sdake | centos7:latest has not changed in 5 weeks | 00:01 |
---|---|---|
sdake | bmace you said latest worked for you on your ubuntu vm? | 00:05 |
bmace | well, it worked better, to the point that i had horizon running and was able to almost configure vms.. but that was maybe 3 weeks ago at this point | 00:06 |
bmace | i don't think it was as far back as 5 | 00:06 |
sdake | i see, so totally different bits | 00:07 |
bmace | i'm trying to get things set up on a different ubuntu box i have here. | 00:07 |
bmace | see if it breaks the same way | 00:07 |
bmace | and it is a bare-metal box so when i want to do the ol work it will compile faster, etc, than if i did all the work in a vm | 00:08 |
*** jtriley has joined #kolla | 00:11 | |
*** jtriley has quit IRC | 00:15 | |
*** sdake_ has joined #kolla | 00:45 | |
*** sdake has quit IRC | 00:49 | |
*** zhiwei has joined #kolla | 01:00 | |
*** sdake has joined #kolla | 01:33 | |
*** sdake_ has quit IRC | 01:37 | |
bmace | sdake fyi, re-build a mint / ubuntu environment and got to the same point, so going to try to rebuild the images rather than pulling them from docker-hub and see if i can work through the issue | 01:38 |
sdake | bmace I got interrupted by two other tasks | 01:40 |
sdake | but i'm back with you now | 01:40 |
sdake | the build has finished on my end | 01:40 |
sdake | preparing to debug | 01:40 |
bmace | i'm gone for a few, supper and such. i expect i'll be back a bit later. | 01:40 |
*** erkules_ has joined #kolla | 01:40 | |
bmace | will be happy to help in any way i can on my return :) | 01:40 |
*** erkules has quit IRC | 01:43 | |
*** sdake has quit IRC | 01:48 | |
*** erkules_ has quit IRC | 01:50 | |
*** erkules_ has joined #kolla | 01:51 | |
*** sdake has joined #kolla | 01:52 | |
*** tobe has joined #kolla | 02:03 | |
openstackgerrit | weiyu proposed stackforge/kolla: ansible libary passing error arguments detach and recreate to project.up(). https://review.openstack.org/187136 | 02:36 |
*** dims___ has quit IRC | 02:48 | |
SamYaple | morning | 02:55 |
bmace | good "morning" sam :) | 02:55 |
*** bmace has quit IRC | 03:10 | |
openstackgerrit | Steven Dake proposed stackforge/kolla: Make crux compatible with delorean packaging https://review.openstack.org/187442 | 03:19 |
sdake | bmace enjoy ^ | 03:19 |
SamYaple | sdake: i tlaked to larsks about crux packaging for debian. he brough up the fact it is already pip installable from the git repo so thats good enough from me | 03:20 |
SamYaple | he mentioned he may sackage it for PyPI but i told him not to worry about debian/ubuntu | 03:21 |
SamYaple | package* | 03:21 |
sdake | I love that our CI is actually doing something now | 03:28 |
sdake | amazing | 03:29 |
*** bmace has joined #kolla | 03:41 | |
*** dims_ has joined #kolla | 03:48 | |
*** dims_ has quit IRC | 03:53 | |
*** ryansb has quit IRC | 04:41 | |
*** ryansb has joined #kolla | 04:49 | |
*** ryansb has quit IRC | 04:49 | |
*** ryansb has joined #kolla | 04:49 | |
*** sdake has quit IRC | 05:07 | |
*** tobe has quit IRC | 05:08 | |
*** ryansb_ has joined #kolla | 05:22 | |
*** ryansb_ has quit IRC | 05:22 | |
*** ryansb_ has joined #kolla | 05:22 | |
*** ryansb has quit IRC | 05:24 | |
*** ryansb_ is now known as ryansb | 05:24 | |
*** tobe has joined #kolla | 05:37 | |
*** loth has quit IRC | 05:55 | |
*** loth has joined #kolla | 06:08 | |
*** tobe has quit IRC | 06:13 | |
*** gfidente has quit IRC | 06:14 | |
openstackgerrit | Martin André proposed stackforge/kolla: Removes Host Cleanup from Container Cleanup Script https://review.openstack.org/182493 | 06:32 |
*** dims_ has joined #kolla | 06:37 | |
*** tobe has joined #kolla | 06:38 | |
*** dims_ has quit IRC | 06:42 | |
openstackgerrit | Martin André proposed stackforge/kolla: Add docker-compose-all wrapper script https://review.openstack.org/177095 | 06:43 |
openstackgerrit | Martin André proposed stackforge/kolla: Remove tools/cleanup-containers https://review.openstack.org/174193 | 06:43 |
*** inc0 has joined #kolla | 06:56 | |
*** athomas has joined #kolla | 07:18 | |
inc0 | howdy folks?:) | 08:14 |
