |
| *** JoseMello has joined #openstack-kolla | 17:44 | |
| inc0_ | aaand deploy failed | 17:45 |
| inc0_ | aaaand I know why | 17:46 |
| Mech422 | grr...I have a really wierd ovs error where its not adding the physical interface to the ovs bridge on my dev nodes.. | 17:46 |
| Mech422 | ubuntu ovs is SUCH a PITA | 17:46 |
| *** inc0 has quit IRC | 17:46 | |
| inc0_ | so I'll tweak it up and restart transmission | 17:46 |
| inc0_ | also make this ansible tweaks | 17:46 |
| inc0_ | we forgot to add /etc/hosts | 17:47 |
| sean-k-mooney | add? | 17:48 |
| inc0_ | so rabbitmq can only work with hostnames | 17:49 |
| inc0_ | so all of controller nodes have to resolve it's hostname | 17:49 |
| inc0_ | each others* | 17:49 |
| britthouser | doh! | 17:49 |
| sean-k-mooney | ok i taught i had that in my kolla-host playbook | 17:49 |
| sean-k-mooney | oh wait they need to have the hostnames of all other controler nodes too? | 17:50 |
| inc0_ | yeah | 17:50 |
| inc0_ | every controller has to resolve every other controller | 17:50 |
| inc0_ | and make it api_interface | 17:51 |
| sean-k-mooney | ok well thats a change i need to make in the kolla-host playbook too. ill do that | 17:51 |
| britthouser | I've seen that with Ceph too....at least with ceph_ceploy | 17:51 |
| inc0_ | kk | 17:51 |
| britthouser | everyone needs to resolve the hostname of the MON ips | 17:51 |
| inc0_ | first outcome of our little testing! | 17:52 |
| britthouser | w00t! | 17:52 |
| inc0_ | also we should add this precheck to well..precheck | 17:52 |
| *** fragatin_ has quit IRC | 17:52 | |
| sean-k-mooney | inc0_: should i add entries for all nodes into /etc/hosts? | 17:53 |
| sean-k-mooney | i had that origiinally but took it out | 17:53 |
| inc0_ | yeah | 17:53 |
| *** fragatina has joined #openstack-kolla | 17:53 | |
| *** yee_ has joined #openstack-kolla | 17:55 | |
| britthouser | inc0 and pbourke I'm starting to wonder if we'd really get any gains by reformatting with different profiles? | 18:00 |
| *** Guest52040 has quit IRC | 18:00 | |
| britthouser | at least not until step3 | 18:00 |
| inc0_ | http://youtu.be/RNZMtym5x1c next try | 18:00 |
| britthouser | Lets just keep it simple and keep everyone on generic unless they are ceph | 18:00 |
| inc0_ | well we need lvm for cinder tho | 18:01 |
| *** shardy is now known as shardy_afk | 18:03 | |
| inc0_ | it's faster now | 18:04 |
| sean-k-mooney | inc0_: wow that does look faster | 18:04 |
| sean-k-mooney | what did you do? | 18:04 |
| inc0_ | these ansible tweaks | 18:05 |
| inc0_ | so we can acutally put this into your playbook | 18:05 |
| openstackgerrit | Duane DeCapite proposed openstack/kolla: Change cleanup to remove as cleanup is a misnomer https://review.openstack.org/351869 | 18:06 |
| inc0_ | what I did is to add 2 config options to ansible | 18:06 |
| *** jemcevoy has quit IRC | 18:06 | |
| *** michauds has quit IRC | 18:06 | |
| inc0_ | pipelining=True and forks=100 | 18:06 |
| sean-k-mooney | inc0_: ah cool | 18:06 |
| sean-k-mooney | could you set those in the global.yml or somting along those lines | 18:07 |
| inc0_ | hostnames, again | 18:07 |
| sean-k-mooney | rabiit again? | 18:07 |
| inc0_ | ah. | 18:08 |
| inc0_ | it checks hostname not fqdn | 18:09 |
| inc0_ | here we go again | 18:10 |
| inc0_ | sooo we need to add this to prechecks and remove from deploy | 18:11 |
| inc0_ | as it is really a precheck | 18:11 |
| *** daneyon has joined #openstack-kolla | 18:11 | |
| inc0_ | and also add this to kolla-host | 18:11 |
| *** daneyon has quit IRC | 18:16 | |
| inc0_ | nova done | 18:24 |
| inc0_ | I think we'll get less than 20min | 18:24 |
| *** yee_ has quit IRC | 18:24 | |
| Daviey | sdake: hey, you would mind readding your +2 (& +W) for visit https://review.openstack.org/337559 (it got lost because someone changed the commit message) | 18:25 |
| *** yee_ has joined #openstack-kolla | 18:26 | |
| sdake | Daviey looking | 18:27 |
| *** senk has joined #openstack-kolla | 18:27 | |
| sdake | Daviey acked | 18:30 |
| inc0_ | damn | 18:30 |
| inc0_ | DAMN! | 18:30 |
| inc0_ | more than 20min | 18:30 |
| inc0_ | 20min0.618s | 18:31 |
