openstackgerrit | Kai Qiang Wu(Kennan) proposed stackforge/kolla: Make doc more readable and easy to follow https://review.openstack.org/216175 | 00:10 |
---|---|---|
*** weiyu has joined #kolla | 00:51 | |
*** sdake_ has quit IRC | 00:52 | |
*** erkules_ is now known as erkules | 00:58 | |
*** erkules has quit IRC | 00:58 | |
*** erkules has joined #kolla | 00:58 | |
Kennan | hi SamYaple: there? | 01:13 |
*** vinkman has left #kolla | 01:18 | |
*** erkules_ has joined #kolla | 01:32 | |
*** erkules has quit IRC | 01:35 | |
*** tummy has joined #kolla | 01:40 | |
*** erkules_ has quit IRC | 01:52 | |
*** erkules_ has joined #kolla | 01:54 | |
*** dims__ has quit IRC | 01:55 | |
*** Haomeng has joined #kolla | 01:56 | |
*** cfarquhar has quit IRC | 01:57 | |
*** jpeeler has quit IRC | 01:57 | |
*** Haomeng|2 has quit IRC | 01:58 | |
*** jpeeler has joined #kolla | 02:04 | |
*** jpeeler has quit IRC | 02:04 | |
*** jpeeler has joined #kolla | 02:05 | |
*** achanda has joined #kolla | 02:11 | |
*** jtriley has quit IRC | 02:29 | |
*** tummy has quit IRC | 02:35 | |
*** achanda has quit IRC | 02:50 | |
*** klint has joined #kolla | 02:58 | |
*** achanda has joined #kolla | 03:01 | |
*** unicell has quit IRC | 03:08 | |
SamYaple | Kennan: I am now! | 03:13 |
*** jruano has quit IRC | 03:13 | |
*** loth has quit IRC | 03:21 | |
*** sdake has joined #kolla | 03:22 | |
sdake | yo | 03:23 |
SamYaple | hey sdake | 03:24 |
SamYaple | we need https://review.openstack.org/#/c/216395/ sdake to print out the summary stuff properly | 03:25 |
sdake | hey sam | 03:25 |
sdake | did you sleep? :) | 03:25 |
SamYaple | barely | 03:25 |
SamYaple | at work now | 03:25 |
SamYaple | gate is almost done :) | 03:26 |
SamYaple | small difference between hp and rax | 03:26 |
SamYaple | once thats sorted its all good | 03:26 |
*** unicell has joined #kolla | 03:27 | |
*** diogogmt has quit IRC | 03:34 | |
*** sdake_ has joined #kolla | 04:06 | |
*** sdake has quit IRC | 04:10 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 04:14 |
*** diga has joined #kolla | 04:17 | |
*** sdake_ is now known as sdake | 04:18 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 04:19 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - gate hacking https://review.openstack.org/216005 | 04:23 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix regression in build-script https://review.openstack.org/216395 | 04:28 |
*** weiyu has quit IRC | 04:35 | |
Kennan | hi SamYaple: | 04:52 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Ceph container https://review.openstack.org/214706 | 04:52 |
Kennan | for my environment, I want to check should I follow this | 04:53 |
Kennan | ansible-playbook -i inventory/all-in-one -e @/etc/kolla/globals.yml -e @/etc/kolla/passwords.yml site.yml | 04:53 |
Kennan | or | 04:53 |
Kennan | sudo ./tools/kolla-ansible deploy | 04:53 |
Kennan | it has different deploy methods | 04:53 |
SamYaple | coolsvap: can you review https://review.openstack.org/#/c/216463/ | 04:53 |
Kennan | I did not know which is the right one | 04:53 |
SamYaple | Kennan: they both do the same thing | 04:54 |
SamYaple | i always use the ansible-playbook command | 04:54 |
SamYaple | i would recommend it for now, but either one will work | 04:54 |
Kennan | OK SamYaple: Thanks let me try this | 04:54 |
*** unicell1 has joined #kolla | 04:55 | |
*** unicell has quit IRC | 04:56 | |
Kennan | Hi SamYaple: | 04:56 |
Kennan | FATAL: all hosts have already failed -- aborting | 04:56 |
Kennan | glue/centos-binary-kolla-ansible:latest not found\"},\"error\":\"Error: image kollaglue/centos-binary-kolla-ansible:latest not found\"}\r\n"], | 04:56 |
sdake | its so glorious to see our gate passing all the time for builds now! | 04:56 |
SamYaple | sdake: youre welcome | 04:56 |
sdake | change always to missing | 04:56 |
SamYaple | Kennan: have you built the images? | 04:56 |
sdake | Kennan in globals.yml make that change | 04:56 |
Kennan | yes, I did run it | 04:56 |
Kennan | INFO:__main__:=========================== | 04:57 |
Kennan | ERROR:__main__:gnocchi-a Failed with status: error | 04:57 |
Kennan | ERROR:__main__:gnocchi-s Failed with status: error | 04:57 |
Kennan | ERROR:__main__:ceilomete Failed with status: error | 04:57 |
Kennan | just this failed | 04:57 |
SamYaple | Kennan: oh you arent using a registry. thats why | 04:57 |
Kennan | other images seems ok | 04:57 |
SamYaple | listen to sdake | 04:57 |
SamYaple | he has the correct suggestion | 04:57 |
SamYaple | docker_pull_policy: "missing" | 04:57 |
SamYaple | ^^ in the globals.yml | 04:57 |
openstackgerrit | Merged stackforge/kolla: Make doc more readable and easy to follow https://review.openstack.org/216175 | 04:58 |
SamYaple | sdake: the rax gate has an 80gb ephemeral | 04:58 |
SamYaple | the hp gate has a 500gb | 04:58 |
SamYaple | O.o | 04:58 |
sdake | samyaple can you translate https://review.openstack.org/#/c/216369/1/docker_templates/base/Dockerfile.j2 to english for me plz | 05:00 |
SamYaple | sdake: i mention what im doing in the commit message? | 05:00 |
sdake | oh let me read that | 05:00 |
sdake | moment | 05:00 |
SamYaple | its not detailed but i dont have time to write a novel | 05:00 |
SamYaple | on the history of python | 05:00 |
SamYaple | for centos and debian | 05:00 |
SamYaple | which i had to read | 05:01 |
SamYaple | to | 05:01 |
SamYaple | figure | 05:01 |
SamYaple | this | 05:01 |
SamYaple | out | 05:01 |
SamYaple | anyway, it only affects ubuntu | 05:01 |
sdake | damn hit +a accidentally | 05:02 |
* sdake needs more caffeine | 05:02 | |
sdake | sam i took away the =a, but not sure what will happen | 05:02 |
sdake | you may have to rebase your patch because of my error | 05:02 |
SamYaple | lok it started merging gate jobs | 05:02 |
SamYaple | awesome | 05:03 |
sdake | mistakes happen | 05:03 |
openstackgerrit | Merged stackforge/kolla: Ubuntu - Haproxy https://review.openstack.org/215924 | 05:03 |
sdake | infra did that to me the other day | 05:03 |
sdake | so so glorious to see the gate jobs stable | 05:04 |
SamYaple | im just super happy that change was as clean as it was | 05:04 |
sdake | now its time to raise the bar! | 05:04 |
SamYaple | python be da bomb | 05:04 |
SamYaple | hey i have to update all the source sections for all the templates sdake | 05:05 |
*** weiyu_ has joined #kolla | 05:05 | |
SamYaple | at the same time imgoing to ajdust the docker files for formating to keep the mconsistent | 05:05 |
SamYaple | i think we have standardized at this point | 05:06 |
SamYaple | without discussion thankfully | 05:06 |
sdake | this change looks wrong for centos binary https://review.openstack.org/#/c/216026/10/docker/common/horizon/config-external.sh | 05:07 |
sdake | did you test only centos source? | 05:07 |
sdake | to cut to the chase which combos of horizon havey oubuid and tested? :) | 05:08 |
SamYaple | sdake: how would that be wrong for centos binary? | 05:08 |
sdake | look at line 1 and 2 | 05:08 |
sdake | that is where the container is expecting those files to be copied i think | 05:08 |
SamYaple | that didnt change... | 05:09 |
coolsvap | SamYaple, yes looking at https://review.openstack.org/#/c/216463/ | 05:09 |
* sdake needs a redbull | 05:09 | |
SamYaple | the bottom section is probably wrong | 05:09 |
SamYaple | if you tell me the centosbinary site filename i can change it (or it could be a different patch?) i dont know | 05:09 |
SamYaple | ill try the cnetos images again tonight | 05:10 |
SamYaple | it was breaking on the base for me | 05:10 |
SamYaple | coolsvap: thanks! | 05:10 |
sdake | samyaple https://review.openstack.org/#/c/216463/ breaks the gate | 05:10 |
sdake | lets do centos binary as a different patch set if it is untessted | 05:10 |
sdake | could yu tell me which combos you have tested? | 05:10 |
SamYaple | neither centos source or binary build for me anymore so ive tested neither on the latest patchsets | 05:11 |
SamYaple | i dont know if either ever worked | 05:11 |
SamYaple | source i hacked to install at one point | 05:11 |
SamYaple | for centos | 05:11 |
SamYaple | and i have verified the _files_ are going to the same location | 05:11 |
SamYaple | for both centos and ubuntu | 05:11 |
SamYaple | coolsvap: ignore that patch | 05:12 |
SamYaple | pep8 error | 05:12 |
coolsvap | yes | 05:12 |
coolsvap | pep8 is there | 05:12 |
sdake | the build gates are failing too for whatever reason | 05:12 |
sdake | ok good with verificatino files going to same location | 05:12 |
Kennan | Hi SamYaple: it seens kolla setup (one VM with 2 Nics configure eth0, eth1) make my ssh lost connection | 05:12 |
Kennan | did it work for that case ? | 05:12 |
SamYaple | sdake: yea but the functional-f21 works | 05:13 |
Kennan | how do you configure that ? | 05:13 |
SamYaple | Kennan: are you giving eth1 to neutron? or eth0? | 05:13 |
sdake | functional-f21 uses build-all-docker-images | 05:13 |
SamYaple | oh | 05:13 |
SamYaple | right | 05:13 |
sdake | kennan eth1 should be unconfigured entirely | 05:13 |
Kennan | hi SamYaple: same as doc | 05:13 |
Kennan | etwork_interface: "eth0" | 05:13 |
Kennan | neutron_external_interface: "eth1" | 05:13 |
Kennan | eth1 is my public ssh ip | 05:14 |
Kennan | I can not ping it now | 05:14 |
sdake | eth0 needs to be your public ssh ip | 05:14 |
sdake | eth0 = management netowrk eth1=unconfigured network connected to external net | 05:14 |
Kennan | hi sdake: yes, seems some difference, my eth0 has 10.** ip, and eth1 has 9.** ip | 05:15 |
sdake | samyaple so you have tested centos source and ubuntu source? | 05:15 |
Kennan | is it needed for eth0 public ip ? | 05:15 |
Kennan | managment network seems not need public ip | 05:16 |
SamYaple | Kennan: you can always use eth0 as your neutron_external_interface | 05:16 |
SamYaple | i dont know your network configuration | 05:16 |
sdake | i have eth0 connected to the internet and my local network | 05:16 |
SamYaple | sdake: yea but not centos full stack deploy, just file validation | 05:16 |
sdake | i have eth1 connected to a local switch | 05:16 |
sdake | samyaple ok so just to confirm the only thing that brings up a horizon ui atm is ubuntu source? | 05:17 |
SamYaple | the only thing i have tested the full stack on, yes | 05:17 |
sdake | rogerthanks | 05:17 |
Kennan | sdake: for simple, if my eth0 is private 10.** ip and eth1 is internet public ip | 05:18 |
sdake | samyaple note that horizon centos binary did work with compose | 05:18 |
Kennan | what network configuration should I configure in kolla ? | 05:18 |
sdake | make neutron net eth0, make the other eth1 | 05:18 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 05:19 |
SamYaple | sdake: that means nothing for ansible! i have to lay down the web config | 05:19 |
Kennan | sdake: do you mean network_interface = eth0 neutron_external_interface eth1 ? | 05:19 |
sdake | right just pointing it it did work | 05:19 |
sdake | kennan no i just said the exact opposite | 05:19 |
Kennan | ok sdake: | 05:20 |
Kennan | let me try | 05:20 |
sdake | you will need to reboot | 05:20 |
sdake | your system is snaufed | 05:20 |
sdake | charlie foxtrotted | 05:20 |
sdake | etc | 05:20 |
SamYaple | ha | 05:22 |