mandre | hi there | 08:17 |
mandre | I just noticed the functional gate lets broken build script links pass, bug reported here https://bugs.launchpad.net/kolla/+bug/1460959 | 08:20 |
openstack | Launchpad bug 1460959 in kolla "Functional gate should fail on broken build link" [Critical,New] | 08:20 |
mandre | jpeeler ^ | 08:20 |
*** erkules_ is now known as erkules | 08:29 | |
*** erkules has quit IRC | 08:29 | |
*** erkules has joined #kolla | 08:29 | |
*** dims_ has joined #kolla | 09:26 | |
*** dims_ has quit IRC | 09:31 | |
openstackgerrit | weiyu proposed stackforge/kolla: ansible libary passing error arguments detach and recreate to project.up(). https://review.openstack.org/187505 | 09:37 |
openstackgerrit | weiyu proposed stackforge/kolla: ansible libary passing error arguments detach and recreate to project.up(). https://review.openstack.org/187136 | 09:41 |
*** pdb has quit IRC | 09:41 | |
*** pdb has joined #kolla | 09:42 | |
*** weiyu has joined #kolla | 09:49 | |
*** zhiwei has quit IRC | 09:57 | |
*** mstachow has joined #kolla | 10:25 | |
mstachow | morning :) | 10:37 |
SamYaple | morning mstachow | 10:39 |
mstachow | How are You today :D ? | 10:39 |
SamYaple | not great tbh | 10:45 |
SamYaple | my servers mainboard is no more | 10:45 |
openstackgerrit | weiyu proposed stackforge/kolla: fix ansible libary passing error arguments. https://review.openstack.org/187136 | 10:49 |
mstachow | that's sad to hear that | 10:49 |
mstachow | I'm in hope that You'll get another one soon | 10:49 |
weiyu | hi Sam Yaple,I fix my commit message | 10:50 |
SamYaple | weiyu: i already approved it :) | 10:50 |
weiyu | thanks | 10:50 |
SamYaple | mstachow: yea its a dual socket board and ive been having asome issues with numa stuff lately so i just bought 2 single sockets boards to replace it | 10:51 |
SamYaple | sholud be here tomorrow | 10:51 |
SamYaple | mandre: can we get https://review.openstack.org/#/c/187136/ merged in? | 10:54 |
SamYaple | needs another +2 | 10:54 |
*** dims_ has joined #kolla | 10:55 | |
*** inc0 has quit IRC | 11:02 | |
*** inc0 has joined #kolla | 11:05 | |
*** inc0 has quit IRC | 11:13 | |
*** inc0 has joined #kolla | 11:13 | |
*** weiyu has quit IRC | 11:14 | |
*** rhallisey has joined #kolla | 11:24 | |
*** ChanServ sets mode: +o rhallisey | 11:35 | |
*** sdake has joined #kolla | 11:49 | |
*** dwalsh has joined #kolla | 11:50 | |
*** inc0 has quit IRC | 11:50 | |
sdake | morning | 11:52 |
*** sdake_ has joined #kolla | 11:56 | |
*** sdake has quit IRC | 11:57 | |
rhallisey | morning | 12:02 |
*** tobe has quit IRC | 12:07 | |
*** sdake has joined #kolla | 12:15 | |
*** sdake has quit IRC | 12:16 | |
*** sdake_ has quit IRC | 12:17 | |
*** dwalsh_ has joined #kolla | 12:24 | |
*** inc0 has joined #kolla | 12:26 | |
SamYaple | morning rhallisey sdake | 12:27 |
*** sdake has joined #kolla | 12:32 | |
*** shardy_ has joined #kolla | 12:32 | |
*** sdake has quit IRC | 12:34 | |
*** shardy has quit IRC | 12:34 | |
openstackgerrit | Merged stackforge/kolla: fix ansible libary passing error arguments. https://review.openstack.org/187136 | 12:36 |
*** shardy_ has quit IRC | 12:38 | |
*** shardy has joined #kolla | 12:39 | |
mandre | hey SamYaple, weiyu's just got his first patch merged! | 12:39 |
mandre | the gate failed again with the same error that Steve reported on the ML | 12:40 |
*** sdake has joined #kolla | 12:40 | |
*** athomas has quit IRC | 13:04 | |
*** sdake_ has joined #kolla | 13:09 | |
*** sdake has quit IRC | 13:13 | |
*** sdake has joined #kolla | 13:18 | |
*** loth has quit IRC | 13:20 | |
*** sdake_ has quit IRC | 13:21 | |
*** loth has joined #kolla | 13:23 | |
*** loth has quit IRC | 13:25 | |
*** athomas has joined #kolla | 13:26 | |
*** loth has joined #kolla | 13:38 | |
*** jtriley has joined #kolla | 13:49 | |
*** dims_ has quit IRC | 13:51 | |
*** dims_ has joined #kolla | 13:51 | |
inc0 | sdake, hello, https://blueprints.launchpad.net/kolla/+spec/install-from-source this bp is about installing openstack from trunk? | 13:51 |