| sean-k-mooney | oh i was just going to say that looks like it succeded | 18:31 |
| sdake | inc0 was that af ull130 node deploy? | 18:31 |
| rhallisey | inc0_, done | 18:31 |
| inc0_ | yeah | 18:31 |
| inc0_ | sdake, | 18:31 |
| rhallisey | wow | 18:31 |
| sdake | wow | 18:31 |
| sdake | hotness | 18:31 |
| sdake | the 1$ queeetion is does it work :) | 18:31 |
| *** narasimha_SV has quit IRC | 18:32 | |
| sean-k-mooney | inc0_: didnt you say that a 64 node deploy took 25. those ansible tweaks seem to have helped alot | 18:32 |
| inc0_ | yeah | 18:32 |
| Daviey | sdake: thanks sir | 18:32 |
| sean-k-mooney | since rhallisey did a 5min build and that was a 20 min deploy that less then half an hour if you have the configs ready | 18:32 |
| inc0_ | yup | 18:33 |
| sdake | but we dont know how long bare metla takes? | 18:33 |
| inc0_ | on a pretty damn big cluster | 18:33 |
| Daviey | inc0_: what is the ansible server system load like? | 18:33 |
| inc0_ | sdake, much much longer...single node boots longer then this full deploy | 18:33 |
| inc0_ | Daviey, didint check, will check next time | 18:33 |
| sean-k-mooney | inc0_: how long did the cobbler install take? | 18:34 |
| Daviey | inc0_: Locally, i've seen it > 10.00.. on a medium sized VM | 18:34 |
| Daviey | Not sure if it is poor ansible tuning or something else | 18:34 |
| sdake | what is being deployed inc0 - compute kit? | 18:34 |
| inc0_ | Daviey, this is pretty powerful metal tho | 18:34 |
| inc0_ | sdake, yup, as default as it gets | 18:35 |
| inc0_ | vm is running | 18:36 |
| openstackgerrit | Merged openstack/kolla: Add Watcher default settings https://review.openstack.org/337559 | 18:36 |
| sdake | so answer is yes | 18:36 |
| *** gfidente has quit IRC | 18:36 | |
| inc0_ | and vm is pinging | 18:37 |
| inc0_ | damn my voice sounds horrible | 18:37 |
| sdake | inc0_ how many control/storage/compute nodes? | 18:37 |
| inc0_ | 4 controllers | 18:38 |
| inc0_ | shitload compute | 18:38 |
| inc0_ | some storage | 18:38 |
| britthouser | w00t! | 18:39 |
| rhallisey | inc0_, live stream gone | 18:39 |
| sean-k-mooney | https://youtu.be/RNZMtym5x1c | 18:39 |
| sean-k-mooney | you can still watch it back after the fact | 18:40 |
| inc0_ | yeah | 18:40 |
| sdake | is the storage doing anything? | 18:40 |
| sdake | any cinder or ceph configured? | 18:40 |
| inc0_ | let's record the shit out of kolla | 18:40 |
| inc0_ | I don't think so | 18:40 |
| inc0_ | cinder was disabled | 18:40 |
| sdake | so the storage nodes were a noop | 18:40 |
| sean-k-mooney | sdake: if they were just storage then yes | 18:41 |
| sean-k-mooney | inc0_: were the stoarge nodes also compute nodes? | 18:41 |
| inc0_ | no | 18:41 |
| inc0_ | but there are only few of storage | 18:41 |
| britthouser | Just a reminder I'm travel for work next week, but I'll check back and see where OSIC stuff is a week from monday, and pitch in where I can. | 18:42 |
| inc0_ | fuck, I'm proud of our project, and I mean really, really proud | 18:42 |
| britthouser | Yeah this is amazing results | 18:42 |
| *** senk has quit IRC | 18:44 | |
| sean-k-mooney | will i put the video link into the etherpad? | 18:44 |
| sdake | sean-k-mooney sure | 18:44 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla: Introduce a script to launch ovsdb-server process https://review.openstack.org/351784 | 18:45 |
| *** zhubingbing has joined #openstack-kolla | 18:46 | |
| inc0_ | sdake, 101 compute | 18:47 |
| zhubingbing | hi，who can help me，I don't know why he appears this Merge Conflict ；https://review.openstack.org/#/c/351027/ | 18:49 |
| patchbot | zhubingbing: patch 351027 - kolla - Add aodh role | 18:49 |
| openstackgerrit | David Wang proposed openstack/kolla-kubernetes: Added support for template variables https://review.openstack.org/351408 | 18:53 |
| sean-k-mooney | zhubingbing: i would guess a change has merged to master that your change is in conflict with | 18:55 |
| sean-k-mooney | zhubingbing: can you try rebaseing to master? | 18:55 |
| zhubingbing | no | 19:02 |
| zhubingbing | i try | 19:03 |
| zhubingbing | root@ip-172-31-22-78:/kolla# git rebase master | 19:05 |
