| pabelanger | jeblair: Hmm, which other job do you see it on. So far I am only seeing it with tox/pre.yaml but the execption appears to move around. I noticed it on a tox/docs.yaml too | 00:01 |
|---|---|---|
| pabelanger | but, agree. maybe they are not related | 00:02 |
| jeblair | pabelanger: oh, do they always show up together? | 00:02 |
| pabelanger | I thought they did, but looking at 4899d90de7c842b0a00134a46b835da9 is is a different location | 00:03 |
| pabelanger | warning seems to always be with tox/pre.yaml | 00:03 |
| pabelanger | I'll have to play with it more | 00:05 |
| jeblair | pabelanger, mordred: i think the log error is because of this: | 00:07 |
| jeblair | 2017-06-10 00:04:02.752414 | Selecting previously unselected package mysql-server-5.7. | 00:07 |
| jeblair | (Reading database ... 26703 files and directories currently installed.) | 00:07 |
| jeblair | 2017-06-10 00:04:02.768363 | Preparing to unpack .../mysql-server-5.7_5.7.18-0ubuntu0.16.04.1_amd64.deb ... | 00:07 |
| jeblair | i'm not sure how we ended up with that line... | 00:07 |
| jeblair | possibly some \r trickery? | 00:08 |
| jeblair | yes | 00:09 |
| jeblair | 2017-06-10 00:03:59.265564 | (Reading database ... ^M .... is the actual line | 00:10 |
| pabelanger | ah, neat | 00:10 |
| jeblair | i don't see what about that could cause that error though | 00:14 |
| jeblair | this is probably about the time we need to re-enable the 'zuul-executor verbose' command. | 00:15 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Enable verbose/unverbose zuul-executor commands https://review.openstack.org/472868 | 00:20 |
| clarkb | jeblair: mordred I dropped a simple native python version of the pid discovery in the change. I have it working locally but I think it needs much more error handling | 00:34 |
| clarkb | anyways I too need to weekend so won't get that cleaned up until later | 00:34 |
| *** jamielennox is now known as jamielennox|away | 00:39 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul feature/zuulv3: executor: add support for custom ansible_port https://review.openstack.org/468710 | 01:19 |
| *** jamielennox|away is now known as jamielennox | 01:44 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Add support for custom ssh port https://review.openstack.org/468752 | 01:56 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Add webapp port and listen_address configuration https://review.openstack.org/472128 | 02:20 |
| *** jamielennox is now known as jamielennox|away | 02:36 | |
| *** jamielennox|away is now known as jamielennox | 02:40 | |
| *** bhavik1 has joined #zuul | 05:05 | |
| *** bhavik1 has quit IRC | 05:09 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Implement a static driver for Nodepool https://review.openstack.org/468624 | 06:50 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Extend Nodepool configuration syntax to support multiple drivers https://review.openstack.org/468751 | 06:50 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Collect request handling implementation in an OpenStack driver https://review.openstack.org/468750 | 06:50 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Abstract Nodepool provider management code https://review.openstack.org/468749 | 06:50 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Abstract Nodepool request handling code https://review.openstack.org/468748 | 06:50 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Implement an OpenContainer driver https://review.openstack.org/468753 | 06:50 |
| *** yolanda has quit IRC | 07:49 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Implement an OpenContainer driver https://review.openstack.org/468753 | 08:07 |
| *** tobiash has quit IRC | 08:34 | |
| *** tobiash has joined #zuul | 09:35 | |
| *** bhavik1 has joined #zuul | 09:57 | |
| openstackgerrit | Ricardo Carrillo Cruz proposed openstack-infra/nodepool feature/zuulv3: Expose key-name parameter on cloud-images https://review.openstack.org/472895 | 11:13 |
| openstackgerrit | Ricardo Carrillo Cruz proposed openstack-infra/nodepool feature/zuulv3: Expose key-name parameter on cloud-images https://review.openstack.org/472895 | 11:19 |
| rcarrillocruz | hey folks, ^ to inject keypairs on cloud images | 11:19 |
| rcarrillocruz | i tested on my rdo cloud | 11:19 |
| rcarrillocruz | pabelanger, mordred ^ | 11:19 |
