| *** dimsum__ has joined #kolla | 00:02 | |
| *** weiyu has quit IRC | 00:02 | |
| *** zhiwei has joined #kolla | 01:11 | |
| *** thumpba has joined #kolla | 01:19 | |
| *** dayou has joined #kolla | 01:20 | |
| *** alisonh has quit IRC | 01:24 | |
| *** alisonh has joined #kolla | 01:30 | |
| *** erkules has joined #kolla | 01:32 | |
| *** weiyu has joined #kolla | 01:33 | |
| *** erkules_ has quit IRC | 01:34 | |
| *** thumpba has quit IRC | 01:35 | |
| *** banix has joined #kolla | 01:55 | |
| *** thumpba has joined #kolla | 01:57 | |
| *** zhiwei has quit IRC | 01:59 | |
| *** zhiwei has joined #kolla | 02:05 | |
| *** weiyu has quit IRC | 02:07 | |
| *** thumpba_ has joined #kolla | 02:09 | |
| *** apmelton_ has joined #kolla | 02:10 | |
| *** mandre_ has joined #kolla | 02:13 | |
| *** thumpba has quit IRC | 02:15 | |
| *** apmelton has quit IRC | 02:15 | |
| *** mandre has quit IRC | 02:15 | |
| *** mandre_ is now known as mandre | 02:15 | |
| *** weiyu_ has joined #kolla | 02:18 | |
| *** banix has quit IRC | 02:18 | |
| *** dayou has quit IRC | 02:27 | |
| *** thumpba_ has quit IRC | 02:28 | |
| *** slagle_ has joined #kolla | 02:29 | |
| *** slagle has quit IRC | 02:34 | |
| *** dayou has joined #kolla | 02:39 | |
| *** klint has joined #kolla | 02:40 | |
| *** jbweber_ has joined #kolla | 03:05 | |
| *** clayton has quit IRC | 03:08 | |
| *** jbweber has quit IRC | 03:08 | |
| *** clayton has joined #kolla | 03:14 | |
| *** sdake has joined #kolla | 03:14 | |
| sdake | SamYaple i tried your suggeston of the auth protocol, it fails with 404 | 03:15 |
|---|---|---|
| kjelly | Hi guys, I use `tools/build.py --push -T 64 --tag 20150914 -b ubuntu -t source` to build iamge. And use `ansible-playbook -i inventory/all-in-one -e @/etc/kolla/globals.yml -e @/etc/kolla/passwords.yml site.yml -vvvv` to deploy openstack. the deployment will fail because there are no tables in keystone database. Does anyone have the same problem. | 03:22 |
| sdake | whats with the --tag operation? | 03:22 |
| kjelly | sdake: I use `--tag` to add tag to docker image. | 03:23 |
| sdake | ar eyou setting that tag in the globalas.yml file? | 03:24 |
| sdake | if not its probably pulling lateset rfrom the docker hug | 03:24 |
| kjelly | sdake: yes | 03:24 |
| sdake | i would suggest buliding without tag | 03:25 |
| kjelly | I have `openstack_release: "20150914"` in my global.yml | 03:25 |
| sdake | aer you pushing to a local registry? | 03:26 |
| kjelly | sdake: I push to docker hub. So I also assign namespace. | 03:27 |
| kjelly | use -n | 03:27 |
| sdake | you elft out the -n operation in your build.py command | 03:27 |
| sdake | fwiw last time i ttried using build.py to push to the registry i had to mdoify the code to use secure authentication | 03:28 |
| SamYaple | sdake: you did something wrong then | 03:29 |
| kjelly | sdake: How to build local registry ? I found the document https://docs.docker.com/registry/ . But I don't know how to let build.py push image to my registry. | 03:29 |
| SamYaple | kjelly: do you know how to push docker images normally? | 03:29 |
| sdake | samyaple paste your diff | 03:30 |
| SamYaple | kjelly: im not sure what your knowlege of docker is | 03:30 |
| sdake | saying "you ddi it wrong" is completely nohelpful | 03:30 |
| kjelly | SamYaple: I know how to push image to docker hub. | 03:30 |
| SamYaple | kjelly: have you done it to a private registry before? | 03:30 |
| kjelly | SamYaple: I use `docker run -d -p 5000:5000 --name registry registry:2` to build private registry. | 03:31 |
| SamYaple | locally you really shouldn't be using v2 registry | 03:31 |
| sdake | kjelly are youd oing that on thesame host your deploying to? | 03:31 |
| SamYaple | its slow and has a bunch of issues | 03:32 |
| sdake | keystone also uses port 5000 | 03:32 |
| SamYaple | but besides that to push to a locally registry with docker is a bit wierd | 03:32 |