| zhubingbing | Current branch bp/BP-sahara is up to date. | 19:05 |
| zhubingbing | it look so fine | 19:05 |
| *** egonzalez90 has quit IRC | 19:05 | |
| Daviey | rhallisey: hey.. I've been experimenting with co-locating compute and storage on the same nodes. I see on the OSIC pad you are down for this.. Does it interest you? | 19:06 |
| irtermite | inc0_: someone needs to bump their terminal size before tmuxing. ;) | 19:06 |
| pbourke | yo | 19:06 |
| pbourke | i saw some good news :) | 19:07 |
| pbourke | trying to watch the video, wont load though | 19:07 |
| irtermite | watching now | 19:07 |
| sean-k-mooney | zhubingbing: you need to do a get fetch then git rebase origin/master | 19:07 |
| zhubingbing | ok i try | 19:08 |
| sean-k-mooney | zhubingbing: git rebase master will just rebase to your local master which is likly behind upstream | 19:08 |
| zhubingbing | may be I git used improperly | 19:09 |
| inc0_ | pbourke, irtermite made few remarks about your terminal size...and I must agree:P | 19:09 |
| pbourke | ha | 19:10 |
| sean-k-mooney | inc0_: it was a little hard to see /read | 19:10 |
| irtermite | *sad trombone* | 19:11 |
| irtermite | pbourke: inc0_: http://www.tecmint.com/record-and-replay-linux-terminal-session-commands-using-script/ | 19:11 |
| irtermite | ;) | 19:11 |
| zhubingbing | git rebase --continue | 19:11 |
| zhubingbing | [detached HEAD 059c55c] Add sahara ansible roles | 19:11 |
| zhubingbing | 19 files changed, 415 insertions(+ | 19:11 |
| zhubingbing | next, i need user git review command? | 19:11 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add sahara ansible roles https://review.openstack.org/351294 | 19:12 |
| sean-k-mooney | zhubingbing: yes if you have resoved and commited any conflit then do git review | 19:12 |
| sean-k-mooney | i see you already have :) | 19:12 |
| zhubingbing | ok thanks | 19:12 |
| zhubingbing | I should be good to learn Git | 19:13 |
| rhallisey | Daviey, sure | 19:13 |
| zhubingbing | thank you sean-k | 19:13 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add aodh role https://review.openstack.org/351027 | 19:17 |
| zhubingbing | thank you once again sean-k-mooney -) | 19:17 |
| *** dwalsh has joined #openstack-kolla | 19:18 | |
| sean-k-mooney | zhubingbing: no worries hopefully that will reslove your merge conflict | 19:18 |
| zhubingbing | I hope this submission will not appear again merge conflict | 19:21 |
| zhubingbing | - -) | 19:21 |
| *** sdake has quit IRC | 19:26 | |
| *** sdake has joined #openstack-kolla | 19:28 | |
| sean-k-mooney | inc0_: pbourke. i just got my kolla-host playbook to pass precheck on centos and ubuntu16.04 is there a way to make /run mounted shared on ubuntu 14.04 that will survie a reboot? | 19:28 |
| sdake | sean-k-mooney there is | 19:28 |
| sdake | sean-k-mooney but i dont know how to do so | 19:28 |
| inc0_ | so I'm ok with requiring 16.04 from newton on | 19:29 |
| sean-k-mooney | the last i checked the docs only had the runtime way but ill look it up | 19:29 |
| inc0_ | by requiring I mean "if you have 14, still works but we're not going to specifically support it" | 19:30 |
| sean-k-mooney | inc0_: well 14.04 wont be supported with newton by cannonical so i think that makes sense | 19:30 |
| inc0_ | yeah, so don't try too hard to fix it | 19:30 |
| inc0_ | it's a non-blocker for me | 19:30 |
| sdake | inc0_ i htink we have jointly made that decision already but we should clarify via a vote on the ml | 19:30 |
| sean-k-mooney | well im fixing it because we are using it in the osic cluster more then anything else | 19:30 |
| sean-k-mooney | ill put in the runtime fix for now which is in our docs | 19:31 |
| inc0_ | sdake, yeah, that being said, we need people to move our containers to 16.04 | 19:33 |
| inc0_ | I think it was ccesario who did this | 19:33 |
| inc0_ | but it kinda died out and I haven't seen him around | 19:33 |
| sean-k-mooney | inc0_: do we know if anyting needs to be changed? | 19:34 |
| inc0_ | well sources list for sure | 19:34 |
| *** dwalsh has quit IRC | 19:34 | |
| inc0_ | apart from that hopefully nothing | 19:34 |
| inc0_ | but well..."hopefully" | 19:34 |