SamYaple | sdake the gate was working before.... but not like you would expect :) | 05:23 |
SamYaple | it just _happened_ to work | 05:23 |
SamYaple | lol | 05:23 |
SamYaple | awesome | 05:23 |
*** dolpher has joined #kolla | 05:25 | |
sdake | coolsvap please place priority on reviewing this and children: https://review.openstack.org/#/c/211286/ | 05:29 |
openstackgerrit | Merged stackforge/kolla: Ubuntu - Sync directories https://review.openstack.org/216369 | 05:30 |
coolsvap | sdake, yes, i am looking at the horizon change | 05:30 |
sdake | coolsvap thx bro :) | 05:32 |
sdake | kennan did that get you rolling? | 05:32 |
Kennan | sdake: I need rebuild my instance, as it can not connect | 05:32 |
sdake | your running in an openstack vm? | 05:33 |
sdake | man lots of -1s in the review queue :( | 05:33 |
Kennan | yes sdake | 05:33 |
sdake | and your eth0 is what your private network? | 05:34 |
sdake | and eth1 is your public network ? | 05:34 |
Kennan | yes, you can say that, eth1 is with floating ip(internet access) | 05:34 |
sdake | eth1 in this case is your managment network then | 05:35 |
sdake | so you need neutron net set to eth0 | 05:35 |
sdake | and the othernetwork set to eth1 | 05:35 |
sdake | you have them backwards | 05:35 |
sdake | in yoru globals.yaml file | 05:35 |
sdake | you also need to set a libvirt override | 05:35 |
sdake | to allow vm in vm operation | 05:35 |
sdake | this is unfortunately undocumented | 05:36 |
Kennan | ok sdake: I found kolla have only neutron_external_interface and network_interface | 05:36 |
coolsvap | SamYaple, can you please address a little comment https://review.openstack.org/#/c/216026/10/docker_templates/horizon/Dockerfile.j2 | 05:36 |
Kennan | I would set eth1 with network_interface | 05:36 |
sdake | neutron_external_interface = eth0, network_interface = eth1 | 05:36 |
sdake | you will be unable to connect to your vms,but atleast you will be able to see them boot ;) | 05:37 |
SamYaple | coolsvap: sure ill look in a minute! | 05:37 |
Kennan | sdake: not sure what's the libvirt you talked about | 05:37 |
Kennan | what need to change ? | 05:37 |
sdake | samyaple can tell you the override | 05:37 |
sdake | you set a value in /etc/kolla/nova.conf | 05:37 |
sdake | its like | 05:37 |
sdake | [libvirt] | 05:37 |
sdake | something = qemu | 05:37 |
Kennan | ok sdake: so all you kolla env is bare metal ? | 05:38 |
sdake | its standard run in vm nova.conf setting | 05:38 |
sdake | right | 05:38 |
sdake | kolla master does not autodetect if its running in vm or not | 05:38 |
sdake | you have to override that | 05:38 |
SamYaple | /etc/kolla/nova/nova-compute.conf | 05:38 |
Kennan | I did not have extra baremetal resource | 05:38 |
sdake | samyaple is correct | 05:38 |
Kennan | so need to use VM to try kolla | 05:38 |
sdake | you can use a vm | 05:38 |
sdake | to play around | 05:38 |
sdake | i run kolla in vms all the time - works fine | 05:39 |
sdake | (using virutalbox) | 05:39 |
sdake | but you have your network setup backwards :) | 05:40 |
sdake | its unfortunate this point isn't obvious | 05:40 |
sdake | our docs obviously need attention | 05:40 |
sdake | upstream drinking from firehose atm | 05:40 |
*** dolpher1 has joined #kolla | 05:42 | |
sdake | also your eth0 can't be configured | 05:42 |
sdake | so unconfgure it | 05:42 |
sdake | we can tackle init-runonce when you get a deploy working ;) | 05:43 |
*** ashishb has joined #kolla | 05:43 | |
*** dolpher has quit IRC | 05:45 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 05:45 |
SamYaple | sdake: please test that ^^ | 05:45 |
SamYaple | it works for me | 05:45 |
SamYaple | if it works for pip instal lfor you | 05:45 |
SamYaple | it should work for the gate too | 05:45 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 05:46 |
SamYaple | or rather, that one | 05:46 |
sdake | your wish is my command | 05:47 |
sdake | samyaple http://ur1.ca/njp9l -> http://paste.fedoraproject.org/258829/14404818 | 05:51 |
sdake | sudo python setup.py install | 05:51 |
sdake | the gate doesn't use kolla-build | 05:53 |
SamYaple | sdake: i know why that is broken | 05:53 |
sdake | the last patch worked with kolla-build | 05:53 |
SamYaple | hold on | 05:53 |
sdake | coolsvap wtb +2s or -1s plz | 05:56 |
sdake | https://review.openstack.org/#/c/216026/10 | 05:57 |
sdake | its hard to work on centos binary of horizon if we have a bunch of dependent patches in gerrit | 05:57 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 05:58 |
SamYaple | sdake: alright try it now | 05:58 |
sdake | ok moment | 05:58 |
SamYaple | this time with less dirty hack! | 05:58 |
SamYaple | coolsvap: i responded to you in https://review.openstack.org/#/c/216026/ | 06:00 |
SamYaple | let me know if you need more info | 06:00 |
sdake | that seems to work samyaple | 06:01 |
SamYaple | well see what the gate has to say | 06:01 |
sdake | samyaple i agree with coolsvap the code is hard to comprehend :) | 06:01 |
sdake | i even ran kolla-build from / and it worked properly :) | 06:01 |
SamYaple | btw soure install is broken for kolla-build until you tell me where build.ini is going | 06:01 |
sdake | ya let me think about that when i haven't been up for 20 hours ;) | 06:02 |
SamYaple | sounds like a plan | 06:02 |
SamYaple | if you place it in /usr/shade/kolla it iwll work as-is :) | 06:02 |
sdake | there is no shade directory ;) | 06:02 |
SamYaple | i still need to add an override options for the buildini conf | 06:02 |
*** dolpher has joined #kolla | 06:02 | |
SamYaple | yea yea you got the point :) | 06:02 |
sdake | kennan did you get bck to starting again? | 06:03 |
sdake | kennan really keen to get you rolling | 06:03 |
Kennan | sdake: as the rebuild instance means full clean, so it going to build image step now. | 06:03 |
sdake | i think /usr/share/kolla is a good spot for it | 06:03 |
Kennan | build image is slow | 06:03 |
SamYaple | sdake: then it works as is, just needs a new patch | 06:04 |
sdake | take 15 min for me | 06:04 |
SamYaple | to move it i think | 06:04 |
sdake | we can install it i think | 06:04 |
sdake | no idea how build.py works ;) | 06:04 |
SamYaple | what do you mean? | 06:04 |
*** dolpher1 has quit IRC | 06:04 | |
SamYaple | is it not clear? | 06:04 |
sdake | ya pretty much just not an expert on that part of python | 06:04 |
SamYaple | i have a function that will return a proper base directory | 06:05 |
SamYaple | i 'guess' that base directory by some common factors | 06:05 |
sdake | sorry not build.py | 06:05 |
SamYaple | oh | 06:05 |
sdake | i meant setuppy | 06:05 |
SamYaple | oh yea | 06:05 |
sdake | setup.py | 06:05 |
SamYaple | i get you there | 06:05 |
* sdake need m0ar redbull | 06:05 | |
sdake | really more keen to get docs publishing rolling | 06:06 |
sdake | its a resource we are not taking advantage of | 06:07 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 06:07 |
SamYaple | sdake: and DocImpact | 06:07 |
SamYaple | we could have a docs team dude | 06:07 |
sdake | i left some comments in that patch SamYaple | 06:07 |
SamYaple | lots of people lik that stuff ive seen | 06:07 |
SamYaple | which patch? | 06:08 |
sdake | the gate patch | 06:08 |
sdake | lsat revision | 06:08 |
SamYaple | i responded before | 06:08 |
sdake | ok | 06:08 |
SamYaple | can I change a patch owner? | 06:08 |
SamYaple | i hate working on patches that arent in my queue | 06:08 |
SamYaple | its hard to keep track of | 06:09 |
SamYaple | ha! | 06:09 |
SamYaple | i can see the gate jobs early! | 06:09 |
sdake | yes but its typically considered rude ;) | 06:09 |
SamYaple | i didnt know that | 06:09 |
SamYaple | sdake: how? | 06:10 |
SamYaple | also obviously only with permission | 06:10 |
SamYaple | im a stickler for that | 06:10 |
sdake | google something appropriate | 06:10 |
sdake | its like the first link I think :) | 06:10 |
SamYaple | no | 06:10 |
SamYaple | owner | 06:10 |
sdake | kennan let me know when you get to deploy i'll google that override for you :) | 06:10 |
SamYaple | not author | 06:10 |
SamYaple | difference | 06:11 |
sdake | oh owner | 06:11 |
sdake | not sure on that | 06:11 |
SamYaple | author is at the git level | 06:11 |
SamYaple | owner is gerrir level | 06:11 |
sdake | dont know what the difference is :) | 06:11 |
Kennan | thanks sdake: waiting for build image | 06:11 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 06:12 |
SamYaple | sdake do you know about http://status.openstack.org/zuul/ | 06:13 |
SamYaple | ? | 06:13 |
sdake | yes of course | 06:13 |
SamYaple | i did too but i didnt realize i could see the logs early | 06:13 |
sdake | ya I use that extensively during gate bringup | 06:13 |
SamYaple | woulda been nice to know | 06:13 |
sdake | didn't realize you weren't aware of it or i would have offered the inf o:) | 06:13 |
SamYaple | well i was aware of the site, not hte functinality | 06:14 |
sdake | its pretty cool to watch during a busy time | 06:14 |
sdake | crazy to see the amount of stuff zuul is serializing | 06:14 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 06:16 |
SamYaple | sdake: all gate jobs are building with that latest patch for pip install :)( | 06:16 |
openstackgerrit | Merged stackforge/kolla: Ubuntu - Horizon https://review.openstack.org/216026 | 06:18 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 06:23 |
SamYaple | coolsvap: https://review.openstack.org/#/c/211286/ | 06:24 |
SamYaple | last horizon patch! | 06:24 |
SamYaple | w00t | 06:24 |
*** unicell1 has quit IRC | 06:24 | |
coolsvap | SamYaple, yes was waiting for the dependent patch to merge | 06:24 |
sdake | coolsvap two reviews for you plz | 06:24 |
sdake | https://review.openstack.org/#/c/211286/ | 06:24 |
SamYaple | sdake: alreayd hit him up | 06:24 |
*** unicell has joined #kolla | 06:25 | |
sdake | https://review.openstack.org/#/c/216463/ | 06:25 |
SamYaple | so centos fedora doesnt come with mkfs.xfs stuff | 06:27 |
SamYaple | awesome! | 06:27 |
SamYaple | well centos does | 06:27 |
SamYaple | fedora doesnt | 06:27 |
sdake | fwiw I haen't added fedora gates | 06:27 |
sdake | just ubuntu trusty | 06:27 |
SamYaple | yea but i dont want to break our current gate | 06:27 |
sdake | oh right | 06:28 |
sdake | i am removing that gate shortly | 06:28 |
SamYaple | i know but still | 06:28 |
sdake | but probably not in the next week | 06:28 |
SamYaple | and it should just change to a fedora build gate i think | 06:28 |
SamYaple | we have the fedora code now | 06:28 |
sdake | did someone fix fedora? | 06:28 |
SamYaple | no i mean gate code | 06:29 |
SamYaple | but this way you wont have to do another patch later | 06:29 |
SamYaple | we can fix fedora | 06:29 |
sdake | yup | 06:29 |
sdake | you dont have to argue :) | 06:29 |
openstackgerrit | Merged stackforge/kolla: Add ansible support for Horizon https://review.openstack.org/211286 | 06:29 |
sdake | its all good | 06:29 |
SamYaple | w00t | 06:29 |
SamYaple | ubuntu merged | 06:29 |
SamYaple | now to clean up the installs | 06:29 |
sdake | how does one go about retagging an image and pushign to a local registry | 06:30 |
SamYaple | comeon ive told you this | 06:30 |
SamYaple | docker tag/docker push | 06:30 |
sdake | i dont know the exact syntax for my local registry | 06:31 |
sdake | going to test right now | 06:31 |
