| *** mbound has quit IRC | 00:06 | |
| sdake | my hand made patch is failing | 00:08 |
|---|---|---|
| sdake | don't get it | 00:08 |
| * sdake fires patch | 00:08 | |
| *** blahRus has quit IRC | 00:08 | |
| rhallisey | what's failing? | 00:14 |
| SamYaple | fix it | 00:14 |
| rhallisey | is neutron passing though? | 00:14 |
| *** cloudnau_ has quit IRC | 00:17 | |
| *** banix has joined #kolla | 00:17 | |
| openstackgerrit | Merged openstack/kolla: Add Ansible support for Ironic https://review.openstack.org/219747 | 00:18 |
| SamYaple | sdake: got all my gear in | 00:19 |
| sdake | nice | 00:19 |
| sdake | time to get it hooked up | 00:19 |
| SamYaple | its installed in one server | 00:19 |
| sdake | did you spin it up? | 00:19 |
| SamYaple | testing for fitness before second server install | 00:19 |
| SamYaple | yea | 00:19 |
| sdake | seem faster? | 00:20 |
| SamYaple | i see the hardware, so good sign | 00:20 |
| SamYaple | building now | 00:20 |
| sdake | it is much faster for me | 00:20 |
| sdake | but iw as running without btrfs before | 00:20 |
| sdake | and now I am running with btrfs | 00:20 |
| sdake | i'm going to make two parottions on my nvm device and put my registry on it as well | 00:21 |
| sdake | just haven't got there yet | 00:21 |
| SamYaple | yea im going to carve up some for docker, some for ceph | 00:22 |
| SamYaple | my registry is backed by ceph | 00:22 |
| sdake | i've been thining of buying 3 more instead of new minidells | 00:23 |
| sdake | and doubling up my minidells for ceph with the 400gb ssd | 00:23 |
| SamYaple | of the 750? | 00:23 |
| sdake | yup | 00:23 |
| sdake | for 1.2 tb ceph storage | 00:23 |
| SamYaple | do you have 3 nodes already? | 00:23 |
| sdake | yup | 00:23 |
| sdake | but only 256gb disk | 00:23 |
| sdake | in each one | 00:23 |
| sdake | if i knew these 750s were coming out i would have got a 1tb disk in them | 00:23 |
| sdake | but instead went for the 256gb ssd | 00:24 |
| *** dims has joined #kolla | 00:24 | |
| sdake | i figured out where setuptools is coming from | 00:24 |
| sdake | pbr is installed in the base image | 00:24 |
| sdake | and that pulls in setuptools 0.9.8 | 00:25 |
| sdake | so we can revert mandre's patch and remove pbr | 00:25 |
| sdake | that should fix it all up | 00:25 |
| sdake | the binary build is reworked to not require pbr | 00:25 |
| sdake | becaue setuptools is a disaster in python version # wise | 00:26 |
| sdake | SamYaple do you think that would work ? | 00:26 |
| sdake | (the 400gb ssd thing) | 00:26 |
| SamYaple | sdake: kinda. remember ceph +3 replicas youll only have 400gb of space (assuming you dont split it) | 00:27 |
| sdake | oh that sucks | 00:28 |
| SamYaple | i would recommend the ssds as a cachetier for read/write | 00:28 |
| sdake | i thought it would like stripe or something | 00:28 |
| sdake | like raid 0 | 00:28 |
| SamYaple | then you can squeeze and additional 400GB or so out | 00:28 |
| SamYaple | or maybe not | 00:28 |
| SamYaple | either way cephs advantage is backing spinning disks | 00:28 |
| SamYaple | you should do that | 00:29 |
| SamYaple | ssd speed backing several tb of storage | 00:29 |
| sdake | do which? | 00:29 |
| SamYaple | read/write cache | 00:29 |
| sdake | oh | 00:29 |
| sdake | i dont have 1tb disks in the machine tho | 00:29 |
| sdake | like i said live and learn :( | 00:29 |
| SamYaple | can you put some in? | 00:29 |
| sdake | if i thorw out my 256gb ssd | 00:29 |
| sdake | only 1 disk slot | 00:29 |
| sdake | minidell | 00:29 |
| SamYaple | damn | 00:30 |
| SamYaple | yea i mean 3 400gb ssd isnt bad for ceph | 00:30 |
| SamYaple | remember ceph is cow so you can launch a bunch of instances on it | 00:30 |
| SamYaple | but youll probably end up with no space | 00:30 |
| sdake | i could buy 3 1tb disks | 00:30 |
| sdake | and 3 750s | 00:31 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Neutron Agents json support https://review.openstack.org/228611 | 00:31 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Fix oslo.log https://review.openstack.org/228675 | 00:31 |
| SamYaple | 1tb dedicated entirely to ceph? or spilt? | 00:31 |
| sdake | 150gb / disk, 850gb ceph | 00:31 |
| SamYaple | that would work well i think | 00:31 |
| sdake | with 100gb docker /var/lib/docker 300gb ceph cache | 00:31 |
| sdake | the biggest disk I can get is 1tb | 00:32 |
| sdake | because they require 2.5inch disks | 00:32 |
| SamYaple | you would get ~1.5TB usable ceph with ~300GB of ssd read/write cache | 00:33 |
| sdake | is ceph hard to setup? | 00:33 |
| SamYaple | i dont think so, no. but there is a learning curve | 00:33 |
| SamYaple | i can help you out with the setup | 00:33 |
| sdake | i need to see if I have two pcie slots | 00:33 |
| sdake | we need a guide for it in kolla i think | 00:33 |
| sdake | minidells a bit constrained on space may not have an 8x slot | 00:34 |
| sdake | or 16x, whtaever it is it needs | 00:34 |
| SamYaple | i can only help so much :) | 00:34 |
| SamYaple | im going to write all this computr stuff off on my taxes | 00:35 |
| sdake | why is nova started before nuetron | 00:38 |
| sdake | shouldn't it be the other way around | 00:38 |
| SamYaple | no | 00:39 |
| SamYaple | so its what i expected. builds are about the same time sdake | 00:39 |
| SamYaple | ~12 minutes | 00:39 |
| sdake | that is how long my builds take now | 00:40 |
| sdake | before it was 1 hour | 00:40 |
| SamYaple | yea thats how mine were before too | 00:40 |
| sdake | must be btfs | 00:40 |
| SamYaple | but i have extra bandwidth now too | 00:40 |
| sdake | btrfs | 00:40 |
| SamYaple | going to play with the threads now that i have extra bandwidth | 00:40 |
| sdake | my single node aio takes 9 minutes to deploy over gige | 00:41 |
| sdake | which is pretty solid | 00:41 |
| sdake | 3 node takes 12 minutes | 00:41 |
| SamYaple | sounds right | 00:41 |
| SamYaple | all my 3 nodes take about 12 | 00:41 |
| sdake | want stats on 100 node | 00:41 |
| SamYaple | i did a deploy from docker hub in 17 minutes | 00:41 |
| SamYaple | i only have 45mb down | 00:42 |
| sdake | to 3 nodes from hub | 00:42 |
| sdake | ? | 00:42 |
| SamYaple | yes | 00:42 |
| sdake | nice | 00:42 |
| SamYaple | i wasnt saturating bandswidth either | 00:42 |
| sdake | failed: [minime-02] => {"changed": true, "failed": true} | 00:42 |
| sdake | msg: Docker API Error: Conflict. The name "neutron_agents" is already in use by container 0bc3c21d7399. You have to delete (or rename) that container to be able to reuse that name. | 00:42 |
| * sdake ughs | 00:42 | |
| SamYaple | never seen that before | 00:42 |
| SamYaple | shouldn't it destroy and recreate that? | 00:43 |
| SamYaple | thats what it does for me... | 00:43 |
| sdake | i did a cleanup-containers prior | 00:43 |
| SamYaple | does that work for you? | 00:43 |
| SamYaple | it broke once for me but i couldnt replicate it | 00:43 |
| sdake | dunno | 00:43 |
| sdake | i'm doing another deploy | 00:43 |
| sdake | if its bust rm -rf /var/lib/docker inc | 00:43 |
| SamYaple | no its btrfs | 00:43 |
| SamYaple | you cant jsut do that | 00:43 |
| SamYaple | it has subvolumes | 00:43 |
| sdake | oh | 00:43 |
| SamYaple | easiest thing to do is a reformat in that case | 00:44 |
| * sdake groans | 00:44 | |
| sdake | oh, I am using overlayfs | 00:44 |
| sdake | so I can do that | 00:44 |
| SamYaple | you also cant builds :) | 00:44 |
| sdake | (this is on the ssd storage) | 00:44 |
| sdake | yeah these are deploy targets only | 00:44 |
| SamYaple | oh cool. let me know how that goes | 00:44 |
| sdake | i have another machine dedicated to building (bigiron) | 00:44 |
| SamYaple | very curious | 00:44 |
| *** loth1 has quit IRC | 00:45 | |
| SamYaple | i have done a few deploys on overlayfs, no issues | 00:45 |
| SamYaple | but i mainly test against btrfs | 00:45 |
| sdake | ya i've done it a bunch | 00:45 |
| sdake | hundreds of deploys on overlayfs | 00:45 |
| *** loth has joined #kolla | 00:46 | |
| sdake | i created an rc2 milestone | 00:46 |
| sdake | 24 bugs | 00:46 |
| sdake | does the ceph support in kolla back glance/cinder/swift? | 00:47 |
| sdake | or is more work needed | 00:47 |
| SamYaple | sdake: it does, but its also not configurable | 00:51 |
| SamYaple | like you can back _just_ glance cinder or nova | 00:51 |
| sdake | i guess i dont understand, you eman each service isn't indivudally configurable, its all or nothing? | 00:53 |
| SamYaple | if you enable ceph it will configure glance nova and cinder to go to ceph | 00:53 |
| SamYaple | there is no way to change that behaviour (i guess the augment file technically could) | 00:54 |
| sdake | that sounds good | 00:54 |
| sdake | what about swift? | 00:54 |
| SamYaple | swift can't be backed by ceph quite like that directly | 00:54 |