| rcarrillocruz | alhtough | 11:31 |
| rcarrillocruz | wait, i think i got confused with the current nodepool docs | 11:31 |
| rcarrillocruz | the current doc point key-name to be an image-name | 11:31 |
| rcarrillocruz | image param sorry | 11:31 |
| rcarrillocruz | but code shows key-name is now a provider param | 11:32 |
| rcarrillocruz | so putting it on provider should just work | 11:32 |
| * rcarrillocruz goes check | 11:32 | |
| *** bhavik1 has quit IRC | 11:32 | |
| mordred | rcarrillocruz: yah. I believe pool-label is what you want | 12:26 |
| mordred | the image section just describes the image to create - the label describes how to boot something | 12:27 |
| rcarrillocruz | yah, just made it work | 12:27 |
| rcarrillocruz | Starting Execute cloud user/final scripts... | 12:28 |
| rcarrillocruz | ci-info: +++++++++++++++++Authorized keys from /home/ubuntu/.ssh/authorized_keys for user ubuntu++++++++++++++++++ | 12:28 |
| rcarrillocruz | ci-info: +---------+-------------------------------------------------+---------+---------------------------------+ | 12:28 |
| rcarrillocruz | ci-info: | Keytype | Fingerprint (md5) | Options | Comment | | 12:28 |
| rcarrillocruz | ci-info: +---------+-------------------------------------------------+---------+---------------------------------+ | 12:28 |
| rcarrillocruz | ci-info: | ssh-rsa | 23:c7:0b:25:fa:28:99🆎2b:77🇨🇨57:19:dd:e5:ca | - | ricardo.carrillo.cruz@gmail.com | | 12:28 |
| rcarrillocruz | ci-info: +---------+-------------------------------------------------+---------+---------------------------------+ | 12:28 |
| rcarrillocruz | the current docs confused me :/ | 12:28 |
| rcarrillocruz | sorry for the noise | 12:28 |
| mordred | :) | 12:28 |
| mordred | and \o/ | 12:28 |
| rcarrillocruz | btw, super cool this, i was very worried nodepool was just a diskimage game, thankfully pabelanger put me on track with this new pool thing | 12:29 |
| rcarrillocruz | which is not the same as the old snapshot thing, but allwos me to do cloud images | 12:29 |
| rcarrillocruz | btw, i'm going to put up a nodepool for network CI this week, will show to my peers | 12:29 |
| rcarrillocruz | prolly put a nightly job with ansible for our integration tests | 12:29 |
| rcarrillocruz | our nodepool will be made of ios, arista eos, juniper | 12:30 |
| rcarrillocruz | lol, interesting use case here | 12:30 |
| mordred | rcarrillocruz: oh - you were able to configure a pool with an existing cloud image? | 12:38 |
| mordred | rcarrillocruz: that was a thing we've talked about enabling, but I didn't think we'd actually done the work to enable that | 12:39 |
| rcarrillocruz | yeah, problem i had is that i did not know how to put key-name to injcet it | 12:39 |
| rcarrillocruz | putting it on the provider->pool->label did the trick | 12:39 |
| mordred | cool - I didn't realize it would look for an image in the cloud by name - I thought the image name was referring to an image nodepool knew about in zk | 12:39 |
| rcarrillocruz | well, that's other thing | 12:40 |
| rcarrillocruz | if i put the cloud-image by name | 12:40 |
| rcarrillocruz | i get failure | 12:40 |
| rcarrillocruz | by ID i don't | 12:40 |
| rcarrillocruz | so i guess a patch is needed here | 12:40 |
| mordred | oh. fascinating | 12:40 |
| rcarrillocruz | i'll investigate next week | 12:40 |
| mordred | yah - that's definitely a use-case people have expressed wanting | 12:40 |
| mordred | so if we're one patch away from it being a thing, that's great | 12:41 |
| rcarrillocruz | yeah, i don't think it would be further away, ID works, so... | 12:41 |
| mordred | neat | 12:41 |
| mordred | rcarrillocruz: so you put the openstack id in the diskimage field of the label? | 12:47 |
| mordred | rcarrillocruz: OH - wait - "cloud_image"is what you use on the label for images already in the cloud | 13:02 |
| mordred | rcarrillocruz: did you see the "cloud-images" section of the docs? | 13:04 |
| rcarrillocruz | Yah, cloud-image | 13:24 |
| rcarrillocruz | Now if I put the name attribute as image name | 13:24 |
| rcarrillocruz | It fails | 13:24 |
| rcarrillocruz | With id works | 13:24 |
| pabelanger | Nice, I was planning trying cloud-images, but haven't yet | 13:26 |
| rcarrillocruz | I get a shade exception about invalid field attribute xenial | 13:26 |