SamYaple | its all the same syntax | 06:31 |
SamYaple | whats your regsitry? | 06:31 |
sdake | 192.168.1.100:4000 | 06:31 |
*** unicell has quit IRC | 06:32 | |
*** unicell1 has joined #kolla | 06:32 | |
SamYaple | assuming oyu have 192.168.1.100:4000/kollaglue/ubuntu-source-horizon ..... | 06:32 |
SamYaple | docker tag 192.168.1.100:4000/kollaglue/ubuntu-source-horizon 192.168.1.100:4000/kollaglue/centos-binary-horizon | 06:32 |
SamYaple | docker push 192.168.1.100:4000/kollaglue/centos-binary-horizon | 06:32 |
SamYaple | remember you can build only horizon locally now | 06:32 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 06:33 |
SamYaple | ~/kolla/tools/build.py -n 192.168.1.100:4000/kollaglue --base ubuntu --type source --template -T1 horizon | 06:33 |
sdake | samyaple i dont undertand that lat point | 06:33 |
SamYaple | sdake: ^ | 06:33 |
sdake | ok let me c&p that in | 06:33 |
SamYaple | that will just build you the horizon image | 06:33 |
SamYaple | then oyu can run the first two commands to retag it and push it | 06:34 |
sdake | horizon uses only port 80? | 06:35 |
SamYaple | dont have ssl support anywhere in kolla yet | 06:35 |
SamYaple | ssl requires a few pieces | 06:35 |
SamYaple | one step at a time | 06:35 |
sdake | yes but the console log stuff uses some other port right? | 06:35 |
SamYaple | no | 06:35 |
SamYaple | it shouldnt | 06:35 |
SamYaple | ah crap | 06:36 |
SamYaple | that ansible code wont work for you | 06:36 |
SamYaple | since you use centos/ubuntu frankenstien ness | 06:36 |
SamYaple | it says 'hey you use centos' and lays down httpd file | 06:36 |
sdake | does it detect that from the type and base? | 06:37 |
SamYaple | yea | 06:37 |
sdake | oh i see the problem | 06:37 |
SamYaple | https://github.com/stackforge/kolla/blob/master/ansible/roles/horizon/templates/horizon.conf.j2#L1 | 06:37 |
sdake | my base is centos | 06:37 |
SamYaple | no its that line from above | 06:37 |
SamYaple | itll lay down a cionfig telling it to log to httpd | 06:37 |
SamYaple | it wil probably crash it | 06:37 |
SamYaple | try it out htough! | 06:37 |
SamYaple | hey did oyu know you dont have t orename the image? | 06:38 |
SamYaple | you can push the ubuntu named image | 06:38 |
SamYaple | in globals.yml specify 'horizon_image_full: 192.168.1.100:4000/kollaglue/ubuntu-source-horizon:latest' | 06:39 |
SamYaple | that will override the image it deploys | 06:39 |
SamYaple | no image retagging hackness | 06:39 |
*** SourabhP has joined #kolla | 06:40 | |
*** SourabhP_ has joined #kolla | 06:41 | |
sdake | is horizon default to enabled? | 06:43 |
SamYaple | yup | 06:44 |
*** SourabhP has quit IRC | 06:45 | |
*** SourabhP_ is now known as SourabhP | 06:45 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 06:46 |
SamYaple | sdake: ^^ that just passed all the gates! | 06:46 |
SamYaple | it works for oyu with pip | 06:46 |
sdake | ship it | 06:46 |
SamYaple | and it works in the same way current master works | 06:46 |
sdake | coolsvap ^^ | 06:46 |
SamYaple | mandre had a comment | 06:46 |
SamYaple | i fixed it so its got the gates again | 06:47 |
SamYaple | but it was a syntax thing, not a functional thing | 06:47 |
SamYaple | so itll pass the gate again | 06:47 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 06:49 |
sdake | samyaple is this the correct syntax? [sdake@bigiron kolla]$ tools/build.py -n 192.168.1.100:4000/kollaglue --base ubuntu --type source --template -T1 horizon --push | 06:49 |
SamYaple | looks good to me | 06:49 |
sdake | odd it prints no message about pushing | 06:50 |
SamYaple | sdake: i think i have teh api wrong for it | 06:50 |
SamYaple | i dont know though | 06:51 |
SamYaple | i never dug into it | 06:51 |
SamYaple | docker may just not return text as a response and i would have to do other things to check | 06:51 |
SamYaple | I have a improve logging patch | 06:51 |
SamYaple | as well as a remove teir patch | 06:51 |
SamYaple | because man does that cause some fights | 06:51 |
sdake | ya diags | 06:52 |
sdake | always brings out the best in people ;-) | 06:52 |
SamYaple | lol | 06:52 |
SamYaple | at least we are all passionate | 06:52 |
*** SourabhP has quit IRC | 06:52 | |
*** SourabhP has joined #kolla | 06:53 | |
SamYaple | that gate patch will be good to go as soon as it hits an hp host and i can test that... | 07:02 |
*** SourabhP has quit IRC | 07:02 | |
*** achanda has quit IRC | 07:10 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 07:11 |
*** inc0 has joined #kolla | 07:11 | |
inc0 | good morning | 07:13 |
SamYaple | morning inc0 | 07:13 |
*** achanda has joined #kolla | 07:14 | |
openstackgerrit | Merged stackforge/kolla: Make Kolla installable via PIP https://review.openstack.org/216341 | 07:17 |
sdake | sup inc0 | 07:17 |
SamYaple | gate is now offically broken... | 07:17 |
inc0 | ow? why is that? | 07:18 |
SamYaple | patch that just merged :) | 07:18 |
SamYaple | the next merging patch fixes it | 07:18 |
inc0 | hmm...nothing to review | 07:19 |
inc0 | damn | 07:19 |
coolsvap | inc0, :) | 07:19 |
inc0 | but I see lots of new names in review queue, that's a good sign | 07:20 |
*** shardy has joined #kolla | 07:21 | |
coolsvap | yes | 07:22 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - gate hacking https://review.openstack.org/216005 | 07:26 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 07:28 |
*** achanda has quit IRC | 07:30 | |
Kennan | hi sdake: the containers run now | 07:36 |
sdake | kennan nice | 07:37 |
sdake | did you set the libvirt override? | 07:37 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 07:37 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - gate hacking https://review.openstack.org/216005 | 07:37 |
openstackgerrit | Merged stackforge/kolla: Modified build.py to run from installed location https://review.openstack.org/216463 | 07:39 |
Kennan | sdake: I did not set nova related, maybe I not know if env work or not now :) | 07:39 |
Kennan | I just run that playbooks not find error | 07:39 |
sdake | what did you set where | 07:39 |
sdake | oh you didn't set | 07:39 |
sdake | sorry bleary eyed | 07:39 |
sdake | you need to set an override to run virt on virt | 07:40 |
sdake | [libvirt] | 07:40 |
sdake | virt_type = qemu | 07:40 |
sdake | run sudo tools/cleanup-containers -fv | 07:40 |
sdake | edit /etc/kolla/nova/nova-compute.conf (iirc) | 07:40 |
sdake | put that [librt] and virt_type = qemu in there | 07:40 |
sdake | rerun playbooks | 07:41 |
sdake | sudo tools/kolla-ansible deploy | 07:41 |
sdake | kennan run time sudo tools/kolla-ansible deploy | 07:41 |
Kennan | sdake: /etc/kolla/nova-compute/nova.conf | 07:42 |
sdake | see what the performance is on deploy | 07:42 |
Kennan | is this ? | 07:42 |
sdake | i am curious | 07:42 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Update Haproxy for Horizon https://review.openstack.org/216582 | 07:42 |
sdake | thats the one kennan | 07:42 |
Kennan | not find libvirt on this file | 07:42 |
Kennan | maybe it has default values for kvm ? | 07:42 |
sdake | you need to write it in | 07:43 |
sdake | new content | 07:43 |
sdake | it defaults to kvm yes | 07:43 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Fix regression in build-script https://review.openstack.org/216395 | 07:44 |
Kennan | ok sdake: I am trying with delete containers as you said | 07:45 |
SamYaple | alright can i get some reviews on this to get the gate printing correclty again? https://review.openstack.org/#/c/216395/ | 07:45 |
SamYaple | now that we have merged the other stuff | 07:45 |
Kennan | and would use ansible-playbook , (seems kolla-ansible deploy is same) | 07:45 |
SamYaple | sdake: https://review.openstack.org/#/c/216582/ | 07:47 |
SamYaple | thats your horizon ansible haproxy patch | 07:47 |
SamYaple | multinode wil lwork but the auth and tokens arent shared yet | 07:48 |
sdake | shouldnt that port be configurable? | 07:48 |
SamYaple | yea but it isnt yet | 07:48 |
SamYaple | its got some reasons why | 07:49 |
SamYaple | i want to get all distros working first | 07:49 |
sdake | file bug plz so we dont lose track of it | 07:49 |
SamYaple | will update blueprint | 07:49 |
sdake | wfm | 07:49 |
mandre | good morning | 07:53 |
SamYaple | hey mandre ! | 07:53 |
mandre | it's good to be back, I should be more available on irc now | 07:54 |
sdake | yo mandre | 07:55 |
sdake | so samyaple just deployed horizon, it doesn't work | 07:55 |
SamYaple | how so? | 07:55 |
sdake | econn refused | 07:55 |
sdake | not restarting | 07:55 |
SamYaple | multinode? | 07:55 |
SamYaple | oh right youre not running with that haproxy patch right? | 07:56 |
sdake | only listening on 192.168.1.100 | 07:56 |
sdake | not on .148 | 07:56 |
SamYaple | i dont know what those ips are | 07:56 |
SamYaple | is 148 the vip? | 07:56 |
sdake | 148 = my kolla internal | 07:56 |
sdake | 100 is my machine | 07:56 |
sdake | ya 148 is vip | 07:56 |
SamYaple | yea haproxy patch man | 07:56 |
SamYaple | i said it ^^^^ | 07:56 |
sdake | can someone approe that plz | 07:56 |
SamYaple | https://review.openstack.org/#/c/216582/ | 07:56 |
SamYaple | you can use it on 100 for right now sdake | 07:57 |
SamYaple | it should work there | 07:57 |
sdake | pokey but works | 07:57 |
SamYaple | pokey as in problems? | 07:58 |
SamYaple | or just wierd you have tohit 100?> | 07:58 |
Kennan | hi sdake: do you know where can find OS_PASSWORD ? | 07:59 |
SamYaple | password.yml Kennan | 07:59 |
SamYaple | that holds all your passwords | 07:59 |
Kennan | ok | 08:00 |
openstackgerrit | Steven Dake proposed stackforge/kolla: DO NOT MERGE - Workaround for horizon https://review.openstack.org/216592 | 08:00 |
sdake | KENNAN its password | 08:00 |
SamYaple | sdake: thats teh same as my patch | 08:00 |
SamYaple | i dont get it | 08:00 |
SamYaple | oh nvm | 08:01 |
sdake | did you submit a patch for that? | 08:01 |
SamYaple | sorry wrong copy buffer | 08:01 |
sdake | mandre or coolsvap can you guys ack https://review.openstack.org/#/c/216582/ plz | 08:02 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Setup the gate for cross-distro support https://review.openstack.org/213397 | 08:02 |
inc0 | sdake, SamYaple about horizon...shouldn't we go for https? | 08:03 |
SamYaple | inc0: yea thats coming | 08:03 |
sdake | inc0 baby steps | 08:03 |
sdake | get it working on centos binary first | 08:03 |
SamYaple | we dont do ssl anywhere right now | 08:03 |
SamYaple | we need memcached or some other mechanism and things that haven't been tested yet | 08:04 |
Kennan | sdake: found /centos-binary-rabbitmq has issue, cause neutron can not communicate with that | 08:04 |
SamYaple | the whole point of installing haproxy 1.5+ is because of its ssl support too | 08:04 |
Kennan | INFO - This container has already been configured; Refusing to copy new configs | 08:04 |
Kennan | ERROR: epmd error for host cookup: address (cannot connect to host/port) | 08:04 |
SamYaple | is the port already in use Kennan ? | 08:05 |
sdake | kennan what base os are you running on? | 08:05 |
Kennan | centos 7.1 | 08:05 |
SamYaple | 4369 i think is the port epmd tries to bind | 08:05 |
sdake | netstat -nl | grep 4369 | 08:06 |
SamYaple | netstat? what is this the dark ages? | 08:06 |