sdake | inc0 or stable branches more likely | 13:52 |
inc0 | pip maybe? | 13:52 |
inc0 | we should make branch configurable if we want to use this | 13:52 |
sdake | samyaple is the man to talk to on that blueprint | 13:53 |
inc0 | SamYaple, still here?:) | 13:53 |
jpeeler | as long as the branch is configurable, could set default to stable branch, but allow building from master. that would rock | 13:57 |
inc0 | that's the plan | 13:57 |
inc0 | we'll talk with SamYaple tomorrow (I guess he's already off) and ask about status | 13:58 |
jpeeler | mandre: i'm looking at the gate build links issue. but i don't know how the underlying bash scripts are reporting up to date if the build wasn't even attempted | 13:58 |
*** sdake_ has joined #kolla | 14:00 | |
sdake_ | ideally we would gate on master as well as stable branches jpeeler | 14:01 |
sdake_ | but lets focus on master gating and get that right first | 14:01 |
sdake_ | jpeeler re gate build links issue - got a bug link? | 14:02 |
jpeeler | i was talking about something else in my earlier comment. anyway, here's the bug https://bugs.launchpad.net/kolla/+bug/1460959 | 14:02 |
openstack | Launchpad bug 1460959 in kolla "Functional gate should fail on broken build link" [Critical,New] | 14:02 |
sdake_ | we probably need a whitelist of what should be built | 14:03 |
sdake_ | taht would fix this problem | 14:03 |
*** sdake has quit IRC | 14:03 | |
jpeeler | what? | 14:04 |
sdake_ | i prefer your blacklist approach though (what doesn't build) | 14:04 |
sdake_ | since it lowers gate maintenance | 14:04 |
sdake_ | so if you can sort out how to fix the actual bug in the build that would be ideal :) | 14:04 |
jpeeler | the problem has to be in the bash docker build scripts, the python test just verifies the output. but i'll investigate | 14:05 |
sdake_ | well the bash docker build script returns failure if a build fails | 14:05 |
jpeeler | it didn't though, that's the bug | 14:05 |
sdake_ | the master build script will print out a list of pas/fail builds | 14:05 |
sdake_ | i see | 14:05 |
sdake_ | well those scripts are black magic, but mandre isnt' around at this time of day | 14:06 |
sdake_ | you guys are 12 hours apart on tzs i think :( | 14:06 |
jpeeler | yeah i still don't have a good feel for where people are yet | 14:06 |
sdake_ | i can catch him late at night | 14:07 |
*** sdake_ has quit IRC | 14:07 | |
jpeeler | where's daneyon these days? | 14:08 |
*** sdake has joined #kolla | 14:08 | |
*** mstachow has quit IRC | 14:24 | |
*** v_b has quit IRC | 14:41 | |
*** v_b has joined #kolla | 14:42 | |
*** jruano has joined #kolla | 14:46 | |
*** athomas has quit IRC | 14:52 | |
*** v_b has left #kolla | 14:52 | |
*** vbelogrudov has joined #kolla | 14:53 | |
*** blahRus has joined #kolla | 14:53 | |
*** vbelogrudov has quit IRC | 14:57 | |
*** vbel has joined #kolla | 14:57 | |
*** bmace_ has joined #kolla | 15:14 | |
*** bmace has quit IRC | 15:17 | |
SamYaple | inc0: yes | 15:20 |
inc0 | afternoon:) akwasnie and me are looking at this bp https://blueprints.launchpad.net/kolla/+spec/install-from-source | 15:21 |
inc0 | can we take on this one? | 15:22 |
SamYaple | inc0: yea implementation just needs to be discussed | 15:23 |
SamYaple | I have a way I do it in yaodu, but it may not be prefer here. i just have a few "requirements" on what the end result would be | 15:23 |
*** vinkman has joined #kolla | 15:23 | |
inc0 | I think Ala is out, we were thinking about making a script install.sh | 15:24 |
inc0 | and just pull git branch based on envs and then do pip install -e . | 15:25 |
SamYaple | yes, let me post the yaodu build dockerfile, thats basically what i do | 15:25 |
SamYaple | one moment | 15:25 |
SamYaple | https://github.com/SamYaple/yaodu/blob/master/ansible/roles/docker_build/templates/ubuntu/keystone/Dockerfile.j2 | 15:26 |
sdake | maybe someday it makes sense to build with ansible, but lets not tackle that today pls :) | 15:26 |
sdake | I have enough ulcers ;) | 15:26 |
SamYaple | sdake: the build would be in the container | 15:26 |
SamYaple | the source build | 15:27 |