| SamYaple | so thats not involved at all | 00:54 |
| *** shakamunyi has joined #kolla | 00:56 | |
| *** dolpher has joined #kolla | 00:56 | |
| sdake | boy deleting the docker dir takes forever | 00:56 |
| SamYaple | it shouldnt | 00:57 |
| SamYaple | somethings probably going on | 00:58 |
| sdake | nah its done | 00:58 |
| sdake | took about 2 minutes | 00:58 |
| *** mbound has joined #kolla | 01:00 | |
| SamYaple | im going to do sr-iov for my instances traffic | 01:05 |
| *** mbound has quit IRC | 01:06 | |
| rhallisey | sdake, I'll check in tmr to make sure everything work | 01:09 |
| * rhallisey build overnight | 01:09 | |
| sdake | tc elections open | 01:10 |
| *** dolpher has quit IRC | 01:10 | |
| *** rhallisey has quit IRC | 01:11 | |
| SamYaple | ive got you ryan. you can bum off my machines once liberty lands and i run a stable kolla | 01:11 |
| SamYaple | sdake: did you run? | 01:11 |
| sdake | considering | 01:12 |
| sdake | elections close the 1st | 01:12 |
| SamYaple | list of those running? | 01:12 |
| sdake | nobody so far | 01:12 |
| sdake | crickets | 01:12 |
| SamYaple | i should do it for the giggles | 01:12 |
| sdake | "I'm sam and I'm here to win!" | 01:12 |
| *** erkules has joined #kolla | 01:13 | |
| sdake | i saw one guy run | 01:13 |
| SamYaple | "there is no 'I' in team, but there is in 'win'" | 01:13 |
| sdake | lol | 01:13 |
| *** erkules_ has quit IRC | 01:16 | |
| *** unicell has quit IRC | 01:33 | |
| *** tfukushima has joined #kolla | 01:42 | |
| sdake | well in the last 6 hours rdo has updated oslo.log to include that bug fix | 01:42 |
| * sdake growls | 01:42 | |
| SamYaple | lulz | 01:43 |
| SamYaple | source will still be affected so all is not lost | 01:44 |
| sdake | it is unclear, i need to build that and check it out | 01:45 |
| sdake | but they probably didn't change the minversion up to the bug fix | 01:46 |
| SamYaple | cool my direct connect 10gb works great with dedicated routes to the other servers | 01:46 |
| sdake | nice | 01:46 |
| SamYaple | so if the 10gb link is up, all the connections will be 10gb between servers | 01:46 |
| SamYaple | if not, it will be 1gb | 01:47 |
| SamYaple | no ip exists on the 10gb link bridges | 01:47 |
| *** jtriley has joined #kolla | 01:58 | |
| *** banix has quit IRC | 02:03 | |
| *** dims has quit IRC | 02:09 | |
| *** tfukushima has quit IRC | 02:25 | |
| *** achanda has joined #kolla | 02:25 | |
| *** tfukushima has joined #kolla | 02:30 | |
| *** klint has joined #kolla | 02:42 | |
| *** loth has quit IRC | 02:47 | |
| *** dolpher has joined #kolla | 02:51 | |
| *** jtriley has quit IRC | 02:56 | |
| *** achanda has quit IRC | 03:09 | |
| *** dims has joined #kolla | 03:09 | |
| *** dims has quit IRC | 03:16 | |
| *** Kennan2 is now known as Kennnan | 03:45 | |
| *** alisonh has quit IRC | 03:51 | |
| *** cloudnau_ has joined #kolla | 03:52 | |
| *** dolpher has quit IRC | 04:01 | |
| *** tpot has joined #kolla | 04:05 | |
| openstackgerrit | Steven Dake proposed openstack/kolla: Neutron Agents json support https://review.openstack.org/228611 | 04:07 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make oslo.log work correctly on RDO platforms https://review.openstack.org/228715 | 04:07 |
| *** alisonh has joined #kolla | 04:09 | |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make oslo.log work correctly on RDO platforms https://review.openstack.org/228715 | 04:10 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Neutron Agents json support https://review.openstack.org/228611 | 04:10 |
| *** diga has joined #kolla | 04:12 | |
| SamYaple | sdake: that patch needs to run on all the distros for source i believe.. | 04:15 |
| sdake | ya good point | 04:15 |
| sdake | thanks i'll fix that | 04:15 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make oslo.log work correctly on RDO platforms https://review.openstack.org/228715 | 04:16 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Neutron Agents json support https://review.openstack.org/228611 | 04:16 |
| SamYaple | hey that neutron-agents one... that wont work like that | 04:16 |
| SamYaple | it needs to change, like alot | 04:17 |
| sdake | what do you propose | 04:17 |
| sdake | it does work as is | 04:17 |
| sdake | ryan wrote it in a "I didn't test it' sort of way | 04:17 |
| sdake | I dont see any other options | 04:17 |
| sdake | but i'm getting ready to hit the sack so feel free to taek it over | 04:18 |
| SamYaple | well the legit way to do it would be import and override the set_configs.py | 04:18 |
| SamYaple | but ill poke it | 04:18 |
| SamYaple | ill get it all worked out | 04:18 |
| sdake | thanks | 04:18 |
| SamYaple | if you say it works i might leave it | 04:18 |
| SamYaple | but ill test it | 04:18 |
| sdake | so set_configs.py runs in each contianer it races | 04:18 |
| sdake | and deletes files and adds files | 04:18 |
| sdake | and crashes basically | 04:18 |
| sdake | so only oen set_config can be run | 04:18 |
| SamYaple | yes cuase of supervisord | 04:18 |
| sdake | vms still dont start | 04:18 |
| SamYaple | yea ill hammer it out | 04:18 |
| sdake | but there are no errors out of the agents that i see | 04:19 |
| SamYaple | they would be getting logged to syslog (which is broken :P) | 04:19 |
| sdake | no syslog works | 04:19 |
| sdake | probably not with that particular ocntainer tho | 04:19 |
| SamYaple | oh you redid the patch for me | 04:19 |
| SamYaple | cool | 04:19 |
| sdake | it was just deleting two lines | 04:20 |
| sdake | i had already had apatch merged earlier today | 04:20 |
| sdake | but it was broke since rdo works with a --no-cache build | 04:20 |
| SamYaple | but https://review.openstack.org/#/c/228448/ is still needed, yes? | 04:20 |
| sdake | no clue didn't even look at taht | 04:21 |
| sdake | wier it wasn't showing up in my review queue | 04:21 |
| SamYaple | you submitted a patch for it though | 04:22 |
| sdake | for what? | 04:23 |
| sdake | i'm bleary eyed dude been up since 2am | 04:23 |
| sdake | sorry if i seem dense :) | 04:23 |
| sdake | i fixed the backtrace that you ahd this morning pointed out | 04:23 |
| SamYaple | nah go to be. ill test all tahts in the queue, fix whats needed | 04:24 |
| tpot | Hiho - not sure whether this is the best place for this. I'm trying to build the liberty-3 containers from scratch and get this error for the neutron container: | 04:24 |
| tpot | Error: Package: 1:python-neutron-8.0.0-dev321.el7.centos.noarch (delorean) | 04:24 |
| tpot | Requires: python-pecan >= 1.0.0 | 04:24 |
| tpot | Installed: python-pecan-0.8.3-1.el7.noarch (@openstack-kilo) | 04:24 |
| tpot | python-pecan = 0.8.3-1.el7 | 04:24 |
| tpot | Available: python-pecan-0.4.5-2.el7.noarch (epel) | 04:24 |
| tpot | python-pecan = 0.4.5-2.el7 | 04:24 |
| sdake | ya | 04:25 |
| sdake | we fixed a bug | 04:25 |
| sdake | use master | 04:25 |
| sdake | or wait for rc1 in about 8 hours | 04:25 |
| tpot | heh | 04:26 |
| SamYaple | tpot: we will have rc1 in a few hours | 04:26 |
| tpot | thanks chaps | 04:26 |
| SamYaple | you could also deploy from source | 04:26 |
| sdake | master is very close to what rc1 will be | 04:26 |
| SamYaple | that should work as is | 04:26 |
| sdake | ya from source will work | 04:26 |
| tpot | oh I didn't think of trying from source - could be fun | 04:26 |
| sdake | tpot we didn't release with taht bug,bu tthe upstream repos changed around where they stored htings | 04:26 |
| sdake | so it broke after we released | 04:26 |
| sdake | they changed their dependency storage repos | 04:27 |
| sdake | brb | 04:27 |
| *** sdake_ has joined #kolla | 04:28 | |
| *** itsuugo has joined #kolla | 04:28 | |
| *** achanda has joined #kolla | 04:29 | |
| sdake_ | sam note if you rework that patch, i want to get rid of config-external entirely | 04:31 |
| sdake_ | i dont want two interfaces into the containers | 04:31 |
| sdake_ | so keep that in mind as you do your work :) | 04:31 |
| SamYaple | config internal is removed entirely i think... | 04:31 |
| *** sdake has quit IRC | 04:31 | |
| sdake_ | essentially I want json into the containers in all containers | 04:31 |
| SamYaple | yes | 04:31 |
| sdake_ | ok | 04:32 |
| sdake_ | read the log before youbegin from today | 04:32 |
| SamYaple | im also going to dedup all of our start.sh so we add it in a single time in the base ;) | 04:32 |
| sdake_ | we had a long discussion about it | 04:32 |
| sdake_ | the irc lo that is | 04:32 |
| sdake_ | log that is | 04:32 |
| SamYaple | yup | 04:32 |
| *** unicell has joined #kolla | 04:35 | |
| sdake_ | SamYaple if you can get to the point where we can deploy vms via heat, I'dcall that victory ;) | 04:35 |
| sdake_ | then the code base willbe back in working shape | 04:35 |
| SamYaple | cool. will test | 04:36 |