SamYaple | ss is the way to go | 08:06 |
Kennan | $ sudo netstat -nl | grep 4369 | 08:06 |
Kennan | get nothing | 08:06 |
sdake | kennan what was yoru deploy time | 08:06 |
Kennan | sdake: I forgot to time the ansible playbook instrutions | 08:07 |
Kennan | maybe serveral mins | 08:07 |
sdake | kennan type that error into google plz | 08:07 |
Kennan | sure | 08:07 |
sdake | kennan your the first person to report that error | 08:08 |
*** jmccarthy has quit IRC | 08:08 | |
sdake | is that hte neutron contianer or rabbit container | 08:08 |
Kennan | rabbit container (logs) | 08:08 |
Kennan | rabbit has issue which cause neutron cli not work | 08:08 |
sdake | fwiw i have seen that error before a long time ago | 08:09 |
sdake | iirc it was because hostname wasn't installed | 08:09 |
sdake | check google | 08:09 |
*** jmccarthy has joined #kolla | 08:09 | |
sdake | wtb AIO gating | 08:09 |
Kennan | I will ask google :) | 08:09 |
SamYaple | hey another core reviewing https://review.openstack.org/#/c/212830/ would be fantastic | 08:10 |
SamYaple | would love to remove that cruft | 08:10 |
sdake | speculation is that master is broken by recent ubuntu changes | 08:10 |
SamYaple | sdake: i only adjusted the templates | 08:10 |
SamYaple | so thats doubtful... | 08:10 |
sdake | ya iirc its because a tool isn't installed that rabbit needs | 08:11 |
sdake | or hostname isnt' set in /etchosts | 08:11 |
sdake | or something like that | 08:11 |
sdake | that is what it is | 08:11 |
sdake | hostname wasn't set in /etc/hosts | 08:11 |
SamYaple | sdake: https://github.com/stackforge/kolla/commit/fe2e454f267ea023d3e3447d743c340d93e96539 | 08:11 |
SamYaple | thats all that was changed | 08:11 |
SamYaple | its possible the rabbitmq.config is breaking it.... but thats not related to epmd | 08:12 |
SamYaple | and its straight from rabbitmq site | 08:12 |
SamYaple | and i tested it... | 08:12 |
SamYaple | but nevertheless possible | 08:12 |
sdake | perhaps something with /etc/hosts changed in the rabbitmq contianer? | 08:13 |
sdake | I remember putting in a workaround for that specific problem | 08:13 |
SamYaple | not by my hands | 08:13 |
sdake | and if you type erro into google it comes back with an exct workaround | 08:13 |
sdake | kennan are you building with --templates? | 08:14 |
sdake | or without? | 08:14 |
Kennan | just run this | 08:14 |
Kennan | tools/build.py -T 1000 | 08:14 |
vbel | good morning/evening | 08:14 |
SamYaple | without templates | 08:14 |
SamYaple | morning vbel | 08:14 |
vbel | is it possible at all to see build process via python? | 08:14 |
vbel | just wonder what is going on :) | 08:15 |
SamYaple | vbel: like realtime output? | 08:15 |
vbel | yes | 08:15 |
SamYaple | you cna with single threaded | 08:15 |
SamYaple | -T1 | 08:15 |
SamYaple | if you want parallel it will spit out all the diffent threads at teh same time | 08:15 |
SamYaple | hard to read | 08:15 |
vbel | great, thanks!!! | 08:15 |
sdake | nice my kolla_external_address is showing up | 08:15 |
sdake | in horizon | 08:16 |
SamYaple | yea its set as an endpoint | 08:16 |
SamYaple | thats the whole point | 08:16 |
SamYaple | `openstack endpoint list` | 08:16 |
sdake | yes i understand how it works now thanks :) | 08:18 |
vbel | How do I specify my own registry? like localhost:5000 | 08:20 |
*** zhiwei has joined #kolla | 08:20 | |
vbel | for pushing from build.py | 08:20 |
SamYaple | vbel: in the configs or when building? | 08:20 |
vbel | building | 08:21 |
sdake | dont use port 5000 it will collide with keystone | 08:21 |
sdake | i use port 400 | 08:21 |
sdake | 4000 | 08:21 |
vbel | yes, I know, I use 9999 on hosts | 08:21 |
vbel | :) | 08:21 |
vbel | thanks | 08:21 |
SamYaple | ~/kolla/tools/build.py -n 192.168.31.12:8082/kollaglue --base ubuntu --type source | 08:21 |
SamYaple | thats what i use vbel | 08:21 |
SamYaple | (obviously you dont have to build ubuntu) | 08:22 |
Kennan | hi SamYaple: | 08:22 |
Kennan | http://paste.openstack.org/show/426825/ | 08:22 |
vbel | ok, so this is part of namespace then. Thanks! | 08:22 |
Kennan | I hit neutron issue in kolla containers | 08:22 |
Kennan | very strange | 08:22 |
sdake | kennan we know about that one | 08:22 |
sdake | what about the rabbitmq one | 08:22 |
*** dolpher has quit IRC | 08:22 | |
SamYaple | Kennan: thats teh centos binary issue sdake | 08:22 |
sdake | you did some kind of workaround and got er working? | 08:22 |
SamYaple | there is some issue with delorean centos binary packages | 08:23 |
sdake | samyaple yes nothing we can do about busted upstream | 08:23 |
* SamYaple uses source | 08:23 | |
SamYaple | :) | 08:23 |
Kennan | http://paste.openstack.org/show/426826/ | 08:23 |
Kennan | this is about rabbitmq | 08:23 |
Kennan | I just configured /etc/hosts with ip and hostname, it seems ok now | 08:23 |
SamYaple | Kennan: i dont know what to tell you about that | 08:23 |
SamYaple | sdake seems ot know more | 08:24 |
Kennan | the strange thing is that before I configured that, both hostname and hostanme -f can show | 08:24 |
sdake | i am certain i had a workaround for that in the codebsae at one time | 08:24 |
SamYaple | as you can see by my linked commit, i did not remove it | 08:24 |
Kennan | hi SamYaple: I want to know what's the proper os can work with kollla containers | 08:24 |
Kennan | as neutron container not work | 08:24 |
sdake | rabbitmq had a bunch of code removal recently iirc | 08:24 |
SamYaple | Kennan: I use ubuntu-source | 08:24 |
Kennan | I can not verify functions in network | 08:24 |
SamYaple | it works great | 08:24 |
Kennan | SamYaple: ubunutu-source ? 15.04 or 14.04 ? | 08:25 |
Kennan | I remember you said 14.04 with 3.19 kernel | 08:25 |
Kennan | ? | 08:25 |
sdake | SamYaple i am pretty sure i saw a patch go by that removed like 50% of rabbitmq.s start.sh | 08:25 |
sdake | maybe that was mariadb | 08:25 |
sdake | kennan please file a bug | 08:26 |
sdake | and link to channel | 08:26 |
sdake | on the rabbitmq bug | 08:26 |
Kennan | hi sdake: maybe not kolla bug, I noticed | 08:27 |
Kennan | that VM have two ips (if I not configured in /etc/hosts specifcly) | 08:27 |
SamYaple | Kennan: that is correct | 08:28 |
Kennan | it would default hostname with ip about private ip | 08:28 |
Kennan | if I configrued public ip with such hostname in /etc/hosts | 08:28 |
Kennan | that can rabbimq work | 08:28 |
Kennan | it seems kolla depends some for rabbitmq host configuration | 08:28 |
SamYaple | sdake: that was removed way back in config-internal | 08:29 |
SamYaple | it was never used in config-external | 08:29 |
SamYaple | ever | 08:29 |
SamYaple | found the code youre talking about htough | 08:29 |
sdake | wierd i wonder why my machine works | 08:29 |
SamYaple | https://github.com/stackforge/kolla/commit/1a145d8a010d05762a0101bb96517eeb5cb52e29 | 08:29 |
SamYaple | config-rabbitmq.sh | 08:29 |
Kennan | rabbitmq-env.conf used public ip (which I think kolla config) | 08:29 |
SamYaple | but again, that code was never ever used in config-external (i wrote the config external piece) | 08:29 |
Kennan | so it need to use such ip for the hostname | 08:29 |
Kennan | if use other ip, the rabbitmq-env.conf would think wrong (i think) | 08:30 |
sdake | - # work around: | 08:30 |
sdake | - # https://bugs.launchpad.net/ubuntu/+source/rabbitmq-server/+bug/653405 | 08:30 |
openstack | Launchpad bug 653405 in rabbitmq-server (Ubuntu Oneiric) "rabbitmq-server fails to start if hostname is unresolvable or has changed since first starting" [High,Fix released] - Assigned to Marc Cluet (lynxman) | 08:30 |
sdake | - HOSTNAME=`/usr/bin/hostname -s 2> /dev/null || /bin/hostname 2> /dev/null` | 08:30 |
sdake | - echo "${RABBITMQ_SERVICE_HOST} ${HOSTNAME}" > /etc/hosts | 08:30 |
*** dolpher has joined #kolla | 08:30 | |
sdake | well bedtime for me | 08:30 |
sdake | horizon works nice job samyaple | 08:31 |
sdake | kennan please file a bug | 08:31 |
Kennan | ok sdake: I would find another OS to make it work now | 08:31 |
*** diga has quit IRC | 08:31 | |
Kennan | as centos neutron not work | 08:31 |
sdake | kennan build from source | 08:31 |
Kennan | really sad | 08:31 |
SamYaple | Kennan: ubuntu-source is the smallest bestest one right now! | 08:31 |
sdake | what is sad? | 08:31 |
Kennan | yes, SamYaple: is it a Cloud oS? | 08:31 |
sdake | delorean releases every 3 days | 08:31 |
SamYaple | Kennan: no its perfect. build from source is perfect for openstack in docker containers | 08:31 |
sdake | neutron cats merged a busted patch | 08:32 |
openstackgerrit | Merged stackforge/kolla: Update Haproxy for Horizon https://review.openstack.org/216582 | 08:32 |
sdake | and delorean snapshotted right then | 08:32 |
sdake | samyaple pleae relay commands to build from source | 08:32 |
SamYaple | ~/kolla/tools/build.py --base ubuntu --type source | 08:32 |
*** weiyu_ has quit IRC | 08:33 | |
SamYaple | ~/kolla/tools/build.py --base ubuntu --type source --template | 08:33 |
SamYaple | you have to use --template for ubuntu | 08:33 |
Kennan | SamYaple: if I understand correctly, it means ubuntu 14.04, I need either change docker configuration for btrfs | 08:33 |
sdake | kennan after you do that you need to change centos to ubuntu and binary to source | 08:33 |
Kennan | or complier kernel with 3.19 | 08:33 |
sdake | in globals.yml | 08:33 |
*** weiyu has joined #kolla | 08:33 | |
sdake | kennan just build from source on centos - may or may not work ;) | 08:34 |
SamYaple | Kennan: you dont need to compile the 3.19 kernel | 08:34 |
SamYaple | its in the repo | 08:34 |
Kennan | OK. let me try your /kolla/tools/build.py --base ubuntu --type source --template | 08:34 |
Kennan | let me verify if it work or not | 08:34 |
sdake | kennan dont forget to change globals.yml | 08:35 |
SamYaple | Kennan: apt-get install linux-image-generic-lts-vivid | 08:35 |
SamYaple | thats teh 3.19 kernel | 08:35 |
SamYaple | for trusty | 08:35 |
*** weiyu has quit IRC | 08:35 | |
coolsvap | sdake, minor comment in the compose patch, https://review.openstack.org/#/c/212830/16/compose/glance-api-registry.yml | 08:37 |
coolsvap | also if you can remove WIP since its ready for review :) | 08:37 |
sdake | VOLUMES need to be dealt with | 08:38 |
* sdake groans | 08:38 | |
SamYaple | sdake: in the next patch | 08:38 |
openstackgerrit | Merged stackforge/kolla: Fix regression in build-script https://review.openstack.org/216395 | 08:38 |
SamYaple | the compose stuff has been busted anyway | 08:38 |
SamYaple | we can merge this | 08:38 |
SamYaple | then fix compose and VOLUMES | 08:38 |
SamYaple | i mean the compose files are hard broke right now | 08:39 |
Kennan | hi SamYaple: do you know when use that build from source, what need to change for globals.vml ? according to sdake said | 08:39 |
Kennan | does any doc cover this ? | 08:39 |
coolsvap | even if its busted the image name is wrong in only glance compose file | 08:39 |
sdake | im still going to bed :) | 08:40 |
sdake | night | 08:40 |
* coolsvap brb | 08:40 | |
sdake | thanks for looking at it tho coolsvap | 08:40 |
*** weiyu_ has joined #kolla | 08:40 | |
coolsvap | sdake, if we are going to merge it i will update the change | 08:40 |
SamYaple | why not just remove compose from the change? | 08:41 |
