| sdake | rhallisey almost done reviewing your patch | 00:00 |
|---|---|---|
| rhallisey | k | 00:00 |
| sdake | jenkins out to lucnh | 00:01 |
| *** achanda has quit IRC | 00:01 | |
| *** alisonh has joined #kolla | 00:03 | |
| sdake | ok comments are done rhallisey | 00:05 |
| sdake | when is a new version coming ;) | 00:05 |
| *** dims_ has joined #kolla | 00:06 | |
| rhallisey | tmr maybe | 00:06 |
| sdake | you going to commit to doing the whole source base ;) | 00:07 |
| *** dims has quit IRC | 00:07 | |
| rhallisey | ya I can do the whole thing | 00:08 |
| rhallisey | just want to have a template done | 00:08 |
| sdake | lets get the job done on one of them | 00:08 |
| sdake | we can possibly dsitribute the wor around | 00:08 |
| sdake | to speed things up | 00:08 |
| sdake | the reference implementation of config.py needs love | 00:08 |
| sdake | i'd like to see the json as a list rather then dictionary | 00:08 |
| sdake | if that is possible | 00:09 |
| rhallisey | kk | 00:10 |
| sdake | the goal is to remove files | 00:10 |
| sdake | and remove cruft | 00:11 |
| sdake | that is why we are doing this | 00:11 |
| sdake | the side benefit is multiple customized config files | 00:11 |
| sdake | pleasa try to address all my comments :) | 00:11 |
| sdake | use click "done" in the review when done with each one :) | 00:11 |
| sdake | i'd really like our contributors to start responding to reviewers comments | 00:12 |
| sdake | its hard to know what got fixed | 00:12 |
| sdake | brb | 00:12 |
| *** dims has joined #kolla | 00:12 | |
| *** sdake_ has joined #kolla | 00:13 | |
| *** blahRus has quit IRC | 00:13 | |
| *** dims_ has quit IRC | 00:15 | |
| *** sdake has quit IRC | 00:16 | |
| rhallisey | sdake_, kk | 00:17 |
| sdake_ | amazing kolla queue down to 1 page | 00:18 |
| rhallisey | ya | 00:18 |
| *** ccrouch1 has quit IRC | 00:20 | |
| *** weiyu_ has joined #kolla | 00:23 | |
| *** diogogmt has joined #kolla | 00:23 | |
| *** openstack has joined #kolla | 00:35 | |
| sdake_ | which part dont you understand the bindmount ? | 00:35 |
| sdake_ | take nova-compute as an eaple | 00:36 |
| sdake_ | /opt/kolla/configs/nova-compute/nova.conf is bindmounted | 00:36 |
| sdake_ | with /opt/kolla/config.json | 00:36 |
| sdake_ | /opt/kolla/configs/cnfig.json would have tobe bindmounted | 00:36 |
| sdake_ | basically it needs to end in the same directory as nova-compute on start of the nova-compute container on the host | 00:36 |
| *** alisonh has joined #kolla | 00:38 | |
| rhallisey | sdake_, I still don't understand your point. When will this fail? | 00:38 |
| sdake_ | if you remove config.json from the dockerfile where does it com from | 00:39 |
| sdake_ | i really don't want config.json in the dockerfile | 00:39 |
| *** diogogmt has quit IRC | 00:39 | |
| rhallisey | ok that's your point | 00:39 |
| rhallisey | gotcha | 00:39 |
| *** diogogmt has joined #kolla | 00:40 | |
| *** sdake has joined #kolla | 00:56 | |
| *** sdake_ has quit IRC | 00:59 | |
| *** weiyu_ has quit IRC | 01:08 | |
| Kennan | hi kolla guys | 01:27 |
| rhallisey | Kennan, hello | 01:27 |
| sdake | sup knenan | 01:27 |
| Kennan | I tried found a ubuntu and updated kernel to 3.19 | 01:27 |
| Kennan | I build from source | 01:27 |
| Kennan | tools/build.py --base ubuntu --type source --template -T 35 | 01:28 |
| Kennan | as this | 01:28 |
| Kennan | and found | 01:28 |
| sdake | right | 01:28 |
| Kennan | ERROR:__main__:cinder-ba Failed with status: parent_error | 01:28 |
| Kennan | ERROR:__main__:cinder-ap Failed with status: parent_error | 01:28 |
| Kennan | ERROR:__main__:cinder-vo Failed with status: parent_error | 01:28 |
| Kennan | ERROR:__main__:cinder-ba Failed with status: error | 01:28 |
| Kennan | ERROR:__main__:kolla-ans Failed with status: error | 01:28 |
| Kennan | ERROR:__main__:cinder-sc Failed with status: parent_error | 01:28 |
| sdake | that works according to SamYaple :) | 01:28 |
| Kennan | those images failed to build | 01:28 |
| sdake | yes i ontice cinder is failing to build as well | 01:28 |
| sdake | i htink someone broke it | 01:28 |
| sdake | just ignor eit, cinder is off by default | 01:28 |
| sdake | you need kolla-ansible | 01:28 |
| Kennan | ok. anyway, let me try verify it other containers work or not | 01:29 |
| sdake | try adding kolla-ansible to your build command | 01:29 |
| sdake | that should get it done | 01:29 |
| sdake | you absolutely must have kolla-ansible | 01:29 |
| Kennan | hi sadke: not sure where to add it ? | 01:30 |
| *** erkules has joined #kolla | 01:30 | |
| Kennan | hi sdake: for kolla-ansible | 01:31 |
| Kennan | what's it for build images? not clear about it | 01:32 |
| Kennan | it seems ./tools/kolla-ansible deploy is for deploy containers | 01:32 |
| sdake | tools/build.py --base ubuntu --type source --template -T 35 kolla_ansible | 01:32 |
| sdake | that will rebuild kolla ansible i think ;) | 01:32 |
| *** dolpher has joined #kolla | 01:32 | |
| sdake | run with -T 1 | 01:32 |
| sdake | as well so you can see what fialed | 01:32 |
| sdake | and paste to a pastebin the log of the fialure | 01:32 |
| sdake | upstream may be busted | 01:33 |
| sdake | sam been jamming patches in at a fast rate | 01:33 |
| sdake | i notice our gate i failing now too on from source builds | 01:33 |
| *** erkules_ has quit IRC | 01:33 | |
| Kennan | ok sdake: let me try clean old images and build again | 01:34 |
| Kennan | will try use your instructions here | 01:34 |
| Kennan | tools/build.py --base ubuntu --type source --template -T 35 kolla_ansible | 01:34 |
| sdake | dont clean old images | 01:34 |
| sdake | keep your images | 01:34 |
| sdake | just reuibd kolla_ansible | 01:34 |
| sdake | that is one rason we added the regex building feature | 01:35 |
| Kennan | ok | 01:35 |
| Kennan | INFO:__main__:Successfully built images | 01:35 |
| Kennan | INFO:__main__:========================= | 01:35 |
| Kennan | INFO:__main__:Images that failed to build | 01:35 |
| Kennan | INFO:__main__:=========================== | 01:35 |
| Kennan | is kolla_ansible right here? seems not any images have such name | 01:36 |
| sdake | try \*ansible\* | 01:37 |
| Kennan | oh, seems all invalid expression | 01:38 |
| Kennan | raise error, v # invalid expression | 01:39 |
| Kennan | it seesm kolla-ansible is OK | 01:42 |
| Kennan | not kolla_ansible :0 | 01:42 |
| sdake | run sudo docker images | grep ansible and paste results | 01:42 |
| *** weiyu_ has joined #kolla | 01:42 | |
| sdake | i don't know regex very well unfortuantely | 01:42 |
| sdake | maybe you an google how to do aw ildcard for *ansible* :) | 01:43 |
| *** zhiwei has joined #kolla | 01:44 | |
| Kennan | hi sdake: I have found the issue, seems code bug | 01:49 |
| Kennan | I would send bug fix if other not fix it | 01:49 |
| sdake | we take git reviews :) | 01:50 |
| sdake | please file a bug first | 01:50 |
| sdake | and uses Closes-Bug: #XYZ at end of commit message | 01:50 |
| Kennan | sure :) | 01:50 |
| *** Haomeng has joined #kolla | 01:56 | |
| *** Haomeng|2 has quit IRC | 01:59 | |
| openstackgerrit | Kai Qiang Wu(Kennan) proposed stackforge/kolla: Fix the wrong instruction for pip https://review.openstack.org/216924 | 02:02 |
| sdake | kennan thanks for the contribution bro! | 02:04 |
| Kennan | it is my pleasure sdake: | 02:04 |
| *** unicell has quit IRC | 02:08 | |
| *** achanda has joined #kolla | 02:23 | |
| *** jasonsb has quit IRC | 02:28 | |
| *** loth has quit IRC | 02:28 | |
| *** ccrouch has joined #kolla | 02:37 | |
| *** achanda has quit IRC | 02:38 | |
| *** dims has quit IRC | 02:46 | |
| *** klint has joined #kolla | 02:52 | |
| SamYaple | hey sdake | 02:55 |
| sdake | yosam | 02:55 |
| SamYaple | did you see my build.py patch? | 02:55 |
| SamYaple | should make everyone stop complaining | 02:55 |
| sdake | not sure i did | 02:56 |
| sdake | say rhallisey hada good idea for config..json worth a review | 02:56 |
| sdake | his code is pretty trrible :) | 02:56 |
| sdake | just be warned | 02:56 |
| SamYaple | https://review.openstack.org/#/c/216691/ | 02:56 |
| SamYaple | sdake: yea i looked at that review | 02:56 |
| SamYaple | i thought I was handling that issue in RC1 | 02:56 |
| SamYaple | we would also need to tweak common.sh to make it compatile until all images are converted | 02:57 |
| SamYaple | good start though | 02:57 |
| sdake | you are samyaple | 02:57 |
| sdake | i think he is concerned it isn't done today right now :) | 02:57 |
| SamYaple | anyway check out the review i linked | 02:58 |
| sdake | ya let me finish with slideware | 02:58 |
| sdake | almost done | 02:58 |
| SamYaple | also makes builds run 20% faster on my machine! | 02:58 |
| sdake | i looked at it briefly looked like alot of change | 02:58 |
| sdake | and went "meh" :) | 02:58 |
| sdake | what preisely doe it fix | 02:58 |
| SamYaple | it is, had to rewrite a big chunk of code to remove the build tiering system | 02:58 |
| SamYaple | basically each parent knows about its children now and will add its children to the work queue once it has been processed | 02:59 |
| sdake | so how do you uild base ahead of somethig nelse? | 02:59 |
| SamYaple | rather than a tiering system | 02:59 |
| SamYaple | ^^ | 02:59 |
| sdake | as long as base builds ahead of osmethin else wfm | 02:59 |
| sdake | but the gate is failing | 02:59 |
| sdake | there is a patch from kennan which may or may not fix it | 02:59 |
| SamYaple | i see the queue initially with images who's parent is None | 02:59 |
| SamYaple | what do you mean the gate is failing? | 03:00 |
| sdake | cool let me dirnk a red bull and i'll have a look :) | 03:00 |
| sdake | source gate is not building | 03:00 |
| SamYaple | it failed on cinder for that patch | 03:00 |
| sdake | cinder fials to bulid | 03:00 |
| sdake | for which patch | 03:00 |
| SamYaple | yea not related to this patch though | 03:00 |
| SamYaple | https://review.openstack.org/#/c/216691/ | 03:00 |
| SamYaple | remove tiering | 03:00 |
| sdake | i saw that | 03:00 |
| sdake | i mean which cinder patch is busted | 03:00 |
| SamYaple | yea | 03:01 |
| SamYaple | i have teh setup-gate working! | 03:01 |
| SamYaple | going to push a final patch and recheck 5-10 times | 03:01 |
| sdake | recheck doen't work if it passes the gate I dont think | 03:01 |
| SamYaple | it does | 03:01 |
| *** rhallisey has quit IRC | 03:02 | |
| SamYaple | the build tiering patch also cleans up logs fixes regex and properly fixes the test_build.py | 03:02 |
| SamYaple | rather than ignoring the issue like the regex patch did | 03:02 |
| openstackgerrit | Merged stackforge/kolla: Fix the wrong instruction for pip https://review.openstack.org/216924 | 03:03 |
| sdake | is heat going to somehow magically land on my desk before 8am :) | 03:03 |
| SamYaple | its going to land in the review queue | 03:04 |
| SamYaple | send me a sinple heat template btw for me to test | 03:04 |
| sdake | in working shape? | 03:04 |
| SamYaple | like a brownbag of commands | 03:04 |
| Kennan | hi guys. I noticed the change in /etc/kolla/nova-compute/nova.conf manully can not help. as every time run anisble deploy would override it. any suggestions? | 03:04 |
