| *** sgw has quit IRC | 00:12 | |
| *** sgw has joined #starlingx | 00:42 | |
| mariocarrillo | Hi, does anyone know how often is created the pike wheels tarball? | 00:43 |
|---|---|---|
| sgw | mariocarrillo: good question, I think we moved to master, so I am not sure if we are still building the pike wheels, but I would defer to slittle1 but they are gone for the day | 00:44 |
| *** zhubx has quit IRC | 00:53 | |
| *** zhubx has joined #starlingx | 00:53 | |
| *** OutBackDingo has joined #starlingx | 00:59 | |
| *** ijolliffe has quit IRC | 01:18 | |
| mariocarrillo | sgw: I have checked in different output build directories and I could not found them | 01:23 |
| *** jsun3 has joined #starlingx | 02:28 | |
| *** jsun3 has quit IRC | 02:36 | |
| *** cheng1 has joined #starlingx | 02:40 | |
| *** cheng1 has quit IRC | 03:20 | |
| *** cheng1 has joined #starlingx | 03:55 | |
| *** cheng1_ has joined #starlingx | 06:22 | |
| *** changcheng has quit IRC | 06:25 | |
| *** cheng1 has quit IRC | 06:25 | |
| *** changcheng has joined #starlingx | 06:28 | |
| *** cheng1_ has quit IRC | 09:55 | |
| *** cheng1_ has joined #starlingx | 10:07 | |
| *** cheng1_ has quit IRC | 11:27 | |
| *** changcheng has quit IRC | 12:14 | |
| *** changcheng has joined #starlingx | 12:14 | |
| *** ijolliffe has joined #starlingx | 12:32 | |
| *** cheng1_ has joined #starlingx | 12:39 | |
| *** sdinescu has quit IRC | 12:42 | |
| *** sdinescu has joined #starlingx | 12:42 | |
| mthebeau | I found a "build wheels" script in build-tools but I don't see that we build that on a regular basis, and the CENGN mirror is not presenting any... rpm -qp --info gnocchi-wheels-4.2.5-1.tis.1.noarch.rpm: | 12:57 |
| mthebeau | Description :Contains python wheels for gnocchi | 12:57 |
| mthebeau | Was there a change in approach perhaps? | 12:57 |
| mthebeau | There's a couple of starlingx-discuss reports for STX_build_wheels back in february. There's a request for better documentation, January 25 from sgw. ;) | 13:03 |
| mthebeau | I'll stop poking until slittle1 and/or dpenney get a chance to respond | 13:04 |
| sgw | Is there a way to rebuild the mock builder chroots? | 13:14 |
| *** cheng1_ has quit IRC | 13:19 | |
| dpenney | pike/master names are migrating to stable/dev to be a little clearer to folks about usage. Both build streams will continue to be built, as some images build only in pike (renamed stable). pike/master has caused confusion about when and where these should be built. | 13:35 |
| dpenney | both pike/master should have been built every week, but I believe the CENGN server was only setup to build master | 13:36 |
| dpenney | Scott was working yesterday to setup the stable/dev builds, and both should run | 13:36 |
| mthebeau | is there such a thing as a "wheels tarball" dpenney? | 13:37 |
| dpenney | yes. For each build stream, stable and dev (formerly pike and master), there is a stx-centos base image, a wheels tarball, and images | 13:38 |
| dpenney | the wiki has not been updated yet for the migration to stable/dev, but will be once all the pieces are in place and will be followed with starling-xdiscuss email | 13:39 |
| dpenney | wiki: https://wiki.openstack.org/wiki/StarlingX/Containers/BuildingImages | 13:39 |
| mthebeau | mariocarrillo was asking about where that tarball is found (or rather, how often it is built) | 13:39 |
| mthebeau | 2019-Mar-26 00:21:57 | 13:40 |
| dpenney | yes, as I mentioned, it should have been built alongside master each week, but due to a misunderstanding, the pike wheels tarball and images had not been built recently | 13:40 |
| mthebeau | cool I'm going to look for an email for mario and copy the link and "due to a misunderstanding, the pike wheels tarball..." text to him | 13:42 |
| dpenney | going forward, the "containers build" from CENGN should include both stable and dev, i believe will still only run once a week | 13:42 |
| *** kenyis709 has joined #starlingx | 13:48 | |
| mthebeau | sgw, re rebuilding the mock build chroots... the instructions I have for cleaning and deleting a build are "build-pkgs --clean && rm -rf <build_dir>". That's all I have to offer on the subject. The chroots are owned by mock and usually you can't rm those... so the build-pkgs --clean must be cleaning them(?) | 14:11 |