SamYaple | compose is all busted to hell anyway | 08:41 |
SamYaple | its not going to be working anytime soon | 08:41 |
sdake | becuase eeveryone bitched when i didn't adjut compose | 08:41 |
SamYaple | well compose is broke now | 08:41 |
SamYaple | i dont want to hold up this patch anylonger | 08:41 |
sdake | i agree it shoud be removed from the repo entirely imo | 08:41 |
SamYaple | ill submit a new compose patch to fix things | 08:41 |
SamYaple | lets just merge this | 08:41 |
SamYaple | its 16 patchsets deep | 08:41 |
SamYaple | and 12 days | 08:42 |
SamYaple | this is what we need to merge | 08:42 |
sdake | mostly because of the first patch | 08:42 |
pbourke | morning all | 08:42 |
sdake | still going to bed | 08:42 |
SamYaple | morning pbourke! | 08:42 |
sdake | night :) | 08:42 |
pbourke | was out yesterday | 08:42 |
SamYaple | coolsvap: lets merge it as is ill submit the compose patch | 08:42 |
coolsvap | SamYaple, alright | 08:42 |
SamYaple | coolsvap: compose is all busted and needs way more love | 08:42 |
pbourke | sdake: night! | 08:42 |
*** shardy_ has joined #kolla | 08:43 | |
coolsvap | sdake, night | 08:43 |
pbourke | SamYaple: for you comment on the swift config.yml - I tried with with_items loop but ansible doesn't seem to like that when using include? | 08:43 |
coolsvap | SamYaple, done! | 08:44 |
* coolsvap brb | 08:44 | |
SamYaple | you used to be able to pbourke, did that change? | 08:44 |
SamYaple | what do you mean doesnt like? | 08:44 |
*** shardy has quit IRC | 08:44 | |
pbourke | SamYaple: it explicitly bombs with an error something like "we don't do this anymore for includes" | 08:44 |
pbourke | SamYaple: I did google it friday evening, let me see if I can find the links again | 08:45 |
SamYaple | its in v2 | 08:45 |
openstackgerrit | Merged stackforge/kolla: WIP: Remove data containers and adjust compose files https://review.openstack.org/212830 | 08:46 |
sdake | pbourke SamYaple please sort out getting swift merged keeping in mind I want to retain 1.9.z compatibility in our playbooks.... | 08:48 |
sdake | i'd like it merged today if possible | 08:48 |
sdake | TIA! | 08:48 |
*** shardy_ has quit IRC | 08:48 | |
SamYaple | pbourke: is swift in a functional state right now? | 08:48 |
pbourke | SamYaple: yes | 08:48 |
SamYaple | well come back to this issue then because it needs to be a different config per service for update reasons | 08:49 |
*** shardy has joined #kolla | 08:49 | |
SamYaple | if you update it once it wont know to restart al lthe services | 08:49 |
SamYaple | pbourke: it looks like youll need a rebase but im fine then | 08:49 |
SamYaple | ill poke at it later | 08:49 |
pbourke | SamYaple: ok, I'll push a rebase and log a bug for the above | 08:49 |
SamYaple | thanks pbourke | 08:52 |
SamYaple | can you make sure to retest after the rebase? | 08:52 |
SamYaple | weve had alot of changes | 08:52 |
*** shardy has quit IRC | 08:52 | |
*** shardy has joined #kolla | 08:54 | |
SamYaple | pbourke: kolla is not pip install-able | 08:54 |
SamYaple | now* | 08:54 |
pbourke | whats the reason? | 08:55 |
SamYaple | what do you mean? | 08:55 |
pbourke | on friday we were discussing how to make kolla a python package | 08:55 |
pbourke | are you saying that's not possible? | 08:56 |
SamYaple | i corrected myself | 08:56 |
SamYaple | 08:55 < SamYaple> pbourke: kolla is not pip install-able | 08:56 |
SamYaple | 08:55 < SamYaple> now* | 08:56 |
pbourke | ah :) | 08:56 |
SamYaple | the patches merged | 08:56 |
pbourke | great! | 08:56 |
SamYaple | also horizon is working (at least on ubuntu-source) | 08:56 |
SamYaple | merged | 08:56 |
SamYaple | i can one click install ubuntu-source openstack and get horizon and instances booting | 08:57 |
pbourke | wow | 08:57 |
pbourke | so no JS errors on horizon? | 08:57 |
SamYaple | no i fixed all that | 08:57 |
pbourke | fantastic | 08:58 |
SamYaple | just a pathing issue | 08:58 |
SamYaple | its kinda confusing what its asking for | 08:58 |
SamYaple | so its easy to messup the path | 08:58 |
SamYaple | and then there is openstack_dashboard and openstack-dashboard | 08:58 |
SamYaple | both needed in diferent places :) | 08:58 |
pbourke | yeah stuff confused me | 08:58 |
pbourke | *that stuff | 08:58 |
SamYaple | indeed | 08:58 |
pbourke | that's really good news though | 08:59 |
SamYaple | i never tackled it in yaodu ofr a reason | 08:59 |
SamYaple | btw, this is where kolla pulls away from yoadu! | 08:59 |
SamYaple | horizon is the first ansible thing merged that yaodu didnt have | 08:59 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216621 | 09:04 |
*** weiyu_ has quit IRC | 09:05 | |
*** weiyu_ has joined #kolla | 09:05 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216622 | 09:06 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216623 | 09:06 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216624 | 09:07 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216629 | 09:22 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 09:22 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216631 | 09:22 |
inc0 | https://www.youtube.com/watch?v=EF0lVn3n12I good tune | 09:23 |
SamYaple | i prefer https://youtu.be/8XFBUM8dMqw | 09:24 |
pbourke | is anyone still getting this rabbitmq permission denied error | 09:24 |
pbourke | has been hanging around since we changed to single data image :? | 09:25 |
SamYaple | pbourke: are yo uusing master? | 09:25 |
SamYaple | vbel fixed that issue | 09:25 |
pbourke | that's what I thought | 09:25 |
SamYaple | it was a problem with ansible | 09:25 |
pbourke | yeah using master it's weird | 09:25 |
*** akwasnie has joined #kolla | 09:25 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216631 | 09:26 |
pbourke | rm -f on containers seems to have done the trick | 09:27 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216629 | 09:27 |
SamYaple | pbourke: yea you would have to rebootstrap rabbit | 09:27 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 09:27 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 09:31 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216635 | 09:33 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216636 | 09:33 |
*** shardy_ has joined #kolla | 09:33 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216637 | 09:34 |
*** shardy has quit IRC | 09:35 | |
*** weiyu_ has quit IRC | 09:37 | |
vbel | pbourke: check what is inside of the container, /opt/kolla/config-rabbit.sh | 09:38 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hackingd https://review.openstack.org/216635 | 09:38 |
*** shardy_ has quit IRC | 09:39 | |
*** shardy has joined #kolla | 09:39 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: sDO NOT MERGE - Gate Hacking https://review.openstack.org/216636 | 09:41 |
*** weiyu has joined #kolla | 09:44 | |
vbel | common guys, where is build.py? :) | 09:45 |
vbel | ../kolla/cmd/build.py: No such file or directory | 09:45 |
SamYaple | vbel: kolla/kolla/cmd/build.py is where it was moved to | 09:45 |
SamYaple | you can stil laccess it in tools/build.py | 09:46 |
inc0 | I like that about kolla | 09:46 |
SamYaple | whats that? | 09:46 |
inc0 | I'm off for night and dir structure is completely different when I wake up | 09:46 |
inc0 | it's full of surprises | 09:46 |
SamYaple | its cause it is a pip packges now :P | 09:46 |
SamYaple | it had to change | 09:46 |
SamYaple | i made sure to preserve existing structure | 09:47 |
SamYaple | all the commands should still work the same | 09:47 |
SamYaple | but sdake wanted t orush it in for reasons | 09:47 |
inc0 | I don't have problems with that, it just happends from time to time | 09:48 |
SamYaple | yea i know what you mean | 09:48 |
*** dims__ has joined #kolla | 09:48 | |
*** dims__ has quit IRC | 09:48 | |
SamYaple | its why im always here | 09:48 |
SamYaple | cant be missin out | 09:48 |
*** dims__ has joined #kolla | 09:49 | |
vbel | SamYaple: no cmd after checkout, also github site has no cmd in kolla root | 09:52 |
SamYaple | vbel: i never said kolla root | 09:52 |
SamYaple | 09:46 < SamYaple> vbel: kolla/kolla/cmd/build.py is where it was moved to | 09:53 |
vbel | SamYaple: found it! thanks :) | 09:53 |
Kennan | hi SamYaple: | 09:53 |
Kennan | I am tring build from source on centos VM now | 09:54 |
Kennan | tools/build.py --base ubuntu --type source --template | 09:54 |
Kennan | is it ? | 09:54 |
SamYaple | that looks right | 09:54 |
SamYaple | keep in mind no one has ever tested building ubuntu images on centos | 09:54 |
SamYaple | just fyi | 09:54 |
Kennan | hi SamYaple: yes seems have issue, just hang | 10:00 |
Kennan | INFO:__main__:Processing: base | 10:00 |
Kennan | not any output until now | 10:00 |
SamYaple | Kennan: its fine | 10:00 |
SamYaple | the logging needs some work, but its processing | 10:01 |
SamYaple | it will spit out output at the end of each image biuld | 10:01 |
Kennan | SamYaple: did you suggest build from source ? if I understand correctly | 10:01 |
Kennan | that's why I tried this way :) | 10:01 |
SamYaple | I did, but I run on ubuntu not centos | 10:01 |
SamYaple | ive never build ubuntu source from centos | 10:02 |
*** pbourke has quit IRC | 10:03 | |
Kennan | oh SamYaple: maybe centos VM is not good choice now :0 | 10:04 |
Kennan | if centos even not work with build from source | 10:05 |
SamYaple | yea that was my point. It should work, its just never been tested | 10:05 |
Kennan | I would choose ubuntu MV | 10:05 |
Kennan | VM | 10:05 |
SamYaple | I only use ubuntu from source at this point, so i can attest to that working | 10:05 |
*** pbourke has joined #kolla | 10:05 | |
SamYaple | all the rest i havent fully tetsed in over a week | 10:05 |
Kennan | ok SamYaple: so ubuntu OS build source for ubuntu image ? | 10:06 |
SamYaple | Thats what i test on, yes | 10:06 |
Kennan | thanks guy | 10:06 |
*** mandre has quit IRC | 10:07 | |
coolsvap | pbourke, https://review.openstack.org/#/c/216295/ needs changes | 10:11 |
*** mandre has joined #kolla | 10:20 | |
Kennan | hi SamYaple: | 10:33 |
Kennan | did build from source have logs ? | 10:33 |
Kennan | it just hang on INFO:__main__:Processing: base | 10:33 |
Kennan | I did not know what happened | 10:33 |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 10:34 |
SamYaple | Kennan: if yo uwant to see output, use `tools/build.py --base ubuntu --type source --template -T1` | 10:36 |
*** shardy_ has joined #kolla | 10:37 | |
*** shardy has quit IRC | 10:38 | |
SamYaple | vbel: i like it! | 10:40 |
SamYaple | good job | 10:40 |
SamYaple | a few comments for you | 10:40 |
vbel | thanks! | 10:41 |
*** zhiwei has quit IRC | 10:41 | |
*** shardy_ has quit IRC | 10:43 | |
SamYaple | vbel: to geenerate the exact same tarball each time with the git clone will be tricy | 10:43 |
SamYaple | ticky* | 10:43 |
SamYaple | tricky* | 10:43 |
*** shardy has joined #kolla | 10:43 | |
vbel | yes, since it ever-growing thing | 10:44 |
SamYaple | youll have to change the timestamps for the files internally and maybe other things | 10:44 |
SamYaple | well no you could do a shallow clone | 10:44 |
*** vbel has quit IRC | 10:44 | |
SamYaple | just keep in mind, git cloning will not be able to use the cache as is | 10:44 |