| openstackgerrit | zhubingbing proposed openstack/kolla: Add Barbibcan container dockerfile https://review.openstack.org/351822 | 19:35 |
| sean-k-mooney | ya i gues it would be worth a try. | 19:35 |
| *** zhubingbing has quit IRC | 19:40 | |
| sean-k-mooney | sdake: redarding the storage driver on centos. does aufs work? | 19:42 |
| Mech422 | inc0_: if I get my current project situated, I can play with ubuntu 16.04 - we're an ubuntu shop | 19:42 |
| inc0_ | would appreciate it | 19:42 |
| inc0_ | I'd gladly take it but I'm overloaded | 19:43 |
| *** yee_ has quit IRC | 19:43 | |
| Mech422 | inc0_: yeah - no promises, but I'll try to give it a shot | 19:43 |
| Mech422 | inc0_: also, looking at scale.io was added to my marching orders | 19:43 |
| *** dave-mccowan has quit IRC | 19:45 | |
| *** dave-mccowan has joined #openstack-kolla | 19:45 | |
| Daviey | sean-k-mooney: I think you are supposed to use overlayfs in CentOS. AIUI, AUFS support isn't even in RH/Centos | 19:46 |
| *** mliima has quit IRC | 19:46 | |
| inc0_ | also ubu16.04 doesn't like aufs I think | 19:46 |
| inc0_ | I run overlay there | 19:46 |
| Daviey | Mech422: I'm strong on Ubuntu, but I don't have the cycles. I can help with adhoc tasks, but i certainly couldn't own it | 19:47 |
| sean-k-mooney | Daviey: i was going to have overlay be the default but sdake said he did not think it was stable enough for production | 19:47 |
| Daviey | inc0_: aufs can die in a fire. | 19:47 |
| Mech422 | inc0_: oh? I've never actually setup a FS driver - I just use the default docker installed one | 19:47 |
| inc0_ | well, it's still better than devicemapper | 19:47 |
| sdake | Mech422 thats painful :) | 19:47 |
| inc0_ | oh my Mech422 | 19:47 |
| Daviey | sean-k-mooney: I'd prefer the suggestion of setting up btrfs via lvm for centos :) | 19:47 |
| inc0_ | Mech422, do docker info | 19:48 |
| sdake | the suggestion was raw btrfs | 19:48 |
| inc0_ | docker info | grep Storage Driver | 19:48 |
| Mech422 | last I heard, btrfs still wasn't quite stable ? | 19:48 |
| inc0_ | it's ok afaik | 19:48 |
| inc0_ | it's in kernel now | 19:48 |
| sean-k-mooney | i would prefer btrfs too but that is a little tricky if your os image is not installed with a btrfs filesystem | 19:48 |
| sdake | lets not compound that perception with lvm :) | 19:48 |
| Mech422 | inc0_: err..I don't have a docker machine up atm - I'm rebuilding cluster for old style deployment | 19:48 |
| Daviey | btrfs is the best of the options. It was announced today that RAID5/6 support was flakey, but the core stuff for what we want is fine | 19:49 |
| inc0_ | and I think xenial actually installs btrfs on server | 19:49 |
| sdake | sean-k-mooney make a /var/lib/docker that is btrfs | 19:49 |
| sdake | that is what i do | 19:49 |
| inc0_ | Mech422, if it's devicemapper | 19:49 |
| inc0_ | change it. | 19:49 |
| inc0_ | cleanse it with fire | 19:49 |
| sean-k-mooney | sdake: how do i do that without mounting a seperate partition there or a loopback file? | 19:49 |
| Daviey | sdake: LVM makes it simpler if you ask me... considering the default install patern for rpm distros is to use LVM.. We just need to tack on the end of that | 19:49 |
| Daviey | If we do that, it becomes almost as simple as loopback | 19:49 |
| sdake | if it works it wfm :) | 19:49 |
| sdake | i never test with lvm | 19:50 |
| sdake | i use raw btrfs | 19:50 |
| *** dave-mccowan has quit IRC | 19:50 | |
| *** gfidente has joined #openstack-kolla | 19:50 | |
| inc0_ | well you need a device for btrfs | 19:50 |
| inc0_ | that's the issue | 19:50 |
| Mech422 | inc0_: looks like ubuntu 14.04 defaults to aufs driver | 19:50 |
| sean-k-mooney | sdake: i use raw brtfs too but i also install my os that way | 19:50 |
| Daviey | well lvm itself is irrelevant.. just a different way of presenting the volume to btrfs | 19:50 |
| Mech422 | inc0_: at least, thats whats running on my home setup | 19:50 |
| inc0_ | aufs is kinda ok | 19:51 |
| inc0_ | I think older versions of docker defaulted to devicemapper | 19:51 |
| inc0_ | anyway, if you ever see devicemapper | 19:51 |
| inc0_ | burn it | 19:51 |