| SamYaple | it requires the registry to be _in_ the image name | 03:32 |
| SamYaple | so you can do something like this (this is what i use): | 03:32 |
| SamYaple | ~/kolla/tools/build.py -n 192.168.31.12:8082/kollaglue --base ubuntu --type source --push | 03:32 |
| SamYaple | that would push to a reistry at 192.168.31.12:8082 | 03:32 |
| openstackgerrit | Merged stackforge/kolla: Properly check failed status https://review.openstack.org/220808 | 03:33 |
| kjelly | I see. I try the way to build images. And deploy again. | 03:35 |
| SamYaple | make sure to update your globals.yml to point to the private registry | 03:35 |
| kjelly | SamYaple: Ok | 03:36 |
| SamYaple | sdake: my guess is you probably just used password deferred option and thats why it didn't work for you | 03:42 |
| kjelly | sdake: The machine which is used for installing openstack is diffent from the machine which is used for running docker registry. | 03:42 |
| *** sdake_ has joined #kolla | 03:42 | |
| sdake_ | kjelly mke sure not ot run the registry on the same port as keystone | 03:43 |
| SamYaple | kjelly: what you are doing is fine | 03:44 |
| SamYaple | kjelly: what sdake is getting at is 5000 will conflict with the default keystone port but if your registry is a different host youre fine | 03:44 |
| *** sdake has quit IRC | 03:44 | |
| kjelly | But I always failed to deploy openstack using ansible because there are no table in keystone database :( | 03:45 |
| sdake_ | your pulling from the docker hub not your local registry | 03:46 |
| sdake_ | docker hub = docker inc's registry | 03:46 |
| sdake_ | that will fail probably as you are seeing | 03:46 |
| sdake_ | if your deploy target (whre keystone runs) and the registry confict on ports, nothingwill work | 03:47 |
| SamYaple | sdake_: that is not true. | 03:47 |
| kjelly | sdake_: What's different between docker hub and local privary registry ? | 03:48 |
| SamYaple | i run with registry on port 5000 | 03:48 |
| SamYaple | it only matters if your deploy host _is_ your registry | 03:48 |
| sdake_ | yes i am pretty sure that is what i just said | 03:48 |
| SamYaple | its not but ok | 03:48 |
| sdake_ | kjelly docker hub = a big registry site run by docker in | 03:48 |
| sdake_ | local registry is when you run docker run registry | 03:49 |
| sdake_ | or whatever the command is | 03:49 |
| sdake_ | prviate registry is a docker hub registry that only you have access to | 03:49 |
| sdake_ | SamYaple can you pleae paste a diff of whatever code your running to gte heat to work | 03:50 |
| SamYaple | sure | 03:50 |
| kjelly | sdake_: Ok | 03:50 |
| sdake_ | tx | 03:50 |
| SamYaple | http://paste.fedoraproject.org/264314/97303144/ | 03:50 |
| kjelly | SamYaple, sdake_ : thanks | 03:51 |
| sdake_ | i ee, you left in the clients keystone | 03:51 |
| sdake_ | i took that out | 03:51 |
| sdake_ | let me try with that | 03:51 |
| sdake_ | what does that config option do precisely? | 03:51 |
| sdake_ | the defferred auth thing | 03:51 |
| SamYaple | do you know what trusts are or why they were implemented? because whats broke here is trusts | 03:52 |
| SamYaple | it has nothing to do with that deprecation warning you are seeing | 03:52 |
| sdake_ | i know trusts are busted and i have a vgue understanding of what they are | 03:52 |
| SamYaple | deferred_auth_method defaults to using trusts | 03:53 |
| sdake_ | i know we need them implemented | 03:53 |
| sdake_ | got it thanks | 03:53 |
| SamYaple | since we arent setting them up, we should be using them | 03:53 |
| SamYaple | the client_keystone bits are only because i only set a v2.0 endpoint in the database | 03:54 |
| SamYaple | im fixing that now so we got full v3 | 03:54 |
| sdake_ | roger | 03:54 |
| SamYaple | it keeps breaking things to have this v2.0 endpoint there | 03:54 |
| sdake_ | is there some advantage to having a v2.0 endpoint? | 03:54 |
| SamYaple | not that i can think of | 03:55 |
| *** dimsum__ has quit IRC | 03:55 | |