*** vbel has joined #kolla | 10:44 | |
*** jruano has joined #kolla | 10:44 | |
SamYaple | we can merge this now though | 10:44 |
SamYaple | (im mean once its reviewed) | 10:44 |
*** britthouser has quit IRC | 10:45 | |
SamYaple | we can fix the cache at a later date once we have thought it through | 10:45 |
*** britthouser has joined #kolla | 10:47 | |
*** pbourke has quit IRC | 11:28 | |
*** pbourke has joined #kolla | 11:28 | |
*** dolpher has quit IRC | 11:34 | |
*** rhallisey has joined #kolla | 11:35 | |
*** Haomeng|2 has joined #kolla | 11:39 | |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 11:41 |
*** Haomeng has quit IRC | 11:42 | |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 11:44 |
*** weiyu has quit IRC | 11:51 | |
*** shardy_ has joined #kolla | 11:58 | |
*** shardy has quit IRC | 12:00 | |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 12:04 |
*** shardy_ has quit IRC | 12:04 | |
*** shardy has joined #kolla | 12:05 | |
coolsvap | rhallisey, if you can rebase https://review.openstack.org/#/c/216295/ on the recent build.py changes, its ready for merge | 12:06 |
SamYaple | coolsvap: can you remove you -2 on ryans patch | 12:06 |
SamYaple | yea that | 12:06 |
rhallisey | I did | 12:06 |
rhallisey | SamYaple, thanks for fixed the summary output. I was wondering what was going on there | 12:06 |
coolsvap | sorry for that | 12:06 |
rhallisey | coolsvap, no problem | 12:06 |
inc0 | hey, guys, -2 is usually "we don't want it here" | 12:07 |
inc0 | and it means death of a patchset | 12:07 |
SamYaple | inc0: its also used to hard block | 12:07 |
SamYaple | which we need inc0 | 12:07 |
inc0 | -1 from core should be enough | 12:07 |
SamYaple | in this ase it was so the pip stuff could merge | 12:07 |
rhallisey | no worries coolsvap you haven't gotten the new core onboarding from steak | 12:07 |
SamYaple | inc0: no it needed a hard blcok so it didnt merge | 12:07 |
SamYaple | it can merge with -1 | 12:07 |
SamYaple | core or not | 12:07 |
inc0 | but it won't merge without +2 +w | 12:08 |
coolsvap | pbourke, did approve it, so i did a hard block | 12:08 |
SamYaple | sure but if a core doesnt see that -1 then ic can accidentally merge, hence the -2 to ensure it doesnt inc0 | 12:08 |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 12:08 |
inc0 | well, guys...don't merge something with -1 from core then;) -2 usually is death of patchset, and people will percieve it as such | 12:09 |
coolsvap | rhallisey, i think it will fail pep8 | 12:09 |
SamYaple | inc0: i will -2 to block for the patch we were waiting on | 12:10 |
SamYaple | that is normal behaviour | 12:10 |
rhallisey | coolsvap, k let me confirm | 12:10 |
inc0 | coolsvap, I don't think it will...I actually confirmed that yesterday | 12:10 |
coolsvap | rhallisey, i already did | 12:10 |
inc0 | unless something bad changed since then | 12:10 |
coolsvap | ./tools/build.py:35:1: H306 imports not in alphabetical order (requests.exceptions.connectionerror, jinja2) | 12:11 |
coolsvap | import jinja2 | 12:11 |
coolsvap | ^ | 12:11 |
inc0 | yeah, but that's problem of whole script | 12:11 |
SamYaple | inc0: no it isnt | 12:11 |
SamYaple | its not in alphabetical order | 12:11 |
SamYaple | r is after j | 12:11 |
inc0 | I mean import order is pep8-broken | 12:12 |
inc0 | requests are in middle of imports for example | 12:12 |
SamYaple | the patchset is out of order | 12:12 |
inc0 | and it shouldn't be | 12:12 |
SamYaple | that is form this patchset | 12:12 |
coolsvap | SamYaple, i agree with inc0, actually | 12:12 |
SamYaple | https://review.openstack.org/#/c/216295/12/kolla/cmd/build.py | 12:12 |
SamYaple | thats in this patchset though | 12:12 |
SamYaple | i dont understand | 12:13 |
inc0 | requests is there for long time | 12:13 |
coolsvap | but this patchset should is not meant to doeal with it | 12:13 |
SamYaple | requests is not in the base patchset | 12:13 |
coolsvap | i will send a patchset which will make more sense to what we are talking about | 12:13 |
SamYaple | that is literally added for this patchset | 12:13 |
SamYaple | im sorrt, the ConnectionError from requests* | 12:13 |
SamYaple | import requests is still up top | 12:13 |
coolsvap | something of this sort is recommendation http://docs.openstack.org/developer/hacking/#import-order-template | 12:13 |
inc0 | SamYaple, https://github.com/stackforge/kolla/blob/master/kolla/cmd/build.py#L25 | 12:14 |
SamYaple | thats not what pep8 is complaining about inc0 | 12:14 |
inc0 | but it is pep8-broken anyway | 12:14 |
SamYaple | its complaining about L34 | 12:14 |
SamYaple | how so is the question | 12:14 |
inc0 | requests is 3rd party library | 12:14 |
inc0 | normal pep8 tool doesn't check that | 12:14 |
SamYaple | now thats fair | 12:15 |
inc0 | requests is 3rd party lib and it should be below stdlib imports | 12:15 |
coolsvap | ordering is misplaced | 12:15 |
inc0 | also, https://github.com/stackforge/kolla/blob/master/kolla/cmd/build.py#L30 is wrong | 12:15 |
inc0 | from imports should be below normal imports | 12:15 |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 12:15 |
rhallisey | k there we go | 12:16 |
SamYaple | inc0: you are wrong there | 12:16 |
SamYaple | http://docs.openstack.org/developer/hacking/#import-order-template | 12:16 |
rhallisey | let me know where you want requirements. I'm not picky at all | 12:16 |
* rhallisey moves on | 12:16 | |
coolsvap | rhallisey, :) | 12:17 |
inc0 | yeah, I give you that, from imports aren't specified explicitly in pep8 | 12:19 |
rhallisey | brb | 12:20 |
*** rhallisey has quit IRC | 12:21 | |
SamYaple | inc0: i guess its just interpretation then | 12:21 |
SamYaple | but flake8 does check that | 12:21 |
inc0 | yeah it does | 12:23 |
inc0 | anyway, requests are in wrong place anyway:P | 12:23 |
SamYaple | agreed | 12:23 |
vbel | import-order-template :))) somewhat confusing when it comes to the lines with project imports | 12:23 |
SamYaple | i hadn't noticed it was 3rd party in the wrong spot | 12:23 |
*** chandankumar has joined #kolla | 12:32 | |
SamYaple | alright guys | 12:41 |
SamYaple | who is ready for another round of build.py!?! | 12:41 |
SamYaple | inc0 you up for it? | 12:41 |
*** weiyu_ has joined #kolla | 12:41 | |
SamYaple | im ready for some critic | 12:41 |
inc0 | we have this as topic for tomorrows meeting | 12:42 |
inc0 | let's discuss that there | 12:42 |
SamYaple | no this is an incomping patchset | 12:42 |
SamYaple | ;) | 12:42 |
inc0 | well, you know I have issues with it:P let me finish rsyslog first tho | 12:43 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove tiered building in build.py https://review.openstack.org/216691 | 12:44 |
SamYaple | there ya go inc0 | 12:44 |
SamYaple | tear it up | 12:44 |
SamYaple | find_parents is most likely the area of discussion fyi | 12:45 |
inc0 | well...I see you didn't remove debug ifs I guess;) | 12:46 |
inc0 | line 392 | 12:47 |
SamYaple | i think i already commented on that :D | 12:47 |
inc0 | you know that if self.images would be dict we wouldn't need parents dict? | 12:49 |
SamYaple | thats not true at all | 12:50 |
SamYaple | im adding to the parents dict and rechecking | 12:50 |
SamYaple | this is honeslty the best of both worlds that we talkedabout yesterday | 12:50 |
SamYaple | self.images doesnt need to be a dict, but we can still do the dit lookup like you wanted | 12:50 |
inc0 | using different dict;) | 12:51 |
SamYaple | because this wouldnt work with self.images even if it were a dict inc0 | 12:51 |
SamYaple | look at it again | 12:51 |
inc0 | ofc it wouldn't but you could implement normal tree with it easily | 12:52 |
SamYaple | im confused, are stil lsaying self.images should be a dict? | 12:53 |
SamYaple | this was teh only use case where oy uwanted that functionality and we have it | 12:53 |
SamYaple | and i agree with it for sure, just not in self.images since it would be pointless | 12:53 |
* SamYaple is trying to make inc0 happy :( | 12:53 | |
SamYaple | i worked on this all night to solve your issues with the biuld script | 12:54 |
inc0 | http://paste.openstack.org/show/427027/ | 12:55 |
SamYaple | no inc0 | 12:56 |
SamYaple | it has ot be the object | 12:56 |
SamYaple | not the name | 12:56 |
inc0 | well...it doesn't if we change that above | 12:57 |
SamYaple | no im telling you children has to be the object | 12:57 |
inc0 | but that only means we remove key from there | 12:57 |
SamYaple | i inject that data into the queue | 12:57 |
inc0 | it works | 12:57 |
SamYaple | brb im cool with a simpleir change | 12:58 |
SamYaple | will check inin a few | 12:58 |
pbourke | inc0: would you have a quick example of a regex to build multiple images? | 12:58 |
pbourke | inc0: tried build.py 'keystone|galera' | 12:58 |
pbourke | but no good | 12:58 |
inc0 | pbourke, build.py neutron ceilometer | 12:59 |
pbourke | ah | 12:59 |
pbourke | thanks | 12:59 |
*** dwalsh has joined #kolla | 12:59 | |
*** klint has quit IRC | 13:00 | |
inc0 | SamYaple, how to debug kolla_ansible container | 13:03 |
inc0 | ? | 13:03 |
inc0 | endpoint times out and I wanna know why | 13:03 |
vbel | https://review.openstack.org/#/c/216655/ - pep8 does not complain, I will be happy to build kilo with it ;) | 13:08 |
*** weiyu_ has quit IRC | 13:10 | |
vincent_vdk | hi, is it possible to run the network service spread over the 3 control nodes when deploying multinode? | 13:15 |
inc0 | vincent_vdk, have you tried to simply adding 3 nodes in network in ansible inventory? | 13:17 |
inc0 | should work | 13:17 |
vincent_vdk | no, but i would like to use the control nodes for that | 13:17 |
vincent_vdk | but not sure it will work because the example inventory uses separate hosts | 13:18 |
*** jtriley has joined #kolla | 13:21 | |
*** dims__ has quit IRC | 13:24 | |
*** dims__ has joined #kolla | 13:25 | |
*** diogogmt has joined #kolla | 13:30 | |
*** ashishb has quit IRC | 13:31 | |
*** absubram has joined #kolla | 13:32 | |
pbourke | is anyone having an issue where mariadb is trying to bootstrap on all nodes | 13:33 |
*** absubram has quit IRC | 13:35 | |
*** dims has joined #kolla | 13:36 | |
*** dims__ has quit IRC | 13:36 | |
*** dims_ has joined #kolla | 13:39 | |
*** dims__ has joined #kolla | 13:41 | |
*** dims has quit IRC | 13:42 | |
*** dims_ has quit IRC | 13:45 | |
*** macjack has joined #kolla | 13:48 | |
*** inc0 has quit IRC | 13:48 | |
*** macjack has quit IRC | 13:48 | |
*** shardy_ has joined #kolla | 13:50 | |
*** shardy has quit IRC | 13:51 | |
*** thumpba has joined #kolla | 13:53 | |
pbourke | seems to be around permissions again | 13:55 |
*** shardy_ has quit IRC | 13:55 | |
*** shardy has joined #kolla | 13:56 | |
*** ccrouch has quit IRC | 13:56 | |
*** jruano has quit IRC | 13:57 | |
*** daneyon_ has quit IRC | 13:58 | |
SamYaple | did inc0 leave? | 13:58 |
*** daneyon has joined #kolla | 13:58 | |
*** shardy has quit IRC | 14:00 | |
*** shardy has joined #kolla | 14:01 | |
SamYaple | pbourke: regex do not work | 14:01 |
SamYaple | i was trying to talk to him the hother day | 14:01 |
SamYaple | it only supports word matches | 14:01 |
SamYaple | liek you cant do just '^base$' | 14:02 |
pbourke | hmm | 14:02 |