| Daviey | inc0_: I thought all default to dm? | 19:51 |
| sean-k-mooney | ubuntu 16.04 is back to devicemapper by default | 19:51 |
| inc0_ | duh | 19:51 |
| sean-k-mooney | inc0_: devicemapper is fine if you use a dedicated disk | 19:51 |
| Daviey | when I was at Canonical, we tried to negotiate with docker to stop using aufs | 19:52 |
| sean-k-mooney | inc0_: its the loopback file that kills performance | 19:52 |
| inc0_ | uf you use dedicated disk use btrfs | 19:52 |
| inc0_ | aufs can die too, as soon as overlay becomes a thing | 19:52 |
| inc0_ | when you have overlay, no need to aufs at all | 19:52 |
| Daviey | overlay is a thing.. it's mainline now | 19:52 |
| inc0_ | I know | 19:53 |
| inc0_ | I run overlay in 16 | 19:53 |
| Mech422 | our stuff is mainly for QA lab - so I'd rather be 'slow and stable' then 'fast with lots of edge conditions' | 19:53 |
| Mech422 | actual vm performance is not a huge deal for QA | 19:53 |
| inc0_ | well, I don't think you'd hit stability issues with btrfs | 19:54 |
| Mech422 | thats why I've shied away from btrfs...I keep hearing its awesome...till its not | 19:54 |
| Daviey | My understanding is that loopback with docker with deteriorate massively over time | 19:54 |
| inc0_ | what you'll get tho is 10x build boost | 19:54 |
| Daviey | but i don't have anything to substantiate that | 19:54 |
| sean-k-mooney | well the variable i added to the kolla-host playbook just take the storage driver name so you can use any of them | 19:54 |
| Mech422 | inc0_: hmm - might have to give it a try then | 19:54 |
| inc0_ | overlay is ceph go-to fs | 19:55 |
| inc0_ | so it's a good idea too | 19:55 |
| Daviey | Mech422: well you aren't storing anything persistent on the docker engine anyway are you? | 19:55 |
| Daviey | Mech422: Surely using ceph for storage.. the docker mount is largely ephemeral ? | 19:55 |
| Mech422 | Daviey: emm - not using docker in prod yet, so no... | 19:55 |
| Daviey | Mech422: I mean, for your kolla deployment | 19:55 |
| Mech422 | Daviey: umm - mysqld rabbitmq etc etc ? | 19:56 |
| Daviey | If the docker storage engine catches fire, you can blow it away and rebuild without data loss. right? | 19:56 |
| Daviey | Right, but they are clustered | 19:56 |
| Daviey | The only thing that isn't is mongodb, right? | 19:56 |
| Mech422 | inc0_: when did ceph go with overlay? I thought it was btrfs/xfs recommended ? | 19:56 |
| Daviey | Mech422: Sorry, i was at cross-purpose.. I mean your client data, such as kvm volumes | 19:56 |
| sean-k-mooney | Daviey: we dont configure our docker volumes to use a prestent backend so without clusering you would lose data | 19:56 |
| *** ppowell has joined #openstack-kolla | 19:57 | |
| Mech422 | Daviey: yeah - client data is all ceph backed | 19:57 |
| Daviey | sean-k-mooney: I mean, what permanent data needs to be in docker? | 19:57 |
| *** dave-mccowan has joined #openstack-kolla | 19:57 | |
| Daviey | A failed docker storage driver, is the same as the node catching fire. Just rebuild it. | 19:57 |
| sean-k-mooney | your mysql db for one | 19:57 |
| Daviey | sean-k-mooney: but that is clustered | 19:57 |
| *** kbaegis has quit IRC | 19:58 | |
| Daviey | So, unless you see 3 docker storage engines failing concurrently? | 19:58 |
| *** kbaegis has joined #openstack-kolla | 19:58 | |
| sean-k-mooney | yes hence why i said without clusering. i dont think there is anything that would be lost with a default kolla deploy except local heka logs | 19:59 |
| Mech422 | Daviey: I though btrfs/ceph/zfs all had issues when disks near capacity....which is something I'd expect to happen on all 3 nodes pretty much at the same time ? | 19:59 |
| Daviey | This is interesting reading for RH/Centos 7, and overlayfs - https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/7.2_Release_Notes/technology-preview-file_systems.html | 19:59 |
| inc0_ | sean-k-mooney, well, logs lands in named volume | 19:59 |
| inc0_ | db is on named volume | 19:59 |
| Daviey | Mech422: So, to counter that.. My 3 control nodes are on different configurations... two are virtualised and one is bare metal | 20:00 |