| SamYaple | its been depercated for like 2 years now | 03:55 |
| SamYaple | errr 2 cycles* | 03:55 |
| sdake_ | samyaple check out docs.openstack.org/developer | 03:55 |
| sdake_ | and tlel me if anything pops out at you | 03:55 |
| SamYaple | nice | 03:56 |
| sdake_ | i was pretty sure that change would get a -2 ;) | 03:56 |
| sdake_ | never knkow unitlyou try | 03:56 |
| SamYaple | well you were just pointing ot more docs.openstack.org documentation | 03:57 |
| sdake_ | yup | 03:57 |
| sdake_ | i also informed the other deployment project of it | 03:57 |
| sdake_ | so they can add inks there | 03:57 |
| sdake_ | i am going to buy a 400gb pcie ssd | 04:08 |
| sdake_ | 1gb/sec writes | 04:08 |
| sdake_ | for my docker partition | 04:08 |
| SamYaple | dont do it | 04:08 |
| SamYaple | its not worth it | 04:08 |
| SamYaple | serious | 04:08 |
| sdake_ | why not worth it | 04:08 |
| SamYaple | i tested an 'on-metal' rackspace instance | 04:09 |
| SamYaple | its the real hardware | 04:09 |
| SamYaple | it did not make it as fast as that price would make you think | 04:09 |
| sdake_ | 400 bucks is not exaactly deadly to me :) | 04:09 |
| sdake_ | plus i am out of hard disk slots in my workstation | 04:09 |
| sdake_ | and also out of disk space | 04:10 |
| SamYaple | i agree, but don't expcet the world if you already build in 10-15 minutes (like i do) | 04:10 |
| sdake_ | it takes me 70 minute to build for some reason | 04:10 |
| SamYaple | its not a bad choice, just dont get your expectations too high | 04:10 |
| sdake_ | minuts | 04:10 |
| sdake_ | with a push | 04:10 |
| SamYaple | wtf | 04:10 |
| SamYaple | oh yea | 04:10 |
| SamYaple | well thatll still happen | 04:10 |
| sdake_ | i am msotly ding it for more disk space | 04:10 |
| sdake_ | not more performance | 04:10 |
| SamYaple | not a bad choice then | 04:10 |
| sdake_ | but more perfomrance = icing on cake | 04:11 |
| SamYaple | i know it wasnt a great improvement over _my_setup_ but it may be over yours. idk how your disks are laid out | 04:11 |
| sdake_ | i have 4 ssds in raid 1 0 setup | 04:12 |
| SamYaple | the type of ssds matter | 04:12 |
| sdake_ | intel datcenter models | 04:12 |
| sdake_ | dc3500s iirc | 04:12 |
| SamYaple | yea those are know for endurance over a long time, not bursty speeds | 04:12 |
| sdake_ | thats why i bought em | 04:13 |
| *** weiyu_ has quit IRC | 04:20 | |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Make heat stack-create work https://review.openstack.org/220765 | 04:24 |
| openstackgerrit | Steven Dake proposed stackforge/kolla: Make heat stack-create work https://review.openstack.org/220765 | 04:25 |
| *** vinkman has quit IRC | 04:47 | |
| *** dimsum__ has joined #kolla | 04:55 | |
| *** dimsum__ has quit IRC | 05:01 | |
| *** Haomeng has joined #kolla | 05:01 | |
| *** Haomeng|2 has quit IRC | 05:03 | |
| *** weiyu_ has joined #kolla | 05:27 | |
| *** weiyu_ has quit IRC | 05:32 | |
| *** diga has joined #kolla | 05:44 | |
| *** weiyu_ has joined #kolla | 06:20 | |
| *** chandankumar has joined #kolla | 06:46 | |
| *** shardy has joined #kolla | 07:31 | |
| *** tfukushima has joined #kolla | 07:31 | |
| *** inc0 has joined #kolla | 07:57 | |
| inc0 | good morning | 07:57 |
| *** weiyu_ has quit IRC | 08:00 | |
| SamYaple | morning inc0 | 08:25 |
| *** weiyu_ has joined #kolla | 08:25 | |
| SamYaple | inc0: im not sure were going to get the logging done the way we have discussed before RC1 | 08:26 |
| *** weiyu_ has quit IRC | 08:26 | |
| SamYaple | and im tired of fighting about it | 08:26 |
| *** athomas has joined #kolla | 08:26 | |
| inc0 | well, I'll do data container | 08:26 |
| SamYaple | with no rsyslog, just files? | 08:26 |
| inc0 | wait now, hold on | 08:26 |
| inc0 | I haven't been looking over review in weekend | 08:27 |
| inc0 | where did anyone say anything about "no rsyslog, do files"? | 08:28 |