sdake | I get what your saying, lets just not do ansible build atm, our build scripts work and we are working on getting the gating giong, and I don't want to disrupt the gating which is our highest priority atm :) | 15:27 |
SamYaple | sdake: you dont get what im saying then | 15:27 |
SamYaple | different build | 15:28 |
SamYaple | this is source container talk | 15:28 |
SamYaple | inc0: basically a tarball is loaded into the container, i update the pip req based on latest from global-req stable branch and pip installed | 15:28 |
SamYaple | tarball being the source of course | 15:28 |
SamYaple | in reality the install can be down in about 3 lines without introducing git as a dep into the container | 15:29 |
SamYaple | s/down/done/ | 15:29 |
SamYaple | what I would really like to be able to do is point to pull in ref_changes from a patch as well | 15:30 |
SamYaple | that would be awesome | 15:30 |
SamYaple | would involve git though | 15:30 |
SamYaple | sdake: my mainboard blew up so my server is down down :( | 15:31 |
inc0 | yeah, I don't think git dependency is that much of an issue | 15:31 |
SamYaple | i dont either inc0, my phrasing was poor | 15:31 |
inc0 | and this will enable whole lot of flexibility | 15:31 |
sdake | the problem with a git dependency is that many private clouds don't have external connectivty | 15:31 |
SamYaple | thats why i like the tarball sdake | 15:32 |
sdake | agree | 15:32 |
SamYaple | we can combine both approaches by git cloning before building the container | 15:32 |
SamYaple | then producing a tarball to consume in the build | 15:32 |
SamYaple | this would allow refchanges as well | 15:32 |
sdake | that wfm, build should have external connectivity hopefully ;) | 15:32 |
sdake | although I don't know if that is actually the case | 15:32 |
inc0 | I was thinking about installing on build, nto after deploy | 15:33 |
SamYaple | yea thats what were talking about inc0 | 15:33 |
SamYaple | fir the build script for source can be configured (and it can without modification) to pull down the appropriate branch and checkout ref_changes then we just generate a tarball locally before building the container | 15:35 |
SamYaple | spit it out and consume the tarball while building the container | 15:35 |
vbel | git is elegant solution though - "git clone && checkout && pip install -r requirements && python setup.py install" just works | 15:35 |
inc0 | but compute nodes has to be able to do git clone, and that's an issue | 15:35 |
sdake | morning vbel | 15:35 |
SamYaple | inc0: why? | 15:35 |
vbel | good morning :) | 15:35 |
SamYaple | all git cloning would be done before container building | 15:36 |
inc0 | that was comment to vbel's comment;) | 15:36 |
*** barra204 has joined #kolla | 15:36 | |
SamYaple | ah :) | 15:36 |
*** shakamunyi has joined #kolla | 15:36 | |
SamYaple | I want to talk about the pip requirements | 15:37 |
SamYaple | does anyone have a reasonable objection to updating the req based on the head for the stable branch matching the build verison? | 15:38 |
SamYaple | https://github.com/openstack/requirements/blob/master/global-requirements.txt | 15:38 |
SamYaple | thos reqs only get more strict as problems occur | 15:38 |
vbel | SamYaple: why not in Dockerfile? It allows caching if you chain commands like this. Build requires connectivity anyway for pypi | 15:38 |
SamYaple | vbel: then the dependancy is inside the container and you have to pass variables into the dockerfile at build all for information your dont need at runtime | 15:39 |
*** openstackgerrit has quit IRC | 15:42 | |
*** openstackgerrit has joined #kolla | 15:43 | |
*** vinkman has quit IRC | 15:44 | |
*** vinkman has joined #kolla | 15:44 | |
inc0 | sdake, would you kindly assign me to this one? https://blueprints.launchpad.net/kolla/+spec/install-from-source | 15:44 |
inc0 | SamYaple, there is another idea - we can install it in virtualenv and tarball whole env | 15:45 |
*** absubram has joined #kolla | 15:45 | |
inc0 | so we won't hit dependency problem after deployment as well | 15:46 |