| sean-k-mooney | inc0_: yes but that named volume is not preseted so if you have local storage issue its gone | 20:00 |
| Daviey | Mech422: so yes, if they are identical.. i agree | 20:00 |
| Mech422 | Daviey: mine are all basically clones | 20:00 |
| Daviey | Mech422: then doesn't matter what storage driver you are using, you are risking an outage with simple disk filling up | 20:02 |
| Mech422 | Daviey: true - but I can monitor absolute disk usage, I can't monitor "how much does this want to not freak out" | 20:03 |
| Mech422 | Daviey: as that tends to be variable depending on cluster usage | 20:03 |
| Daviey | meh, i think for our usage btrfs is pretty well tested. | 20:03 |
| Daviey | We aren't pushing the envelope or using advanced RAID functionality where it is weaker | 20:04 |
| Mech422 | Daviey: I'll give it a shot when I have a chance...I just have way too many other things that are higher priority atm | 20:04 |
| Daviey | :).. find that 25th hour in the day | 20:06 |
| Mech422 | Daviey: btw - now canonical is shipping zfs, I wonder how uptake/usage of the 2 will differ | 20:06 |
| Mech422 | zfs seemed a lot more 'battle tested', but btrfs has a really nice feature set, and a lot of followers | 20:07 |
| inc0_ | so on that note, I wonder if you could use rbd for named volume in docker | 20:07 |
| Daviey | Mech422: yeah! (I use zfs on ubuntu home server, been rock solid for 3 years) | 20:07 |
| Daviey | inc0_: funny you say that... I almost did a few mins ago.. but thought it might be a bit too insidious | 20:07 |
| inc0_ | why is that? | 20:08 |
| Daviey | inc0_: it isn't an entirely silly idea. Makes the node more stateless and removes the issue we have right now of setting up the filesystem | 20:08 |
| Daviey | it isn't like the container has high write throughput | 20:09 |
| sean-k-mooney | inc0_: ya i looked at that breifly | 20:10 |
| *** kbaegis has quit IRC | 20:11 | |
| inc0_ | Daviey, that's what we'll be doing in k8s | 20:11 |
| sean-k-mooney | inc0_: the issue is with the ceph containers | 20:11 |
| inc0_ | yeah, it has to be external ceph | 20:11 |
| inc0_ | or ceph logs would be only ones not on ceph | 20:11 |
| inc0_ | because well...egg chicken?;) | 20:11 |
| sean-k-mooney | inc0_: i said it before but it might be worth executing the ceph role form kolla-host and then using external ceph with kolla ansible at some point | 20:12 |
| sean-k-mooney | inc0_: the ceph logs would still have to be stored on the host though | 20:14 |
| inc0_ | well, they will go through heka to elastic | 20:14 |
| inc0_ | anyway, ceph is not stateless at all | 20:14 |
| inc0_ | as osd really don't like to be moved around | 20:14 |
| inc0_ | doesnt | 20:14 |
| *** yee_ has joined #openstack-kolla | 20:17 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla: Updated from global requirements https://review.openstack.org/351901 | 20:22 |
| *** michauds has joined #openstack-kolla | 20:25 | |
| openstackgerrit | Patrick Powell proposed openstack/kolla: Add Nova sanity check https://review.openstack.org/347416 | 20:26 |
| *** mewald has quit IRC | 20:29 | |
| *** yee_ has quit IRC | 20:29 | |
| openstackgerrit | Patrick Powell proposed openstack/kolla: Add Heat sanity check https://review.openstack.org/350071 | 20:30 |
| Mech422 | inc0_: sdake: that reminds me - now pbourke is back, we should figure out the sgdisk thing in the ceph startup | 20:30 |
| Mech422 | its hard coded to look for /usr/sbin/sgdisk but for ubuntu its /sbin/sgdisk, and sdake doesn't want to use shell path to find it | 20:30 |
| *** dwalsh has joined #openstack-kolla | 20:33 | |
| Mech422 | personally, I think thats what the shell path is for - if someone owns your containers, you have bigger issues - but we could just test existance of /usr/sbin/sgdisk and fallback to /sbin/sgdisk if it doesn't exist | 20:34 |
| Daviey | inc0_: Sorry, when i said stateless.. i meant the docker storage, not the osd's | 20:34 |
| inc0_ | yeah, but we deploy ceph as well | 20:34 |
| Mech422 | just not sure where other distros put the binary ? could end up with a bunch of ladder logic looking for it ? | 20:34 |