| inc0 | sdake want's data container, and I'm completely ok with it, it makes sense | 08:28 |
| inc0 | but who said "no rsyslog?"oO | 08:29 |
| SamYaple | how are you going to share the /dev/log device? | 08:29 |
| SamYaple | you cant do that through a data container | 08:29 |
| inc0 | but sdake said he was ok that rsyslog will be an effective data container | 08:29 |
| SamYaple | he doesnt know what he is talking about | 08:29 |
| inc0 | so it won't be data container per se, it will be rsyslog container | 08:30 |
| inc0 | so pretty much what we have now if we make bindmount optional | 08:30 |
| SamYaple | if we arent using the rsyslog process, which we wouldn't be, then we arent using the rsyslog container | 08:30 |
| SamYaple | inc0: no | 08:30 |
| SamYaple | how can you share /dev/log | 08:30 |
| SamYaple | you cant do that in a container | 08:30 |
| SamYaple | that requires writing stuff to the host | 08:31 |
| inc0 | well, I'm creating my own /dev/log as seen in review | 08:31 |
| inc0 | and it will be /var/lib/kolla | 08:31 |
| inc0 | usr/lib* | 08:31 |
| SamYaple | no thats not whats happening | 08:31 |
| inc0 | I just need any place in host to add completely removable /dev/log | 08:32 |
| inc0 | I did that in /tmp | 08:32 |
| SamYaple | which is what sdake is against | 08:32 |
| SamYaple | thats my point | 08:32 |
| inc0 | no, I don't think he is against that | 08:32 |
| SamYaple | he most certainly is | 08:32 |
| inc0 | what he is against is to bindmount /var/log as default | 08:32 |
| SamYaple | yea thats no issue | 08:32 |
| SamYaple | binding a containers dev to some location on the host isn't good though | 08:32 |
| SamYaple | is what his point was | 08:32 |
| inc0 | https://review.openstack.org/#/c/215062/11/ansible/roles/common/tasks/start.yml look at line 43 | 08:33 |
| SamYaple | inc0: i am well aware of what you are doing and how you are doing it | 08:34 |
| inc0 | it seems contradictory to line 21 tho | 08:34 |
| inc0 | I mean his comments | 08:34 |
| inc0 | sdake_, are you awake? | 08:34 |
| SamYaple | thats because he doesnt know what he is talking about | 08:34 |
| SamYaple | this was my point before | 08:34 |
| SamYaple | either way there is stil lalot to do (like splitting out each service into its own file) and you dont seem to be doing that | 08:34 |
| inc0 | yeah, let's sort out this one first | 08:35 |
| SamYaple | well im going to hammer out all of the logging to files because i can assure you thats where we will land on this because i will not fight about this anymore | 08:36 |
| *** jmccarthy has quit IRC | 08:36 | |
| *** jmccarthy has joined #kolla | 08:36 | |
| inc0 | meh | 08:36 |
| SamYaple | also how i feel | 08:37 |
| inc0 | that will mess up with centralized logging later on | 08:37 |
| SamYaple | nah itll be fine | 08:37 |
| SamYaple | just double reads/writes | 08:37 |
| SamYaple | annoying but it fits what sdake wants | 08:37 |
| SamYaple | so meh | 08:37 |
| inc0 | and having to worry about disk space everywhere | 08:37 |
| *** weiyu_ has joined #kolla | 08:37 | |
| SamYaple | yep | 08:37 |
| SamYaple | im in your camp here bud. but we do have a deadline fast approaching | 08:38 |
| SamYaple | so well revisit in mitaka when we start central logging | 08:38 |
| inc0 | maybe | 08:38 |
| inc0 | sooo...rsyslog is merged | 08:38 |
| inc0 | and we won't use it right? | 08:39 |
| SamYaple | as it stands right now, i doubt it | 08:39 |
| SamYaple | unless we kinda repurpose it and throw in supervisord and logrotate | 08:39 |
| inc0 | let me ask sdake one more timne | 08:39 |
| inc0 | because that's contradictions | 08:39 |
| inc0 | and that means there is some misunderstanding in place | 08:39 |
| SamYaple | go for it, but make sure to explain that the containers /dev is written to the host filesystem | 08:39 |
| inc0 | I will | 08:40 |
| inc0 | I think what sdake means is that he want's /dev/log disapear when we remove container | 08:40 |