| dpenney | I think you can do a mock --clean, passing the correct -r config file and probably some other option for cleaning the various parallel build mockroots... slittle1? | 14:12 |
| mthebeau | there are a lot of mock.*clean functions in build-tools/build-rpms* | 14:15 |
| sgw | dpenney: the "correct -r config file", from the /localdisk/loadbuild/swold/starlingx/std/configs/swold-starlingx-tis-r5-pike-std/ directory? | 14:15 |
| mthebeau | find $MY_WORKSPACE/configs/$MY_BUILD_ENVIRONMENT* -name "$MY_BUILD_ENVIRONMENT*b[0-9]*.cfg" | 14:17 |
| dpenney | yeah, that looks right.... you can do things like "mock -r <cfgfile> --shell" to poke around, for example... I wasn't sure where the parallel build files were, but that looks right | 14:17 |
| mthebeau | yes, that looks the same | 14:17 |
| dpenney | but the build-pkgs --clean may do that already... that's a question for slittle1 | 14:18 |
| mthebeau | if all then # wipe all ... mock_clean... else # only what we built... mock_partial_clean | 14:26 |
| mthebeau | if TARGETS == "" then ALL == true | 14:27 |
| mthebeau | so I think 'yes', build-pkgs --clean ends up calling mock -r on all the cfg files found in that find command | 14:30 |
| *** kenyis709 has quit IRC | 14:58 | |
| mthebeau | sgw, I am reproducing FileNotFoundError on my build server. | 15:00 |
| *** kenyis709 has joined #starlingx | 15:00 | |
| dpenney | did you upgrade mock on your system? check its dependencies | 15:01 |
| mthebeau | "yes" I recently created my build server from internal wiki instructions; I promptly ran a yum update command. The deps for mock on my system say python >= 2.7; I checked those previously; the red hat bugzilla applies to me in this case. | 15:04 |
| *** sgw has quit IRC | 15:05 | |
| mthebeau | ps the error was absent from my first build because I was using build-avoidance | 15:05 |
| *** kenyis709 has quit IRC | 15:15 | |
| mthebeau | I tried yum downgrade... but there was a conflict with mock-core-configs... so I removed mock and mock-core-configs and installed mock-1.2.17-1.el7.centos - the only alternative to mock-1.4.14-2.el7... | 15:25 |
| mthebeau | will see that happens | 15:25 |
| dpenney | interesting that the dependencies for the centos mock-1.4.14 say python 2.7, if this is a python3 thing. I misread that when you noted the dependency earlier | 15:29 |
| *** ijolliffe has quit IRC | 15:32 | |
| *** ijolliffe has joined #starlingx | 15:32 | |
| *** jsun3 has joined #starlingx | 15:40 | |
| mthebeau | Ya. We don't have to open a stx bug, right? Since it's a build server thing. | 15:52 |
| * dtroyer scans scrollback after 3 hour meeting block is over… | 15:56 | |
| dtroyer | I have a question regarding building pike vs master referenced above… | 15:57 |
| dtroyer | is anyone still pulling from starlingx-staging/stx-nova master branch? | 15:57 |
| dtroyer | I want to change that to be a mirror of upstream Nova master… if something is still using it we should sync up | 15:58 |
| *** lcastell has joined #starlingx | 16:00 | |
| *** lcastell has quit IRC | 16:01 | |
| *** sgw has joined #starlingx | 16:13 | |
| mariocarrillo | dpenney: thanks :) | 16:19 |
| dpenney | dtroyer: the stx-nova container image is currently building only in master or dev build, from https://github.com/openstack/nova.git ... Once stein is released, we'll setup stx-nova (and others) stable build images | 16:25 |
| dtroyer | dpenney: Thanks. So it is safe for me to start that work on stx-nova and make master a mirror of upstream master? | 16:26 |
| dpenney | I think so, but let me check something... Al has been working on removing openstack pieces from the host | 16:27 |
| dtroyer | ok, no rush, I'm just getting things ready | 16:27 |
| dpenney | Al has removed the openstack packages from the ISO, but not yet from the build (over concerns around wheels, etc, and not wanting to break image builds). He's going to raise a WIP review for removing everything from the build, and I'll verify the wheels/images build ok from that) | 16:37 |