| inc0_ | discussion was, how to use ceph we deploy for our own deployment;) | 20:35 |
| *** Jeffrey4l has quit IRC | 20:35 | |
| *** Jeffrey4l has joined #openstack-kolla | 20:36 | |
| Mech422 | inc0_: I like the idea of 'pre-deploying' ceph, then just always pretending ceph is 'external' | 20:37 |
| Mech422 | inc0_: it gives a single code path, without having to do internal vs external checks | 20:38 |
| *** jtriley has quit IRC | 20:38 | |
| *** shardy_afk is now known as shardy | 20:38 | |
| inc0_ | yeah, that's the idea | 20:38 |
| *** ppowell has quit IRC | 20:42 | |
| Mech422 | Daviey: btw, have you noticed any problems rebooting ubuntu boxes if you radically change /etc/network/interfaces ? | 20:43 |
| Mech422 | Daviey: I use ansible to convert from a linux networking to openvswitch network setup, and it always seems to get stuck on reboot trying to shutdown the network | 20:43 |
| *** Satya__ has quit IRC | 20:44 | |
| Mech422 | Daviey: its understandable its confused that interfaces disappeared/changed/whatever - but annoying it prevents the reboot | 20:44 |
| Daviey | Mech422: what do you mean by radically ? | 20:44 |
| Daviey | Mech422: i don't think i've ever used ovs on the bare metal | 20:45 |
| Mech422 | Daviey: ahh | 20:45 |
| *** banix has quit IRC | 20:45 | |
| Daviey | Mech422: why are you OVS'ing the bare metal? | 20:46 |
| Mech422 | Daviey: because I can ? | 20:46 |
| Mech422 | Daviey: production uses ovs bridges, bonds, vlans, etc | 20:46 |
| Mech422 | all on bare metal | 20:46 |
| Daviey | ah, i see | 20:47 |
| Mech422 | Daviey: its just a pet peeve - my foreman preseeds setup 'normal' linux networking - I just need to use that 25th hour in the day to create an ovs based preseed | 20:48 |
| Daviey | Mech422: what connects to the ovs bridge? | 20:48 |
| Mech422 | Daviey: err..in what way ? production has a bonded pair on one bridge, along with vlans...then I have an OVSPatch to an integration bridge for Openstack | 20:49 |
| Daviey | Mech422: I suppose i am wondering if you are using it for more than the neutron bridge? | 20:50 |
| Mech422 | Daviey: yes - since eth0/eth1 are bonded, the host mgmt interface is a vlan via ovs | 20:51 |
| *** dwalsh has quit IRC | 20:52 | |
| Mech422 | Daviey: along with vlans for ceph/database replication, etc | 20:52 |
| openstackgerrit | Merged openstack/kolla-kubernetes: Add doc for configuring private docker registry https://review.openstack.org/350785 | 20:52 |
| Mech422 | Daviey: (this was a 'vanilla' openstack liberty install, so it doesn't use kolla standards... ) | 20:52 |
| Daviey | Mech422: So i use a bonded interface with vlan, and let neutron create the bridge to bring it into openstack | 20:53 |
| Mech422 | Daviey: cool - that part works for me :-) | 20:53 |
| *** daneyon has joined #openstack-kolla | 20:54 | |
| Mech422 | Daviey: its just annoying my dev. boxes don't reboot if I play fast and loose with /etc/network/interfaces | 20:54 |
| Mech422 | I kinda think reboot should be "reboot!" (reboot dammit!) | 20:54 |
| Mech422 | ipmi reboot fixes it though | 20:54 |
| Daviey | Mech422: yeah.. weird.. no idea.. but you are right | 20:56 |
| Daviey | Mech422: anything left on any of the tty's when it is wedged? | 20:57 |
| *** dwalsh has joined #openstack-kolla | 20:57 | |
| *** daneyon has quit IRC | 20:58 | |
| *** shardy has quit IRC | 20:58 | |
| openstackgerrit | David Wang proposed openstack/kolla-kubernetes: Added support for template variables https://review.openstack.org/351408 | 20:59 |
| Mech422 | Daviey: normal console shutdown messages then I think it get stuck on 'Shutting down networking' or something to that effect | 21:02 |
| Mech422 | Daviey: its understandable - but it really should timeout and just finish rebooting | 21:03 |
| openstackgerrit | sean mooney proposed openstack/kolla: add baremetal role to install kolla deps. https://review.openstack.org/325631 | 21:04 |
| openstackgerrit | sean mooney proposed openstack/kolla: This change allow setting kernel commandline args https://review.openstack.org/351412 | 21:04 |
| *** bmace has joined #openstack-kolla | 21:04 | |
| *** yee_ has joined #openstack-kolla | 21:04 | |