| SamYaple | yea | 03:04 |
| sdake | in the demos directory | 03:04 |
| SamYaple | Kennan: update it from the deploy host | 03:04 |
| sdake | kennan that should not be the case | 03:04 |
| SamYaple | sdake: it is | 03:04 |
| SamYaple | he is talking about editing the target host | 03:05 |
| sdake | ya don't edit the target host | 03:05 |
| sdake | edit the deploy host | 03:05 |
| Kennan | hi, I used all-in-one case | 03:05 |
| sdake | in that case editing that file should be fine | 03:05 |
| Kennan | and I changed /etc/kolla/ to use qemu | 03:05 |
| SamYaple | no sdake it wont | 03:05 |
| sdake | did you examine the internals of the container? | 03:05 |
| SamYaple | Kennan: if you want a permanat change /etc/kolla/config/nova/nova-compute.conf | 03:05 |
| sdake | i am pretty sure that is what file i said | 03:05 |
| sdake | kennan but if its not my apologies | 03:05 |
| Kennan | no issue. But could you guys tell me how to edit that to use qemu ? | 03:06 |
| Kennan | I did not have kvm enabled in VM | 03:06 |
| sdake | Modify /etc/kolla/nova/nova-compute.conf | 03:06 |
| sdake | Add the override: | 03:06 |
| sdake | [libvirt] | 03:06 |
| sdake | virt-type=qemu | 03:06 |
| sdake | itmay be virt_type | 03:06 |
| sdake | need to check the intranets | 03:06 |
| Kennan | sdake: it can not work. as rm containers, and deploy again, it would back to old | 03:07 |
| sdake | virt_type | 03:07 |
| Kennan | so the virt_type can not be kept | 03:07 |
| sdake | kennan that is not how it is supposed to work | 03:07 |
| sdake | /etc/kolla/nova/nova-compute.conf right? | 03:08 |
| SamYaple | /etc/kolla/config/nova/nova-compute.conf | 03:08 |
| Kennan | sdake: it is | 03:08 |
| SamYaple | sdake: wrong file | 03:08 |
| Kennan | it is /etc/kolla/nova-compute/nova.conf | 03:08 |
| SamYaple | /etc/kolla/config/nova/nova-compute.conf | 03:08 |
| sdake | i thought /etc/kolla/nova-compute/nova.conf was the default override file? | 03:08 |
| SamYaple | no, never has been | 03:08 |
| SamYaple | /etc/kolla/config/nova/nova-compute.conf | 03:08 |
| Kennan | ok SamYaple | 03:09 |
| Kennan | let me try that | 03:09 |
| sdake | groan that is confusing | 03:09 |
| SamYaple | only on AIO | 03:09 |
| SamYaple | we went over thisw | 03:09 |
| Kennan | hi | 03:09 |
| Kennan | -rw-r--r-- 1 root root 0 Aug 26 02:03 /etc/kolla/config/nova/nova-compute.conf | 03:09 |
| Kennan | ? | 03:09 |
| SamYaple | Kennan: correct | 03:09 |
| Kennan | it is zero size | 03:09 |
| SamYaple | you put your override settings in there | 03:09 |
| Kennan | ok | 03:09 |
| sdake | kennan my apologies dude | 03:09 |
| sdake | still learning my way around the syste m;-) | 03:09 |
| Kennan | no issue :) sdake: | 03:10 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 03:10 |
| *** jasonsb has joined #kolla | 03:13 | |
| openstackgerrit | Kai Qiang Wu(Kennan) proposed stackforge/kolla: Fix doc for sudo case https://review.openstack.org/216941 | 03:17 |
| SamYaple | sdake: both centos and fedora are working for builds in both rax and hpcloud servers! | 03:18 |
| SamYaple | properly using the dedicated ephemeral disk as well | 03:19 |
| Kennan | SamYaple: is it build from source work ? | 03:19 |
| Kennan | for binary, i am not sure(as I hit issue yesterday) | 03:20 |
| SamYaple | Kennan: no this is for the gating system for commits. We are setting up the gates to do ubuntu and centos binary and source installs | 03:20 |
| Kennan | OK SamYaple: just want to learn what's gate system like, for example, all is ubuntu OS ? or | 03:21 |
| SamYaple | right now we have 3 build gates, 1 fedora (binary only) 2 centos (binary and source) | 03:22 |
| SamYaple | once my patch merges we will have another ubuntu source gate | 03:22 |
| Kennan | Good to know, what's public cloud os look like ? I mean when hpcloud or rax allocate servers | 03:22 |
| Kennan | is it VM or bare metal? | 03:23 |
| SamYaple | vm | 03:23 |
| SamYaple | you can go launch them by hand in the differnet clouds | 03:23 |
| SamYaple | they are pretty vanilla images | 03:23 |
| Kennan | Ok SamYaple: final question, what's OS for the VM ? and version (I choose ubuntu 14.04 with 3.19 kernel as you suggested :) | 03:23 |
| Kennan | I failed on centos 7.1 before | 03:24 |
| SamYaple | for the gate? | 03:24 |
| Kennan | yes | 03:24 |
| SamYaple | ubuntu 14.04 3.13 kernel for the ubuntu builds, centos 7 3.10 kernel for centos builds | 03:24 |
| SamYaple | I cant control the kernel in the gate | 03:25 |
| Kennan | ok. SamYaple: I remember ubuntu 14.04 3.13 kernel need with btrfs for docker ? | 03:25 |
| SamYaple | yes | 03:25 |
| SamYaple | or a statically built docker | 03:26 |
| Kennan | seems btrfs needs some manual configuraion in docker side, perhap I need read your code to learn more details | 03:26 |
| SamYaple | Kennan: yea, but just normal you can do `docker daemon -s btrfs` and it would work | 03:26 |
| Kennan | oh? I thought it would have a separate disk and resizefs with btrfs | 03:27 |
| sdake | what does --base-tag do? | 03:27 |
| SamYaple | Kennan: it does | 03:27 |
| SamYaple | just after yo uset that up | 03:27 |
| SamYaple | sdake: https://github.com/stackforge/kolla/blob/master/docker_templates/base/Dockerfile.j2#L1 | 03:28 |
| Kennan | ok SamYaple: if I understand your point clearly, you need have related disk resizefs with btrfs, and then can run with docker daemon -s btrfs | 03:28 |
| Kennan | right? | 03:28 |
| SamYaple | Kennan: yes | 03:28 |
| SamYaple | format disk to btrfs, mount at /var/lib/docker, start docker with `docker daemon -s btrfs` | 03:29 |
| Kennan | ok. I would check your gate code, and try to understand how you did that to make ubuntu 14.04 with 3.13 kernel work | 03:29 |
| Kennan | SamYaple: did you have such link for gate code your changes? | 03:29 |
| SamYaple | i do | 03:30 |
| SamYaple | https://review.openstack.org/#/c/213397/29/tests/setup_centos.sh | 03:30 |
| Kennan | thanks | 03:30 |
| *** achanda has joined #kolla | 03:30 | |
| *** achanda has quit IRC | 03:35 | |
| *** diga has joined #kolla | 03:40 | |
| *** unicell has joined #kolla | 03:52 | |
| openstackgerrit | Zhou Zheng Sheng proposed stackforge/kolla: Speedup Git Cloning Ansible Repository https://review.openstack.org/216948 | 03:54 |
| *** unicell has quit IRC | 04:07 | |
| *** unicell has joined #kolla | 04:07 | |
| *** chandankumar has joined #kolla | 04:22 | |
| *** weiyu_ has quit IRC | 04:30 | |
| *** ashishb has joined #kolla | 04:30 | |
| *** ashishb has quit IRC | 04:45 | |
| *** chandankumar has quit IRC | 04:54 | |
| *** achanda has joined #kolla | 05:01 | |
| *** weiyu_ has joined #kolla | 05:21 | |
| *** dims has joined #kolla | 05:36 | |
| *** dims has quit IRC | 05:42 | |
| *** weiyu_ has quit IRC | 05:46 | |
| openstackgerrit | Merged stackforge/kolla: Speedup Git Cloning Ansible Repository https://review.openstack.org/216948 | 06:00 |
| *** weiyu has joined #kolla | 06:01 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add Ansible support for Heat https://review.openstack.org/205528 | 06:04 |
| mandre | yo kolla | 06:23 |
| sdake | yo mandre | 06:23 |
| sdake | nice to see all the new contributors ;-) | 06:24 |
| sdake | and users as well | 06:24 |
| sdake | we finally made something useful ;) | 06:24 |
| mandre | the chan got really busy while i was out | 06:25 |
| mandre | amazing | 06:25 |
| *** jpeeler has quit IRC | 06:26 | |
| mandre | the review queue is almost empty, what's up with SamYaple? | 06:26 |
| *** vern has left #kolla | 06:29 | |
| *** chandankumar has joined #kolla | 06:30 | |
| SamYaple | yea what sup with me? | 06:37 |
| SamYaple | what did i do? | 06:37 |
| sdake | mandre you have a whoel new code base to learn - better get busy ;-) | 06:38 |
| SamYaple | I feel like I have done something wrong. What did I do wrong mandre? | 06:38 |
| *** achanda has quit IRC | 06:40 | |
| mandre | SamYaple: you used to fill the review queue with patches | 06:46 |
| SamYaple | oh it was filled yesterday | 06:46 |
| SamYaple | its just empty while i do things for sdake | 06:47 |
| SamYaple | mandre: https://review.openstack.org/#/c/216691/ https://review.openstack.org/213397 | 06:47 |
| SamYaple | both of those are ready to go | 06:47 |
| SamYaple | and could use serious review | 06:47 |
| SamYaple | they are gate affecting things | 06:47 |
| *** mandre has quit IRC | 06:47 | |
| sdake | we need unit tests for build.py | 06:49 |
| SamYaple | sdake: agreed, not in that patch though | 06:50 |
| SamYaple | large enough laready | 06:50 |
| sdake | ya todo for later | 06:50 |
| *** jpeeler has joined #kolla | 06:53 | |
| SamYaple | sdake: https://review.openstack.org/#/c/216691/ https://review.openstack.org/213397 | 06:55 |
| SamYaple | both need reviews | 06:55 |
| sdake | i just reviewed one o fthem | 06:55 |
| sdake | hte build.py one | 06:56 |
| sdake | boy that was hard to reiew | 06:56 |
| sdake | one small nitpick | 06:56 |
| *** inc0 has joined #kolla | 06:56 | |
| sdake | why is cinder failing to build from source | 06:56 |
| * sdake groans | 06:56 | |
| SamYaple | sdake: you broke it a while ago | 06:57 |
| SamYaple | i told you that | 06:57 |
| sdake | hrm must have missed that | 06:57 |
| SamYaple | the sed lines you added in are dangerous _and_ broke source | 06:58 |
| *** dolpher has quit IRC | 06:59 | |
| inc0 | good morning | 07:00 |
| sdake | samyaple reviewed | 07:00 |
| inc0 | SamYaple, now I can calmly review your code, sorry for waiting | 07:00 |
| sdake | so on dangerous side lets table that discussion until tomorrow ;) | 07:00 |
| sdake | on breaking from source - bummer :( | 07:01 |
| SamYaple | inc0: thanks man! | 07:01 |
| sdake | samyaple my understanding is kernel is resonsible for device # allocation not udev | 07:01 |
| SamYaple | i added your code suggestion, even though the while loop is way more effiecent, your code is more readable | 07:01 |
| inc0 | I'm a wedding planner, so afternoons are everything but calm for me;) | 07:01 |
| sdake | udev just creates the devices as directed by the kernel after device allocation occurs | 07:01 |
| SamYaple | sdake: i am aware and ill explain what its broken and bad | 07:02 |
| sdake | inc0 you serously plan weddings? | 07:02 |
| SamYaple | but not right now | 07:02 |
| inc0 | sdake, just one | 07:02 |
| inc0 | my own | 07:02 |
| sdake | ya we got more important shit to do SamYaple | 07:02 |
| sdake | lets argue about it later ;) | 07:02 |
| sdake | inc0 my consolences! | 07:03 |
| inc0 | haha | 07:03 |
| SamYaple | sdake: its not an agruement, its a requirement ill point you to the breakdown of the out-of-sync udev stuff | 07:03 |
| sdake | cool well can we discuss it tomorrow | 07:03 |
| sdake | i rally need to stay focused on what i'm doing :) | 07:03 |
| sdake | promise i'll have all time in world for you tomorrow for this topic | 07:04 |
| inc0 | I'm glad you've added tests Sam, thanks | 07:05 |
| SamYaple | inc0: ive also made it so you can use regex proper like "^base$" or "keystone|^cinder-api$" | 07:07 |
| inc0 | but normal build.py neutron works? | 07:08 |
| sdake | inc0 on eway to find out :) | 07:08 |
| SamYaple | inc0: correct | 07:08 |