| dpenney | so the staging nova repo is currently used in the build, but doesn't need to be. In summary. :) | 16:38 |
| dtroyer | ok, thanks. I'll hold off on the repo changes until that merges | 16:38 |
| dpenney | Al's WIP review: https://review.openstack.org/#/c/648180/ | 16:42 |
| *** ijolliffe has quit IRC | 16:43 | |
| *** kenyis709 has joined #starlingx | 16:44 | |
| *** ijolliffe has joined #starlingx | 16:44 | |
| *** kenyis709 has quit IRC | 16:54 | |
| dpenney | heads up: it looks like the master / dev image builds are currently broken for stx-aodh and stx-heat, due to some upstream stuff. Someone purged croniter 0.3.27 from pypi, which is in the upper-constraints built into the loci/requirements image. I'll either need to update our master-wheels.cfg to specify the newer version, or wait for loci/requirements to rebuild to pick up the new upper-constraints.txt | 17:00 |
| dpenney | see https://github.com/openstack/requirements/commit/77e9d3524f74d20c2a8de9651304601277e2d048 | 17:00 |
| dpenney | our master wheels tarball is built by downloading the loci/requirements wheels tarball and adding to or modifying it | 17:03 |
| *** kenyis709 has joined #starlingx | 17:24 | |
| *** kenyis709 has quit IRC | 18:12 | |
| dpenney | sgw: https://review.openstack.org/648201 is created to update base wheels to handle the master/dev build failures due to upstream croniter module change | 18:13 |
| mthebeau | I have a happy build with downgraded mock. | 18:59 |
| *** ijolliffe has quit IRC | 19:00 | |
| *** kenyis709 has joined #starlingx | 20:20 | |
| sgw | dumb question time: I am trying to debug a duplex installation with libvirt, controller-0 is booted and I think is operational when the controller-1 node boots, it's failing to get a response from DHCP and PXE fails | 20:25 |
| dpenney | in the case of virtualbox, we have to configure it so that the boot goes over a specific NIC - the mgmt interface | 20:26 |
| dpenney | if you think the DHCP request is going out over the correct interface, you could check on controller-0 if it's received (DHCP logs in daemon.log, I believe) or use tcpdump to trace for the incoming packet | 20:27 |
| sgw | dpenney: I looked at tcpdump and can see the bootp request come from the -1 node, I can also see activity (UDP) from the -0 node, so I am assuming they are talking, I will check the daemon.log (that helps) | 20:28 |
| dpenney | you can also check to make sure dnsmasq is running on controller-0: sudo sm-dump |grep dnsmasq | 20:30 |
| sgw | dpenny: yup enabled-active | 20:32 |
| *** kenyis709 has quit IRC | 20:58 | |
| *** kenyis709 has joined #starlingx | 21:00 | |
| mthebeau | hey dpenney; do our containers have a common base container? | 21:11 |
| dpenney | well, I think what you're asking is whether they have a common base layer, and yes. The images we build start from a stock centos image and build up from that | 21:12 |
| dpenney | we build a base stx-centos first, on top of stock centos, and then use that as the base for the service images | 21:13 |
| mthebeau | docker pull starlingx/stx-centos:dev-20190325T013001Z | 21:17 |
| mthebeau | woot; ok thanks | 21:17 |
| sgw | yeah, that's a thing! | 21:19 |
| mthebeau | the layers listed in docker image inspect are shasums... | 21:20 |
| mthebeau | (wondering if there's a way to get the names from shasums | 21:21 |
| dpenney | not sure... I haven't looked into that | 21:21 |
| *** ijolliffe has joined #starlingx | 21:23 | |
| mthebeau | ya, google is not offering easy clues. In anycase; | 21:29 |
| mthebeau | ./build-tools/build-docker-images/stx-centos/Dockerfile.master | 21:30 |
| mthebeau | ./build-tools/build-docker-images/stx-centos/Dockerfile.pike | 21:30 |
| mthebeau | starlingx/stx-centos:dev-20190325T013001Z | 21:30 |
| *** kenyis709 has quit IRC | 21:36 | |
| *** kenyis709 has joined #starlingx | 22:31 | |
| *** zhubx has quit IRC | 22:37 | |
| *** zhubx007 has joined #starlingx | 22:37 | |
| *** kenyis709 has quit IRC | 22:40 | |
| *** lcastell has joined #starlingx | 22:46 | |
| *** sgw has quit IRC | 23:54 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!