| sean-k-mooney | i still need to add the host name generate but the kolla-host playbook should now work on ubuntu 14.04,16.04 and centos7.2 | 21:05 |
| Daviey | sean-k-mooney: I'm not sure i'm a fan of it.. but Ubuntu MAAS uses this for hostname generation, https://pypi.python.org/pypi/petname | 21:07 |
| *** egonzalez90 has joined #openstack-kolla | 21:07 | |
| sean-k-mooney | Daviey: oh that is not what i ment | 21:08 |
| sean-k-mooney | i need to populate the ip to hostname mapping in /etc/hosts for each of the contolernodes | 21:08 |
| sean-k-mooney | otherwise rabbitmq will fail | 21:08 |
| Daviey | sean-k-mooney: Oh... that | 21:09 |
| *** JoseMello has quit IRC | 21:10 | |
| sean-k-mooney | i had code to do it in my first commit https://review.openstack.org/#/c/325631/1/ansible/roles/baremetal/templates/hosts.j2 but i removed it | 21:10 |
| sean-k-mooney | i need to look at marks suggtion and see if i can use it | 21:11 |
| *** sdake has quit IRC | 21:12 | |
| patchbot | sean-k-mooney: patch 325631 - kolla - add baremetal role to install kolla deps. | 21:12 |
| *** inc0_ has quit IRC | 21:13 | |
| *** yee_ has quit IRC | 21:13 | |
| *** michauds has quit IRC | 21:13 | |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla: Introduce a script to launch ovsdb-server process https://review.openstack.org/351784 | 21:15 |
| *** yee_ has joined #openstack-kolla | 21:16 | |
| openstackgerrit | Prithiv proposed openstack/kolla: Dockerfiles for Networking-SFC Container https://review.openstack.org/344930 | 21:18 |
| openstackgerrit | Prithiv proposed openstack/kolla: Dockerfiles for Networking-SFC Container https://review.openstack.org/344930 | 21:20 |
| *** dwalsh has quit IRC | 21:21 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Add missing project_yaml parameter to role metadata https://review.openstack.org/351928 | 21:22 |
| *** jtriley has joined #openstack-kolla | 21:23 | |
| openstackgerrit | Prithiv proposed openstack/kolla: Dockerfiles for Networking-SFC Container https://review.openstack.org/344930 | 21:24 |
| *** michauds has joined #openstack-kolla | 21:26 | |
| openstackgerrit | Christian Berendt proposed openstack/kolla: Enable CADF events in Keystone and introduce kolla_enable_enhanced_security https://review.openstack.org/349006 | 21:31 |
| *** ad_rien_ has joined #openstack-kolla | 21:42 | |
| *** yee_ has quit IRC | 21:51 | |
| *** egonzalez90 has quit IRC | 21:52 | |
| *** yee_ has joined #openstack-kolla | 21:53 | |
| *** sdake has joined #openstack-kolla | 21:54 | |
| *** jtriley has quit IRC | 21:56 | |
| sdake | 50f in the morning 115 in the afternoon | 21:57 |
| sdake | brutal temperature chane | 21:57 |
| sdake | change | 21:57 |
| sdake | inc0 ping | 22:05 |
| sdake | guess not around | 22:05 |
| *** signed8b_ has joined #openstack-kolla | 22:06 | |
| *** signed8bit has quit IRC | 22:09 | |
| *** ad_rien_ has quit IRC | 22:16 | |
| *** gfidente has quit IRC | 22:22 | |
| wirehead_ | yeah, I can get away without the AC today. Nice. | 22:24 |
| *** dave-mccowan has quit IRC | 22:24 | |
| *** ljjjustin has quit IRC | 22:33 | |
| *** ljjjustin has joined #openstack-kolla | 22:37 | |
| *** yee_ has quit IRC | 22:40 | |
| *** banix has joined #openstack-kolla | 22:41 | |
| *** imcsk8 has quit IRC | 22:50 | |
| *** imcsk8 has joined #openstack-kolla | 22:50 | |
| openstackgerrit | Revanth proposed openstack/kolla: Customizations for kolla-toolbox https://review.openstack.org/351946 | 22:54 |
| *** signed8b_ is now known as signed8bit_Zzz | 22:56 | |
| openstackgerrit | Ken Wronkiewicz proposed openstack/kolla-kubernetes: Fix Nova bootstrap https://review.openstack.org/346452 | 23:08 |
| *** rhallisey has quit IRC | 23:13 | |
| *** sdake has quit IRC | 23:31 | |
| *** srwilkers has joined #openstack-kolla | 23:43 | |
| *** michauds has quit IRC | 23:44 | |
| *** ljjjustin has quit IRC | 23:45 | |
| openstackgerrit | David Wang proposed openstack/kolla-kubernetes: Add Ceph support for Mariadb https://review.openstack.org/350312 | 23:46 |
| *** ljjjustin has joined #openstack-kolla | 23:49 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!