| openstackgerrit | Hui Kang proposed openstack/kolla: Add consul containers for dynamic configuration and discovery. This implementation includes https://review.openstack.org/228719 | 04:37 |
| sdake_ | wow thats a ton of work | 04:38 |
| sdake_ | wonder if its useful | 04:38 |
| SamYaple | yea that seems tightly coupled and not turn on/off. ill review further | 04:40 |
| *** itsuugo has quit IRC | 04:42 | |
| SamYaple | Hui are you in channel? | 04:43 |
| SamYaple | sdake_: that blueprint hasn't been approved or discussed. well need to have some procedure around this stuff at some point | 04:43 |
| sdake_ | i am aware | 04:46 |
| sdake_ | i dont even know what consul does | 04:46 |
| sdake_ | or if its useful | 04:46 |
| sdake_ | or fits in with our direction | 04:46 |
| sdake_ | but i'd bet it took atleat a month to develop all that code | 04:46 |
| SamYaple | service discovery i believe, like does a service exist | 04:46 |
| sdake_ | such a shame it has a 1 line commit log :( | 04:47 |
| SamYaple | or is a service running rather | 04:47 |
| sdake_ | isn't that what keystone does? | 04:47 |
| SamYaple | yea plus it should be broken into containers commited first then ansible | 04:47 |
| SamYaple | for review purposes | 04:47 |
| *** tfukushima has quit IRC | 04:52 | |
| *** itsuugo has joined #kolla | 04:56 | |
| sdake_ | SamYaple here are the docs https://review.openstack.org/#/c/228719/1/specs/consul-microservice-orchestration.rst | 05:05 |
| SamYaple | yea sdake_ im going through that whole review | 05:11 |
| *** dims has joined #kolla | 05:13 | |
| *** dims has quit IRC | 05:18 | |
| *** cloudnau_ has quit IRC | 05:29 | |
| *** tpot has quit IRC | 05:35 | |
| *** achanda has quit IRC | 05:36 | |
| *** inc0 has joined #kolla | 05:40 | |
| inc0 | good morning | 05:41 |
| SamYaple | morning inc0 | 05:42 |
| inc0 | wassup? | 05:42 |
| *** Haomeng has joined #kolla | 05:42 | |
| SamYaple | not much | 05:45 |
| SamYaple | got my 10gb and pcie ssd | 05:45 |
| inc0 | 10GB? there are 10gb disks? | 05:46 |
| inc0 | I mean, lately | 05:46 |
| SamYaple | 10gb nics | 05:46 |
| inc0 | ah, makes sense | 05:46 |
| SamYaple | pcie ssd 400gb | 05:46 |
| inc0 | do you have ikea's lack? | 05:48 |
| SamYaple | lack? | 05:50 |
| inc0 | http://www.ikea.com/us/en/catalog/products/90302060/ | 05:50 |
| inc0 | a table, lack | 05:50 |
| inc0 | to make a rack...so you have lack rack | 05:51 |
| inc0 | http://i.imgur.com/EpXi2h9.jpg | 05:51 |
| SamYaple | i dont personally no | 05:51 |
| SamYaple | thats awesome | 05:51 |
| inc0 | intentionally or not they created a side table which leg span is exactly what's required to put rack servers there | 05:51 |
| SamYaple | guess im going to ikea | 05:52 |
| inc0 | when I'll rent a house I'm going to buy catalyst and some storage server | 05:52 |
| inc0 | or buy one...whatever, lack is what I'll use | 05:52 |
| *** tfukushima has joined #kolla | 05:53 | |
| SamYaple | indeed. thats great | 05:54 |
| inc0 | and if you want to be more mobile, there is lack version with wheels | 05:55 |
| *** sdake_ is now known as sdake | 05:56 | |
| inc0 | problem is servers tent to be too long to fit there, but you can support them with yet another rack | 05:56 |
| inc0 | lack I mean... | 05:56 |
| *** tfukushima has quit IRC | 05:58 | |
| *** tfukushima has joined #kolla | 06:04 | |
| *** itsuugo has quit IRC | 06:09 | |
| *** itsuugo has joined #kolla | 06:14 | |
| *** itsuugo has quit IRC | 06:16 | |
| *** openstackgerrit has quit IRC | 06:16 | |
| *** openstackgerrit has joined #kolla | 06:16 | |
| *** akwasnie has joined #kolla | 06:24 | |
| *** sdake_ has joined #kolla | 06:33 | |
| *** sdake has quit IRC | 06:36 | |
| SamYaple | sdake_: also centos-source is borked iwth the oslo.log patch | 06:42 |
| SamYaple | i think its missing the patch binary | 06:42 |
| sdake_ | yes | 06:42 |
| sdake_ | i am testing a fix now | 06:42 |
| sdake_ | i just noticed it | 06:42 |
| sdake_ | i tested 3 of the 4 buidls | 06:42 |
| * sdake_ groans | 06:42 | |
| sdake_ | or 2 of the 3 | 06:42 |
| sdake_ | or whatever | 06:42 |
| * sdake_ smacks head against walls | 06:42 | |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make oslo.log work correctly on RDO platforms https://review.openstack.org/228715 | 06:45 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Neutron Agents json support https://review.openstack.org/228611 | 06:45 |
| sdake_ | samyaple were you able to launch a vm with that patch? | 06:45 |
| sdake_ | or just visual inspection? | 06:45 |
| SamYaple | visual. building with your just pushed now | 06:45 |
| sdake_ | for some reason I am not able to launch a vm with heat | 06:45 |
| sdake_ | but i'm not sure if its this code base or som other problem | 06:46 |
| sdake_ | since like 15 things were broken in the last week :) | 06:46 |
| SamYaple | yup | 06:46 |
| SamYaple | but we resolved them | 06:46 |
| SamYaple | we got really bad with our commits | 06:46 |
| sdake_ | well something is sitllbusted | 06:46 |
| SamYaple | myself included | 06:46 |
| sdake_ | yes people are not testing | 06:46 |
| sdake_ | which just pisses eveyrone off because they have to spend days bisecting other poeples work | 06:46 |
| SamYaple | well in my case i was removing files and said "i dont need to test that, we dont call tha anymore" | 06:47 |
| SamYaple | then kneejerk patches to fix kneejerk patches | 06:47 |
| sdake_ | we really need a gate | 06:47 |
| SamYaple | yup | 06:47 |
| SamYaple | larsks: thats right i said kneejerk | 06:47 |
| sdake_ | but until that point we relaly need to test every patch manually :) | 06:47 |
| sdake_ | suspect larsks is in a deep sleep atm | 06:47 |
| sdake_ | he is east coast | 06:47 |
| SamYaple | he'll see it | 06:48 |
| sdake_ | samyapel visual shoudl be good enough with most patches coming from me, but in this case i didn't have a baseline working system to test against | 06:51 |
| SamYaple | im testing it fully | 06:52 |
| SamYaple | i want working rc1 | 06:52 |
| sdake_ | but it is working better then it was previously | 06:52 |
| sdake_ | yes clearly we need rc1 working | 06:52 |
| sdake_ | i am not tagging until i can deploy multinode and get vms coming out of heat | 06:52 |
| sdake_ | there is more work to be done | 06:52 |
| SamYaple | unless its a heat bug | 06:52 |
| SamYaple | ;) | 06:52 |
| sdake_ | ya its not a heat bug :) | 06:52 |
| SamYaple | dejavu | 06:53 |
| SamYaple | im kidding i dont suspect heat | 06:53 |
| sdake_ | its a holy shit we just changed the entire code base with 2 major refactors at once without any testing | 06:53 |
| SamYaple | not entirely true, but fair enough | 06:53 |
| sdake_ | right i know not a major refactor | 06:53 |
| sdake_ | minor refactors :) | 06:53 |
| SamYaple | duuuude | 06:55 |
| SamYaple | these ssds work amazing backing ceph | 06:55 |
| sdake_ | ya they rock | 06:56 |
| sdake_ | best gear i've ever bought | 06:56 |
| SamYaple | the nics i bought are sr-iov capable | 06:57 |
| SamYaple | so ill be testing that gainst kolla :) | 06:57 |
| sdake_ | ok folks, t-2 weeks until release | 07:09 |
| sdake_ | that is 1 bug a day we need to get fixed :) | 07:09 |
| sdake_ | plus all the new ones we find | 07:09 |
| *** sdake_ is now known as sdake | 07:10 | |
| sdake | 2015-09-29 07:13:17.211 1 ERROR nova.compute.manager [instance: c037ca7f-2e97-4eef-abe5-840b5a72feeb] if ret == -1: raise libvirtError ('virDomainCreateWithFlags() failed', dom=self) | 07:14 |
| sdake | 2015-09-29 07:13:17.211 1 ERROR nova.compute.manager [instance: c037ca7f-2e97-4eef-abe5-840b5a72feeb] libvirtError: internal error: process exited while connecting to monitor: Could not access KVM kernel module: Permission denied | 07:14 |
| sdake | 2015-09-29 07:13:17.211 1 ERROR nova.compute.manager [instance: c037ca7f-2e97-4eef-abe5-840b5a72feeb] failed to initialize KVM: Permission denied | 07:14 |
| inc0 | hmm...non-priviledged container? | 07:14 |
| inc0 | first guess | 07:14 |
| sdake | ansible has privileged: True | 07:15 |
| SamYaple | oh i bet i know what that is | 07:15 |
| inc0 | do share Sam please | 07:16 |
| SamYaple | the /dev/kvm node has wrong perms | 07:16 |
| sdake | did someone muck wit hthe nova data container? | 07:16 |
| SamYaple | because libvirt isn't being read correctly | 07:16 |
| *** marun has quit IRC | 07:16 | |
| SamYaple | no but i did remove the chmod since it works for ubuntu with teh libvirtd.conf | 07:16 |
| sdake | [sdake@minime-02 log]$ ls -l /dev/kvm | 07:16 |
| sdake | crw-------. 1 root root 10, 232 Sep 29 01:19 /dev/kvm | 07:16 |
| SamYaple | yup thats it | 07:16 |
| SamYaple | so its it reading libvirtd.conf? | 07:16 |
| sdake | how w ould I know | 07:17 |
| SamYaple | idk | 07:17 |
| SamYaple | where is teh centos libvirt conf? | 07:17 |
| sdake | Running command: /usr/sbin/libvirtd --listen | 07:17 |
| SamYaple | yea thats fine | 07:17 |