| inc0 | but...aren't we writting to host file system in for example db? | 08:40 |
| SamYaple | nope | 08:41 |
| *** weiyu_ has quit IRC | 08:41 | |
| SamYaple | i will have a patch to make all these data containers host bindmountable (option) before RC1 | 08:41 |
| *** weiyu_ has joined #kolla | 08:41 | |
| SamYaple | because right now performance is kill | 08:41 |
| *** shardy_ has joined #kolla | 08:41 | |
| openstackgerrit | Merged stackforge/kolla: Make heat stack-create work https://review.openstack.org/220765 | 08:41 |
| SamYaple | not a priority for me atm though | 08:41 |
| *** shardy has quit IRC | 08:43 | |
| inc0 | I'll address comments in review today | 08:43 |
| inc0 | and later on confront it with sdake | 08:43 |
| SamYaple | curious, how were you planning on addressing the log-per-service issue? | 08:43 |
| inc0 | well, we would have to specify log facility | 08:44 |
| inc0 | per service | 08:44 |
| *** shardy_ has quit IRC | 08:46 | |
| *** shardy has joined #kolla | 08:47 | |
| pbourke | morning all | 08:47 |
| inc0 | good morning | 08:47 |
| pbourke | can I get some eyes on https://review.openstack.org/#/c/220172/ at some stage today please | 08:48 |
| inc0 | back in about an hour | 08:48 |
| SamYaple | how much will you pay me | 08:50 |
| *** weiyu_ has quit IRC | 08:50 | |
| *** weiyu_ has joined #kolla | 08:51 | |
| SamYaple | pbourke: will we need rsyncd anywhere else? | 08:52 |
| pbourke | SamYaple: I dont think so | 08:53 |
| SamYaple | my thoughts as well | 08:53 |
| SamYaple | in that case what are your thoughts on not making it an rsyncd role? | 08:53 |
| SamYaple | just make it another task in the swift role | 08:53 |
| SamYaple | we dont have a libvirt or openvswitch role | 08:54 |
| SamYaple | they are just tasks in nova or neutron | 08:54 |
| *** weiyu_ has quit IRC | 08:54 | |
| *** weiyu has joined #kolla | 08:56 | |
| *** weiyu_ has joined #kolla | 08:58 | |
| *** weiyu has quit IRC | 08:58 | |
| *** dimsum__ has joined #kolla | 08:59 | |
| *** dayou has quit IRC | 09:00 | |
| *** weiyu_ has quit IRC | 09:00 | |
| *** weiyu_ has joined #kolla | 09:00 | |
| *** jasonsb has quit IRC | 09:01 | |
| *** pbourke has quit IRC | 09:05 | |
| *** pbourke has joined #kolla | 09:06 | |
| *** mbound has joined #kolla | 09:06 | |
| pbourke | SamYaple: yeah that makes sense I guess | 09:06 |
| *** dimsum__ has quit IRC | 09:07 | |
| *** jasonsb has joined #kolla | 09:08 | |
| *** weiyu_ has quit IRC | 09:14 | |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Correct name of source archive if reference contains '/' https://review.openstack.org/220542 | 09:14 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Cannot deploy by tag if groups are distributed among hosts https://review.openstack.org/220923 | 09:27 |
| SamYaple | pbourke: looks good otherwise! | 09:28 |
| SamYaple | vbel1: the issue i have with that gather facts review is what happens if the host is inaccessible? | 09:30 |
| *** weiyu_ has joined #kolla | 09:33 | |
| vbel1 | it will proceed, but if that host is missing in task ansible will stop as usual | 09:36 |
| vbel1 | it's normal situation, just like if you run without tags | 09:37 |
| SamYaple | sure but this still breaks --limit | 09:37 |
| SamYaple | I would prefer to use caching since that fixes the --limit issue and this one at the same time | 09:38 |
| SamYaple | we already have an ansible 1.8 requirement so everyone has caching | 09:38 |
| openstackgerrit | Vladislav Belogrudov proposed stackforge/kolla: Cannot deploy by tag if groups are distributed among hosts https://review.openstack.org/220923 | 09:38 |
| vbel1 | yes, it caches but we have to run ansible _twice_ to get cache :) | 09:40 |
| SamYaple | Well the initial ansible run wont work anyway if you lmit it | 09:41 |
| SamYaple | i mean using tags for an initial run gets you.... nothing | 09:41 |
| SamYaple | but the problem with this fact gathering for _all_ hosts every run is when you have 200-1000 nodes itll take forever | 09:41 |