sdake | inc0 whats ure launchpad id | 15:46 |
sdake | inc0? | 15:46 |
inc0 | yes | 15:46 |
inc0 | ah sorry | 15:46 |
inc0 | inc007 | 15:46 |
sdake | looks like you already are:) | 15:46 |
inc0 | allright, thank you:) | 15:47 |
SamYaple | inc0: but then the container grows really large from doing the temp work | 15:47 |
sdake | agree we want small containers ;-) | 15:47 |
inc0 | well, I don't want to make architectural decissions at 6pm;) | 15:49 |
inc0 | I'll look at it tomorrow | 15:49 |
SamYaple | inc0: i dont want to make them at 11AM | 15:49 |
SamYaple | ive been up all night | 15:49 |
sdake | vbel are you thinking of contributing to kola? | 15:49 |
sdake | samyaple 36 hours a day :) | 15:50 |
SamYaple | 25/8 | 15:50 |
inc0 | allright, I'm going home | 15:50 |
inc0 | gn guys | 15:50 |
inc0 | get some sleep SamYaple | 15:50 |
inc0 | cya all tomorrow | 15:50 |
*** sdake_ has joined #kolla | 15:52 | |
vbel | sdake: yes :) | 15:52 |
sdake_ | vbel what are you interested in working on | 15:54 |
sdake_ | we have two efforts underway - container content and ansible deployment | 15:54 |
*** inc0 has quit IRC | 15:55 | |
sdake_ | and a third effort actually, which is gating | 15:55 |
sdake_ | jpeeler is leading the gating effort | 15:55 |
*** athomas has joined #kolla | 15:55 | |
sdake_ | I lead the container content effort | 15:55 |
sdake_ | and myself and SamYaple lead the ansible effort | 15:55 |
*** sdake has quit IRC | 15:56 | |
jpeeler | sdake_: just FYI, i'm sort of blocked on setup issues ATM. i wanted a modern setup and haven't gotten everything working yet | 15:56 |
sdake_ | jpeeler ack on the blocking | 15:57 |
sdake_ | samyaple is working on the single nic solution for kolla in his 36th hour of the day:) | 15:57 |
sdake_ | are you blocked in other ways? | 15:57 |
jpeeler | i migrated to RDO on F22 and haven't gotten networking working yet if that's what you're asking | 15:59 |
jpeeler | if the single nic solution were to come along, i could just forget using openstack | 15:59 |
sdake_ | ya yum upgrade should have a double check option :) | 16:00 |
SamYaple | jpeeler: i was testing it with subinterfaces before my box blew up | 16:00 |
SamYaple | rush ordered replacement parts but they wont all be in until friday | 16:00 |
jpeeler | SamYaple: yeah you said it would be some time, no worries | 16:00 |
SamYaple | we should be able to do single nic without vxlan or vlan interfaces | 16:00 |
sdake_ | I tried to do the job with bridges | 16:01 |
sdake_ | and spent 7 days on it and couldn't get it to work | 16:01 |
sdake_ | hate to admit it, but finally gave up:( | 16:01 |
SamYaple | bridges would require iptables rules :[] | 16:01 |
SamYaple | thats an unhappy face, btw | 16:02 |
sdake_ | here was my work: http://ur1.ca/mpsmv -> http://paste.fedoraproject.org/228130/14332609 | 16:02 |
sdake_ | no iptables rules there | 16:03 |
sdake_ | vbel hopefully I didn't scare you away ;) - if you just want to dig in at your own pace that works for me :) | 16:07 |
sdake_ | vbel if you need pointers on where to contribute feel free to ping me | 16:07 |
SamYaple | sdake_: was talking with vbel in a private channel clarifying some source install things | 16:08 |
sdake_ | oh | 16:08 |
sdake_ | ok - well have those in channel :) | 16:08 |
sdake_ | so we can all learn new stuffz :) | 16:08 |
sdake_ | i learned a whole bunch about source install via your discussion with inc0 | 16:08 |
vbel | sdake: sorry, talking to your colleages :) I like containers in general, keeaplived/haproxy stuff, ndb / mysql cluster, ... many things that I still can understand :) | 16:09 |
sdake_ | cool well there is a mountain of work to do :) | 16:09 |
sdake_ | we definately need a keepalived and haproxy container | 16:09 |
sdake_ | we have an haproxy container already but I dont think it really does the job | 16:10 |
sdake_ | what is ndb? | 16:10 |
SamYaple | unfortunately there is no api to update the haproxy configs on the fly :( on a nonpersistent socket that can do _some_ things | 16:10 |
SamYaple | any changes require a container restart | 16:10 |