| inc0 | I'll give it a try once I'm done with code review;) | 07:08 |
| *** mandre has joined #kolla | 07:10 | |
| mandre | sorry SamYaple, my vps decided to kill itself | 07:11 |
| mandre | i'll do your reviews | 07:11 |
| SamYaple | the more eyes the better | 07:11 |
| SamYaple | both of them heavily affect the gate | 07:11 |
| SamYaple | if they are bad our gate is dead | 07:12 |
| sdake | our gate is nonvoting | 07:12 |
| sdake | so its not dead ;) | 07:12 |
| sdake | setup_gate is only run on image and functional jobs | 07:12 |
| sdake | it just votes -1 ;) | 07:12 |
| SamYaple | well i mean for image building | 07:12 |
| SamYaple | we need t osee if that works for commits | 07:12 |
| SamYaple | voting or not | 07:12 |
| sdake | i could make those jobs eaily voting | 07:13 |
| sdake | but i dont trust the gate enough to do so | 07:13 |
| sdake | we tried that once and the gate was busted for several days | 07:13 |
| sdake | because delorean ate shit | 07:13 |
| SamYaple | i agree it shouldnt be voting, but if things dont build we cant confidently approve patches | 07:14 |
| sdake | if we were to make the gating voting infra would need to mirror delorean | 07:15 |
| inc0 | when I had my build patch queue, each patchset could have different FAIL/SUCCESS combination | 07:15 |
| inc0 | while last one/middle one could have all success | 07:15 |
| sdake | inc0 that is fixed by retry=3 | 07:15 |
| sdake | as the default | 07:15 |
| inc0 | ok | 07:16 |
| *** jmccarthy has joined #kolla | 07:16 | |
| SamYaple | sdake: https://review.openstack.org/#/c/213397/ | 07:16 |
| SamYaple | im not sure what you want me to document | 07:16 |
| SamYaple | that is clear as day, im adding a key | 07:17 |
| SamYaple | its not magic, its a key | 07:17 |
| sdake | a key to what? | 07:17 |
| sdake | i dont get it | 07:17 |
| SamYaple | thats the way yo uadd keys | 07:17 |
| SamYaple | signing key | 07:17 |
| sdake | whos signing key is it? | 07:17 |
| SamYaple | thats the way you add keys for packages to debian | 07:17 |
| SamYaple | for the docker packages.. | 07:17 |
| SamYaple | add the repo add the key update | 07:17 |
| SamYaple | thats the debian way | 07:17 |
| sdake | add a comment | 07:18 |
| sdake | to the review pls | 07:18 |
| SamYaple | i did | 07:19 |
| sdake | ok looking | 07:19 |
| SamYaple | what would you like me to say and where | 07:19 |
| inc0 | SamYaple, do we have situation when we don't build base image? | 07:19 |
| SamYaple | inc0: building only data image i guess | 07:19 |
| inc0 | data image is base from code standpoint | 07:20 |
| SamYaple | inc0: data image is based of of the distro, not the base image | 07:20 |
| inc0 | base image == image with no parent | 07:20 |
| inc0 | that's what I mean | 07:20 |
| SamYaple | the initial images that can be built without dependacies are base and data | 07:20 |
| SamYaple | yup | 07:20 |
| inc0 | I think sooner or later we will have this no-base-rebuild use case | 07:21 |
| SamYaple | i think you mentioned you wanted image building without parents at one point, but the cache takes care of that | 07:21 |
| inc0 | for optimizations | 07:21 |
| SamYaple | the cache handles that | 07:21 |
| inc0 | unless parent changes | 07:21 |
| SamYaple | well the nyou should be rebuilding it all | 07:21 |
| inc0 | which happends, and we don't want to force stuff | 07:21 |
| SamYaple | not just teh child | 07:21 |
| SamYaple | the child wont be valid if its based off a change base | 07:22 |
| sdake | i am not keen to keep adding cli options to build.py ad infinitum | 07:22 |
| inc0 | you should, but I think people should have a choice in this regard anyway | 07:22 |
| sdake | the more options the more comlex | 07:22 |
| SamYaple | im going to have t odisagree there. changes in the base heavily affect al lthe other images | 07:22 |
| sdake | need to keep things simple | 07:22 |
| SamYaple | the cache wil lallow quick/instant rebuilds | 07:22 |
| SamYaple | since i did the set_time thing the cache has always hit for me | 07:22 |
| inc0 | SamYaple, they will, but power user might want to tailor their base themselves | 07:23 |
| SamYaple | inc0: yea so the rebuild all the images | 07:23 |
| SamYaple | otherwise how would they use the base | 07:23 |
| sdake | how do you build just one image with regex? | 07:23 |
| inc0 | if you rebuild and you have your own base, this will override it right? | 07:23 |
| SamYaple | sdake: it biulds the image and its parents sdake | 07:24 |
| sdake | i get that - i mean whta is the regex operation | 07:24 |
| inc0 | sdake, currently you're not, requirements were that this build filtered images AND their parents | 07:24 |
| SamYaple | sdake: build.py cinder-base | 07:24 |
| sdake | i thought so | 07:24 |
| sdake | kennan tried build.py ansible_kolla today and it failed | 07:24 |
| SamYaple | sdake: its kolla-ansible | 07:24 |
| SamYaple | or ansible-kolla | 07:24 |
| sdake | oh well that would explain it | 07:24 |
| SamYaple | but - not _ | 07:24 |
| *** dims has joined #kolla | 07:24 | |
| inc0 | or just ansible | 07:25 |
| inc0 | we don't have more of these | 07:25 |
| SamYaple | that also work | 07:25 |
| inc0 | build script won't try to pull stuff right? | 07:25 |
| sdake | samyaple heat in my future soon ? :) | 07:25 |
| *** weiyu has quit IRC | 07:26 | |
| sdake | runnign out of midnight oil | 07:26 |
| inc0 | so use case: you want to distribute your builds | 07:26 |
| inc0 | one server does base, other servers does other | 07:26 |
| SamYaple | sdake: no, created_time is wierd | 07:26 |
| inc0 | which would be quite cool for constant rebuilds, which people might actually like | 07:27 |
| *** jmccarthy has quit IRC | 07:27 | |
| *** weiyu has joined #kolla | 07:27 | |
| SamYaple | inc0: in that case we are talking about a custom build script | 07:27 |
| SamYaple | go for it | 07:27 |
| sdake | samyaple consult asalkeld :) | 07:27 |
| sdake | itso k to ask for help - god knows I ask 24/7 :) | 07:27 |
| inc0 | SamYaple, if we'll add conditional pulling/not rebuilding, it will be doable | 07:27 |
| SamYaple | sdake: yea i know where the issue is, just pdb in a container is... not so good | 07:27 |
| SamYaple | inc0: that breaks the reliability of the image though | 07:28 |
| sdake | which part of i am not keen to add cli options to build.py wasn't clear? | 07:28 |
| SamYaple | if the base doesnt match it should be rebuild | 07:28 |
| SamYaple | and we are defintely not adding distributed building | 07:28 |
| SamYaple | what a nightmare that would be | 07:28 |
| inc0 | ehh, I'm always for new features, sorry | 07:28 |
| inc0 | :P | 07:28 |
| *** dims has quit IRC | 07:29 | |
| sdake | build.py looks feature complete from my perspective | 07:29 |
| sdake | lets let the community weigh in on what else they need | 07:29 |
| sdake | before we gold plate it further | 07:29 |
| SamYaple | sdake: we still need some git tweaks, like refcommit changes and things | 07:30 |
| SamYaple | but thats options in build.ini not flags | 07:30 |
| sdake | SamYaple agreed | 07:30 |
| sdake | vbel is working on that | 07:30 |
| sdake | after that, and possibly adding something to build stuff for plugins | 07:30 |
| sdake | i dont see much to add | 07:30 |
| sdake | we have a mountain of work to do without refactoring build.py a million times ;) | 07:32 |
| sdake | although i like sam's lsat refactor | 07:33 |
| SamYaple | final remove the bash remnants | 07:34 |
| SamYaple | finally* | 07:34 |
| SamYaple | finnaly use threads and queue correctly! | 07:34 |
| SamYaple | no more tiers! | 07:34 |
| sdake | if it came with a fix for heat you would be my hero for a day :) | 07:34 |
| sdake | maybe even a week | 07:34 |
| sdake | this could use reviews for ryan in the morning to gleen over https://review.openstack.org/#/c/216874/ | 07:37 |
| sdake | it will affect how the whole project works and our work over the next month | 07:37 |
| sdake | so would appreciate wide review :) | 07:37 |
| vbel | good morning/evening :) | 07:38 |
| SamYaple | sdake heat is almost done, just after fixing the first problem, the same issue popped up for different values :/ | 07:39 |
| sdake | hot | 07:39 |
| sdake | paste to #heat would be my next step :) | 07:39 |
| mandre | SamYaple: wrt cross-distro gate setup patch, what's the reason for all these LVM commands for centos and fedora? | 07:40 |
| sdake | mandre he is setting up lvm thin provisioning to speedu p the gating process | 07:41 |
| sdake | atm lvm iuses loopbackby default on centos | 07:41 |
| SamYaple | yea that | 07:41 |
| mandre | maybe worth putting a comment in the code | 07:42 |
| SamYaple | also docker 1.8.1 recommends thinprov versus direct disk | 07:42 |
| SamYaple | mandre: would you mind if i did that in the next patch? | 07:42 |
| SamYaple | this is holding up infra mergeing the ubuntu gate | 07:42 |
| SamYaple | i need that to test the ubuntu code | 07:42 |
| SamYaple | I can add it now if you really want though, or i can promise to add it in the next patch :) | 07:43 |
| mandre | even if you promise, there's still a risk you forget :) | 07:43 |
| mandre | ubuntu doesn't need thin provisioning? | 07:44 |
| SamYaple | oh ye of little faith | 07:44 |
| SamYaple | well for one mandre i cant test ubuntu right now :/ | 07:44 |
| SamYaple | this patch is holding up the gate | 07:44 |
| SamYaple | so i will have to patch ubuntu right after this merges to test teh ubuntu gate | 07:44 |
| sdake | samyaple note the gate patches need a rebase | 07:44 |
| mandre | sorry, i meant there is still a risk a bus run over you | 07:44 |
| SamYaple | jesus dude | 07:45 |
| SamYaple | wtf | 07:45 |
| sdake | mandre uses th ebus argument ;) | 07:45 |
| sdake | samyaple i mean my gate patches | 07:45 |
| SamYaple | oh | 07:45 |
| SamYaple | https://review.openstack.org/#/c/213531/ | 07:45 |
| SamYaple | looks good to me | 07:45 |
| mandre | SamYaple: hold on a second for that PS, i had another comment | 07:46 |
| *** weiyu has quit IRC | 07:46 | |
| SamYaple | yea im patching heat bugs atm take your time | 07:46 |
| mandre | commented in the review | 07:47 |
| sdake | SamYaple i thought i had a docker images gate in there for centos | 07:47 |
| sdake | looks like i dont | 07:47 |
| sdake | or whateveryou removed inthatlast patch | 07:48 |
| * sdake is bleary eyed | 07:48 | |
| sdake | the reason cinder from source doesn't work is lvm isn't instlled | 07:49 |
| SamYaple | sdake: i know | 07:49 |
| SamYaple | i know why the sed failed | 07:50 |
| *** weiyu_ has joined #kolla | 07:50 | |
| sdake | pretty sure cinder from source didn't work previousl y) | 07:50 |
| SamYaple | but it built | 07:51 |
| sdake | SamYaple what is the lvm package on ubuntu | 07:51 |
| SamYaple | lvm2 | 07:52 |
| *** shardy has joined #kolla | 07:52 | |
| SamYaple | apt-get install lvm2 should be enough | 07:52 |
| SamYaple | well apt-get install -y --no-install-recommends lvm | 07:52 |
| SamYaple | well apt-get install -y --no-install-recommends lvm2* | 07:52 |
| sdake | what doe --no-intall-recommends do | 07:55 |
| sdake | does | 07:55 |
| SamYaple | keeps alot of unneeded dependancies out | 07:56 |