| SamYaple | and when you limit you dont even resolve the issue | 09:42 |
| openstackgerrit | Merged stackforge/kolla: Correct name of source archive if reference contains '/' https://review.openstack.org/220542 | 09:47 |
| vbel1 | ok, so we proceed with more complex invocation of ansible to use caching then. It is not enough to hack haproxy, we wil have the situation if users want separate [database] group or anything similar (ceilometer, keystone...) | 09:50 |
| SamYaple | vbel1: it is enough to hack haproxy | 09:50 |
| SamYaple | the database is accessed through the VIP which is statically configured | 09:51 |
| SamYaple | its not fetched via cache | 09:51 |
| SamYaple | oh wait i see what you are saying | 09:51 |
| SamYaple | but yea still enough to change haproxy around | 09:51 |
| SamYaple | ill submit the patch. give me like 15 minutes, ive just been hammering on ceph | 09:51 |
| vbel1 | we had some users that wanted ceilometer and keystone on separate network :) yes, with VIP it can be enough | 09:52 |
| vbel1 | SamYaple: thanks! | 09:52 |
| SamYaple | vbel1: still shouldnt be an issue, but that would require a bit more ansible stuff to make that work | 09:52 |
| SamYaple | you should be able to do that through the inventory variables right now vbel1 | 09:53 |
| vbel1 | yes, that should work. I try to run database separately from controllers, it works well | 09:55 |
| SamYaple | indeed. just like i designed! i often test running a single service (nova-api or something) on a completely different host to make sure it works | 09:56 |
| *** Kennan_Vacation2 has joined #kolla | 10:02 | |
| *** Kennan_Vacation has quit IRC | 10:02 | |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add rsyncd image and related Ansible for Swift https://review.openstack.org/220172 | 10:04 |
| inc0 | back | 10:05 |
| *** weiyu_ has quit IRC | 10:05 | |
| *** weiyu_ has joined #kolla | 10:07 | |
| *** diga has quit IRC | 10:20 | |
| *** akwasnie has joined #kolla | 10:34 | |
| *** zhiwei has quit IRC | 10:36 | |
| *** weiyu_ has quit IRC | 10:37 | |
| openstackgerrit | James McCarthy proposed stackforge/kolla: Add missing Swift services, also ansible configuration https://review.openstack.org/218795 | 10:41 |
| *** weiyu_ has joined #kolla | 10:49 | |
| *** weiyu_ has quit IRC | 11:04 | |
| vbel1 | SamYaple: if haproxy requires knowledge of all hosts we can put 'hosts: !compute' in site.yml and add 'when: host in network' in its tasks | 11:05 |
| *** weiyu has joined #kolla | 11:05 | |
| SamYaple | vbel1: it doesnt require knowlege of all hosts | 11:06 |
| SamYaple | i have the patch just have work being work so i havent tested it yet | 11:06 |
| *** weiyu has quit IRC | 11:06 | |
| vbel1 | SamYaple: hosts: !compute | 11:06 |
| SamYaple | vbel1: that will still pull in a whole bunch of unneeded hosts | 11:07 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Cleanup small todo in Vagrantfile https://review.openstack.org/220953 | 11:08 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add missing node types to Vagrantfile https://review.openstack.org/220955 | 11:11 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Add rsyncd image and related Ansible for Swift https://review.openstack.org/220172 | 11:14 |
| *** dimsum__ has joined #kolla | 11:33 | |
| *** dimsum__ has quit IRC | 11:40 | |
| openstackgerrit | Sam Yaple proposed stackforge/kolla: Gather facts from the hosts before using them https://review.openstack.org/220970 | 12:14 |
| SamYaple | vbel1: https://review.openstack.org/220970 | 12:14 |
| SamYaple | sorry ive been fairly busy at work | 12:14 |
| SamYaple | that patch is more what i had in mind (short of fact caching) | 12:14 |
| pbourke | mandre: I can submit another ps with your change, and add you add co-author. does that work for you? | 12:27 |
| pbourke | mandre: alterantely you can submit an alternative and I'll abandon | 12:28 |