SamYaple | (ie adding new services) | 16:10 |
SamYaple | or hosts | 16:10 |
SamYaple | or removing | 16:10 |
SamYaple | really any changes. | 16:11 |
vbel | ndb cluster is engine behind mysql cluster | 16:11 |
sdake_ | hey sam how do you do orchestration in ansible for things that require multiple steps for our ha containers like mariadb | 16:11 |
SamYaple | ansible has ways to run commands on single hosts and run commands once (but on any host) | 16:11 |
sdake_ | ndb so its like galera? | 16:11 |
SamYaple | i determine if the cluster exists and then check if everyone is up and joined in the cluster (create it if it doesnt exist) | 16:12 |
vbel | sdake_: not for self-advertisement - https://www.youtube.com/watch?v=RJf7cwkytOE | 16:12 |
sdake_ | vbel watching - be back in 10 mins :) | 16:13 |
sdake_ | vbel ahh I noticed you commented on the HA spec | 16:16 |
sdake_ | surprising how many ha rocket scientists we have in kolla community :) | 16:16 |
SamYaple | ha/horizontal scaling is the only reason im doing openstack | 16:17 |
vbel | sdake_: :)))) | 16:17 |
SamYaple | i like it alot | 16:17 |
sdake_ | vbel inc0 and danehans are the other rocket scientists | 16:18 |
sdake_ | I wrote corosync | 16:18 |
sdake_ | so I guess I'm a rocket scientist too | 16:18 |
sdake_ | but that is in the past :) | 16:18 |
*** rhallisey has quit IRC | 16:19 | |
sdake_ | sed through all sources - groan :) | 16:20 |
vbel | despair :) | 16:23 |
sdake_ | nice job on the presentation | 16:23 |
sdake_ | was that at vancouver summit? | 16:23 |
sdake_ | multiple db backends wfm | 16:24 |
sdake_ | that should be easily configurable | 16:24 |
vbel | yes, so called vBrownBag in 106 room, 10 minutes per presentation | 16:24 |
SamYaple | well do all the things! | 16:24 |
sdake_ | we already have config options for network manager | 16:24 |
sdake_ | we will do the things that matter, but what matters to me is different then what matters to others in the community ;) | 16:25 |
SamYaple | once kolla gets solid footing pumping out contaienrs will be half a days work at most | 16:25 |
SamYaple | *new* containers | 16:25 |
sdake_ | we are already there on the containers but packaging the big tent in rdo is a big job | 16:25 |
sdake_ | and I don't think we want to be in that business | 16:25 |
sdake_ | I did it for magnum because I had a demo to deliver | 16:26 |
sdake_ | I think we should leave the rdo community (of which i'm also a participant) to that job | 16:26 |
sdake_ | to do that job taht is | 16:26 |
sdake_ | (package the big tent) | 16:26 |
*** dasm|afk has quit IRC | 16:27 | |
sdake_ | I guess this will naturally lead to feature imparity between source and binary installs | 16:27 |
sdake_ | because packaging from source will be faster | 16:27 |
sdake_ | (not waiting on a dependency) | 16:27 |
sdake_ | oh well more fun to deal with ;-) | 16:28 |
vbel | source base containers would be really great for everyone | 16:28 |
SamYaple | ^^ | 16:28 |
*** sdake has joined #kolla | 16:29 | |
*** dasm|afk has joined #kolla | 16:32 | |
*** sdake__ has joined #kolla | 16:32 | |
*** sdake_ has quit IRC | 16:32 | |
*** rhallisey has joined #kolla | 16:35 | |
*** sdake has quit IRC | 16:36 | |
*** sdake has joined #kolla | 16:36 | |
*** absubram has quit IRC | 16:38 | |
*** absubram has joined #kolla | 16:38 | |
*** sdake__ has quit IRC | 16:40 | |
sdake | I have 8 hours of meetings today | 16:50 |
* sdake groans | 16:50 | |
sdake | jpeeler rhallisey its essential this is merged: https://review.openstack.org/#/c/187442/ | 16:54 |
sdake | master is broken without it | 16:54 |
rhallisey | sdake, yes that was on my to do list | 16:54 |
rhallisey | no one else was using master so I just switched back to kilo :p | 16:55 |
sdake | rhallisey what is the permissions of /dev/pts/ptmx on your cinder test system? | 16:56 |
openstackgerrit | Merged stackforge/kolla: Make crux compatible with delorean packaging https://review.openstack.org/187442 | 16:57 |
rhallisey | sdake, ya I'm fixing that | 16:57 |
sdake | ya your fixing what | 16:58 |