| rcarrillocruz | Cos well, my image is xenial | 13:26 |
| pabelanger | we likely can update nodepool dsvm job too, and use cirros image for test | 13:26 |
| rcarrillocruz | I can share my nodepool yaml | 13:26 |
| rcarrillocruz | Sec | 13:26 |
| rcarrillocruz | http://paste.openstack.org/show/612191/ | 13:27 |
| rcarrillocruz | pabelanger , mordred ^ | 13:27 |
| mordred | rcarrillocruz, pabelanger: oh! I think I see where we're missing something | 13:28 |
| rcarrillocruz | next step, upload an IOS or EOS image | 13:28 |
| rcarrillocruz | and write the ansible scaffolding to create groups per inventory | 13:29 |
| rcarrillocruz | and kick our integration runs | 13:29 |
| rcarrillocruz | it's unfortunate, but we do not run any integration testing for networking | 13:29 |
| rcarrillocruz | except for vyos | 13:29 |
| rcarrillocruz | which is why the intent of this, put up an internal CI, couple it with ARA and run it periodically | 13:30 |
| rcarrillocruz | then move on to zuul and check it out the GH support | 13:30 |
| rcarrillocruz | all this is going to be big for us, most vendors will just use Zuul as 3rd party | 13:30 |
| rcarrillocruz | cos, we can't have every possible gear they have out there or wokring on on our side | 13:31 |
| rcarrillocruz | it's exactly the same thing for neutron 3rd party and cinder 3rd party | 13:31 |
| rcarrillocruz | just for ansible network modules work | 13:31 |
| pabelanger | rcarrillocruz: nice | 13:38 |
| pabelanger | just removed zmq-publishers from ansible role, no longer used | 13:38 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool feature/zuulv3: Support booting cloud-images by name or id https://review.openstack.org/472959 | 13:53 |
| mordred | rcarrillocruz: ^^ that should fix you | 13:53 |
| mordred | rcarrillocruz: also - we have group support for inventories in zuul now | 13:54 |
| mordred | rcarrillocruz: so you hopefully shouldn't need to create any scaffolding for that | 13:55 |
| mordred | rcarrillocruz: oh - unless you're just using nodepool at the moment and not yet using zuul for the job running | 13:55 |
| jeblair | rcarrillocruz, mordred: https://review.openstack.org/458073 is the patch that added it | 14:08 |
| mordred | jeblair: cool- the patch to update it mostly just fixes a little logic error | 14:09 |
| jeblair | ya looking it over now | 14:09 |
| jeblair | lgtm | 14:13 |
| mordred | sweet. rcarrillocruz if you're in a position to see if that fixes you for listing image by name, let us know | 14:15 |
| jeblair | mordred: https://review.openstack.org/472868 is an easy saturday review | 14:16 |
| mordred | jeblair: easy like sunday morning? | 14:16 |
| jeblair | it is easier than going 16-0 in the finals | 14:16 |
| jeblair | that is, apparently, hard. | 14:16 |
| mordred | jeblair: this is what I hear | 14:18 |
| mordred | jeblair: that patch looks fine to me - how do we set verbose to true? | 14:18 |
| jeblair | mordred: "zuul-executor verbose" and "zuul executor unverbose" | 14:18 |
| SpamapS | busy bees in here | 14:19 |
| mordred | ah. neat | 14:19 |
| jeblair | mordred: i believe the commands are still plumbed through the cli | 14:19 |
| mordred | jeblair: oh! yes - I see it now | 14:19 |
| mordred | jeblair: I'm feeling like maybe we should emit debug logs from the zuul_stream callback to the stdout debug all the time but to the build log itself if -vvv is passed - what do you think? | 14:20 |
| mordred | so that we always catch it in tests, but only emit to build logs if we toggle verbose? or do you think we should just set verbose in the tests and not emit to either place unless verbose is toggled? | 14:21 |
| jeblair | mordred: i kind of like our ansible-log/executor-log split now -- if we can have consistent output into ansible log, but then control the verbosity of what ends up in executor log (so that an operator can debug zuul or ansible issues), i think that would be ideal. so then in tests, just 'vvv' all the time. | 14:22 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add shutdown option for zuul_console https://review.openstack.org/472867 | 14:36 |
| mordred | jeblair: cool | 14:36 |
| mordred | jeblair, clarkb: ^^ thereis the clarkb code incorporated into the change - Ihave not added additional error handling | 14:36 |
| mordred | I _did_ add reading /proc/net/tcp6 after testing it on one of our build nodes | 14:37 |