| SamYaple | its why the ubuntu images are less than half the size of centos | 07:56 |
| *** weiyu_ has quit IRC | 08:06 | |
| *** jmccarthy has joined #kolla | 08:07 | |
| inc0 | SamYaple, build.py neutron builds data container as wel | 08:09 |
| inc0 | well | 08:09 |
| SamYaple | inc0: no it doesnt ;) | 08:09 |
| inc0 | Processing: data | 08:10 |
| inc0 | I can see it. | 08:10 |
| inc0 | right now. | 08:10 |
| SamYaple | that doesnt say building does it now? | 08:10 |
| SamYaple | look at the first step in the builder function | 08:11 |
| SamYaple | for clarity, I can move the log position for the Processed message | 08:11 |
| inc0 | I'd say don't log "processing" unless you are actually processing it | 08:11 |
| SamYaple | so it always shows up | 08:11 |
| SamYaple | i am processing it | 08:12 |
| inc0 | I personally thin it's misleading | 08:12 |
| SamYaple | i process it and check if its matched, or if its parents are errord, or if it is source builds or not | 08:12 |
| inc0 | at least make it debug | 08:12 |
| inc0 | rather than info | 08:12 |
| inc0 | and "building" would be info | 08:12 |
| SamYaple | that sfair | 08:12 |
| sdake | inc0 that sounds good to me | 08:12 |
| SamYaple | the logging could be switched around in that whole section | 08:13 |
| SamYaple | right now it doesnt show "Proceeced" if it errors | 08:13 |
| SamYaple | only if it builds correctly | 08:13 |
| SamYaple | not really accurate | 08:13 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 08:13 |
| inc0 | other than that it looks pretty solid | 08:14 |
| inc0 | will improve multithreaded performance | 08:14 |
| SamYaple | yea, its about 20% faster since my threads always have things to process | 08:14 |
| sdake | vbel can we call this done ; https://blueprints.launchpad.net/kolla/+spec/add-proxy-to-dockerfiles | 08:14 |
| SamYaple | but its mostly dependant on disk anyway | 08:14 |
| inc0 | depends on your network | 08:15 |
| inc0 | it can be network-dependant...trust me, I know... | 08:15 |
| SamYaple | well its disk bound for me to build in 12 minutes | 08:15 |
| inc0 | yeah proxy works with headers | 08:15 |
| inc0 | it's something like that for me too...in home | 08:16 |
| vbel | sdake: yes, with small note - I need to put a few unset http_* to kolla-common.sh to clean up it so that openstack services don't go to my proxies :) | 08:16 |
| inc0 | in my office its horrible | 08:16 |
| SamYaple | vbel: that sounds fine | 08:16 |
| sdake | vbel please submit a patch - i want to start closing out blueprints that are close | 08:16 |
| SamYaple | we source that file each time anyway | 08:16 |
| *** akwasnie has joined #kolla | 08:16 | |
| inc0 | vbel, add no_proxy env variable | 08:16 |
| SamYaple | but the issue is that if you ever exec in teh contaienr it will have those values set | 08:17 |
| SamYaple | just keep that in mind | 08:17 |
| sdake | we need to document how to build with proxies | 08:17 |
| sdake | vbel if you want to take that on it would be appreciated as well | 08:17 |
| sdake | i struggled with kennan today to help him through it | 08:17 |
| vbel | yes, will do | 08:17 |
| sdake | lots of people will want to know how to do that | 08:17 |
| vbel | can I use wildcard for no_proxy? | 08:17 |
| inc0 | I'm not sure | 08:18 |
| sdake | from my experience with magnum, atleast 30% of people end up deploying behind a proxy | 08:18 |
| vbel | it is easier to unset things to be sure no one attempts to read them | 08:18 |
| vbel | no_proxy is not very popular and can be dismissed by some apps | 08:18 |
| inc0 | vbel, mind sending a patch that will crate a placeholder? | 08:19 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Install lvm in Cinder from-source containers https://review.openstack.org/217025 | 08:20 |
| vbel | inc: you mean templating kolla-common.sh? Or just unsets? | 08:21 |
| vbel | :) | 08:21 |
| sdake | if we can get two cr's to review that would appreciate it, gate is busted as is | 08:21 |
| inc0 | I don't think we can template kolla-common | 08:22 |
| inc0 | but some code for unset would be nice | 08:22 |
| sdake | i had hoped to get rid of kolla-common.sh | 08:22 |
| inc0 | custom unset I mean | 08:22 |
| inc0 | or custom include | 08:22 |
| vbel | ok | 08:23 |
| sdake | vbel just jam it in there for now | 08:23 |
| SamYaple | vbel: cant you override the docker varilables? | 08:24 |
| sdake | lets get a patch in the queue today pls :) | 08:24 |
| SamYaple | ENV HTTP_PROXY '' | 08:24 |
| SamYaple | ? | 08:24 |
| SamYaple | if so we can add a footer to the end of the final images (cinder-api, glance-registry, etc) | 08:24 |
| sdake | samyaple he uses the include to set up the proxies | 08:24 |
| SamYaple | I am aware | 08:24 |
| SamYaple | so we can add the footer to clean up | 08:25 |
| sdake | kind of messy but atleast we can get rid of kolla-common.sh that way :) | 08:25 |
| SamYaple | yea but the common.sh thing wouldnt work anyway | 08:25 |
| SamYaple | if you docker exec into the image its gonna get wierd | 08:25 |
| *** weiyu_ has joined #kolla | 08:25 | |
| sdake | samyaple what hour is give up hour for heat | 08:25 |
| SamYaple | idk man | 08:26 |
| SamYaple | its a new wierd bug now | 08:26 |
| SamYaple | and i dont know the code base | 08:26 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Templates for rsyslog https://review.openstack.org/215181 | 08:26 |
| sdake | vbel if we start removing /docker now without git references, is it going to damage your work? | 08:27 |
| sdake | samyaple fwiw we tried ENV https_proxy="" | 08:29 |
| sdake | and I dont think that worked | 08:29 |
| sdake | didn't work i mean | 08:29 |
| inc0 | sdake, where it didn't worked? | 08:29 |
| sdake | the environment variable needs to be unset | 08:29 |
| inc0 | you need several places | 08:29 |
| sdake | in vbel's tseting when we POCed this idea | 08:29 |
| SamYaple | thats not how oyu set docker env variables | 08:30 |
| SamYaple | ENV HTTPS_PROXY "" | 08:30 |
| sdake | yaa that didn't work | 08:30 |
| inc0 | sdake, did you add it to /etc/default/docker as well? | 08:30 |
| SamYaple | ok | 08:30 |
| inc0 | SamYaple, https_proxy="" works well | 08:30 |
| inc0 | ENV https... | 08:30 |
| SamYaple | inc0: not in a dockerfile | 08:30 |
| inc0 | SamYaple, in dockerfile, I do it all the time | 08:30 |
| sdake | add what to e/tc/defualt/docker? | 08:31 |
| SamYaple | sdake just said it didnt inc0 | 08:31 |
| sdake | vbel tested it | 08:31 |
| sdake | i didn't test it | 08:31 |
| SamYaple | but if that _does_ whats the problem wit hthe footer? | 08:31 |
| sdake | so i dont know exactly what he did | 08:31 |
| inc0 | add export https_proxy="" there | 08:31 |
| SamYaple | inc0: that wont do it | 08:31 |
| inc0 | this way docker will use proxy for example for image pulling | 08:31 |
| SamYaple | thats not how docker does env variables | 08:31 |
| inc0 | SamYaple, I'm doing it even now | 08:31 |
| SamYaple | export will not work in the dockerfile like oyu think then | 08:32 |
| inc0 | SamYaple, I'm doing it even now | 08:32 |
| inc0 | without that curls and yums doesn't work, with that it does | 08:32 |
| sdake | inc0 the idea is that the running container will nto use the proxy | 08:32 |
| sdake | but the build proxy will | 08:32 |
| sdake | this is vbels requirement | 08:32 |
| inc0 | sdake, yeah, you need to either unset it or add hosts to no_proxy | 08:32 |
| inc0 | I personally use no_proxy and it helps, openstack understands no_proxy | 08:33 |
| sdake | unset it how? | 08:33 |
| inc0 | let me paste you my include header file | 08:33 |
| sdake | sounds good | 08:33 |
| sdake | the only thing we teted that worked was to ENV it in, and unset in start.sh | 08:34 |
| sdake | setting to ="" didn't do the job | 08:34 |
| sdake | but we didn't test ENV "" | 08:34 |
| SamYaple | inc0: that doesnt work like i said | 08:34 |
| SamYaple | i just tested it | 08:34 |
| sdake | we tested start.sh export https_proxy="" | 08:34 |
| inc0 | http://paste.openstack.org/show/428062/ | 08:34 |
| SamYaple | inc0: now try and unset that | 08:35 |
| SamYaple | sdake: ENV HTTPS_PROXY "" | 08:36 |
| SamYaple | that does work however | 08:36 |
| SamYaple | obviously export doesnt work in the dockerfile | 08:36 |
| SamYaple | thats my point | 08:36 |
| sdake | samyaple cool that isn't quite what vbel tested | 08:36 |
| vbel | ENV https - if you do sudo or su you are out of luck | 08:36 |
| sdake | vbel tested export https_proxy="'" in _start.sh_ | 08:36 |
| sdake | not in the dockerfile | 08:37 |
| SamYaple | vbel: I beileve it is broken, but you can set it to empty was my point. i did not test if it works. i trust you :) | 08:37 |
| vbel | I will run tests today :) | 08:37 |
| sdake | ya if tis set to empty - the services think proxy_server ="" | 08:38 |
| vbel | will tell what works here | 08:38 |
| sdake | what there needs to be is an unset docker operation | 08:38 |
| SamYaple | sdake: or build env variables ;) | 08:38 |
| vbel | yes, people requested UNSET many times. One workaround is to squash images :) | 08:38 |
| sdake | ya then we wouldn't need all this hacking | 08:38 |
| sdake | pbourke you around? | 08:39 |
| inc0 | UNSET will be hard as it has to be done on last container | 08:40 |
| inc0 | OR we can do include header and include footer in build script | 08:40 |
| inc0 | and add this to EVERY dockerfile | 08:41 |
| SamYaple | inc0: yea that was my suggestion | 08:41 |
| inc0 | that would work and that would be most clean solution | 08:41 |
| SamYaple | but apparent that wont work | 08:41 |
| SamYaple | vbel: said it still is broken | 08:41 |
| inc0 | let me check plz | 08:41 |
| pbourke | sdake: morning | 08:42 |
| sdake | hey pbourke | 08:42 |
| sdake | whats the word on swift | 08:42 |
| pbourke | sdake: sorry no patch yesterday | 08:42 |
| vbel | I would include header and footer in all templates, not just base. But for current situation I would just add "unset proxy" in common runtime script (kolla-common.sh or whatever we think appropriate) | 08:42 |
| sdake | did the solution involve not rquiring ansible 2.0? | 08:43 |
| pbourke | sdake: we're going to leave it as is for this patch | 08:43 |
| pbourke | sdake: as it does work and will need to investigate a soltuion that works for 1.9 and 2.0 | 08:43 |
| sdake | cool i dont want a hard dep on ansible 2.0 in the base ansible files | 08:43 |
| sdake | i dont mind a hard dep on asinbile in container kolla-ansible | 08:43 |
| pbourke | sdake: I just was making sure the patch still is good when rebased from master | 08:43 |
| pbourke | sdake: and fell over a few things | 08:44 |
| sdake | cool well lets get it up and merged :) | 08:44 |
| pbourke | today is the day | 08:44 |
| SamYaple | yea sdake we can still have a unique swift.conf per service but i wont hold up the patch anyfurther | 08:44 |
| SamYaple | ill write that in later | 08:44 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Install lvm in Cinder from-source containers https://review.openstack.org/217025 | 08:45 |
| vincent_vdk | hmm, can't build gnocchi-api because yum fails | 08:48 |
| vincent_vdk | ah, packages not available | 08:50 |
| sdake | gnocchi has been busted forever | 08:53 |