| SamYaple | but where is the libvirt conf? | 07:17 |
| SamYaple | /etc/libvirt? libvirtd? | 07:17 |
| sdake | 5 seconds plz | 07:18 |
| sdake | /etc/libvirt/libvirtd.conf | 07:18 |
| SamYaple | https://github.com/openstack/kolla/blob/master/ansible/roles/nova/templates/libvirtd.conf.j2 | 07:18 |
| SamYaple | does it match that? | 07:18 |
| sdake | and there is a libvirt.conf | 07:18 |
| SamYaple | centos vs ubuntu may be looking for different default confs | 07:18 |
| inc0 | that would be strange | 07:19 |
| SamYaple | happens all the time | 07:19 |
| inc0 | and it would break long ago | 07:19 |
| SamYaple | look at mariadb | 07:19 |
| SamYaple | not if the change only recently merged (hint: it did) | 07:19 |
| sdake | [sdake@minime-02 etc]$ docker exec nova_libvirt ls /etc/libvirt | 07:19 |
| sdake | libvirt.conf | 07:19 |
| sdake | libvirtd.conf | 07:19 |
| inc0 | yay | 07:20 |
| inc0 | we have a bug | 07:20 |
| sdake | the correct file is libvirt.conf | 07:20 |
| sdake | the one written is libvirtd.conf | 07:20 |
| SamYaple | correct file _for_ _centos_ | 07:20 |
| SamYaple | it works for ubuntu | 07:21 |
| sdake | mind working up a patch for me | 07:21 |
| inc0 | we need conditional in our start script then | 07:21 |
| SamYaple | which is why we need gating | 07:21 |
| sdake | since you are ansible rock star plazz | 07:21 |
| SamYaple | i cant even run centos | 07:21 |
| sdake | ok well can someone give me a test patch plz - i dont know how to do conditional stuff in ansible | 07:21 |
| SamYaple | yea hold on | 07:21 |
| SamYaple | ill gist it | 07:21 |
| sdake | tia | 07:21 |
| inc0 | in our json we'll need dest_path conditional based on distro | 07:21 |
| SamYaple | youll need to test | 07:21 |
| sdake | of course | 07:22 |
| sdake | "MY NAME IS KAHN!" | 07:22 |
| inc0 | guys that's a start script stuff not ansible | 07:22 |
| sdake | well i just killed my deployment | 07:23 |
| SamYaple | inc0: wrong | 07:23 |
| SamYaple | https://gist.github.com/SamYaple/fe04af142be53e0d2648 | 07:23 |
| mlil | Quick question: I get an error where ansible tries to create /etc/kolla. The user it uses has sudo rights, but I suspect this is not enough, right? Should I change the owner of /etc to my user, or pre-create the /etc/kolla directory and let that one be owned by my user? | 07:23 |
| inc0 | hmmm...ok that would work as well | 07:23 |
| SamYaple | mlil: for liberty there is lots of breakage for using non-root user with ansible for ssh | 07:24 |
| SamYaple | mlil: therefore while you _can_ get it to work, we can't really help you. there is alot of factors and ansible is one of them | 07:24 |
| SamYaple | bugs in ansible prevent it from respecting some options | 07:24 |
| SamYaple | pbourke_: if you are around, you have some experince in this area, can you help mlil at all wit hthis? | 07:25 |
| inc0 | mlil, did you tell ansible to use sudo? | 07:26 |
| inc0 | you need to do that in inventory | 07:26 |
| inc0 | explicitly | 07:27 |
| sdake | deploying | 07:27 |
| mlil | SamYaple: Ok. Sounds like I should set ansible user to root then. And no, inc0, I did not know you could/should tell ansible about sudo... But the reason I didn't believe it is that sudo will still ask for password, right? | 07:27 |
| inc0 | mlil, it doesn't have to | 07:27 |
| inc0 | and afair ansible can handle that | 07:27 |
| SamYaple | mlil: ansible respects your .ssh/config file | 07:27 |
| SamYaple | so you can set that up with keys and users and passwords | 07:28 |
| inc0 | SamYaple, he was reffering to making sudo calls | 07:28 |
| inc0 | sudo may ask for pass | 07:28 |
| sdake | just set up passwordless sudo | 07:28 |
| inc0 | mlil, http://docs.ansible.com/ansible/become.html | 07:28 |
| SamYaple | appologies i read that as ssh not sudo :) | 07:29 |
| inc0 | you can use pass for sudo, another user, whatever you like | 07:29 |
| inc0 | only you need to specify that in inv | 07:29 |
| inc0 | brb | 07:29 |
| mlil | inc0: Ok, I'll read up on that link. Thanks! | 07:30 |
| SamYaple | inc0 for prez | 07:32 |
| sdake | 015-09-29 07:33:03.229 1 ERROR nova.virt.libvirt.host File "/usr/lib64/python2.7/site-packages/libvirt.py", line 105, in openAuth | 07:36 |
| sdake | 2015-09-29 07:33:03.229 1 ERROR nova.virt.libvirt.host if ret is None:raise libvirtError('virConnectOpenAuth() failed') | 07:36 |
| sdake | 2015-09-29 07:33:03.229 1 ERROR nova.virt.libvirt.host libvirtError: Failed to connect socket to '/var/run/libvirt/libvirt-sock': No such file or directory | 07:36 |
| *** shardy has joined #kolla | 07:37 | |
| sdake | libvirt is in a restart loop | 07:38 |
| * sdake groans | 07:38 | |
| *** itsuugo has joined #kolla | 07:38 | |
| *** itsuugo has quit IRC | 07:40 | |
| *** exploreshaifali has joined #kolla | 07:40 | |
| sdake | samyaple arey ou sure this syntax is correct? +{% set libvirt_conf = 'libvirtd.conf' if kolla_base_distro in ['ubuntu', 'debian'] else 'libvirt.conf' %} | 07:41 |
| SamYaple | itscopy paste from elsewhere, but just look at the file it produced | 07:41 |
| SamYaple | /etc/kolla/nova-libvirt/ | 07:42 |
| sdake | you mean + "dest": "/etc/libvirt/{{ libvirt_conf }}", | 07:42 |
| sdake | there is no /etc/libvirt directory in the container | 07:42 |
| sdake | the directory is being deleted I suspect | 07:42 |
| SamYaple | does sentos expect /etc/libvirt.conf? or /etc/libvirt/libvirt.conf? | 07:43 |
| SamYaple | centos* | 07:43 |
| sdake | /etc/libvirt/libvirt.conf is whta wa there previously | 07:43 |
| sdake | but now /etc/libvirt directory is gone | 07:43 |
| sdake | for some odd reaosn setconfigs.py deletes /etc/libvirt | 07:43 |
| SamYaple | thats what set_configs.py does | 07:43 |
| SamYaple | it removes the dest | 07:44 |
| SamYaple | so that code isnt right or something | 07:44 |
| SamYaple | what does your config.json say? | 07:44 |
| SamYaple | cat /etc/kolla/nova-libvirt/config.json | 07:44 |
| sdake | { | 07:45 |
| sdake | "command": "/usr/sbin/libvirtd --listen", | 07:45 |
| sdake | "config_files": [ | 07:45 |
| sdake | { | 07:45 |
| sdake | "source": "/opt/kolla/config_files/libvirtd.conf", | 07:45 |
| sdake | "dest": "/etc/libvirt/libvirt.conf", | 07:45 |
| sdake | "owner": "root", | 07:45 |
| sdake | "perm": "0600" | 07:45 |
| sdake | } ] | 07:45 |
| sdake | } | 07:45 |
| SamYaple | and /etc/kolla/nova-libvirt/libvirtd.conf exists? | 07:45 |
| sdake | yup | 07:45 |
| sdake | oh | 07:46 |
| sdake | not sure | 07:46 |
| sdake | i'll check | 07:46 |
| *** jmccarthy has quit IRC | 07:47 | |
| *** jmccarthy has joined #kolla | 07:48 | |
| sdake | yes that file is present | 07:48 |
| SamYaple | willing to buy lint checks | 07:49 |
| *** itsuugo has joined #kolla | 07:49 | |
| SamYaple | i have made some god awful commits lately | 07:51 |
| SamYaple | not being able to test so i dont tet and push | 07:51 |
| SamYaple | cascadding failures | 07:51 |
| SamYaple | horrible | 07:51 |
| sdake | do you have a candidate bug ? | 07:52 |
| sdake | or something for me to run throug hthe grinder | 07:52 |
| SamYaple | for ceph yes | 07:52 |
| * sdake loves being the single point of testing for all kolla! | 07:52 | |
| SamYaple | hey you said 25th and thne werent around | 07:52 |
| SamYaple | it was a freeforall mergefest | 07:52 |
| sdake | intentional ;) | 07:53 |
| SamYaple | lies | 07:53 |
| sdake | so any idea as to how to solv ethis problem? | 07:53 |
| SamYaple | i havent even gotten there yet | 07:53 |
| SamYaple | im working my way to it | 07:53 |
| sdake | note i am not certain the config file is actually libvirtd.conf | 07:53 |
| sdake | let me try running a fresh libvirt container sec | 07:54 |
| sdake | there is a libvirt.conf and libvirtd.conf file | 07:54 |
| sdake | libvirtd.conf looks like the configuration file | 07:55 |
| SamYaple | never again. we need gating asap | 07:55 |
| inc0 | yeah, huge priority | 07:56 |
| SamYaple | its my #1 for mitaka | 07:56 |
| sdake | what is it youchanged, you remoed a chmod? | 07:56 |
| inc0 | I'd add sanity control after each step of deployment as well | 07:57 |
| SamYaple | sdake: yea just throw it back in there | 07:57 |
| sdake | what wa where ? | 07:57 |
| sdake | in start.sh? | 07:57 |
| inc0 | that and remove config_external, that happend too | 07:57 |
| SamYaple | commiting ontop of broken repo because couldn't test :/ | 07:59 |
| SamYaple | we just started peeling back the fixes | 07:59 |
| SamYaple | bad sam! | 07:59 |
| sdake | boy debugging restart loops is annoyign | 08:00 |
| SamYaple | docker_restart_policy: "no" | 08:00 |
| SamYaple | never leave home without it | 08:01 |
| sdake | i need a running container... | 08:01 |
| SamYaple | start it with sleep infiinity and execture start.sh by hand | 08:01 |
| sdake | what was the /dev/kvm chmod set to? | 08:02 |
| sdake | 660? | 08:03 |