rhallisey | I just rebooting and /dev/pts was having issue with libvirt | 16:58 |
rhallisey | c---------. 1 root root 5, 2 Jun 2 12:56 /dev/pts/ptmx | 16:58 |
sdake | libvirt is a root process, Ithink it should be able to access that anyway correct? | 16:59 |
rhallisey | uh let me confirm | 17:00 |
harmw | what should ./tools/update-build-links do? (I'm too lazy to read it) | 17:05 |
openstackgerrit | Harm Weites proposed stackforge/kolla: Add the Designate DNSaaS services. https://review.openstack.org/182799 | 17:06 |
*** daneyon has joined #kolla | 17:06 | |
harmw | SamYaple: sdake please take a (final) look at that :p | 17:06 |
rhallisey | sdake, when you mount /dev in cinder it was causing libvirt to fail to boot instance | 17:07 |
rhallisey | chmod 666 /dev/pts/ptmx fixes that | 17:07 |
rhallisey | *in my cinder patch, /dev in libvirt is the issue | 17:14 |
*** shardy_ has joined #kolla | 17:33 | |
*** shardy has quit IRC | 17:34 | |
harmw | sdake: opinion on galera to replace mariadb? | 17:36 |
*** shardy_ has quit IRC | 17:38 | |
*** shardy has joined #kolla | 17:39 | |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add cinder into the docker compose setup https://review.openstack.org/173507 | 17:48 |
*** daneyon has quit IRC | 17:52 | |
*** akwasnie has quit IRC | 17:52 | |
sdake | harmw no strong feelings as long as galera runs single node | 17:56 |
*** daneyon has joined #kolla | 17:56 | |
sdake | i.e. supports an AIO deploy model without alot of hassle | 17:57 |
harmw | yep, afaik it does | 17:57 |
sdake | sounds like the bindmount of d/ev doesn't quite work as expected rhallisey | 17:57 |
rhallisey | ya | 17:57 |
sdake | i am concerned about making /dev/pts/ptmx 666 on the host for security reasons tho | 17:57 |
sdake | harmw that script updates all the ./build files in the tree to the correct build script | 17:58 |
*** bmace_ has quit IRC | 17:58 | |
rhallisey | have any ideas how to get around that | 17:58 |
sdake | try mode 600 ? | 17:59 |
*** bmace has joined #kolla | 17:59 | |
sdake | on your host what does it start out as (/dev/pts/ptmx) | 17:59 |
bmace | looks like char file set to 000 on mint | 18:01 |
rhallisey | sdake, 000 | 18:01 |
sdake | wonder why libvirt runs without /dev bindmount :) | 18:02 |
*** athomas has quit IRC | 18:04 | |
*** daneyon has quit IRC | 18:06 | |
Slower | sdake: so something in the cinder container is changing the perms on /dev/pts/ptmx? | 18:06 |
sdake | slower no idea what is changing it or what the default is on a fresh system but rhallisey says 000 | 18:09 |
Slower | sdake: I'm seeing 000 too | 18:09 |
*** openstackgerrit has quit IRC | 18:09 | |
Slower | c---------. 1 root root 5, 2 May 30 13:00 ptmx | 18:10 |
*** openstackgerrit has joined #kolla | 18:10 | |
Slower | no selinux? | 18:10 |
rhallisey | yea, no selinux | 18:10 |
harmw | brr, to lazy to read but perhaps this is useful: https://www.kernel.org/doc/Documentation/filesystems/devpts.txt | 18:17 |
Slower | sdake: you shouldn't need /dev/ bind mounted normally, it's just that we need the dynamic devices | 18:17 |
Slower | I almost wonder if we shouldn't be using device-add to dynamically add the new devices to the container | 18:17 |
sdake | bindmounting dev is the way to go | 18:18 |
sdake | i just don't understand why bindmounting results in an eaccess | 18:18 |
Slower | me neither | 18:18 |
Slower | and it's messing up the host too | 18:18 |
Slower | I can't start a new tmux session after that | 18:18 |
Slower | or even a new screen in tmux | 18:18 |
Slower | seems like the container needs to mount it's own pts? | 18:19 |
sdake | does it without cinder? | 18:20 |
sdake | maybe that is a better solution | 18:20 |
sdake | mount -t devpts /dev/pts | 18:20 |
Slower | ya | 18:20 |
rhallisey | lets see.. | 18:23 |
Slower | $ ls -ld /dev/pts │···················· | 18:25 |
Slower | drwxr-xr-x. 2 root root 0 Feb 12 00:19 /dev/pts | 18:25 |
Slower | I think the pts dir is getting changed | 18:25 |
Slower | $ ls -ld pts | 18:25 |
Slower | d---------. 2 root root 0 Jun 2 08:23 pts | 18:25 |
*** vinkman has quit IRC | 18:25 | |
Slower | maybe.. | 18:26 |