| vincent_vdk | :) | 08:53 |
| * sdake would +2 +a a patch to remove gnocchi from code base ;) | 08:53 | |
| SamYaple | it biulds for me! | 08:54 |
| vincent_vdk | can you skip that one when building all the containers | 08:54 |
| vincent_vdk | not for me | 08:54 |
| SamYaple | ubuntu source for me | 08:54 |
| vincent_vdk | ah | 08:54 |
| vincent_vdk | this is on centos/rdo | 08:54 |
| SamYaple | yup | 08:54 |
| SamYaple | delorean is alwaays busted | 08:54 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 08:55 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Templates for rsyslog https://review.openstack.org/215181 | 08:55 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: rsyslog container https://review.openstack.org/214589 | 08:55 |
| sdake | vincent_vdk I would support a patch to remove the yum shit from gnocchi | 08:57 |
| sdake | since we have no ansible work for it | 08:58 |
| sdake | and all it does atm is fail to buidl and cause confusion | 08:58 |
| sdake | if someone fixes it later thats good ;) | 08:58 |
| sdake | i dont want non-building containers in the build | 08:58 |
| sdake | its sloppy and bad for our project | 08:58 |
| sdake | if we need to keep it for from source so be it | 08:59 |
| vincent_vdk | to be honest, i don't even know what it does | 08:59 |
| *** weiyu_ has quit IRC | 08:59 | |
| sdake | its a time series database | 08:59 |
| vincent_vdk | would have to look it up :) | 08:59 |
| sdake | that is as much as I know ;) | 08:59 |
| vincent_vdk | ah | 08:59 |
| vincent_vdk | i don't see the use to add all these kind of things to openstack | 08:59 |
| sdake | i think its a backend replacement for ceilometer | 09:00 |
| sdake | but I am not certain | 09:00 |
| sdake | we are packaging the big tent long term | 09:00 |
| sdake | so get used to it :) | 09:00 |
| sdake | there are 40+ projects in the big tent atm | 09:01 |
| sdake | atleaset 30 of them are clientserver apps | 09:01 |
| *** inc0 has quit IRC | 09:01 | |
| sdake | SamYaple I just had an epiphany that json change is rquired for upgrade from liberty to mitaka | 09:02 |
| *** weiyu has joined #kolla | 09:02 | |
| sdake | samyaple please prove me wrong | 09:02 |
| SamYaple | required? | 09:03 |
| *** inc0 has joined #kolla | 09:05 | |
| sdake | as in how would we do upgrade from liberty to mitaka and bring json with it | 09:05 |
| sdake | without it being specified in the docker files in liberty | 09:06 |
| SamYaple | sdake: mitaka has new containers and playbook changes, whats the issue? | 09:07 |
| *** jmccarthy has quit IRC | 09:08 | |
| *** jmccarthy has joined #kolla | 09:08 | |
| sdake | thinking | 09:09 |
| sdake | been up long time | 09:09 |
| *** jmccarthy has quit IRC | 09:13 | |
| inc0 | SamYaple, about https://review.openstack.org/215062 - common role by default runs on every node? or is it specified somewhere? | 09:14 |
| sdake | samyaple not sure - we will need to revisit this discussion - need to set up one more centos node | 09:14 |
| inc0 | because I'd keep rsyslog as separate logging role, as we'll expand on that - and soon | 09:14 |
| sdake | brb folks time to setup another centos | 09:15 |
| *** weiyu has quit IRC | 09:15 | |
| SamYaple | inc0: no the rsyslog container goes in the common role | 09:16 |
| SamYaple | the logging role will be for other things later | 09:16 |
| SamYaple | like kibana eletic serach etc | 09:16 |
| inc0 | oO | 09:16 |
| inc0 | why not keeping it as separate role? | 09:16 |
| inc0 | it's clean | 09:17 |
| SamYaple | because the logging container runs on all nodes... | 09:18 |
| SamYaple | its a requirement of all services | 09:18 |
| SamYaple | it goes in the common role | 09:18 |
| SamYaple | please dont fight me on the ansible architecture too | 09:18 |
| inc0 | you fight me on python, it's only fair | 09:19 |
| SamYaple | i wrote that python and the ansible | 09:19 |
| inc0 | well, still, I'll challange it | 09:19 |
| SamYaple | go for it | 09:20 |
| inc0 | you can always -2 my patches;) | 09:20 |
| SamYaple | if i cant hold up to challenges i dont have a good point | 09:20 |
| inc0 | or ask me politely to go away and never come back;) | 09:20 |
| SamYaple | jeez thats horrible | 09:20 |
| SamYaple | i like you | 09:20 |
| inc0 | anyway, I'm ok with common role | 09:20 |
| inc0 | just teasing | 09:20 |
| SamYaple | we may butt heads but that isnt a horrible thing man | 09:20 |
| *** jmccarthy has joined #kolla | 09:26 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Implement Ansible role for Swift https://review.openstack.org/215065 | 09:27 |
| *** akscram1 has joined #kolla | 09:28 | |
| *** akscram has quit IRC | 09:28 | |
| pbourke | sdake: ^^ | 09:28 |
| inc0 | guys btw feedback from ops meetup - monitoring | 09:30 |
| inc0 | do we want to create monitoring stack for our containers? | 09:30 |
| SamYaple | like to monitor if the containers are up inc0? | 09:30 |
| inc0 | no, like zabbix | 09:32 |
| SamYaple | im not familiar with that | 09:32 |
| SamYaple | heard the name though | 09:32 |
| inc0 | graphite? | 09:32 |
| inc0 | splunk? | 09:33 |
| inc0 | that sort of stuff | 09:33 |
| SamYaple | yea sure i think that would be good optional addons | 09:33 |
| *** weiyu_ has joined #kolla | 09:33 | |
| sdake | i am wary of adding new blueprints at this time inc0 | 09:40 |
| sdake | but if you want to put into discussion phase that is our typical process :) | 09:40 |
| inc0 | well I was thinking early M rly | 09:40 |
| inc0 | just something I've heard from folks who have been on ops summit | 09:41 |
| inc0 | I had meeting yesterday, and monitoring stack was one of big gaps | 09:41 |
| sdake | pbourke does that work | 09:41 |
| pbourke | sdake: yes | 09:41 |
| sdake | pbourke what is the swift "api" port that i need to expose in my firewall for the swift cli to work | 09:43 |
| pbourke | sdake: 8080 | 09:44 |
| openstackgerrit | Merged stackforge/kolla: Install lvm in Cinder from-source containers https://review.openstack.org/217025 | 09:46 |
| sdake | do directoreis /srv/node/sdb1 and sdb2 and sdb3 need to exist on host pbourke? | 09:46 |
| pbourke | sdake: no that's handled by the data container | 09:46 |
| pbourke | sdake: the only thing you need to supply are the rings | 09:46 |
| *** dims has joined #kolla | 09:48 | |
| *** weiyu_ has quit IRC | 09:54 | |
| sdake | pbourke is the only thing you tried from source or does rdo binary work too? | 09:54 |
| pbourke | sdake: I only tried rdo binary actually | 09:55 |
| sdake | nice | 09:55 |
| sdake | mandre or samyaple can one of you please review https://review.openstack.org/#/c/215065/ | 09:56 |
| sdake | we can fix up problems with followup patches | 09:56 |
| inc0 | sdake, do you sleep? like...ever? | 10:00 |
| sdake | inc0 yes | 10:00 |
| sdake | all next week - sleeping | 10:00 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 10:02 |
| vbel | ^^^ now handles build.py links, kolla-build and tox | 10:04 |
| *** jmccarthy has quit IRC | 10:06 | |
| *** jmccarthy1 has joined #kolla | 10:06 | |
| SamYaple | vbel: tests fine for me | 10:07 |
| SamYaple | well see what the gate has to say, but im good with it thus far | 10:07 |
| *** weiyu_ has joined #kolla | 10:12 | |
| *** weiyu_ has quit IRC | 10:14 | |
| pbourke | vbel: if 'source' in image and 'source' in image['source']: | 10:15 |
| pbourke | is that intentional? | 10:15 |
| pbourke | looks strange :) | 10:15 |
| vbel | yes | 10:15 |
| vbel | we have source type and sometimes source is missing, e.g. for base and data | 10:15 |
| vbel | so one is type and another is location of source from build.ini | 10:16 |
| *** Kennan2 has joined #kolla | 10:16 | |
| *** Kennan has quit IRC | 10:16 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 10:19 |
| vbel | is there online debug output for registry push in build.py? | 10:23 |
| SamYaple | vbel: no the push doesnt seem to be returning anything we are catching vbel | 10:23 |
| SamYaple | its probably looking for the wrong key in return | 10:23 |
| SamYaple | i dont know | 10:23 |
| *** weiyu_ has joined #kolla | 10:24 | |
| vbel | SamYaple: ok, thanks! For some reason I cannot push it - requests.packages.urllib3.exceptions.ReadTimeoutError: UnixHTTPConnectionPool(host='localhost', port=None): Read timed out | 10:24 |
| SamYaple | can you push from the command line | 10:24 |
| pbourke | push was working for me yesterday | 10:25 |
| *** zhiwei has quit IRC | 10:25 | |
| pbourke | with appropriate tag | 10:25 |
| pbourke | vbel: check, port=None | 10:25 |
| sdake | samyaple | 10:40 |
| sdake | i am ready to try multinode for first time | 10:40 |
| sdake | need bit of handholding | 10:40 |
| sdake | my AIO deploys perfectly from registry | 10:40 |
| sdake | do i need to put my hostnames in /etc/hostname? | 10:41 |
| *** vbel has quit IRC | 10:44 | |
| sdake | anyone that has deployed multinode please throw me a bone re how to configure my system | 10:45 |
| sdake | do i need /etc/hostname with ips in it ? | 10:45 |
| *** vbel has joined #kolla | 10:45 | |
| SamYaple | sdake: the hosts should be able to resolve each other | 10:46 |
| SamYaple | so if you have server01 if you be able to ping server02 | 10:46 |
| SamYaple | by name | 10:46 |
| SamYaple | if you have that setup, that should be enouhg | 10:46 |
| inc0 | sdake, I run ansible with ips | 10:46 |
| sdake | even ping the deployment from host? | 10:46 |
| inc0 | sdake, what would prohibit it? deployment host has to ssh to deployed one | 10:48 |
| sdake | inc0 i will try ips first | 10:48 |
| inc0 | so packages must be able to travel both ways | 10:48 |
| sdake | you use ips in the inventory file? | 10:48 |
| inc0 | yup | 10:48 |
| SamYaple | inc0: are you doing multinode _kolla_ with ips in the inventory file? | 10:49 |
| inc0 | I even done 10.10.10.[12-20] as compute nodes | 10:49 |
| inc0 | yup SamYaple | 10:49 |
| SamYaple | cool | 10:49 |
| inc0 | any problem with it? | 10:49 |
| SamYaple | i havent tested that glad that works | 10:49 |
| SamYaple | i would have though galera would break because we use hostnames there | 10:49 |
| sdake | good that takes a load off -d on't have to get my resolver shit working | 10:49 |
| inc0 | SamYaple, it might, I didn't run full HA with it | 10:50 |
| inc0 | one controller and multiple computes | 10:50 |
| inc0 | worth a test tho | 10:51 |
| * sdake groans | 10:51 | |
| SamYaple | sdake: youll need hostnames | 10:51 |
| sdake | samyaple ack | 10:51 |
| SamYaple | make sure they can all resolve each others hostnames and youre good | 10:51 |
| inc0 | SamYaple, if Galera uses /etc/hostname on each node, that might work | 10:51 |
| inc0 | not sure about /etc/hosts | 10:51 |
| SamYaple | inc0: not when server02 cant resolve server01 | 10:51 |
| inc0 | I think having hostnames are safest approach tho, yeah | 10:52 |
| SamYaple | put the the servers in /etc/hosts on each node and youre golden | 10:52 |
| SamYaple | and I only test with hostnames in the inventory, but you might get away with ips there | 10:52 |
| inc0 | yeah, just make one big /etc/hosts and scp it everywhere | 10:52 |
| SamYaple | its only three nodes so i think hell be ok | 10:52 |
