| *** sgw has joined #starlingx | 00:02 | |
| *** dklyle has joined #starlingx | 00:21 | |
| *** david-lyle has joined #starlingx | 00:27 | |
| *** dklyle has quit IRC | 00:30 | |
| *** TxGirlGeek has joined #starlingx | 00:47 | |
| *** TxGirlGeek has quit IRC | 01:03 | |
| *** david-lyle has quit IRC | 01:14 | |
| *** lcastell has quit IRC | 01:22 | |
| *** jaewook_oh has joined #starlingx | 02:15 | |
| *** xxie1[m] has joined #starlingx | 04:30 | |
| *** vmrod25 has quit IRC | 05:49 | |
| *** OutBackDingo has quit IRC | 06:42 | |
| *** OutBackDingo has joined #starlingx | 06:46 | |
| *** OutBackDingo has quit IRC | 07:12 | |
| *** hakhande_ has joined #starlingx | 07:16 | |
| *** OutBackDingo has joined #starlingx | 07:27 | |
| *** gkadam has joined #starlingx | 08:22 | |
| *** niklassoderstrom has joined #starlingx | 09:45 | |
| *** niklassoderstrom has quit IRC | 09:57 | |
| *** niklassoderstrom has joined #starlingx | 09:58 | |
| *** yan0s has joined #starlingx | 11:15 | |
| *** Bhujay has joined #starlingx | 11:28 | |
| *** ijolliffe has quit IRC | 13:02 | |
| *** jaewook_oh has quit IRC | 13:03 | |
| *** sgw has quit IRC | 13:27 | |
| *** ijolliffe has joined #starlingx | 13:37 | |
| *** sgw has joined #starlingx | 13:54 | |
| *** sgw has quit IRC | 14:11 | |
| *** sgw has joined #starlingx | 14:12 | |
| *** Bhujay has quit IRC | 14:28 | |
| *** yan0s has left #starlingx | 14:59 | |
| *** sgw has quit IRC | 14:59 | |
| *** sgw has joined #starlingx | 15:03 | |
| *** dklyle has joined #starlingx | 15:17 | |
| *** hakhande_ has quit IRC | 15:19 | |
| *** TxGirlGeek has joined #starlingx | 15:46 | |
| *** dklyle has quit IRC | 16:02 | |
| *** dklyle has joined #starlingx | 16:24 | |
| *** Bhujay has joined #starlingx | 16:38 | |
| *** Bhujay has quit IRC | 16:39 | |
| *** Bhujay has joined #starlingx | 16:40 | |
| *** Bhujay has quit IRC | 16:41 | |
| *** Bhujay has joined #starlingx | 16:41 | |
| *** Bhujay has quit IRC | 16:42 | |
| *** Bhujay has joined #starlingx | 16:43 | |
| *** Bhujay has quit IRC | 16:44 | |
| *** Bhujay has joined #starlingx | 16:44 | |
| *** Bhujay has quit IRC | 16:45 | |
| *** Bhujay has joined #starlingx | 16:46 | |
| *** Bhujay has quit IRC | 16:50 | |
| *** lcastell has joined #starlingx | 16:55 | |
| *** dklyle has quit IRC | 16:55 | |
| *** dklyle has joined #starlingx | 16:56 | |
| *** dklyle has quit IRC | 17:01 | |
| *** dklyle has joined #starlingx | 17:01 | |
| *** dklyle has quit IRC | 17:13 | |
| *** gkadam has quit IRC | 17:29 | |
| *** TxGirlGeek has quit IRC | 17:35 | |
| *** TxGirlGeek has joined #starlingx | 17:36 | |
| *** dklyle has joined #starlingx | 17:39 | |
| *** dklyle has quit IRC | 17:48 | |
| *** dklyle has joined #starlingx | 18:24 | |
| *** dklyle has quit IRC | 18:29 | |
| *** TxGirlGeek has quit IRC | 18:29 | |
| *** TxGirlGeek has joined #starlingx | 18:35 | |
| *** dklyle has joined #starlingx | 18:48 | |
| *** dklyle has quit IRC | 18:54 | |
| *** mugsie has quit IRC | 18:55 | |
| *** mugsie has joined #starlingx | 18:57 | |
| sgw | dpenney_: Question: Are you running the build-wheel and build-docker scripts natively or inside the docker we use to do the centos build? | 19:05 |
|---|---|---|
| dpenney_ | I run them natively, and build natively. I'm guessing the CENGN build for the docker image runs them natively as well. I haven't tried within the docker setup you use. Are you hitting issues? | 19:12 |
| sgw | dpenney_: Yup! I am working around it, but it will require some changes to the docker image and tb.sh script, since our current image does not have docker in it and we need to modify it to run docker-in-docker (using -v to map the docker.sock file) | 19:23 |
| *** TxGirlGeek has quit IRC | 19:29 | |
| sgw | dpenney_: what other assumptions are made in order to run the wheel/docker build scripts? Do I need to be in the f/stein branches for example? I got the following failure: | 19:35 |
| sgw | Failed to parse name or version from /localdisk/loadbuild/swold/stx/std/build-wheels-centos-pike/base/*.whl | 19:35 |
| dpenney_ | so it would assume the env is setup, like MY_REPO and MY_WORKSPACE, that you've done "build-pkgs" already (which would build the packages that contain any stx wheels). The build-wheel-tarball.sh tool calls build-base-wheels.sh first to build any needed upstream wheels, then get-stx-wheels.sh to retrieve the stx wheels | 19:39 |
| dpenney_ | given that, I guess you'd need to be inside your build docker? | 19:40 |
| sgw | Ok, so maybe my build is not current enough, I will rebuild build-pkgs, will it be possible to just rebuild one container or wheel? | 19:41 |
| dpenney_ | the error seems odd... do you have a std/build-wheels-centos-master/base dir in your $MY_WORKSPACE? | 19:42 |
| dpenney_ | this error comes from build-wheel-tarball.sh after it has already run build-base-wheels.sh, so that dir should contain a set of wheels | 19:43 |