| SamYaple | 660 i think | 08:03 |
| sdake | what is default on ubuntu, 660? | 08:03 |
| sdake | default on centos is 600 | 08:03 |
| SamYaple | it really doesnt matter | 08:03 |
| SamYaple | we run as root | 08:03 |
| SamYaple | it was set to 660 though | 08:03 |
| sdake | why is it not working as root then? | 08:03 |
| sdake | seems odd | 08:03 |
| SamYaple | libvirt uses group kvm | 08:04 |
| SamYaple | hence the 660 | 08:04 |
| SamYaple | thats what the libvrirtd.conf fixes | 08:04 |
| SamYaple | uses root:root | 08:04 |
| SamYaple | not root:kvm | 08:04 |
| sdake | so you think libvirtd.conf isn't beingparsed? | 08:04 |
| SamYaple | idk there are other issue afoot | 08:06 |
| SamYaple | for me | 08:06 |
| SamYaple | and ceph | 08:06 |
| sdake | so the libvirtd.conf file is mode 600 | 08:09 |
| sdake | can libvirtd read it? | 08:09 |
| *** marun has joined #kolla | 08:13 | |
| sdake | SamYaple i am pretty sure deleting all of /etc/libvirt is a bad idea, i think we need /etc/libvirt/nwfilter directory | 08:13 |
| inc0 | I'll try to run basic centos with libvirt to confirm default filename | 08:14 |
| inc0 | or hold on, I think I have access to this one | 08:14 |
| sdake | it has something to do with the ip spoofing xml stuff | 08:14 |
| *** marun has quit IRC | 08:18 | |
| SamYaple | sdake: libvirt seems to be working for me | 08:20 |
| sdake | your permissions are different i think | 08:20 |
| sdake | mine is 600 root:root | 08:20 |
| SamYaple | but i found a bug in the recursive copy in set_configs.py | 08:20 |
| sdake | on host | 08:20 |
| *** athomas has joined #kolla | 08:21 | |
| SamYaple | it works when the file is relative | 08:21 |
| sdake | i am pretty sure youc an't delete /etc/libvirt/nwfilter on centos | 08:21 |
| sdake | on ubuntu maybey ou can | 08:21 |
| sdake | its hard to tell why the damn thing is exiting | 08:22 |
| sdake | because it doesn't print out anything | 08:22 |
| SamYaple | sdake: check the log_data | 08:22 |
| SamYaple | oh nevermind | 08:22 |
| SamYaple | i havent gotten to that | 08:22 |
| sdake | i am redeploying but sure once that finishes | 08:22 |
| SamYaple | are you running btrfs? | 08:22 |
| sdake | overlay | 08:23 |
| SamYaple | :/ | 08:23 |
| sdake | i changed permissions of libvirtd.conf file to 664 and now it works | 08:25 |
| sdake | from 600 | 08:25 |
| sdake | my speuclation was correct | 08:25 |
| SamYaple | do it. thatll be fine for ubuntu | 08:25 |
| SamYaple | patch that | 08:25 |
| sdake | yup | 08:25 |
| sdake | let me test a launch ofa vm tho | 08:25 |
| SamYaple | i have no way to have caught that :( | 08:25 |
| SamYaple | gating | 08:25 |
| mlil | SamYaple, inc0: Just wanna say I got further now that I enabled ansible_sudo:yes and enabled password-less sudo on the machines, thanks for the help! Now I got other errors that I feel responsible for causing... ;-) | 08:26 |
| sdake | /dev/kvm i nthe container is 600 | 08:27 |
| sdake | so i am not sure libvirt will work with it | 08:27 |
| sdake | testing now | 08:27 |
| SamYaple | just throw the chmod back in and we can sort out later | 08:27 |
| *** akwasnie has quit IRC | 08:29 | |
| sdake | ya /dev/kvm needs something with a chmod | 08:31 |
| SamYaple | sdake: https://github.com/openstack/kolla/blob/aebb16fff9f33727aa9ce49120a1c6384544bfdc/docker/nova/nova-libvirt/start.sh#L11-16 | 08:36 |
| SamYaple | just fyi | 08:36 |
| sdake | ya i found it already | 08:36 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Fix ceph after config-external removal https://review.openstack.org/228788 | 08:39 |
| SamYaple | that fixes all of the ceph issues | 08:39 |
| SamYaple | .... tested :/ | 08:40 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make libvirt function on CentOS https://review.openstack.org/228790 | 08:42 |
| *** exploreshaifali has quit IRC | 08:45 | |
| *** erkules has quit IRC | 08:46 | |
| *** erkules has joined #kolla | 08:46 | |
| sdake | i've offocially been awake for 24 hours | 08:48 |
| SamYaple | go to bed | 08:49 |
| SamYaple | ill test your patch on ubuntu | 08:49 |
| *** gfidente has joined #kolla | 08:49 | |
| sdake | testing patch on centos atm | 08:49 |
| sdake | and smoking cigar | 08:49 |
| sdake | then hitting the rcak | 08:49 |
| SamYaple | ive got the whole stream of in queue patches building right now | 08:49 |
| sdake | but thanks for testing on ubuntu | 08:49 |
| sdake | did ryn remove that /dev/kvm permissions setting ? | 08:50 |
| SamYaple | all me | 08:50 |
| sdake | nice ;) | 08:50 |
| SamYaple | i broke alot of shit | 08:50 |
| SamYaple | i feel really bad | 08:50 |
| SamYaple | i was burn-out commiting | 08:50 |
| sdake | not familiar with that term | 08:51 |
| SamYaple | i made it up | 08:51 |
| SamYaple | i am a bit burned out. so my quality was lacking | 08:52 |
| sdake | i feel pretty well rested actually :) | 08:52 |
| SamYaple | three weeks will do that | 08:52 |
| SamYaple | sick or not | 08:52 |
| sdake | ya 3 weeks of rack time | 08:52 |
| sdake | i am tired of sleeping | 08:52 |
| *** Kennnan2 has joined #kolla | 08:54 | |
| sdake | STDAKE-M-K04E:demo sdake$ nova list | grep ACTIVE | wc -l | 08:54 |
| sdake | 8 | 08:54 |
| *** Kennnan has quit IRC | 08:54 | |
| sdake | victory | 08:54 |
| sdake | that 3 patch patch stream gets heat back into operating state | 08:54 |
| SamYaple | cool. im testing that + ceph | 08:55 |
| sdake | yar, I dont hae a ceph setup atm | 08:55 |
| SamYaple | no problem. ill be the guienea pig there | 08:56 |
| *** itsuugo has quit IRC | 09:07 | |
| *** itsuugo has joined #kolla | 09:08 | |
| *** daneyon_ has joined #kolla | 09:11 | |
| *** daneyon has quit IRC | 09:11 | |
| *** jmccarthy has quit IRC | 09:15 | |
| *** jmccarthy has joined #kolla | 09:15 | |
| *** dims has joined #kolla | 09:16 | |
| sdake | looks like delorean is busted atm | 09:19 |
| * sdake growls | 09:19 | |
| sdake | i can't build this openstack-base redux | 09:19 |
| SamYaple | garr | 09:19 |
| *** dims has quit IRC | 09:22 | |
| SamYaple | sdake: https://review.openstack.org/#/c/228788/ | 09:22 |
| SamYaple | thats teh ceph patch | 09:22 |
| SamYaple | it toches set_configs.py | 09:22 |
| SamYaple | damn ceph is zippy | 09:24 |
| *** diga has quit IRC | 09:25 | |
| *** itsuugo has quit IRC | 09:28 | |
| *** itsuugo has joined #kolla | 09:29 | |
| jmccarthy | Hmm (for swift as an example) what technique are folks thinking to get across multiple disks for storage ? Bind mounts ? | 09:36 |
| SamYaple | jmccarthy: it has to be bind mounts for sure | 09:36 |
| SamYaple | we arent mounting the filesystem in the container | 09:37 |
| SamYaple | if the container crash that mount gets wierdly lost | 09:37 |
| SamYaple | so looking at some of the stuff i did in ceph, i think we can make this work a bit dynamically | 09:37 |
| *** diga has joined #kolla | 09:41 | |
| jmccarthy | So bind mounts to the data container, that seems ok .. I suppose there could be differences across hosts, dynamic how ? | 09:42 |
| sdake | [sdake@bigiron kolla]$ docker images |grep libvirt |grep centos-binary | grep latest | 09:43 |
| sdake | 192.168.1.100:4000/kollaglue/centos-binary-nova-libvirt latest a289cbb79fbe About a minute ago 730.6 MB | 09:43 |
| sdake | how big is your libvirt for ubuntu SamYaple | 09:43 |
| sdake | it went from 1gb to 730mb | 09:44 |
| sdake | with the openstack-common layer changes | 09:44 |
| SamYaple | as it should | 09:44 |
| SamYaple | you are welcome | 09:44 |
| SamYaple | root@server01:~# docker images | grep libvirt | 09:44 |
| SamYaple | 192.168.31.12:8082/kollaglue/ubuntu-source-nova-libvirt latest e111edf4464c 34 minutes ago 588.1 MB | 09:44 |
| *** dims has joined #kolla | 09:47 | |
| SamYaple | ubuntu@sy-test01:~$ dd if=/dev/zero of=bigfile.bin bs=1G count=10 | 09:51 |
| SamYaple | 10+0 records in | 09:51 |
| SamYaple | 10+0 records out | 09:51 |
| SamYaple | 10737418240 bytes (11 GB) copied, 20.2222 s, 531 MB/s | 09:51 |
| SamYaple | sdake: ^^^^^^ | 09:51 |
| SamYaple | thats inside an instance | 09:51 |
| SamYaple | through ceph | 09:51 |
| sdake | fast | 09:51 |
| *** tfukushima has quit IRC | 09:54 | |
| *** tfukushima has joined #kolla | 09:56 | |
| inc0 | new disk SamYaple ? | 09:56 |
| inc0 | cmon buy another one like that and make raid 0 out of it | 09:57 |
| SamYaple | inc0: thats ceph with spinning disk storage | 09:57 |
| SamYaple | and a ssd writeback teir | 09:57 |
| SamYaple | numbers from _inside_ the instance | 09:57 |
| inc0 | so it can be even faster | 09:58 |
| inc0 | yeah I know, but its 10gig right? | 09:58 |
| SamYaple | 10gig networking yes | 09:58 |
| inc0 | then probably bottleneck is in disks right? especially if it's spinning as you say | 09:58 |
| inc0 | even with tiered right? | 09:59 |
| jmccarthy | SamYaple: Is some of work done in kolla or in yaodu is it ? | 10:00 |