| inc0 | you can even scp it with ansible;) | 10:52 |
| inc0 | if its 1 controller and 2 computes it should work anyway | 10:53 |
| inc0 | I didn't do multiple controllers with that | 10:53 |
| openstackgerrit | Merged stackforge/kolla: Implement Ansible role for Swift https://review.openstack.org/215065 | 10:53 |
| vbel | running tox and Co still not enough to get thru gate. Can I run gate by hand locally? | 10:56 |
| sdake | samyaple quick setup question | 10:57 |
| sdake | one of my hosts is not reading /etc/hosts first | 10:57 |
| SamYaple | what do you mean? | 10:57 |
| sdake | instead its ranodmly creating ip addresses when i ping | 10:57 |
| SamYaple | wtf? | 10:57 |
| sdake | PING minime-02.broked.org (92.242.140.2) 56(84) bytes of data. | 10:57 |
| sdake | ^C | 10:57 |
| sdake | --- minime-02.broked.org ping statistics --- | 10:58 |
| sdake | 3 packets transmitted, 0 received, 100% packet loss, time 1999ms | 10:58 |
| sdake | minime-02 is 192.168.1.102 | 10:58 |
| sdake | it works from all machines that are not deployment hosts | 10:58 |
| sdake | rather the hosts to deploy from | 10:58 |
| SamYaple | thats a fqdn name youre trying to ping though | 10:58 |
| sdake | basically where I am launching ansible from | 10:58 |
| sdake | ping minime-02 is the command | 10:58 |
| pbourke | vbel: tox -e<name-of-failing-test> | 10:58 |
| sdake | [sdake@bigiron etc]$ ping minime-02 | 10:58 |
| sdake | PING minime-02.broked.org (92.242.140.2) 56(84) bytes of data. | 10:58 |
| SamYaple | sdake: dude im kinda fried at the moment. i would guess your /etc/hosts is setup wrong | 10:59 |
| SamYaple | thats the first stop in name resolution | 10:59 |
| sdake | samyaple i copied it directly from my 3 other working boxes.. | 10:59 |
| SamYaple | idk man | 10:59 |
| * sdake hates computers | 11:00 | |
| *** weiyu_ has quit IRC | 11:00 | |
| vbel | tox does not fail locally. It just fail while gate-kolla-build-images-centos-binary | 11:01 |
| sdake | cooli put a short name in e/tc/hsots and its working | 11:01 |
| pbourke | sdake: i know that feeling | 11:01 |
| sdake | samyaple do i put th short name in the inventory file or the fqdn | 11:02 |
| SamYaple | both if you want | 11:02 |
| SamYaple | shortnames are required | 11:02 |
| *** weiyu_ has joined #kolla | 11:02 | |
| inc0 | sdake, it has to be resolvable, that's all afaik | 11:02 |
| sdake | ok here goes nothign ;) | 11:03 |
| sdake | why isn't the inventory file in /etc btw | 11:03 |
| sdake | just curiosity there | 11:03 |
| inc0 | it's not technically config | 11:04 |
| inc0 | it's more like input to function, same as plays | 11:04 |
| sdake | change ansible_ssh_user=sam to =root? | 11:04 |
| inc0 | if you don't have sam's account;) | 11:04 |
| inc0 | remember to either spread ssh keys or give ansible_ssh_pass | 11:05 |
| sdake | i have full ssh for root an sdake accoutns | 11:05 |
| sdake | from to all machines | 11:05 |
| inc0 | then you can either do user=root or sdake, root is shorter | 11:06 |
| sdake | and have eto sudo for multinode when running playbok? | 11:07 |
| inc0 | ansible knows how to sudo | 11:08 |
| inc0 | afair by default, but let me check | 11:08 |
| pbourke | does anyone have build.py seem to just hang | 11:09 |
| pbourke | even with -T 1, it just sits there with no output seemingly forever | 11:09 |
| sdake | fatal: [minime-one] => {'msg': "AnsibleUndefinedVariable: One or more undefined variables: 'dict object' has no attribute u'ansible_em1'", 'failed': True} | 11:09 |
| pbourke | sdake: did you set your interface in /etc/kolla/globals.yml | 11:09 |
| sdake | to em1 | 11:09 |
| inc0 | every server has this iface? | 11:10 |
| inc0 | pbourke, works for me, | 11:10 |
| sdake | em1 is present across all nodes in the cluster | 11:11 |
| inc0 | and it's up? | 11:11 |
| inc0 | hmm | 11:11 |
| vbel | Sorry, user root is not allowed to execute '/bin/testr run test_build.BuildTestCentosBinaryDocker' as docker ???? | 11:11 |
| sdake | yes its my management network | 11:11 |
| *** jmccarthy1 has quit IRC | 11:12 | |
| vbel | I cannot run sudo -g docker testr run test_build.BuildTestCentosBinaryDocker as myself or root, I am in docker group | 11:12 |
| inc0 | sdake, paste full log please | 11:12 |
| pbourke | vbel: try without sudo? | 11:12 |
| pbourke | vbel: I had that issue before | 11:12 |
| pbourke | removed sudo and it worked, not sure why | 11:12 |
| vbel | I run tox -e | 11:13 |
| vbel | ok, will try | 11:13 |
| pbourke | remove sudo in tox.ini | 11:13 |
| sdake | http://ur1.ca/nk725 -> http://paste.fedoraproject.org/259489/40587594 | 11:13 |
| sdake | there was a bug filed about this | 11:13 |
| sdake | I am running deploy host of f21, running target machines of centos7 | 11:13 |
| sdake | i bet selinux | 11:14 |
| sdake | let me turn that shit off | 11:14 |
| sdake | it is not selinux | 11:15 |
| inc0 | hmm...that's strange | 11:15 |
| inc0 | seems like ansible fails to gather facts | 11:15 |
| inc0 | properly | 11:15 |
| sdake | dependency problem? | 11:15 |
| sdake | it would be fantastic if ansible told me what it was doing | 11:16 |
| sdake | is there anything on these machines to look at? | 11:16 |
| inc0 | can you manually add task debug to print hostvars? | 11:16 |
| sdake | like files | 11:16 |
| inc0 | http://docs.ansible.com/ansible/debug_module.html | 11:16 |
| sdake | inc0 I added -vv in kolla-ansible tool | 11:16 |
| inc0 | you need to specify task with debug | 11:16 |
| inc0 | in play | 11:17 |
| vbel | I can run tox locally but gate fails - http://logs.openstack.org/55/216655/7/check/gate-kolla-build-images-centos-source/ab128b4/console.html | 11:17 |
| inc0 | what ansible was doing is it tried to render haproxy.cfg | 11:17 |
| inc0 | and failed with KeyError on line 23 (or similar) | 11:17 |
| *** jmccarthy has joined #kolla | 11:18 | |
| inc0 | https://github.com/stackforge/kolla/blob/master/ansible/roles/haproxy/templates/haproxy.cfg.j2#L23 | 11:18 |
| inc0 | it seems ansible doens't see your interfaces | 11:18 |
| sdake | does jinja2 need to be installed on the deployed to hosts? | 11:18 |
| pbourke | vbel: I suspect this line: if os.path.basename(script_path) == 'kolla': | 11:18 |
| pbourke | that is not evaluating true on the CI machine | 11:19 |
| inc0 | no, just on deployer | 11:19 |
| sdake | what python code is running on the deployed to host? | 11:19 |
| vbel | pbourke: yes, looks like that | 11:19 |
| inc0 | shade, MySQL-python | 11:20 |
| inc0 | and docker-py | 11:20 |
| inc0 | but I'm not sure any more, I think shade is in kolla_ansible | 11:20 |
| vbel | can one run tox not in kolla folder? | 11:20 |
| sdake | inc0 I mean what python code | 11:20 |
| vbel | pbourke: can one run tox not in kolla folder? | 11:20 |
| inc0 | it's generated by ansible from plays | 11:20 |
| sdake | SamYaple really looking for debug help here - totally stuck | 11:20 |
| sdake | where is it stored on disk | 11:20 |
| sdake | i want to look at it | 11:20 |
| SamYaple | whats going on? | 11:20 |
| pbourke | vbel: not that I know of, there may be some arg to set base dir | 11:21 |
| inc0 | sdake, I don't think this fails on deployed host tho | 11:21 |
| inc0 | I think it fails on rendering templates, and that's deployer host thing | 11:21 |
| vbel | pbourke: I can run tox from kolla, from tests ... | 11:21 |
| sdake | samyaple http://paste.fedoraproject.org/259489/40587594 | 11:21 |
| inc0 | but I don't know where are these pythons...tmp would be my guess | 11:21 |
| pbourke | vbel: its workspace is /home/jenkins/workspace/gate-kolla-build-images-centos-source | 11:22 |
| pbourke | vbel: is that an issue? | 11:22 |
| pbourke | vbel: try renaming your kolla dir to something else and see if it fails | 11:22 |
| pbourke | sdake: do all hosts have an ip assigned to em1 | 11:23 |
| pbourke | sdake: and are interfaces up | 11:23 |
| sdake | pbourke all ip assigned to em1, all interfaces are up | 11:24 |
| vbel | pbourke: yes, it is :) Thanks! Will hack it further | 11:24 |
| *** rhallisey has joined #kolla | 11:24 | |
| pbourke | ansible is refusing to gather info about those interfaces for some reason | 11:25 |
| inc0 | sdake, add this debug please, we'll see what info it actually gathered? | 11:26 |
| *** pbourke has quit IRC | 11:28 | |
| *** pbourke has joined #kolla | 11:28 | |
| pbourke | https://github.com/ansible/ansible/issues/4848#issuecomment-40912176 | 11:29 |
| pbourke | sdake: can you try deploying one service, e.g. rabbitmq | 11:31 |
| pbourke | sdake: then try the rest | 11:31 |
| pbourke | (--tags=rabbitmq) | 11:31 |
| *** weiyu_ has quit IRC | 11:32 | |
| pbourke | actually it tries to access the interface so will probably be same problem | 11:32 |
| inc0 | I have different issue with ansible...instead of prints it does cowsay... oO | 11:32 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 11:34 |
| *** Haomeng|2 has joined #kolla | 11:37 | |
| *** shardy_ has joined #kolla | 11:38 | |
| *** shardy has quit IRC | 11:39 | |
| *** Haomeng has quit IRC | 11:40 | |
| *** shardy_ has quit IRC | 11:43 | |
| *** shardy has joined #kolla | 11:44 | |
| *** dims_ has joined #kolla | 11:48 | |
| *** dims has quit IRC | 11:52 | |
| sdake | mariadb busted - please see this log - multinode deploy http://ur1.ca/nk7g1 -> http://paste.fedoraproject.org/259509/89970144 | 11:53 |
| sdake | need help asap ;) | 11:53 |
| inc0 | sdake, hmm...looks like permissions in /var/lib/mysql? | 11:55 |
| sdake | i see that | 11:56 |
| sdake | however same exact container starts up on a differnet node | 11:56 |
| inc0 | but /var/lib/mysql is a volume | 11:56 |
| inc0 | so it will inherit perms | 11:56 |
| vbel | can be 'data' container change, does it chown it while bootstrapping? | 11:57 |
| SamYaple | remember the mysql folder is empty for new nodes | 11:59 |
| SamYaple | only the bootstrap node (that geenrates the files) needs to chown | 12:00 |
| SamYaple | the rest dont initialze the files | 12:00 |
| SamYaple | glaera does that | 12:00 |
| inc0 | http://michaelheap.com/cowsay-and-ansible/ really... | 12:04 |
| SamYaple | I HATE IT | 12:04 |
| inc0 | lol | 12:05 |
| inc0 | I loled when I saw that.. | 12:06 |
| *** jmccarthy has quit IRC | 12:09 | |
| inc0 | ANSIBLE_COW_SELECTION=random | 12:11 |
| inc0 | :D | 12:11 |
| inc0 | I think I got wh40k titan now | 12:11 |
| *** jmccarthy has joined #kolla | 12:14 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 12:16 |
| *** jasonamyers has joined #kolla | 12:27 | |
| sdake | any of this log helpful to people: http://ur1.ca/nk7qg -> http://paste.fedoraproject.org/259523/05919191 | 12:28 |
| sdake | pbourke ^^ | 12:28 |
| sdake | mariadb is failing to startup multinode | 12:29 |
| sdake | or galera | 12:29 |
| sdake | or whatever the hell it is | 12:29 |
| pbourke | sdake: I had a different maria issue | 12:30 |
| sdake | pbourke one issue is the chown is in the wrong place | 12:30 |
| pbourke | sdake: in the end I hacked my multinode to have maria just run one one node :( | 12:30 |
| pbourke | sdake: yeah it was permission related | 12:30 |
| sdake | it needs to run every time not only on the bootstrap node | 12:30 |
| sdake | i fixed that and rebuilt | 12:30 |