| dpenney_ | you could run build-wheel-tarball.sh in a bash -x command for tracing and send me the output, along with maybe a "find std/build-wheels-centos-master"? | 19:44 |
| sgw | I think you just answered a question, I am on master not f/stien, so I have a build-wheels-centos-pike/base directory that is empty | 19:45 |
| dpenney_ | the branch doesn't really matter... what matters is the --release arg you specify | 19:45 |
| dpenney_ | ie. build-wheel-tarball.sh --os centos --release master | 19:45 |
| *** TxGirlGeek has joined #starlingx | 19:46 | |
| sgw | Iwas going from your README in build-tools/build-docker-images that showed OPENSTACK_RELEASE=pike | 19:46 |
| dpenney_ | since you didn't specify --release master, it defaulted to pike | 19:46 |
| sgw | I did specify it on the command line | 19:46 |
| dpenney_ | all depends on what you want to build | 19:46 |
| sgw | I am going by the README, if I should build master, let me know, I figured the examples in the readme would work | 19:47 |
| dpenney_ | it's fine... I just grabbed the master path from my recent builds... so swap "master" for "pike" in what I've said :) | 19:47 |
| dpenney_ | either should work fine. so building pike, your std/build-wheels-centos-pike/base should have wheels | 19:48 |
| sgw | Ok, and the answer is still that the base dir that you asked about is empty | 19:48 |
| *** dklyle has joined #starlingx | 19:50 | |
| dpenney_ | if you can send me a trace of the build, maybe that will point us in the right direction | 19:51 |
| *** dklyle has quit IRC | 19:53 | |
| sgw | I will first try to rerun build-pkgs as I said this is an older build, so I wonder if it pre-dates added some of the wheel tooling. | 19:55 |
| dpenney_ | that shouldn't really impact the error you're having though | 19:55 |
| dpenney_ | you're not getting to the stage where it's accessing output from build-pkgs | 19:56 |
| sgw | What puts the *.whls in to the base directory, the build-wheels/* scripts? | 19:57 |
| dpenney_ | well... it is... it runs both build-base-wheels.sh and get-stx-wheels.sh by the time of your failure, but build-base-wheels.sh does not rely on any output from build-pkgs | 19:57 |
| dpenney_ | yeah, it's the build-base-wheels.sh that populates the base dir | 19:57 |
| sgw | Ok, let me try -x here first and read scripts. | 19:57 |
| dpenney_ | you should also be able to run build-base-wheels.sh directly, if you want | 19:58 |
| sgw | Actually I did that earlier and it worked | 19:58 |
| dpenney_ | but if it fails, that failure should be caught | 19:58 |
| sgw | Was using it directly to test my docker enablement changes | 19:59 |
| dpenney_ | ok, you could also do that one in a bash -x and we can see what's happening.. or not happening | 19:59 |
| sgw | Right, so just to confirm, it should be the script that puts the *.whls files in base? | 20:00 |
| dpenney_ | within that script, the base output dir is BUILD_OUTPUT_PATH ... that gets passed as a bind-mount to the docker run command that launchs the builder, and that build script puts the files into that bind-mounted dir | 20:01 |
| dpenney_ | docker files and scripts for that builder are in build-wheels/docker ... runs the docker-build-wheel.sh script, which copies the built or downloaded wheels to the mounted base output dir | 20:02 |
| *** happycat68 has quit IRC | 20:19 | |
| *** dklyle has joined #starlingx | 20:20 | |
| *** happycat68 has joined #starlingx | 20:21 | |
| *** happycat68 is now known as tao_liu | 20:23 | |
| *** tao_liu has quit IRC | 20:27 | |
| *** dklyle has quit IRC | 20:29 | |
| *** tao_liu has joined #starlingx | 20:30 | |
| *** tao_liu has quit IRC | 20:32 | |
| *** tao_liu has joined #starlingx | 20:33 | |
| *** dklyle has joined #starlingx | 20:49 | |
| *** marj3 has quit IRC | 20:59 | |
| *** dklyle has quit IRC | 21:10 | |
| mariocarrillo | Hi dpenny_, I wonder, why most of the "output" directories in http://mirror.starlingx.cengn.ca/mirror/starlingx/master/centos/ does not include the directory wheels with the tar package | 21:18 |
| *** dklyle has joined #starlingx | 21:22 | |
| *** dklyle has quit IRC | 21:28 | |
| *** ijolliffe has quit IRC | 21:29 | |
| *** dklyle has joined #starlingx | 21:42 | |
| dpenney_ | the docker images and wheels are built separately and on a different schedule. slittle1 added a hook to the jenkins job to create a latest_docker_image_build symlink to help finding the most recent tarball, but I only see that for the f/stein build here: http://mirror.starlingx.cengn.ca/mirror/starlingx/f/stein/centos/latest_docker_image_build/ | 21:47 |
| dpenney_ | slittle1: was the hook only added for the f/stein image build? | 21:48 |
| *** dklyle has quit IRC | 22:01 | |
| mariocarrillo | dpenney_: thanks | 22:17 |
| *** ijolliffe has joined #starlingx | 23:02 | |
| *** sgw has quit IRC | 23:14 | |
| *** dklyle has joined #starlingx | 23:34 | |
| *** dklyle has quit IRC | 23:40 | |
| *** dklyle has joined #starlingx | 23:54 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!