| mordred | but it works in that state on a node running zuul_console | 14:37 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add shutdown option for zuul_console https://review.openstack.org/472867 | 14:38 |
| mordred | giving clarkb credit in the commit :) | 14:38 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Enable verbose/unverbose zuul-executor commands https://review.openstack.org/472868 | 14:38 |
| jeblair | clarkb, mordred: did any of that code get borrowed from psutil? if so, we should annotate it. | 14:39 |
| jeblair | it certainly has a different shape, so maybe it's all original. | 14:40 |
| jeblair | just wanted to double check | 14:40 |
| jeblair | left comments with the only 2 things i'd do for error handling. | 14:40 |
| mordred | jeblair: yah - my hunch from reading it was that it was clarkb code | 14:44 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add shutdown option for zuul_console https://review.openstack.org/472867 | 14:53 |
| mordred | jeblair: now with error handling | 14:53 |
| jeblair | mordred: 2 comments | 14:55 |
| mordred | ah - good call | 14:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add shutdown option for zuul_console https://review.openstack.org/472867 | 14:56 |
| mordred | jeblair: I have validated that the code still works when it works :) | 14:57 |
| mordred | and neat - I just killed a zuul_console process with it and then verified that it correctly returned an error but not an exception when I tried it again | 14:59 |
| clarkb | jeblair: mordred my code is all original other than verifying scanning /proc that way is sane in psutil | 15:00 |
| clarkb | because pur use case is more constrained it was simpler to just focus on that then make general.solution fit | 15:01 |
| jeblair | clarkb: cool. thought it might be, but just wanted to double check in case it slipped our attention. | 15:04 |
| mordred | clarkb: woot | 15:04 |
| mordred | clarkb: it works great - thanks for figuring that out | 15:04 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add shutdown option for zuul_console https://review.openstack.org/472867 | 15:07 |
| mordred | jeblair, SpamapS: ^^ pep8 fix | 15:07 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool feature/zuulv3: Support booting cloud-images by name or id https://review.openstack.org/472959 | 15:16 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add shutdown option for zuul_console https://review.openstack.org/472867 | 15:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Use display.display for executor debug messages https://review.openstack.org/472840 | 15:54 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Use threads instead of processes in zuul_stream https://review.openstack.org/472850 | 15:54 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Don't wait for forever to join streamer https://review.openstack.org/472839 | 15:54 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Use display.display for executor debug messages https://review.openstack.org/472840 | 15:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Don't wait for forever to join streamer https://review.openstack.org/472839 | 15:56 |
| mordred | jeblair, SpamapS: reorganized that stack so that we don't have the add/remove of terminate() call | 15:56 |
| mordred | and also fixed the [Zuul] prefix thing | 15:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Make logging helper method in zuul_stream https://review.openstack.org/472963 | 16:15 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Only prepend hostname on multi-node plays https://review.openstack.org/472964 | 16:15 |
| * SpamapS reviews a few things then goes back to rusting for pleasure | 16:22 | |
| *** jkilpatr has joined #zuul | 16:29 | |
| mordred | SpamapS: mmm. rusting | 17:26 |
| SpamapS | mordred: tokio is making rustygear so much more comprehensible | 18:03 |
| SpamapS | And I get to drop my _AWFUL_ networking code | 18:04 |
| mordred | SpamapS: what's tokio | 18:38 |
| *** dkranz has quit IRC | 20:17 | |
| SpamapS | mordred: the mio author's long term vision for Rust IO | 20:19 |
| SpamapS | mordred: quite nice separation of codec/protocol from net code. | 20:19 |
| *** dkranz has joined #zuul | 20:21 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!