Slower | hmm | 18:30 |
Slower | ugly | 18:30 |
*** vinkman has joined #kolla | 18:34 | |
*** vinkman has quit IRC | 18:40 | |
sdake | sounds like a docker bug | 18:41 |
sdake | rhallisey go throw rocks at dwalsh plz :) | 18:41 |
rhallisey | it does | 18:41 |
*** daneyon has joined #kolla | 18:43 | |
Slower | ya it kind of seems like you need to special case /dev/pts handling when you bind mount /dev | 18:43 |
*** daneyon has quit IRC | 18:43 | |
sdake | and /dev/shm | 18:43 |
*** daneyon has joined #kolla | 18:44 | |
Slower | yeah | 18:44 |
Slower | I would guess that would just involve making sure you mounted your own copy of it | 18:45 |
Slower | which it doesn't seem to do | 18:45 |
*** openstackgerrit has quit IRC | 18:56 | |
*** openstackgerrit has joined #kolla | 18:56 | |
*** daneyon has quit IRC | 19:04 | |
*** dwalsh has quit IRC | 19:14 | |
*** dwalsh_ has quit IRC | 19:14 | |
*** daneyon has joined #kolla | 19:20 | |
*** loth has quit IRC | 19:26 | |
*** blahRus has quit IRC | 19:27 | |
*** blahRus has joined #kolla | 19:28 | |
*** loth has joined #kolla | 19:29 | |
*** dwalsh has joined #kolla | 19:39 | |
*** blahRus has quit IRC | 19:48 | |
*** blahRus has joined #kolla | 19:48 | |
*** blahRus has quit IRC | 19:54 | |
*** blahRus has joined #kolla | 19:56 | |
*** sdake_ has joined #kolla | 19:59 | |
*** sdake has quit IRC | 20:03 | |
*** openstackgerrit has quit IRC | 20:10 | |
*** openstackgerrit has joined #kolla | 20:10 | |
openstackgerrit | Ryan Hallisey proposed stackforge/kolla: Add cinder into the docker compose setup https://review.openstack.org/173507 | 20:16 |
*** apmelton_away is now known as apmelton | 20:22 | |
loth | sdake_: work is going to be here? https://github.com/stackforge/kolla/tree/master/ansible | 20:23 |
sdake_ | loth yes | 20:37 |
sdake_ | i think /ansible makes perfect sense for a tld name ;) | 20:37 |
loth | ok ill start poking around | 20:37 |
sdake_ | the code there is wrong | 20:51 |
sdake_ | check out yaodu | 20:51 |
loth | hmm, seems its not picking up my docker-compose https://pastebin.mozilla.org/8835479 | 20:51 |
sdake_ | ya its busted | 20:52 |
sdake_ | there is a review in the queue to fix it | 20:52 |
sdake_ | but we are not using docker compose with ansible | 20:52 |
sdake_ | docker compose will be a dev workflow only | 20:52 |
loth | just normal docker driver for ansible? | 20:52 |
loth | module* | 20:52 |
*** jtriley has quit IRC | 21:05 | |
*** bkolabing has joined #kolla | 21:09 | |
*** bkolabing has left #kolla | 21:09 | |
*** shardy has quit IRC | 21:13 | |
loth | sdake_: dont see an env file option here? http://docs.ansible.com/docker_module.html | 21:25 |
sdake_ | right | 21:25 |
loth | will just be one long json env var? | 21:25 |
sdake_ | unclear | 21:25 |
*** jruano has quit IRC | 21:32 | |
*** sdake has joined #kolla | 21:54 | |
loth | sdake_: "The docker module docker_env_builder is already implemented." Is that documented anywhere? | 21:55 |
sdake | its in the review | 21:56 |
sdake | sec | 21:56 |
*** daneyon has quit IRC | 21:56 | |
sdake | https://review.openstack.org/#/c/182168/ | 21:57 |
sdake | this module: | 21:57 |
sdake | https://review.openstack.org/#/c/182168/2/ansible/library/file_to_dict | 21:57 |
*** sdake_ has quit IRC | 21:58 | |
loth | i think i get the general picture now | 22:01 |
*** dwalsh has quit IRC | 22:02 | |
*** absubram has quit IRC | 22:07 | |
*** vinkman has joined #kolla | 22:15 | |
*** vinkman has quit IRC | 22:27 | |
*** dims__ has joined #kolla | 22:31 | |
*** dims_ has quit IRC | 22:34 | |
loth | sdake: it looks like yaodu doesnt use the volume containers, is kolla switching to that method? | 22:49 |
sdake | kolla is sticking to voluem containers | 22:51 |
*** apmelton is now known as apmelton_away | 22:59 | |
*** bradjones has quit IRC | 23:43 | |
*** bradjones has joined #kolla | 23:45 | |
*** diga_ has joined #kolla | 23:46 | |
*** barra204 has quit IRC | 23:57 | |
*** shakamunyi has quit IRC | 23:57 | |
*** diga_ has quit IRC | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!