pbourke | mysqld: Can't create/write to file '/var/lib/mysql/aria_log_control' (Errcode: 13 "Permission denied") | 14:02 |
SamYaple | what image? | 14:02 |
pbourke | this is a symptom of out of date data containers? | 14:02 |
pbourke | I've rebuilt all with --no-cache from master | 14:02 |
SamYaple | it very well could be | 14:02 |
SamYaple | i would recommend removeing your containers and purging all images | 14:03 |
pbourke | have done | 14:03 |
SamYaple | master works for me right now, i just did a quick demo | 14:03 |
SamYaple | but im using ubuntu source | 14:03 |
pbourke | i'll keep at it | 14:03 |
SamYaple | sorry :/ | 14:03 |
*** diogogmt has quit IRC | 14:04 | |
pbourke | not your fault | 14:04 |
pbourke | probably ;) | 14:04 |
SamYaple | with the sick burn ohhhh | 14:04 |
pbourke | sucks though cause I need to get swift rebased | 14:05 |
*** chandankumar has quit IRC | 14:06 | |
vincent_vdk | so if you want to deploy on multiple hosts i guess it's adviced to run your own docker registry | 14:10 |
pbourke | vincent_vdk: yes | 14:15 |
pbourke | vincent_vdk: see the included vagrantfile for an example | 14:15 |
*** ftcpops has joined #kolla | 14:16 | |
vincent_vdk | ik had issues with latest docker and registry < 2.x | 14:18 |
vincent_vdk | couldn't push images | 14:18 |
vincent_vdk | but thanks for the pointer | 14:18 |
SamYaple | do not use docker registry 2.0 locally | 14:18 |
SamYaple | its bad | 14:18 |
SamYaple | and slower | 14:18 |
vincent_vdk | SamYaple: ave you tried 2.1? | 14:18 |
SamYaple | use the 'latest' tag, it still points to the 1.x registry | 14:18 |
*** ccrouch has joined #kolla | 14:18 | |
SamYaple | yea | 14:18 |
*** ccrouch has quit IRC | 14:18 | |
*** ccrouch has joined #kolla | 14:18 | |
SamYaple | all the 2.0 requires compression so its much slower locally | 14:18 |
vincent_vdk | works fine here | 14:19 |
SamYaple | did the fix the duplicate tags pushing up | 14:19 |
SamYaple | if you push the same image twice it will fully upload twice | 14:19 |
SamYaple | or at least it did on the last registry 2 that i ran | 14:19 |
vincent_vdk | ah, would have to check that | 14:20 |
vbel | registry 2.1 failed - I could not push anything to it | 14:21 |
vbel | just buggy image | 14:21 |
vbel | I use registry:2.0 | 14:21 |
SamYaple | yea i would not recommend that | 14:21 |
vbel | registry:2 == registry:2.1 = registry:latest | 14:21 |
SamYaple | latest tag is still 0.9.1 | 14:22 |
SamYaple | for good reason | 14:22 |
vbel | sorry, registry:2 == registry:2.1 != registry:latest | 14:22 |
SamYaple | https://hub.docker.com/_/registry/ | 14:22 |
vincent_vdk | vbel: strange, i had that issue until I updated to 2.1 | 14:23 |
vincent_vdk | i would have to check again | 14:23 |
vbel | vincent_vdk: I am getting error after it successfully pushed to 2.1 | 14:24 |
vincent_vdk | one coould actually use tarballs :) | 14:25 |
vincent_vdk | without layers for prod | 14:25 |
vincent_vdk | and skip the whole registry thing | 14:25 |
vbel | if you want to open stuff (for production) it is the easiest to use docker hub. This registry image is for devops :) | 14:26 |
vincent_vdk | :) | 14:27 |
vincent_vdk | after all the tripleo based setups i'm going to give kolla another try | 14:28 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove tiered building in build.py https://review.openstack.org/216691 | 14:28 |
SamYaple | pbourke: I also included super in here for oyu https://review.openstack.org/#/c/216691/2/kolla/cmd/build.py | 14:30 |
SamYaple | also i changed the 'parent' key reference from a string to an image by creating a seperate key 'parent_name' | 14:31 |
*** ashishb has joined #kolla | 14:38 | |
*** diogogmt has joined #kolla | 14:53 | |
*** ccrouch has quit IRC | 14:54 | |
*** achanda has joined #kolla | 14:54 | |
*** ccrouch has joined #kolla | 14:55 | |
openstackgerrit | Merged stackforge/kolla: Fail faster when docker isn't running https://review.openstack.org/216295 | 14:57 |
pbourke | no such option: --install-option | 14:58 |
pbourke | when building kolla-ansible | 14:58 |
pbourke | looks like a bug | 14:59 |
*** achanda has quit IRC | 15:04 | |
pbourke | SamYaple: can you build kolla-ansible on ubuntu base? | 15:10 |
*** inc0 has joined #kolla | 15:11 | |
inc0 | afternoon | 15:11 |
*** jruano has joined #kolla | 15:11 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove tiered building in build.py https://review.openstack.org/216691 | 15:27 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove tiered building in build.py https://review.openstack.org/216691 | 15:29 |
SamYaple | pbourke: nah its broken | 15:29 |
SamYaple | pbourke: i have to change all the source stuff to unify centos and ubuntu | 15:29 |
SamYaple | inc0: hey! | 15:29 |
SamYaple | updated the patch | 15:29 |
SamYaple | take a look and review away | 15:29 |
SamYaple | pbourke: i have a patchset series that fixes all of that | 15:30 |
*** blahRus has joined #kolla | 15:31 | |
*** shardy_ has joined #kolla | 15:31 | |
*** shardy has quit IRC | 15:32 | |
*** shardy_ has quit IRC | 15:36 | |
*** ashishb has quit IRC | 15:36 | |
*** ccrouch has quit IRC | 15:37 | |
*** ccrouch has joined #kolla | 15:37 | |
*** shardy has joined #kolla | 15:37 | |
*** ccrouch has quit IRC | 15:37 | |
*** chandankumar has joined #kolla | 15:37 | |
*** ccrouch has joined #kolla | 15:37 | |
*** ccrouch has quit IRC | 15:38 | |
*** ccrouch1 has joined #kolla | 15:38 | |
*** rhallisey has joined #kolla | 15:40 | |
inc0 | SamYaple, I can't test it right now, but could you please test ./build.py neutron-agents ? will it build base as well? | 15:43 |
SamYaple | yes | 15:44 |
SamYaple | regex all works | 15:44 |
SamYaple | thats why the parent linking is nice, i walk up the parent chain and set status = 'matched' | 15:44 |
SamYaple | regular regex is still borken though | 15:44 |
SamYaple | you cant use ^base$ | 15:45 |
inc0 | yup, that's price of comfort | 15:45 |
inc0 | but neutron.*base should work | 15:46 |
SamYaple | regex is kinda a bad name for that option then | 15:46 |
SamYaple | yea that works | 15:46 |
*** cfarquhar has joined #kolla | 15:57 | |
*** cfarquhar has quit IRC | 15:57 | |
*** cfarquhar has joined #kolla | 15:57 | |
*** akwasnie has quit IRC | 15:59 | |
*** stvnoyes has joined #kolla | 16:03 | |
*** stvnoyes has left #kolla | 16:03 | |
*** rhallisey has quit IRC | 16:06 | |
*** inc0 has quit IRC | 16:06 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove tiered building in build.py https://review.openstack.org/216691 | 16:08 |
SamYaple | pbourke: i have regex fixed in that patch now | 16:08 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: Remove tiered building in build.py https://review.openstack.org/216691 | 16:09 |
SamYaple | sdake: ping | 16:10 |
*** SourabhP has joined #kolla | 16:15 | |
*** diga has joined #kolla | 16:15 | |
*** SourabhP has quit IRC | 16:20 | |
*** diogogmt has quit IRC | 16:24 | |
*** diogogmt has joined #kolla | 16:25 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 16:28 |
openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Add ability to make tarballs from Git clones https://review.openstack.org/216655 | 16:28 |
*** unicell has joined #kolla | 16:30 | |
*** unicell1 has quit IRC | 16:31 | |
*** unicell has quit IRC | 16:31 | |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 16:32 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 16:35 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 16:38 |
openstackgerrit | Sam Yaple proposed stackforge/kolla: DO NOT MERGE - Gate Hacking https://review.openstack.org/216630 | 16:42 |
*** dwalsh has quit IRC | 16:56 | |
*** diga has quit IRC | 17:02 | |
*** chandankumar has quit IRC | 17:03 | |
*** jmccarthy has quit IRC | 17:05 | |
*** unicell has joined #kolla | 17:11 | |
*** ashishb has joined #kolla | 17:12 | |
*** unicell1 has joined #kolla | 17:13 | |
*** unicell has quit IRC | 17:13 | |
*** achanda has joined #kolla | 17:15 | |
bmace | sdake or SamYaple around? | 17:17 |
*** absubram has joined #kolla | 17:22 | |
pbourke | there is something up with mariadb | 17:31 |
pbourke | :( | 17:31 |
harmw | sdake: pong | 17:32 |
*** jasonsb has quit IRC | 17:33 | |
*** jasonsb has joined #kolla | 17:34 | |
*** jasonsb has quit IRC | 17:38 | |
*** loth has joined #kolla | 17:39 | |
*** ftcpops has quit IRC | 17:45 | |
*** akwasnie has joined #kolla | 17:48 | |
*** shardy_ has joined #kolla | 18:04 | |
*** shardy has quit IRC | 18:05 | |
*** shardy_ has quit IRC | 18:09 | |
*** shardy has joined #kolla | 18:10 | |
*** jasonsb has joined #kolla | 18:12 | |
*** tummy has joined #kolla | 18:12 | |
*** britthouser has quit IRC | 18:17 | |
*** britthouser has joined #kolla | 18:17 | |
*** jpeeler has quit IRC | 18:19 | |
*** jpeeler has joined #kolla | 18:22 | |
*** sdake_ has joined #kolla | 18:25 | |
*** ftcpops has joined #kolla | 18:27 | |
*** sdake has quit IRC | 18:29 | |
*** rhallisey has joined #kolla | 18:31 | |
sdake_ | bmace wound me | 18:50 |
sdake_ | pburke yes kennan reported tht earlier | 18:50 |
harmw | sdake_: pong #2 | 18:51 |
*** chandankumar has joined #kolla | 18:52 | |
harmw | and remove the _ :> it hurts | 18:52 |
*** jasonsb has quit IRC | 18:52 | |
*** jasonsb has joined #kolla | 18:53 | |
*** daneyon_ has joined #kolla | 18:53 | |
sdake_ | harmw shoot | 18:54 |
*** sdake_ is now known as sdake | 18:54 | |
harmw | no you, you pinged earlier :) | 18:54 |
*** daneyon has quit IRC | 18:55 | |
sdake | i ws probably looking for a review but i don't reall | 18:57 |
*** jruano has quit IRC | 19:02 | |
*** ccrouch1 has quit IRC | 19:02 | |
*** ccrouch has joined #kolla | 19:03 | |
ftcpops | Hi Kolla - first kolla question for me. I'm trying to deploy kolla on a CentOS 7 VM behind a proxy. I think I have the proxy setup correct, but I can't seem to get past this error: ERROR:__main__:The command '/bin/sh -c yum install -y https://rdoproject.org/repos/openstack-kilo/rdo-release-kilo.rpm' returned a non-zero code: 1 | 19:12 |
*** daneyon_ has quit IRC | 19:12 | |
ftcpops | ANy ideas? | 19:12 |
sdake | yes you need to use a proxy override in the build process | 19:12 |
*** daneyon has joined #kolla | 19:12 | |
*** daneyon has quit IRC | 19:12 | |
sdake | this is undocumented as we just added it few days ago | 19:12 |
ftcpops | Okay thanks. I'm glad I didn't just overlook something obvious | 19:13 |
*** sdake_ has joined #kolla | 19:15 | |
sdake_ | tbh i'm not quite sure how to do the job, but it has to do with the build operation but it involves kolla-build -i | 19:17 |
sdake_ | -i INCLUDE_HEADER, --include-header INCLUDE_HEADER | 19:17 |
sdake_ | Path to custom file to be added at beginning of base | 19:17 |
sdake_ | Dockerfile | 19:17 |
sdake_ | basically I think you put ENV http_proxy=proxyserver | 19:18 |
sdake_ | and one for pip as well | 19:18 |
sdake_ | i don't know the exact details, vbel is your man there | 19:18 |
harmw | https_proxy= probably | 19:18 |
sdake_ | right | 19:18 |
ftcpops | okay, I'm poking around | 19:18 |
*** sdake has quit IRC | 19:18 | |
harmw | I still think that patched dockerd is cleaner, instead of inserting extra files :p | 19:18 |
sdake_ | its totally not documented, i'm not entirely sure how it works but it solves oracle's use case | 19:18 |