| SamYaple | jmccarthy: at this point all yaodu things are in Kolla :) | 10:00 |
| jmccarthy | Trying to get a better understanding of where this might be going =) | 10:01 |
| SamYaple | basically i have the ceph storage dynamically looked up and bootstrapped and all | 10:01 |
| SamYaple | it looks for partitions labeled 'KOLLA_CEPH_OSD_BOOTSTRAP' to bootstrap them | 10:01 |
| SamYaple | the admin must set that label explicitly | 10:02 |
| SamYaple | per drive | 10:02 |
| SamYaple | then it formats and bootstraps the disk | 10:02 |
| SamYaple | then kolla looks for labels 'KOLLA_CEPH_DATA' and mounts those and launches teh osd (disk daemons) per disk | 10:02 |
| SamYaple | swift can be done similiarly | 10:02 |
| sdake | that sounds pretty cool SamYaple | 10:05 |
| jmccarthy | Has that merged ? | 10:05 |
| SamYaple | jmccarthy: indeed | 10:05 |
| SamYaple | its all working | 10:05 |
| SamYaple | on the fly boostraps in new disks as well ;) | 10:05 |
| SamYaple | just run the playbooks it will pick them up | 10:05 |
| jmccarthy | I gotta check it out =) ! | 10:06 |
| SamYaple | it actually has a few advantages over how ceph actually deploys | 10:06 |
| SamYaple | im writing docs for it but im so bad with docs | 10:06 |
| sdake | just write them badly and I'll improve them :) | 10:06 |
| SamYaple | Kennnan2: if i do a raw text dump can you docify it? | 10:06 |
| sdake | i'll do that sam | 10:06 |
| sdake | kennan2 is notnative english speaker | 10:07 |
| SamYaple | ok ill do a raw text dump then | 10:07 |
| SamYaple | fair enough. but i can assure you he can english better than me | 10:07 |
| sdake | ya his english is good | 10:07 |
| sdake | but i'm reallyg ood at docs minus the typos :) | 10:07 |
| sdake | here is the openstack-common image sizes: http://ur1.ca/nvlyu -> http://paste.fedoraproject.org/272579/21361144 | 10:09 |
| sdake | for centos-binary | 10:09 |
| SamYaple | far more reasonable | 10:10 |
| SamYaple | ubuntu still beats them, but thats because of --no-install-recommends | 10:10 |
| SamYaple | as time goes on i bet as people want new things those numbers even out | 10:10 |
| SamYaple | (ubuntu will always win though) | 10:10 |
| openstackgerrit | Merged openstack/kolla: Make oslo.log work correctly on RDO platforms https://review.openstack.org/228715 | 10:12 |
| openstackgerrit | Merged openstack/kolla: Neutron Agents json support https://review.openstack.org/228611 | 10:12 |
| openstackgerrit | Merged openstack/kolla: Make libvirt function on CentOS https://review.openstack.org/228790 | 10:13 |
| openstackgerrit | Merged openstack/kolla: Remove config-external for rsyslog https://review.openstack.org/228448 | 10:13 |
| *** marun has joined #kolla | 10:14 | |
| pbourke_ | mlil: hi, if you're still having trouble give me a shout and I'll see if I can help | 10:14 |
| SamYaple | sdake: just need your vote and i think we can tag https://review.openstack.org/#/c/228788/ | 10:14 |
| sdake | i want to add this openstack common layer | 10:15 |
| sdake | but ya | 10:15 |
| sdake | reviewingnow | 10:15 |
| sdake | groan the gate is broken ? | 10:15 |
| SamYaple | hmm while i also want to see that added sdake i feel that might have the potential for unforseen breakage | 10:15 |
| SamYaple | but ill follw you | 10:15 |
| SamYaple | delorean is yea | 10:15 |
| SamYaple | but this only touches ceph ansible files (and set_configs.py) | 10:15 |
| SamYaple | review on the bases of set_configs.py | 10:16 |
| SamYaple | basis* | 10:16 |
| sdake | ya i dont mind waiting until after rc1 | 10:18 |
| sdake | as long as we can merge it in the next couple weeks | 10:18 |
| sdake | because it will make kolla quite a bit better | 10:18 |
| sdake | I dont know why we waited until the last minut eto do rhel but I was out of action for most of this cycle unfortunately :( | 10:18 |
| sdake | rather centos not rhel | 10:18 |
| *** marun has quit IRC | 10:19 | |
| openstackgerrit | Merged openstack/kolla: Fix ceph after config-external removal https://review.openstack.org/228788 | 10:20 |
| sdake | nice thing about jamming it into rc1 is unforseen breakage iill be resolved by rc2 which is th 2nd | 10:20 |
| sdake | only 3-4 days | 10:20 |
| SamYaple | ubuntu@sy-test01:~$ dd if=/dev/zero of=bigfile.bin bs=1G count=10 | 10:22 |
| SamYaple | 10+0 records in | 10:22 |
| SamYaple | 10+0 records out | 10:22 |
| SamYaple | 10737418240 bytes (11 GB) copied, 14.3469 s, 748 MB/s | 10:22 |
| SamYaple | LOOK! LOOK AT THEM NUMBERS | 10:22 |
| sdake | do a 200gb write :) | 10:22 |
| SamYaple | thats 748 MegaBYTES | 10:22 |
| SamYaple | ok | 10:22 |
| SamYaple | gotta launch a new instance | 10:22 |
| SamYaple | no space :p | 10:22 |
| jmccarthy | SamYaple: Oh ! Ok this isn't using any data containers then ? | 10:22 |
| sdake | i'm crious what you get with that size of data output | 10:22 |
| sdake | data containers couldn't produce that kind of io | 10:23 |
| SamYaple | you get 100gb cause i dont want to create a new flavor | 10:24 |
| jmccarthy | I was sort of thinking it might go data container with the stuff bind mounted to it, no ? | 10:24 |
| jmccarthy | sdake: Not sure I see the difference, couldn't services use data container as they currently do, but use bind mounts instead of volumes ? (thinking out loud) | 10:27 |
| sdake | bind mounts are jut slower | 10:27 |
| SamYaple | no data conatienrs are slower than bindmounts | 10:27 |
| SamYaple | ceph is using all bindmounts remember | 10:27 |
| jmccarthy | Maybe I'm saying it wrong, a data container that has a bind mount vs using bind mount directly is noticeably slower ? | 10:29 |
| SamYaple | jmccarthy: i know what you are saying, and no that wont be anyslower than direct bindmount | 10:30 |
| sdake | yup i woudl expect so | 10:30 |
| SamYaple | no it wont because of how docker shares volumes | 10:30 |
| SamYaple | it would still be as fast as bindmount | 10:30 |
| SamYaple | bindmount==fastest | 10:30 |
| *** pbourke_ has quit IRC | 10:34 | |
| *** pbourke_ has joined #kolla | 10:34 | |
| jmccarthy | Well I'm sort of thinking for multiple disks, for swift an option could be re-jig the swift_data container with bind mounts (vs volumes that it has now) - no changes needed really in any other containers (generate rings accordingly) - job done, no ? | 10:35 |
| SamYaple | ubuntu@sy-test01:~$ dd if=/dev/zero of=bigfile.bin bs=1G count=140 | 10:35 |
| SamYaple | 140+0 records in | 10:35 |
| SamYaple | 140+0 records out | 10:35 |
| SamYaple | 150323855360 bytes (150 GB) copied, 205.496 s, 732 MB/s | 10:35 |
| sdake | wow smoking fast | 10:35 |
| sdake | expected like 300mb/sec ;) | 10:36 |
| SamYaple | and i can use fstrim once i delete that file to few up that 150GB btw | 10:37 |
| SamYaple | it sends discard t oceph and recovers that space in the cluster | 10:38 |
| jmccarthy | (aside from launching multiple servers i.e. per disk - that's another thing) | 10:38 |
| SamYaple | jmccarthy: yea maybe. my thought was to trash the swift data container and force people to use loopback devices for testing | 10:39 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make a openstack-common container for CentOS variants https://review.openstack.org/228831 | 10:39 |
| SamYaple | realworld the datacontainer will never be useful | 10:39 |
| SamYaple | and as long as you can test without having to have a physical disk, i dont think it matters if its a few extra steps (creating the loop device) | 10:40 |
| SamYaple | sdake: are you mocing naything form centos source in that patch? | 10:40 |
| sdake | data container is useful for some small non-perfomrnat things | 10:40 |
| sdake | like rabbitm qdata | 10:40 |
| sdake | and mariadb data | 10:40 |
| SamYaple | sdake: i meant just for swift | 10:40 |
| SamYaple | the otehrs are cool | 10:40 |
| sdake | oh right | 10:40 |
| SamYaple | sdake: are you mocing naything form centos source in that patch? like al lthe build tools should be in openstack-common | 10:41 |
| SamYaple | moving* | 10:41 |
| Kennnan2 | SamYaple: just notice your message, sorry | 10:41 |
| Kennnan2 | what can I do ? | 10:41 |
| sdake | all the build tools are in opentack-common for source | 10:41 |
| sdake | but not binary | 10:41 |
| sdake | i didn't touch how from source works for the mosst part | 10:41 |
| sdake | only from binary | 10:41 |
| sdake | which only neds git and pip | 10:41 |
| sdake | (for kolla-ansible and neutron) | 10:42 |
| SamYaple | Kennnan2: haha its ok. I was asking you to do work. I have a raw text dump for how to use ceph with Kolla and i dont really know the formatting to make that docs friendly | 10:42 |
| SamYaple | sdake: but they arent, im looking at them | 10:42 |
| sdake | testing binary atm | 10:42 |
| SamYaple | oh so just binary first. thats fine | 10:43 |
| sdake | i trimmed the gat | 10:43 |