| *** dimsum__ has joined #kolla | 12:30 | |
| mandre | pbourke, sure you can submit a new patch | 12:30 |
| mandre | i'm setting up a new machine and trying to see if i can run vagrant with libvirt :) | 12:31 |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 12:31 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Cleanup small todo in Vagrantfile https://review.openstack.org/220953 | 12:33 |
| *** dimsum__ has quit IRC | 12:34 | |
| openstackgerrit | Michal Jastrzebski (inc0) proposed stackforge/kolla: Ansible bits to rsyslog https://review.openstack.org/215062 | 12:41 |
| inc0 | bbl when sdake wakes up | 12:44 |
| inc0 | cyas | 12:44 |
| *** inc0 has quit IRC | 12:44 | |
| vbel1 | SamYaple: thanks! Will check now | 12:45 |
| vbel1 | SamYaple: so we use hosts: all-that-needed and when: in haproxy :) Fully agree | 12:47 |
| vbel1 | with this approach | 12:47 |
| *** jtriley has joined #kolla | 12:51 | |
| SamYaple | vbel1: correct | 12:52 |
| SamYaple | its lowest hosts gauranteed | 12:52 |
| SamYaple | downside is we will likely forget to add things from time to time | 12:52 |
| openstackgerrit | Paul Bourke proposed stackforge/kolla: Cleanup small todo in Vagrantfile https://review.openstack.org/220953 | 12:52 |
| vbel1 | SamYaple: I see what you meant before :) | 12:52 |
| SamYaple | upside is no _required_ caching | 12:52 |
| vbel1 | SamYaple: yes, editing that file is for power users | 12:53 |
| SamYaple | ok so youre cool with this approach then vbel1? Want ot make sure it works for everyone | 12:57 |
| vbel1 | SamYaple: yes, this was in my head as well, I thought it would be more difficult to implement. But it is not, looks great for me. | 12:59 |
| SamYaple | cool. yea its deciving to this about because we only have to include the api hosts, not _all_ hosts | 13:00 |
| *** dayou has joined #kolla | 13:00 | |
| *** klint has quit IRC | 13:00 | |
| pbourke | SamYaple: vbel1: are we sure this bug is related to haproxy only? | 13:01 |
| pbourke | I've seen it when trying to deploy rabbitmq | 13:01 |
| vbel1 | pbourke: does rabbit want to know about other services? Probably not, so this is ok in my opinion. | 13:02 |
| pbourke | vbel1: i added the db nodes and now seeing it when trying to copy the configs over: http://paste.openstack.org/show/448668/ | 13:03 |
| pbourke | (ndb configs) | 13:04 |
| vbel1 | but we have to run haproxy - to avoid direct contact between other groups, only thru VIP | 13:04 |
| SamYaple | pbourke: im sure the issue that the bug is for was solved with that patch, but there may be other things? | 13:04 |
| SamYaple | i dont know what ndbcluster is pbourke | 13:04 |
| SamYaple | but you will likely need to do a simliar thing to the hosts that i did with the haproxy section | 13:04 |
| pbourke | SamYaple: yeah we currently use it in place of galera, just trying to point out that like you say there may be other places the same fix is needed | 13:05 |
| SamYaple | pbourke: not in Kolla there isn't | 13:05 |
| SamYaple | do you have the ansible code for these tasks anywhere i can view them pbourke? | 13:06 |
| SamYaple | ill happily do a quick audit | 13:06 |
| vbel1 | pbourke: you have some hosts in [database] probably that are not discovered | 13:06 |
| *** erkules has quit IRC | 13:08 | |
| *** erkules has joined #kolla | 13:08 | |
| *** banix has joined #kolla | 13:10 | |
| *** dayou has quit IRC | 13:22 | |
| *** dayou has joined #kolla | 13:24 | |
| *** banix has quit IRC | 13:26 | |
| *** dimsum__ has joined #kolla | 13:34 | |
| *** dimsum__ has quit IRC | 13:40 | |
| *** mlil has quit IRC | 13:43 | |
| *** chandankumar has quit IRC | 13:46 | |
| *** inc0 has joined #kolla | 13:53 | |
| inc0 | hello, sdake_ around? | 13:55 |
| *** jtriley has quit IRC | 14:08 | |
| *** dimsum__ has joined #kolla | 14:23 | |
| *** inc0 has quit IRC | 14:36 | |
| *** inc0 has joined #kolla | 14:42 | |
| *** dimsum__ has quit IRC | 14:46 | |