sdake_ | which is precisely building behind a proxy server firewall setup | 19:19 |
ftcpops | heh | 19:19 |
ftcpops | proxies are a pain | 19:19 |
sdake_ | harmw kolla community is not delivering forked versions of docker ;-) | 19:20 |
harmw | no lol | 19:20 |
*** jasonsb has quit IRC | 19:21 | |
sdake_ | ftcpops i'd document it but i don't have a proxy server setup environment to do so | 19:22 |
sdake_ | recommend creating a file in ~/include.docerfile | 19:22 |
sdake_ | and putting in that ENV https_proxy=yourproxyserver | 19:22 |
sdake_ | and putting ENV pip_proxy (or whatever the pp proxy command is=yur proxy server | 19:22 |
* sdake_ needs more minerals | 19:23 | |
*** sdake_ is now known as sdake | 19:23 | |
sdake | and using kolla-build -i ~/include.dockerfile | 19:25 |
*** dims__ has quit IRC | 19:25 | |
*** dims has joined #kolla | 19:26 | |
*** achanda has quit IRC | 19:28 | |
sdake | vbel is your man for that question he is in europe | 19:28 |
sdake | i'll ask him to submit a docs change so we dont all struggle with that one ftcpops :) | 19:28 |
ftcpops | okay, I'm still plugging away | 19:28 |
ftcpops | :D | 19:28 |
bmace | hey sdake, was looking at the node_templates_directory env_var. is the intent that it is the location stuff is put that is consumed by the containers and the "empty" stuff is in /etc/kolla on the ansible / deployer node? | 19:30 |
sdake | templates directory is consumed by ansible iirc | 19:31 |
sdake | it belongs in /var/lib/kolla imo not /usr/share/kolla | 19:31 |
sdake | but yes /etc/kolla directory is consumed by deployer mode | 19:31 |
bmace | consumed on the deploy side? | 19:31 |
sdake | bmace i believe so yes | 19:31 |
* sdake wtbs a document titled "questioned-asked-a-milssion-plus-one-times.rst ;) | 19:32 | |
bmace | well, if the templates dir is on the deploy side, how does it get populated? /etc/kolla is easy enough, you just copy kolla/etc/kolla/* recursively into /etc/kolla | 19:32 |
sdake | bmace ansible creates it | 19:32 |
bmace | i did a search for node_templates_directory to see if it was documented anywhere and it isn't. | 19:33 |
*** mfalatic has quit IRC | 19:33 | |
sdake | templates dir is only consumed on the deployed to node | 19:33 |
bmace | kk, that is fine then. | 19:33 |
sdake | at least that is my understanding | 19:33 |
sdake | they are not meant to be user editable i know that ;) | 19:34 |
*** jasonsb has joined #kolla | 19:36 | |
*** ccrouch has quit IRC | 19:41 | |
*** ccrouch has joined #kolla | 19:45 | |
*** ccrouch has quit IRC | 19:50 | |
*** ccrouch1 has joined #kolla | 19:50 | |
ftcpops | I think I'm getting somewhere. I created an include_proxy.dockerfile and used -i to bring it in. I also had to add --template and --no-cache to get it to move past this (I hope) | 19:51 |
ftcpops | Thanks for the pointers! | 19:51 |
*** chandankumar has quit IRC | 19:56 | |
*** ashishb has quit IRC | 20:01 | |
*** jasonsb has quit IRC | 20:02 | |
*** jasonsb has joined #kolla | 20:03 | |
*** achanda has joined #kolla | 20:11 | |
sdake | ftcops cool did that work ? :) | 20:12 |
sdake | you shouldnt have ot use -no-cache every time | 20:12 |
ftcpops | it's still going | 20:14 |
ftcpops | It's on a pretty wimpy VM | 20:14 |
*** tummy has quit IRC | 20:37 | |
sdake | fwiw build on bare metal takes about 15 min with 100mbit network | 20:43 |
sdake | on vm over wireless it will take longer - especiallygoing through a proxy | 20:43 |
*** daneyon has joined #kolla | 20:56 | |
*** achanda has quit IRC | 21:15 | |
*** tummy has joined #kolla | 21:17 | |
*** achanda has joined #kolla | 21:17 | |
*** jtriley has quit IRC | 21:29 | |
sdake | ftcpops if you have benchmarks on your setup feel free to add them to the quicktart guide to set peoples expectations around build time | 21:33 |
ftcpops | :) | 21:33 |
ftcpops | Still going | 21:34 |
sdake | bummer over an hour | 21:34 |
sdake | once you getfinished make sure to use the correct interface for the network_interface and neutron_interface or you could break your vm :) | 21:34 |
*** sdake_ has joined #kolla | 21:45 | |
ftcpops | ok :) | 21:47 |
*** sdake has quit IRC | 21:48 | |
*** thumpba has quit IRC | 21:48 | |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Replace config-external with a JSON file for Neutron-OVS https://review.openstack.org/216874 | 21:50 |
rhallisey | sdake_, ^ let me know if you like that model | 21:50 |
sdake_ | rhallisey its gotta work for ansible dude :) | 21:51 |
rhallisey | why wouldn't that work | 21:51 |
sdake_ | you don't change the ansible files? | 21:52 |
*** shardy has quit IRC | 21:52 | |
rhallisey | no | 21:52 |
rhallisey | because if I did then it probably wouldn't work for tripleo | 21:52 |
sdake_ | looking at patch | 21:52 |
sdake_ | moment | 21:53 |
rhallisey | this will run the service bases on what volumes are present in the container | 21:53 |
rhallisey | eh that made no sense | 21:53 |
rhallisey | this will run the service based on what configs are mounted into the container | 21:53 |
sdake_ | so initially patch looks good | 21:53 |
sdake_ | how does config.json get into the container? | 21:53 |
rhallisey | just put it in Dockerfile | 21:54 |
sdake_ | so tripleo will modify dockerfiles? | 21:54 |
rhallisey | we would modify json file | 21:54 |
sdake_ | right i mean json file | 21:54 |
rhallisey | build it and push it somewhere | 21:54 |
rhallisey | and pull it down from kolla | 21:54 |
rhallisey | or where ever | 21:55 |
rhallisey | this json file can be a massive list | 21:55 |
rhallisey | we just let people add to it | 21:55 |
rhallisey | the container will run based on what's present | 21:55 |
sdake_ | your patch doesn't break ansible deploy? | 21:55 |
rhallisey | testing it now just a sec | 21:55 |
rhallisey | it shouldn't | 21:55 |
sdake_ | i'd like config.py in common dir | 21:55 |
sdake_ | with symlink :) | 21:55 |
rhallisey | ya | 21:56 |
rhallisey | for sure | 21:56 |
rhallisey | I agree | 21:56 |
rhallisey | this is just test patch sorta | 21:56 |
sdake_ | ok so I am +1 on this change but would like other CR reviews | 21:56 |
sdake_ | we really need to fix it properly to have a stable abi | 21:56 |
rhallisey | ya cool | 21:56 |
rhallisey | just looking for opinions | 21:56 |
rhallisey | since you were around I figured I'd ask :) | 21:56 |
sdake_ | this is much better then ian's patch :) | 21:57 |
sdake_ | or your patch that ian submitted | 21:57 |
sdake_ | or whoever wrote that first thing that I almost -2'ed ;) | 21:57 |
sdake_ | config-external needs to stay though | 21:57 |
rhallisey | ya we both just came up with it for a quick fix | 21:58 |
sdake_ | it is what decide beetween copy once and copy always | 21:58 |
rhallisey | oh hold on I missed somthing... | 21:58 |
rhallisey | ya you're right except I handle that.. | 21:58 |
rhallisey | somewhere.. | 21:59 |
rhallisey | hold just a second | 21:59 |
rhallisey | oh looks like docker_templates/base/kolla-common is linked elsewhere | 22:00 |
sdake_ | busy with meeting bbiaf | 22:00 |
sdake_ | in 1 hour i'll be back and we can finish | 22:00 |
*** tummy has quit IRC | 22:01 | |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Replace config-external with a JSON file for Neutron-OVS https://review.openstack.org/216874 | 22:01 |
rhallisey | ok that's what it should look like | 22:01 |
*** sdake_ is now known as sdake | 22:01 | |
rhallisey | sdake, works with ansible | 22:14 |
sdake | rhallisey still in meeting i'll get back to you | 22:14 |
rhallisey | sdake, kk I'll brb | 22:15 |
sdake | what would make it ideal is if ansible laid down that json file | 22:15 |
sdake | even if it was hard coded to begin with | 22:15 |
rhallisey | sdake, kk | 22:15 |
*** absubram has quit IRC | 22:15 | |
rhallisey | brb | 22:16 |
*** rhallisey has quit IRC | 22:16 | |
ftcpops | INFO:__main__:Successfully built images | 22:19 |
ftcpops | INFO:__main__:========================= | 22:19 |
ftcpops | INFO:__main__:Images that failed to build | 22:19 |
ftcpops | INFO:__main__:=========================== | 22:19 |
ftcpops | Is that good or bad? | 22:19 |
*** achanda has quit IRC | 22:19 | |
*** achanda has joined #kolla | 22:21 | |
*** sdake_ has joined #kolla | 22:21 | |
*** sdake has quit IRC | 22:24 | |
*** sdake_ is now known as sdake | 22:24 | |
*** akwasnie has quit IRC | 22:47 | |
*** rhallisey has joined #kolla | 22:49 | |
*** ftcpops has quit IRC | 23:00 | |
*** tummy has joined #kolla | 23:05 | |
*** unicell1 has quit IRC | 23:06 | |
*** tummy has quit IRC | 23:07 | |
*** daneyon has quit IRC | 23:18 | |
*** tummy has joined #kolla | 23:18 | |
sdake | rhallisey ping i'm alive again | 23:25 |
*** bitblt has joined #kolla | 23:28 | |
*** bitblt has quit IRC | 23:28 | |
*** blahRus has quit IRC | 23:31 | |
*** blahRus has joined #kolla | 23:31 | |
rhallisey | sdake, hey | 23:32 |
sdake | rhalisey i meant confg.py should be in docker/common/base/config.py | 23:34 |
sdake | and should probably be written in such a way as to replace set_configs | 23:34 |
sdake | config.py will become our new json decoder | 23:35 |
rhallisey | arn't we deprecating that? | 23:35 |
sdake | that is a peice of the json blueprint | 23:35 |
*** unicell has joined #kolla | 23:35 | |
sdake | drepcating what? | 23:35 |
rhallisey | docker/* | 23:35 |
rhallisey | docker directory | 23:35 |
sdake | all the docker/common files are going to move to docker_templates | 23:35 |
rhallisey | ok | 23:35 |
sdake | at some point | 23:35 |
sdake | but its a common file to be used across projects | 23:35 |
rhallisey | ya agreed | 23:35 |
sdake | rather across containers | 23:36 |
rhallisey | this is just the first pass that will show the logic | 23:36 |
sdake | if we are going to write a json decoder I wan't it written properly ;) | 23:36 |
rhallisey | sure thing | 23:36 |
sdake | and I want it to partse ose.nviron(KOLLA_CONFIG_STRATEGY) | 23:36 |
sdake | os.environ | 23:37 |
sdake | so kolla_common.sh goes byebye | 23:37 |
sdake | long term | 23:37 |
sdake | or myabve even short term :) | 23:37 |
sdake | i'm sure sam will have some feedback but I am not opposed to taking a incremental approach to implementing this work | 23:38 |
*** diogogmt has quit IRC | 23:38 | |
sdake | call config.py set_config.py as well | 23:38 |
sdake | or configs | 23:38 |
sdake | its plural ;) | 23:38 |
sdake | i will leave comments in the review | 23:39 |
rhallisey | kk thx | 23:39 |
sdake | rhallisey are you opposed to a co-author on this patch | 23:39 |
sdake | i think sam will probbly want to rewrite your python code knowing sam ;) | 23:40 |
rhallisey | eh idc | 23:40 |
rhallisey | what's wrong with it lol | 23:40 |
rhallisey | or are you just warning me | 23:40 |
sdake | its not pep8 compliant for one | 23:40 |
rhallisey | ya I know | 23:41 |
rhallisey | ill fix that | 23:41 |
rhallisey | i was talking about makor things | 23:41 |
*** tummy has quit IRC | 23:41 | |
sdake | my first suggestoin and i'd make this now | 23:42 |
sdake | is to set the blueprint appropriately in the commit log | 23:42 |
sdake | before sam has a look at it :) | 23:43 |
sdake | oh i see you have nice :) | 23:44 |
*** alisonh has quit IRC | 23:44 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!