| sdake | i didn't do anything with source I dont think | 10:43 |
| SamYaple | can you put that in the commit? | 10:43 |
| SamYaple | because you put implements and it sounds like youve got another commit coming | 10:43 |
| Kennnan2 | ok. let me know if need any help further. BTW, I tried https://review.openstack.org/#/c/218144/ | 10:43 |
| sdake | thats the last one | 10:43 |
| Kennnan2 | to add kolla in requirements family | 10:43 |
| SamYaple | sdake: that doesnt implement openstack-common though | 10:43 |
| sdake | isn't it Implements for the final one? | 10:43 |
| SamYaple | right... my point is it isnt done | 10:44 |
| sdake | what else is needed? | 10:44 |
| SamYaple | https://github.com/openstack/kolla/blob/master/docker/base/Dockerfile.j2#L228-242 | 10:44 |
| SamYaple | that moves to openstack common, or at least some of it | 10:44 |
| sdake | oh right | 10:44 |
| sdake | i did onlybinary | 10:45 |
| sdake | i forgot about source there | 10:45 |
| sdake | source is more complex | 10:45 |
| SamYaple | yea i was saying that but you are really tired | 10:45 |
| SamYaple | its actually pretty straight forward | 10:45 |
| SamYaple | if its gc or dev libraries, opentsack-common it goes | 10:45 |
| SamYaple | if its for building it leaves the base | 10:45 |
| openstackgerrit | Steven Dake proposed openstack/kolla: Make a openstack-common container for CentOS variants https://review.openstack.org/228831 | 10:46 |
| SamYaple | you can probably collapes the source and binary sections in the base file after that move | 10:46 |
| sdake | well i'm not doing source prior to rc1 | 10:46 |
| SamYaple | perhaps we should not fly to close to the sun and tag the current working? | 10:46 |
| SamYaple | your call, none of this affects ubuntu | 10:47 |
| sdake | we can but rc2 releaes on the 2nd | 10:47 |
| SamYaple | wait rc1 and then rc2 within a day or two? | 10:47 |
| sdake | thats just how the schedule worked out | 10:47 |
| SamYaple | ok | 10:48 |
| sdake | STDAKE-M-K04E:demo sdake$ nova list | grep ACTIVE | wc -l | 10:48 |
| sdake | 8 | 10:48 |
| SamYaple | well like i said dont touch ubuntu so no complaint here | 10:48 |
| sdake | because we are so late with rc1 | 10:48 |
| sdake | i can wait and just tag wht we have | 10:48 |
| sdake | and we can put in to rc2 | 10:48 |
| sdake | i am going to bed though | 10:48 |
| sdake | night :) | 10:48 |
| SamYaple | sounds good | 10:48 |
| SamYaple | well hold back commits and tag | 10:48 |
| sdake | feel free to review please | 10:49 |
| sdake | i'll tag when i wake up | 10:49 |
| sdake | adn also I'llbe branching | 10:49 |
| sdake | so msater willbe open for new work | 10:49 |
| *** sdake_ has joined #kolla | 10:50 | |
| *** sdake has quit IRC | 10:54 | |
| *** exploreshaifali has joined #kolla | 11:03 | |
| *** shardy_ has joined #kolla | 11:03 | |
| *** tfukushima has quit IRC | 11:05 | |
| *** shardy has quit IRC | 11:05 | |
| *** shardy_ has quit IRC | 11:09 | |
| *** shardy has joined #kolla | 11:09 | |
| *** itsuugo has quit IRC | 11:11 | |
| SamYaple | inc0: you around? | 11:15 |
| inc0 | yeah | 11:15 |
| SamYaple | VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE,VIR_MIGRATE_PERSIST_DEST,VIR_MIGRATE_TUNNELLED | 11:15 |
| jmccarthy | SamYaple: Wondering (for ceph) about why having containers work out for themselves in bootstrap vs doing configuration via ansible ? | 11:15 |
| SamYaple | performance wise, what should i be using there? | 11:15 |
| SamYaple | jmccarthy: can you give an example? | 11:15 |
| inc0 | SamYaple, I'd remove VIR_MIGRATE_TUNNELED | 11:17 |
| inc0 | so you move communication from libvirt->libvirt to qemu->qemu | 11:17 |
| SamYaple | right, im reading http://libvirt.org/migration.html right now | 11:17 |
| SamYaple | im unclear on the UNDEFINE_SOURCE and PERSIST_DEST | 11:18 |
| inc0 | undefine_source means that it will undefine base vm | 11:18 |
| inc0 | so remove domain | 11:18 |
| SamYaple | oh ok that i get | 11:18 |
| inc0 | I'm not sure about persist | 11:18 |
| jmccarthy | SamYaple: In ansible ? I suppose somewhere per host disk/partition break down would be in inventory or somewhere, and it could setup bind mounts then when starting containers ? | 11:18 |
| SamYaple | jmccarthy: its magic! | 11:19 |
| SamYaple | no inventory file holding that info | 11:19 |
| SamYaple | dynamically fetched and done all up right | 11:19 |
| jmccarthy | Well not currently, but it's an approach I mean | 11:19 |
| SamYaple | no it is currently | 11:19 |
| SamYaple | oh you mean swift | 11:19 |
| jmccarthy | I mean no inventory file holding that info currently for swift | 11:20 |
| SamYaple | right. yea i am against that setup | 11:20 |
| SamYaple | its big and klunky | 11:20 |
| SamYaple | you get to 200 nodes with 20 disks each and you have 4000 entries... | 11:20 |
| SamYaple | pita | 11:20 |
| SamYaple | we can do it dynamically | 11:20 |
| SamYaple | i just need to play with it | 11:20 |
| jmccarthy | Hmm ok yea I know what you mean, just wondering about possible support complications | 11:21 |
| *** itsuugo has joined #kolla | 11:22 | |
| SamYaple | inc0: "The --persist flag will cause a configuration file to be created on the destination host after a successful migration." | 11:22 |
| SamYaple | jmccarthy: i currently support an implementation with each disk specified, i would take dyanmic anyday | 11:22 |
| *** tfukushima has joined #kolla | 11:23 | |
| SamYaple | inc0: looks like tunneled is default now https://github.com/openstack/nova/blob/1bbf51e5da545e287b8fe3339155c094c0eb11ca/nova/virt/libvirt/driver.py#L162 | 11:24 |
| inc0 | SamYaple, it is, it's more flexible | 11:25 |
| SamYaple | im testing performance iwth nad without | 11:25 |
| SamYaple | did you see my ceph numbers for disk writes? | 11:25 |
| inc0 | not every hypervisor support hypervisor->hypervisor communication | 11:25 |
| inc0 | yeah, looks nice | 11:25 |
| SamYaple | thats on top of 16TB of storage | 11:25 |
| inc0 | I'm about to test out ceph on our hardware as well | 11:26 |
| SamYaple | 2x400GB ssd for cache, 4x4TB spinning disk | 11:26 |
| inc0 | do we support dedicated disk for journal? | 11:26 |
| SamYaple | we dont _support_ anything :P | 11:26 |
| inc0 | ok, can I do it? | 11:27 |
| SamYaple | but no. thats not baked into the scripts | 11:27 |
| inc0 | so that would make things even better probably | 11:27 |
| SamYaple | no | 11:27 |
| inc0 | how so? | 11:27 |
| SamYaple | because ive just got the writeback layer as cache | 11:27 |
| SamYaple | bascially instead of the ssd being the journal its a writeback cache | 11:27 |
| inc0 | well ok | 11:27 |
| SamYaple | more effiecent | 11:27 |
| inc0 | ow? I thought it's better to have joutnals separated on dedicated fast disks | 11:28 |
| SamYaple | sure... if you dont use a cache tier | 11:28 |
| SamYaple | with a writeback cachetier you are better of performance wise throwing those ssd into a cache | 11:28 |
| SamYaple | but that is dependant on cluster configuration and version of ceph blah blah | 11:29 |
| inc0 | well you don't lose ssd for just journal and journals will be ssd-fast | 11:29 |
| *** dwalsh_ has joined #kolla | 11:30 | |
| inc0 | cool, I didn't know that | 11:30 |
| inc0 | but again, I'm storage ignorant | 11:30 |
| *** dwalsh has joined #kolla | 11:30 | |
| SamYaple | cache tiers are newer and newer == more possible bugs. but its been out for several releases and its pretty solid in hammer | 11:31 |
| SamYaple | journal is more traditional, but far worse use of ssd resources | 11:31 |
| SamYaple | you can do both of course | 11:31 |
| inc0 | yeah, but now when I think about it there is little point | 11:32 |
| inc0 | as long as you won't run out of cache | 11:32 |
| SamYaple | well if you run out its just passthrough straight to osd | 11:32 |
| SamYaple | i mean you dont break things | 11:32 |
| SamYaple | its just super slow | 11:32 |
| *** exploreshaifali has quit IRC | 11:32 | |
| SamYaple | basically the cachetier is a pool of ssd only osds | 11:34 |
| *** itsuugo has quit IRC | 11:50 | |
| *** itsuugo has joined #kolla | 12:01 | |
| *** exploreshaifali has joined #kolla | 12:07 | |
| *** tfukushima has quit IRC | 12:12 | |
| *** itsuugo has quit IRC | 12:12 | |
| *** shardy_ has joined #kolla | 12:13 | |
| *** itsuugo has joined #kolla | 12:13 | |
| *** shardy has quit IRC | 12:14 | |
| *** marun has joined #kolla | 12:15 | |
| *** shardy_ has quit IRC | 12:18 | |
| SamYaple | im building ubuntu liberty rc1 images | 12:18 |
| *** shardy has joined #kolla | 12:19 | |
| *** marun has quit IRC | 12:20 | |
| *** dolpher has joined #kolla | 12:24 | |
| *** dwalsh_ has quit IRC | 12:25 | |
| *** dwalsh has quit IRC | 12:25 | |
| openstackgerrit | Merged openstack/kolla: Make a openstack-common container for CentOS variants https://review.openstack.org/228831 | 12:27 |