| pbourke | I haven't seen the above though, let me look more closely | 12:30 |
| sdake | now having this problem | 12:30 |
| pbourke | how did you get around the interface problem | 12:31 |
| inc0 | SamYaple, voluming /dev/log doesn't seem to be working :( I've ran rsyslog container with /dev/log:/dev/log and another kolla_ansible with /dev/log:/dev/log | 12:32 |
| inc0 | logger foo doesn't work | 12:32 |
| *** shardy_ has joined #kolla | 12:34 | |
| *** shardy has quit IRC | 12:35 | |
| inc0 | I'm not sure if we can do without syslogd in container | 12:35 |
| sdake | pbourke i put my 3 machines in the network szection | 12:36 |
| pbourke | sdake: yeah I saw you pm, thanks | 12:36 |
| sdake | samyaple says he understnds problem and will address it | 12:36 |
| *** diga has quit IRC | 12:38 | |
| *** shardy_ has quit IRC | 12:39 | |
| *** shardy has joined #kolla | 12:40 | |
| vbel | yahoo, got thru gate :) https://review.openstack.org/#/c/216655/ | 12:43 |
| *** weiyu has joined #kolla | 12:43 | |
| *** shardy has quit IRC | 12:44 | |
| *** shardy has joined #kolla | 12:45 | |
| *** vbel has quit IRC | 12:58 | |
| *** vbel has joined #kolla | 13:05 | |
| pbourke | vbel: added one small question on your review | 13:05 |
| vbel | yes, name in archive should be keystone-xzcs , so source['dest'] is not what we need - it will be keystone.tar. With arcname=os.path.basename(clone_dir) we will get keystone-master or keystone-2015.1.1 | 13:08 |
| vbel | we have to have service-xxx because we link it inside of Dockerfiles with ln -s /service-* /service | 13:11 |
| *** weiyu has quit IRC | 13:13 | |
| pbourke | vbel: thanks for clarification on archive name | 13:24 |
| pbourke | vbel: nice job | 13:24 |
| vbel | pbourke: thanks a million for the review! | 13:25 |
| *** weiyu has joined #kolla | 13:26 | |
| *** inc0 has quit IRC | 13:28 | |
| *** jtriley has joined #kolla | 13:37 | |
| *** weiyu has quit IRC | 13:42 | |
| *** klint has quit IRC | 13:47 | |
| *** inc0 has joined #kolla | 13:54 | |
| vbel | it looks like build.py cannot find base images if I specify 'oracle' as namespace. But any other string works ('vlad', 'oracle123') | 13:57 |
| vbel | Can anyone try that please? | 13:57 |
| *** Kennan2 has quit IRC | 14:01 | |
| *** diogogmt has quit IRC | 14:01 | |
| SamYaple | vbel: what do you mean? | 14:02 |
| SamYaple | im not sure i follow | 14:02 |
| *** Kennan has joined #kolla | 14:02 | |
| *** ftcpops has joined #kolla | 14:04 | |
| vbel | tools/build.py --tag latest --namespace oracle -b oraclelinux -t source --template --debug -T1 keystone < does not work, but if namespace is localhost:5000/oracle or vlad or oracle123 it works | 14:04 |
| *** diogogmt has joined #kolla | 14:04 | |
| SamYaple | weird, how does it fail? | 14:05 |
| *** Haomeng|2 has quit IRC | 14:05 | |
| vbel | yes, strange. I see Sorted image keystone with parent oracle123/oraclelinux-source-base:latest with oracle123 as name, but not with simple oracle | 14:06 |
| *** unicell1 has joined #kolla | 14:06 | |
| vbel | Could not find parent image from some images. Aborting | 14:06 |
| SamYaple | ah i think i know why | 14:06 |
| *** unicell has quit IRC | 14:06 | |
| SamYaple | we check if the namespace is in the image['fullname'] | 14:06 |
| SamYaple | it will match everything for you | 14:07 |
| SamYaple | because oracle is in oraclelinux | 14:07 |
| SamYaple | i can fix that | 14:07 |
| mandre | SamYaple: no comment about why lvm thin provisioning is necessary in your new patch :( | 14:07 |
| SamYaple | mandre: i had a mental note i was going to take care of it next patch and spaced. I addressed your other comment | 14:08 |
| SamYaple | give me a minute and ill reposst | 14:08 |
| mandre | yeah i saw that | 14:09 |
| vbel | SamYaple: thanks a lot :) | 14:09 |
| SamYaple | vbel: ill post a patch for you to test in a minute | 14:09 |
| SamYaple | mandre: what would you like it to say? | 14:09 |
| SamYaple | without doing this it broke centos | 14:09 |
| mandre | you sure? the current gate setup script doesn't have it i believe | 14:10 |
| SamYaple | mandre: docker 1.8.1 in this patch | 14:10 |
| SamYaple | things changed | 14:10 |
| mandre | i see | 14:10 |
| SamYaple | this is better though anyway | 14:10 |
| SamYaple | its better perforamnce | 14:10 |
| SamYaple | and more expplicit so when we do dind we have a better playground | 14:10 |
| mandre | then just put a comment in the code, it's not obvious if you don't know the reason | 14:11 |
| mandre | if you can link to docker recommentation that'd be even better | 14:11 |
| mandre | any reason why you only do it for centos/fedora and not ubuntu? | 14:12 |
| mandre | older docker? | 14:12 |
| SamYaple | mandre: i dont even have an ubuntu gate to test on | 14:13 |
| SamYaple | but i cant use devicemapper for ubuntu | 14:13 |
| SamYaple | to old linked libraries | 14:13 |
| *** rhallisey has quit IRC | 14:17 | |
| *** diogogmt has quit IRC | 14:18 | |
| *** rhallisey has joined #kolla | 14:19 | |
| SamYaple | nope vbel that wasnt the issue | 14:19 |
| *** thumpba has joined #kolla | 14:19 | |
| *** rhallisey_ has joined #kolla | 14:20 | |
| *** thumpba has quit IRC | 14:20 | |
| SamYaple | vbel: can you file a bug | 14:20 |
| *** thumpba has joined #kolla | 14:20 | |
| *** inc0 has quit IRC | 14:21 | |
| SamYaple | vbel: if self.namespace not in image['parent_name']: | 14:22 |
| SamYaple | image['parent'] = None | 14:22 |
| SamYaple | youre namespace _will_ match your base distro | 14:22 |
| SamYaple | let me think for a moment | 14:22 |
| *** rhallisey has quit IRC | 14:24 | |
| SamYaple | vbel: can you file a bug with that code with this irc conversation | 14:24 |
| SamYaple | i will have to 'think' later, my brain is fried :) | 14:24 |
| SamYaple | also feel free to fix it, but i will if you dont | 14:24 |
| SamYaple | this would also affect me if i used ubun as my namespace, wth base ubuntu | 14:25 |
| vbel | SamYaple: yes, here - https://bugs.launchpad.net/kolla/+bug/1489021 | 14:26 |
| openstack | Launchpad bug 1489021 in kolla "cannot use 'oracle' as namespace for build.py" [Medium,New] | 14:26 |
| SamYaple | awesome thanks vbel | 14:26 |
| SamYaple | if you want to assign it to me ill figure it out for you | 14:26 |
| vbel | ok, thanks! | 14:27 |
| SamYaple | i suggest you just use orac1e | 14:27 |
| SamYaple | :) | 14:27 |
| SamYaple | fixed! | 14:27 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: DO NOT MERGE - Workaround for horizon https://review.openstack.org/216592 | 14:58 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: DO NOT MERGE - Fix up galera https://review.openstack.org/217251 | 14:58 |
| *** sdake has quit IRC | 14:58 | |
| *** daneyon has joined #kolla | 15:01 | |
| *** jmccarthy1 has joined #kolla | 15:08 | |
| *** jmccarthy has quit IRC | 15:08 | |
| *** diogogmt has joined #kolla | 15:15 | |
| *** inc0 has joined #kolla | 15:24 | |
| inc0 | meeting now or in 1hr? | 15:24 |
| pbourke | does it count for stats if you review a patch without voting | 15:27 |
| inc0 | nope | 15:34 |
| mandre | inc0: meeting in 1h i think | 15:34 |
| inc0 | thanks mandre that's what I though | 15:35 |
| inc0 | t | 15:35 |
| mandre | thanks for the reminder btw, i totally forgot we were wednesday | 15:36 |
| SamYaple | pbourke: no it does not | 15:38 |
| pbourke | kk thought as much | 15:38 |
| *** chandankumar has quit IRC | 15:40 | |
| -openstackstatus- NOTICE: restarting gerrit due to a slow memory leak | 15:43 | |
| *** blakec has joined #kolla | 15:56 | |
| *** dims_ has quit IRC | 16:00 | |
| *** dims has joined #kolla | 16:01 | |
| *** unicell1 has quit IRC | 16:04 | |
| *** akwasnie1 has joined #kolla | 16:04 | |
| *** rhallisey_ is now known as rhallisey | 16:09 | |
| *** erkules has quit IRC | 16:13 | |
| *** erkules has joined #kolla | 16:13 | |
| *** akwasnie1 has left #kolla | 16:17 | |
| *** inc0_ has joined #kolla | 16:17 | |
| *** akwasnie1 has joined #kolla | 16:17 | |
| *** inc0_ has quit IRC | 16:17 | |
| *** sdake has joined #kolla | 16:20 | |
| sdake | so tired | 16:21 |
| inc0 | sdake, I bet...you didn't sleep at all did you? | 16:21 |
| SamYaple | sdake: fixed heat | 16:21 |
| SamYaple | sdake https://review.openstack.org/#/c/217058/ | 16:22 |
| SamYaple | after almost 10 hours of pdb | 16:22 |
| sdake | reminder meeting today in 10 mins | 16:22 |
| SamYaple | i was stepping through sqlalchemy to find the issue :/ | 16:22 |
| sdake | samyple hot | 16:23 |
| harmw | meeting time guys | 16:29 |
| *** jmccarthy1 has quit IRC | 16:29 | |
| sdake | kolla meeting now | 16:30 |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: WIP:Replace config-external with a JSON file for Neutron-OVS https://review.openstack.org/216874 | 16:40 |
| *** mfalatic has joined #kolla | 16:44 | |
| *** unicell has joined #kolla | 16:52 | |
| *** achanda has joined #kolla | 16:52 | |
| *** jasonamyers has quit IRC | 16:54 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Add Ansible support for Heat https://review.openstack.org/205528 | 17:02 |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove tiered building in build.py https://review.openstack.org/216691 | 17:05 |
| *** akwasnie has quit IRC | 17:19 | |
| *** daneyon has quit IRC | 17:27 | |
| pbourke | ok im out too guys, will pick up reviews in the morning | 17:33 |
| pbourke | sdake: hope you're still alive after demo ;) | 17:33 |
| inc0 | cya pbourke ++;) | 17:34 |
| sdake | pbourke demo failed duringovs bridge bringup | 17:34 |
| sdake | pbourke ciao | 17:34 |
| inc0 | I'm off as well, have a good night everyone | 17:34 |
| *** inc0 has quit IRC | 17:34 | |
| mandre | same here, good night all | 17:34 |
| rhallisey | pbourke, you're spamming the etherpad with '+' | 17:35 |
| akwasnie1 | bye Michal :) | 17:35 |
| rhallisey | lol | 17:35 |
| sdake | holly christ undo | 17:35 |
| *** akwasnie1 has quit IRC | 17:35 | |
| harmw | lol | 17:35 |
| rhallisey | haha | 17:35 |
| rhallisey | wtf is happening | 17:36 |
| *** loth has joined #kolla | 17:38 | |
| *** achanda_ has joined #kolla | 17:39 | |
| *** achanda has quit IRC | 17:39 | |
| *** dims_ has joined #kolla | 17:41 | |
| *** dims has quit IRC | 17:42 | |
| rhallisey | pbourke, around? | 17:43 |
| SamYaple | wtf guys | 17:46 |
| SamYaple | i levae for 5 minutes | 17:46 |
| *** achanda_ has quit IRC | 17:48 | |
| rhallisey | what is going on in the etherpad | 17:50 |
| SamYaple | pbourke went native it seems | 17:50 |
| *** achanda has joined #kolla | 17:51 | |
| sdake | rhallisey the etherpad is beign spammed by someone | 17:56 |
| rhallisey | :/ | 17:56 |
| sdake | #openstack-infra is investigating | 17:56 |
| sdake | after #openstack-infra investigated, the person stopped | 17:56 |
| sdake | because opentackinra was on to them | 17:56 |
| SamYaple | sdake: it wasnt me! | 17:57 |
| SamYaple | i know thats what you were thinking | 17:57 |
| SamYaple | admit it | 17:57 |
| sdake | no i wasn't thinking that at all | 17:57 |
| rhallisey | I thought one of you guys feel asleep on the keyboard | 17:57 |
| SamYaple | i make jokes. its my thing | 17:57 |