| openstackgerrit | Mick Thomspon proposed stackforge/kolla: Add ansible support for Murano https://review.openstack.org/218894 | 14:47 |
| *** inc0 has quit IRC | 15:01 | |
| *** diga has joined #kolla | 15:03 | |
| *** jtriley has joined #kolla | 15:05 | |
| *** jtriley has quit IRC | 15:10 | |
| *** jtriley has joined #kolla | 15:13 | |
| *** inc0 has joined #kolla | 15:14 | |
| *** jtriley has quit IRC | 15:21 | |
| *** inc0 has quit IRC | 15:30 | |
| *** inc0 has joined #kolla | 15:40 | |
| *** dimsum__ has joined #kolla | 15:46 | |
| *** akwasnie has quit IRC | 15:49 | |
| *** dimsum__ has quit IRC | 15:53 | |
| *** inc0 has quit IRC | 16:13 | |
| *** jtriley has joined #kolla | 16:17 | |
| *** jtriley has quit IRC | 16:25 | |
| *** inc0 has joined #kolla | 16:26 | |
| *** mbound has quit IRC | 16:32 | |
| *** chandankumar has joined #kolla | 16:35 | |
| *** athomas has quit IRC | 16:39 | |
| *** dimsum__ has joined #kolla | 16:43 | |
| *** jtriley has joined #kolla | 16:48 | |
| *** dimsum__ has quit IRC | 16:49 | |
| *** inc0 has quit IRC | 16:57 | |
| *** tfukushima has quit IRC | 16:59 | |
| *** inc0 has joined #kolla | 17:01 | |
| *** inc0 has quit IRC | 17:06 | |
| *** jtriley has quit IRC | 17:09 | |
| *** jtriley has joined #kolla | 17:11 | |
| *** daneyon has joined #kolla | 17:13 | |
| *** daneyon has quit IRC | 17:19 | |
| *** SourabhP has joined #kolla | 17:19 | |
| *** SourabhP_ has joined #kolla | 17:21 | |
| *** diga has quit IRC | 17:22 | |
| *** SourabhP has quit IRC | 17:24 | |
| *** SourabhP_ is now known as SourabhP | 17:24 | |
| *** inc0 has joined #kolla | 17:25 | |
| *** inc0 has quit IRC | 17:29 | |
| *** jtriley has quit IRC | 17:47 | |
| *** SourabhP_ has joined #kolla | 17:49 | |
| *** SourabhP has quit IRC | 17:49 | |
| *** SourabhP_ is now known as SourabhP | 17:50 | |
| *** dimsum__ has joined #kolla | 17:50 | |
| *** dimsum__ has quit IRC | 17:56 | |
| *** inc0 has joined #kolla | 17:56 | |
| inc0 | afternoon | 17:57 |
| inc0 | sdake_, around? | 17:57 |
| inc0 | guess not, cya tomorrow then;) | 17:58 |
| *** inc0 has quit IRC | 18:03 | |
| *** SourabhP has quit IRC | 18:03 | |
| openstackgerrit | Merged stackforge/kolla: Cleanup small todo in Vagrantfile https://review.openstack.org/220953 | 18:11 |
| openstackgerrit | Merged stackforge/kolla: Add ansible support for Murano https://review.openstack.org/218894 | 18:18 |
| *** chandankumar has quit IRC | 18:23 | |
| *** dimsum__ has joined #kolla | 18:34 | |
| *** jtriley has joined #kolla | 18:44 | |
| *** dimsum__ has quit IRC | 18:47 | |
| *** jtriley has quit IRC | 18:48 | |
| *** dayou has quit IRC | 19:12 | |
| *** dayou has joined #kolla | 19:12 | |
| *** dimsum__ has joined #kolla | 19:13 | |
| *** jtriley has joined #kolla | 19:14 | |
| *** jtriley has quit IRC | 19:20 | |
| *** daneyon has joined #kolla | 19:24 | |
| *** daneyon_ has joined #kolla | 19:36 | |
| *** daneyon has quit IRC | 19:39 | |
| *** jtriley has joined #kolla | 20:16 | |
| *** jtriley has quit IRC | 20:21 | |
| *** dimsum__ has quit IRC | 20:36 | |
| *** dimsum__ has joined #kolla | 20:51 | |
| *** sdake_ has quit IRC | 20:55 | |
| *** jtriley has joined #kolla | 20:57 | |
| *** jtriley has quit IRC | 21:02 | |
| *** shardy has quit IRC | 21:31 | |
| *** vbel1 has quit IRC | 21:40 | |
| *** vbel1 has joined #kolla | 21:40 | |
| *** jtriley has joined #kolla | 21:58 | |
| *** jtriley has quit IRC | 22:04 | |
| *** akwasnie has joined #kolla | 22:24 | |
| *** sdake has joined #kolla | 22:53 | |
| *** jtriley has joined #kolla | 22:59 | |
| *** jtriley has quit IRC | 23:04 | |
| *** dayou has quit IRC | 23:20 | |
| *** dayou has joined #kolla | 23:21 | |
| *** dayou has quit IRC | 23:25 | |
| *** sdake has quit IRC | 23:42 | |
| openstackgerrit | Merged stackforge/kolla: Gather facts from the hosts before using them https://review.openstack.org/220970 | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!