| *** diga has quit IRC | 12:27 | |
| *** rhallisey has joined #kolla | 12:29 | |
| *** itsuugo has quit IRC | 12:30 | |
| *** dims_ has joined #kolla | 12:36 | |
| *** dims has quit IRC | 12:40 | |
| *** exploreshaifali has quit IRC | 12:40 | |
| *** dwalsh_ has joined #kolla | 12:44 | |
| *** dwalsh has joined #kolla | 12:44 | |
| *** dolpher has quit IRC | 12:52 | |
| *** dolpher has joined #kolla | 12:52 | |
| *** dims_ has quit IRC | 12:53 | |
| SamYaple | inc0: changing the queue scheduler for the disk in the vm gives me | 12:55 |
| SamYaple | ubuntu@sy-test01:~$ dd if=/dev/zero of=bigfile.bin bs=1G count=10 | 12:55 |
| SamYaple | 10+0 records in | 12:55 |
| SamYaple | 10+0 records out | 12:55 |
| SamYaple | 10737418240 bytes (11 GB) copied, 11.5836 s, 927 MB/s | 12:55 |
| SamYaple | also without TUNNELLED the live migration was much _much_ faster | 12:56 |
| SamYaple | i did one of these large writes while doing a live-migration and it was seemless | 12:56 |
| *** dolpher has quit IRC | 12:58 | |
| *** dims has joined #kolla | 13:01 | |
| *** jruano has joined #kolla | 13:02 | |
| *** shardy_ has joined #kolla | 13:08 | |
| *** shardy has quit IRC | 13:09 | |
| *** jpeeler has quit IRC | 13:10 | |
| *** jpeeler has joined #kolla | 13:13 | |
| *** jpeeler has joined #kolla | 13:13 | |
| *** shardy_ has quit IRC | 13:13 | |
| *** shardy has joined #kolla | 13:14 | |
| *** dolpher has joined #kolla | 13:19 | |
| *** tfukushima has joined #kolla | 13:22 | |
| *** dolpher has quit IRC | 13:25 | |
| inc0 | SamYaple, cool | 13:25 |
| inc0 | 1gbps is nice io | 13:26 |
| *** banix has joined #kolla | 13:33 | |
| SamYaple | inc0: you mean 1gB yes? | 13:34 |
| inc0 | even 1GBps | 13:34 |
| *** klint has quit IRC | 13:36 | |
| SamYaple | even if in normal usage it sustains 300MB in the instance thats great | 13:36 |
| SamYaple | but this wasnt even bursty so it was pretty cool | 13:36 |
| *** dwalsh has quit IRC | 13:40 | |
| *** dwalsh_ has quit IRC | 13:40 | |
| *** banix has quit IRC | 13:41 | |
| *** jtriley has joined #kolla | 13:45 | |
| *** jmccarthy has quit IRC | 13:48 | |
| inc0 | I'm going off - tomorrow I'm giving it a try myself;) | 13:51 |
| inc0 | I have pretty good hardware | 13:51 |
| inc0 | quantas storage servers with 24 intel ssd | 13:51 |
| inc0 | will be fun-fun | 13:51 |
| *** jmccarthy has joined #kolla | 13:51 | |
| *** jpeeler has quit IRC | 13:51 | |
| *** jpeeler has joined #kolla | 13:52 | |
| *** dwalsh has joined #kolla | 13:53 | |
| *** jruano has quit IRC | 13:53 | |
| *** dwalsh_ has joined #kolla | 13:53 | |
| *** dolpher has joined #kolla | 14:01 | |
| *** itsuugo has joined #kolla | 14:05 | |
| *** achanda has joined #kolla | 14:15 | |
| *** marun has joined #kolla | 14:15 | |
| *** marun has quit IRC | 14:20 | |
| *** achanda has quit IRC | 14:22 | |
| *** dolpher has quit IRC | 14:24 | |
| *** inc0 has quit IRC | 14:29 | |
| *** banix has joined #kolla | 14:45 | |
| *** achanda has joined #kolla | 14:51 | |
| *** jasonamyers has joined #kolla | 14:52 | |
| *** jtriley has quit IRC | 14:56 | |
| *** achanda has quit IRC | 14:59 | |
| *** dims has quit IRC | 15:04 | |
| *** jtriley has joined #kolla | 15:06 | |
| *** inc0 has joined #kolla | 15:06 | |
| *** CBR09 has joined #kolla | 15:14 | |
| *** dims__ has joined #kolla | 15:16 | |
| *** inc0 has quit IRC | 15:16 | |
| *** dwalsh has quit IRC | 15:21 | |
| *** dwalsh_ has quit IRC | 15:21 | |
| *** blahRus has joined #kolla | 15:29 | |
| *** dwalsh has joined #kolla | 15:36 | |
| *** dwalsh_ has joined #kolla | 15:37 | |
| *** dwalsh_ has quit IRC | 15:53 | |
| *** dwalsh has quit IRC | 15:53 | |
| pbourke_ | would anyone be able to comment on https://bugs.launchpad.net/bugs/1500858 ? | 15:54 |
| openstack | Launchpad bug 1500858 in kolla "Config changes do not cause a container restart when using COPY_ALWAYS" [Undecided,New] | 15:54 |
| pbourke_ | mainly whether it's intended behaviour or not | 15:54 |
| rhallisey | not sure | 15:55 |
| *** shardy_ has joined #kolla | 15:58 | |
| *** diga has joined #kolla | 15:59 | |
| *** shardy has quit IRC | 16:00 | |
| *** marun has joined #kolla | 16:01 | |
| *** itsuugo has quit IRC | 16:03 | |
| *** shardy_ has quit IRC | 16:04 | |
| *** shardy has joined #kolla | 16:04 | |
| *** marun has quit IRC | 16:06 | |
| CBR09 | Anyone know when RC1 will be release? | 16:06 |
| *** dwalsh_ has joined #kolla | 16:07 | |
| *** dwalsh has joined #kolla | 16:07 | |
| *** diga has quit IRC | 16:09 | |
| *** diga has joined #kolla | 16:10 | |
| rhallisey | CBR09, I think sdake_ is cutting it this morning | 16:11 |
| CBR09 | you mean, rc1 will be release in next few hours? | 16:15 |
| *** CBR09 has quit IRC | 16:24 | |
| *** unicell has quit IRC | 16:24 | |
| *** CBR09 has joined #kolla | 16:30 | |
| *** dims__ has quit IRC | 16:38 | |
| *** dims__ has joined #kolla | 16:39 | |
| rhallisey | CBR09, should be soon | 16:40 |
| *** jruano has joined #kolla | 16:43 | |
| *** inc0 has joined #kolla | 16:57 | |
| *** unicell has joined #kolla | 17:00 | |
| *** CBR09 has quit IRC | 17:01 | |
| *** jruano has quit IRC | 17:03 | |
| *** diga has quit IRC | 17:04 | |
| inc0 | good evening | 17:06 |
| *** marun has joined #kolla | 17:17 | |
| *** jasonamyers has quit IRC | 17:19 | |
| *** jasonamyers has joined #kolla | 17:20 | |
| *** marun has quit IRC | 17:22 | |
| rhallisey | inc0, hey | 17:22 |
| inc0 | wassup? | 17:22 |
| rhallisey | nothing | 17:25 |
| rhallisey | what's up with you? | 17:25 |
| *** itsuugo has joined #kolla | 17:26 | |
| inc0 | playing with ansible, I'm going to deploy ceph tomorrow | 17:26 |
| inc0 | and play with it | 17:26 |
| inc0 | on real hardware, quality stor nodes | 17:26 |
| *** athomas has quit IRC | 17:27 | |
| *** itsuugo has quit IRC | 17:27 | |
| rhallisey | nice dude | 17:27 |
| *** achanda has joined #kolla | 17:29 | |
| *** jruano has joined #kolla | 17:32 | |
| *** coolsvap has quit IRC | 17:33 | |
| *** coolsvap has joined #kolla | 17:37 | |
| *** marun has joined #kolla | 17:38 | |
| *** itsuugo has joined #kolla | 17:39 | |
| openstackgerrit | Hui Kang proposed openstack/kolla: Add consul containers for dynamic configuration and discovery. This implementation https://review.openstack.org/228719 | 17:40 |
| *** exploreshaifali has joined #kolla | 17:40 | |
| *** exploreshaifali has quit IRC | 17:41 | |
| *** loth has joined #kolla | 17:57 | |
| *** dims has joined #kolla | 18:00 | |
| *** dims__ has quit IRC | 18:02 | |
| *** dwalsh_ has quit IRC | 18:11 | |
| *** dwalsh_ has joined #kolla | 18:11 | |
| *** jasonamyers has quit IRC | 18:18 | |
| *** itsuugo has quit IRC | 18:26 | |
| *** itsuugo has joined #kolla | 19:10 | |
| *** itsuugo has quit IRC | 19:27 | |
| *** itsuugo has joined #kolla | 19:28 | |
| inc0 | and I thought it's louder here during US day;) | 19:37 |
| rhallisey | inc0, it usually is | 19:39 |
| rhallisey | are you in texas now? | 19:39 |
| inc0 | no, I'm moving about mid Nov | 19:39 |
| inc0 | after Tokyo | 19:39 |
| rhallisey | cool | 19:39 |
| inc0 | I'm just playing with ansible before sleep, so my dreams will be even more psychodelic | 19:40 |
| *** inc0 has quit IRC | 19:50 | |
| *** jruano has quit IRC | 19:55 | |
| *** dims has quit IRC | 20:28 | |
| *** dims has joined #kolla | 20:29 | |
| *** jruano has joined #kolla | 20:42 | |
| *** tpot has joined #kolla | 20:49 | |
| *** tpot has quit IRC | 20:50 | |
| *** tpot has joined #kolla | 20:50 | |
| *** rhallisey has quit IRC | 20:57 | |
| *** jtriley has quit IRC | 21:05 | |
| *** dwalsh has quit IRC | 21:12 | |
| *** dwalsh_ has quit IRC | 21:13 | |
| *** itsuugo has quit IRC | 21:13 | |
| *** tfukushima has quit IRC | 21:21 | |
| *** shardy has quit IRC | 21:25 | |
| *** jtriley has joined #kolla | 21:28 | |
| *** jruano has quit IRC | 21:29 | |
| *** jtriley has quit IRC | 21:35 | |
| *** devlaps has joined #kolla | 21:52 | |
| *** banix has quit IRC | 22:37 | |
| *** dims_ has joined #kolla | 23:19 | |
| *** dims has quit IRC | 23:22 | |
| *** banix has joined #kolla | 23:28 | |
| *** sdake_ has quit IRC | 23:30 | |
| *** sdake has joined #kolla | 23:30 | |
| *** blahRus has quit IRC | 23:30 | |
| *** dims_ has quit IRC | 23:31 | |
| *** dims_ has joined #kolla | 23:32 | |
| *** maximedev has quit IRC | 23:32 | |
| *** jmccarthy has quit IRC | 23:33 | |
| *** maximedev has joined #kolla | 23:33 | |
| *** jmccarthy has joined #kolla | 23:34 | |
| *** sdake has quit IRC | 23:44 | |
| *** sdake_ has joined #kolla | 23:44 | |
| *** achanda_ has joined #kolla | 23:44 | |
| *** achanda_ has quit IRC | 23:45 | |
| sdake_ | rc1 is tagged and in the wild :) | 23:45 |
| *** sdake_ is now known as sdake | 23:45 | |
| *** achanda has quit IRC | 23:48 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