| sdake | I was thinkign it is osmone in openstack-meeting-4 and someone in openstack-infra | 17:57 |
| SamYaple | uhoh | 17:58 |
| SamYaple | looks like more spam | 17:58 |
| SamYaple | now its copy pasta | 17:58 |
| *** jasonsb has quit IRC | 18:04 | |
| *** absubram has joined #kolla | 18:07 | |
| *** jasonamyers has joined #kolla | 18:12 | |
| *** unicell has quit IRC | 18:14 | |
| *** unicell has joined #kolla | 18:14 | |
| *** achanda has quit IRC | 18:16 | |
| *** openstackgerrit has quit IRC | 18:17 | |
| *** openstackgerrit has joined #kolla | 18:17 | |
| openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Need to import gitapi instead of git https://review.openstack.org/217336 | 18:18 |
| SamYaple | rhallisey: | 18:20 |
| SamYaple | i dont have an error with that script | 18:20 |
| SamYaple | did you pip install -r requirements.txt? | 18:20 |
| rhallisey | oh gitpython | 18:21 |
| rhallisey | that's the module I was looking for | 18:21 |
| rhallisey | I was looking for python-git | 18:22 |
| rhallisey | k abandoned | 18:22 |
| SamYaple | openstack talks are up guys | 18:26 |
| SamYaple | only 2 kolla talks got added | 18:33 |
| SamYaple | actually just one | 18:33 |
| SamYaple | the other just mentions kolla | 18:33 |
| rhallisey | ya just notice | 18:35 |
| rhallisey | any reason kolla only got 1? | 18:35 |
| rhallisey | from the community | 18:35 |
| rhallisey | think that's a bit odd | 18:35 |
| SamYaple | well my subbissions got moved catagories at teh last second, after most had voted | 18:35 |
| SamYaple | :( | 18:35 |
| SamYaple | unlikely the votes were adjusted | 18:35 |
| rhallisey | that kinda sucks | 18:36 |
| SamYaple | indeed | 18:36 |
| rhallisey | I'd images there would've been 3-4 for the community up there | 18:36 |
| rhallisey | 1! | 18:36 |
| SamYaple | i thought that lab was going to be killer, but it was one of the track changes last second | 18:36 |
| SamYaple | well this sucks :( | 18:41 |
| SamYaple | its cool though the talk with sdake will be good | 18:41 |
| *** jasonsb has joined #kolla | 18:44 | |
| rhallisey | maybe kolla is next summit | 18:46 |
| rhallisey | I really thought I would be a think for this one | 18:46 |
| rhallisey | it* | 18:46 |
| SamYaple | to be fair, next summit well have a stable branch and people can test without building images | 18:47 |
| SamYaple | maybe we just drank the kolla kool-aid and it isn't so amazing :) | 18:48 |
| ftcpops | Well, how's that for an introduction to kolla? | 18:48 |
| rhallisey | SamYaple, I think we're drinking fine whiskey | 18:48 |
| ftcpops | I went out for a run, came back to see my keyboard spewing "+" all over your etherpad | 18:48 |
| ftcpops | SORRY ABOUT THAT! | 18:48 |
| rhallisey | lol! | 18:49 |
| ftcpops | I was like, wow, who's going nuts on the '+'s? oh...crap... | 18:49 |
| *** jeblair has joined #kolla | 18:52 | |
| *** dims_ has quit IRC | 18:53 | |
| *** dims has joined #kolla | 18:54 | |
| *** achanda has joined #kolla | 19:16 | |
| *** jasonamyers has quit IRC | 19:25 | |
| *** jasonamyers has joined #kolla | 19:25 | |
| *** achanda has quit IRC | 19:25 | |
| *** daneyon has joined #kolla | 19:29 | |
| *** daneyon has quit IRC | 19:33 | |
| *** daneyon has joined #kolla | 19:48 | |
| *** anteaya has joined #kolla | 19:51 | |
| anteaya | ftcpops: might I suggest in future you select a different etherpad username | 19:51 |
| anteaya | ftcpops: the name you selected combined with a stuck keyboard is an odd combination | 19:52 |
| *** daneyon has quit IRC | 19:54 | |
| *** thumpba has quit IRC | 19:55 | |
| *** daneyon has joined #kolla | 19:55 | |
| ftcpops | "quit being an asshat"? I didn't type that in, I think someone helped me with that one ;-) | 19:55 |
| ftcpops | Although for the Tokyo Summit, I'll make sure to put that on my name badge | 19:57 |
| rhallisey | lol | 20:00 |
| *** jasonsb has quit IRC | 20:01 | |
| *** achanda has joined #kolla | 20:02 | |
| *** jasonsb has joined #kolla | 20:02 | |
| *** achanda has quit IRC | 20:08 | |
| *** daneyon has quit IRC | 20:09 | |
| ftcpops | sdake: sorry again about the etherpad plus spamming :-( | 20:11 |
| ftcpops | I just ran, and timed: kolla/cmd/build.py -T 1000 -i include_proxy.dockerfile --template --no-cache | 20:11 |
| ftcpops | real 304m5.164s | 20:11 |
| ftcpops | user 0m1.970s | 20:11 |
| ftcpops | sys 0m3.291s | 20:11 |
| ftcpops | System is a CentOS 7 VM running on top of Windows via Virtual Box with 4Gb RAM | 20:12 |
| ftcpops | 100Mb Ethernet via Proxy | 20:12 |
| *** jasonamyers has quit IRC | 20:12 | |
| *** jasonamyers has joined #kolla | 20:23 | |
| *** daneyon has joined #kolla | 20:24 | |
| *** achanda has joined #kolla | 20:26 | |
| *** thumpba has joined #kolla | 20:29 | |
| *** jasonsb has quit IRC | 20:33 | |
| *** jasonsb has joined #kolla | 20:36 | |
| anteaya | ftcpops: yeah, maybe locking the screen, seems like you have some helpful housemates | 21:01 |
| *** openstackgerrit has quit IRC | 21:01 | |
| *** openstackgerrit has joined #kolla | 21:02 | |
| *** jasonamyers has quit IRC | 21:03 | |
| *** achanda has quit IRC | 21:07 | |
| *** daneyon_ has joined #kolla | 21:16 | |
| *** daneyon has quit IRC | 21:19 | |
| ftcpops | Has anyone successfully run this command (or similar) from behind a corporate proxy? ansible-playbook -i inventory/all-in-one -e @/etc/kolla/globals.yml -e @/etc/kolla/passwords.yml site.yml | 21:22 |
| rhallisey | ftcpops, I've been having issues myself lately with it | 21:25 |
| rhallisey | ftcpops, what are you running into? | 21:25 |
| ftcpops | PLAY [haproxy] **************************************************************** | 21:25 |
| ftcpops | GATHERING FACTS *************************************************************** | 21:25 |
| ftcpops | ok: [localhost] | 21:25 |
| ftcpops | TASK: [common | Starting Ansible container] *********************************** | 21:25 |
| ftcpops | failed: [localhost] => {"changes": ["{\"status\":\"Pulling repository docker.io/kollaglue/centos-binary-kolla-ansible\"}\r\n", "{\"errorDetail\":{\"message\":\"Error: image kollaglue/centos-binary-kolla-ansible:latest not found\"},\"error\":\"Error: image kollaglue/centos-binary-kolla-ansible:latest not found\"}\r\n"], "failed": true, "status": ""} | 21:25 |
| ftcpops | msg: Unrecognized status from pull. | 21:25 |
| ftcpops | FATAL: all hosts have already failed -- aborting | 21:25 |
| *** absubram has quit IRC | 21:26 | |
| ftcpops | but when I run: docker run -it kollaglue/centos-binary-kolla-ansible bash it works | 21:26 |
| rhallisey | ftcpops, do $ docker images | 21:26 |
| rhallisey | what's the tag for that image | 21:26 |
| rhallisey | it's looking for latest | 21:26 |
| jpeeler | my guess is that you need to set docker_pull_policy: "missing" | 21:28 |
| ftcpops | sorry, I don't know how to find the tag :( | 21:28 |
| ftcpops | jpeeler: when I set it to missing, I get this: | 21:29 |
| ftcpops | TASK: [haproxy | Starting keepalived container] ******************************* | 21:29 |
| ftcpops | failed: [localhost] => {"changed": true, "failed": true} | 21:29 |
| ftcpops | msg: Docker API Error: Cannot start container 62830bd04a38b9bb306649041d0028d1db1b2494d49b8cf8ebfb6d0b2c366db6: [8] System error: exec format error | 21:29 |
| ftcpops | FATAL: all hosts have already failed -- aborting | 21:29 |
| jpeeler | what does "docker ps" show? i don't think i've seen that error before | 21:30 |
| ftcpops | $ docker ps | 21:31 |
| ftcpops | CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES | 21:31 |
| ftcpops | efab539742c5 kollaglue/centos-binary-kolla-ansible:latest "/bin/sleep infinity" 2 minutes ago Up 2 minutes kolla_ansible | 21:31 |
| jpeeler | use a pastebin please - that wasn't too bad though | 21:32 |
| ftcpops | ok :) | 21:32 |
| jpeeler | ok pastebin docker ps -a, docker logs <failed haproxy id> | 21:32 |
| jpeeler | i mean if this is an all in one install, you may be able to just skip haproxy like rhallisey does, right ryan? | 21:33 |
| rhallisey | exec format error O.o | 21:33 |
| ftcpops | http://paste.openstack.org/show/428763/ | 21:34 |
| rhallisey | I'm guessing it something in /etc/globals.yml | 21:34 |
| rhallisey | I'm not sure what that iss | 21:35 |
| jpeeler | uh, did that docker ps output get truncated? | 21:36 |
| jpeeler | it doesn't show the keepalived container exited, so maybe kill it and rerun? | 21:36 |
| ftcpops | okay, I'll kill them all and rm them and will start over | 21:36 |
| jpeeler | there's a stupid script "cleanup-containers" | 21:37 |
| jpeeler | if you haven't done anything but kolla containers, use that | 21:37 |
| ftcpops | beautiful. Thanks, jpeeler | 21:38 |
| ftcpops | yeah, it's a VM I'm using just for kolla | 21:38 |
| *** jtriley has quit IRC | 21:40 | |
| *** thumpba has quit IRC | 21:40 | |
| *** thumpba has joined #kolla | 21:40 | |
| *** shardy has quit IRC | 21:46 | |
| *** thumpba has quit IRC | 21:47 | |
| *** thumpba has joined #kolla | 21:49 | |
| *** thumpba has quit IRC | 21:51 | |
| ftcpops | Okay, here's a clean run with docker_pull_policy: always ==> http://paste.openstack.org/show/428765/ | 21:51 |
| ftcpops | Here's a clean run with docker_pull_policy: missing ==> http://paste.openstack.org/show/428766/ | 21:52 |
| rhallisey | ftcpops, have you built the container fairly recently? | 21:54 |
| ftcpops | this morning | 21:54 |
| rhallisey | master may just be broken | 21:54 |
| *** dims_ has joined #kolla | 21:56 | |
| rhallisey | ftcpops, I'm rebuilding now will try again soon | 21:58 |
| ftcpops | okay thanks rhallisey! I have to go coach baseball soon, but I'll check back. Thanks for all your help and sorry again for all the plusses (so embarrassing) | 21:59 |
| rhallisey | ftcpops, no problem have fun! It was just an accident no worries! | 21:59 |
| *** dims has quit IRC | 21:59 | |
| *** achanda has joined #kolla | 22:06 | |
| *** ftcpops has quit IRC | 22:06 | |
| *** SourabhP has joined #kolla | 22:06 | |
| *** diogogmt has quit IRC | 22:08 | |
| *** sdake_ has joined #kolla | 22:15 | |
| *** sdake has quit IRC | 22:18 | |
| *** dims_ has quit IRC | 22:24 | |
| *** dims has joined #kolla | 22:24 | |
| *** achanda_ has joined #kolla | 22:28 | |
| *** achanda has quit IRC | 22:29 | |
| sdake_ | whitepaper on container tech and kolla mentioned quite a bit from foundation: http://www.openstack.org/assets/pdf-downloads/Containers-and-OpenStack.pdf | 22:30 |
| rhallisey | sdake_, except only 1 talk from the community got accepted | 22:31 |
| rhallisey | cool though | 22:31 |
| sdake_ | ya not alot of slots | 22:31 |
| sdake_ | you sid one other tlak mentioning kolla got accepted that was what? | 22:32 |
| rhallisey | it was a talk about how ebay was looking to use it | 22:32 |
| rhallisey | it's a user_story | 22:32 |
| rhallisey | they mention kolla | 22:32 |
| sdake_ | which one | 22:33 |
| sdake_ | link plz | 22:33 |
| *** dims has quit IRC | 22:34 | |
| *** sdake_ is now known as sdake | 22:41 | |
| *** rhallisey has quit IRC | 22:41 | |
| *** achanda has joined #kolla | 23:18 | |
| *** achanda_ has quit IRC | 23:18 | |
| *** achanda has quit IRC | 23:38 | |
| *** blakec has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!