| sfbender | Fabien Boucher created software-factory/sf-ci master: Fix repoxplorer functional test for last repoxplorer version https://softwarefactory-project.io/r/15123 | 01:54 |
|---|---|---|
| sfbender | Tristan de Cacqueray created software-factory/sf-config master: cauth: prevent deployment issue when running on external host https://softwarefactory-project.io/r/15124 | 01:58 |
| sfbender | Tristan de Cacqueray created software-factory/sf-config master: cgit: fix multinode deployment without gerrit https://softwarefactory-project.io/r/15125 | 01:58 |
| sfbender | Tristan de Cacqueray created software-factory/sf-config master: config-update: replace fetch by synchronize https://softwarefactory-project.io/r/15126 | 01:58 |
| sfbender | Tristan de Cacqueray created software-factory/managesf master: configurations: remove the git clone depth argument https://softwarefactory-project.io/r/15127 | 02:35 |
| sfbender | Tristan de Cacqueray created software-factory/sf-config master: cgit: add smart http git transport configuration https://softwarefactory-project.io/r/15128 | 03:42 |
| sfbender | Tristan de Cacqueray created software-factory/sf-config master: cgit: enable local-cgit repo to be configured https://softwarefactory-project.io/r/15129 | 03:42 |
| *** raukadah has quit IRC | 05:45 | |
| *** chandankumar has joined #softwarefactory | 05:46 | |
| *** jangutter has joined #softwarefactory | 07:23 | |
| sfbender | Tristan de Cacqueray created software-factory/sf-config master: cgit: do not require component when config locations are remote https://softwarefactory-project.io/r/15130 | 07:35 |
| jangutter | tristanC: thanks very much for your help on Friday, after a set of reboots everything is working perfectly! | 07:57 |
| tristanC | jangutter: excellent, that's good to hear | 07:57 |
| tristanC | jangutter: i'm pretty sure it was related to the fast-forward-upgrade from 3.0 to 3.2 | 07:58 |
| jangutter | tristanC: yep. Also, our 3.0 install was "unorthodox" to say the least. I did a backup -> recover -> FF upgrade. | 07:58 |
| tristanC | jangutter: from now on, using "sfconfig --update" should just work automagically | 07:58 |
| jangutter | tristanC: I'm working through the papercuts left over. I notice https://softwarefactory-project.io/r/#/c/10569/ regressed, unfortunately. | 07:59 |
| jangutter | tristanC: manually adding rev 1 of that patch works around the issue again. | 07:59 |
| tristanC | jangutter: arg, indeed, the patch somehow dissapeared. looking for what happened | 08:06 |
| jangutter | tristanC I found it. | 08:06 |
| jangutter | tristanC: https://softwarefactory-project.io/r/#/c/13569/ <--- this "reverted" it. | 08:07 |
| tristanC | jangutter: my bad, let me revert that chunk | 08:08 |
| sfbender | Tristan de Cacqueray created software-factory/sf-config master: base: ensure DocumentRoot is set to /var/www https://softwarefactory-project.io/r/15134 | 08:09 |
| sfbender | Jan Gutter created software-factory/sf-config master: Set DocumentRoot explicitly in logserver conf https://softwarefactory-project.io/r/15135 | 08:11 |
| jangutter | Whoops. | 08:11 |
| jangutter | tristanC: I was worried that DocumentRoot should be /var/www/html in certain circumstances. | 08:13 |
| tristanC | jangutter: that works for me, we'll merge your version | 08:15 |
| jangutter | tristanC: another one that should be a clean backport to 3.2 is: https://softwarefactory-project.io/r/#/c/15131/ | 08:15 |
| jangutter | tristanC: but other than that, things went more smoothly than I expected. I switched from a nodepool/openstack vm based jobs to runC based jobs and things are working pretty well. | 08:17 |
| jangutter | testinfra does fail when the nodepool/runC bits don't start up 100%, but that was easy to fix | 08:19 |
| tristanC | jangutter: thanks for the backport, once our cloud provider is working we'll land it in 3.2 | 08:22 |
| tristanC | jangutter: what was the testinfra issue with runC? | 08:23 |
| jangutter | tristanC: no worries, I'll try to shake out as much as I can. Thanks a lot for all the work you guys are doing and we're shamelessly using :-p | 08:23 |
| jangutter | tristanC: it's not a biggie: I just noticed that testinfra fails if test nodes haven't been spun up by nodepool yet. | 08:24 |
| jangutter | tristanC: since one of the tests is "are there running containers". If nodepool is slow or something, the test might fail even though the runC node is running "perfectly". | 08:25 |
| jangutter | tristanC: but at that stage, software-factory is up and running, and cleaning out the stale nodes on nodepool made everything work. | 08:27 |
| jangutter | tristanC: specifically - https://github.com/softwarefactory-project/sf-config/blob/master/testinfra/test_hypervisor-runc.py#L19 <--- if nodepool hasn't spawned the nodes, this'll fail even if runC is happy. | 08:29 |
| jangutter | tristanC: I'm not sure how to "fix" the test though. | 08:30 |
| tristanC | jangutter: oh right, i think the fix is to remove the min-ready from the default configuration and remove that test | 08:35 |
| tristanC | jangutter: instead, we could check that the provider got started properly | 08:35 |
| *** jpena|off is now known as jpena | 08:56 | |
| sfbender | Merged DLRN master: Remove unneeded "git reset" in the run method https://softwarefactory-project.io/r/15116 | 09:40 |
| *** jangutter_ has joined #softwarefactory | 11:07 | |
| *** jangutte_ has joined #softwarefactory | 11:09 | |
| *** jangutter has quit IRC | 11:09 | |
| *** jangutter_ has quit IRC | 11:11 | |
| *** jangutte_ is now known as jangutter | 11:25 | |
| jangutter | Hi, weird question. I've been using rh-python35-python-shade and ansible on software-factory 3.1... any alternatives you could suggest? | 12:09 |
| sfbender | Javier Peña created dlrnapi_client master: Fix exception handling https://softwarefactory-project.io/r/15138 | 12:37 |
| *** jpena is now known as jpena|lunch | 12:38 | |
| *** jpena|lunch is now known as jpena | 13:31 | |
| *** mhu has quit IRC | 13:36 | |
| *** mhu has joined #softwarefactory | 13:37 | |
| *** rfolco is now known as rfolco|pto | 13:50 | |
| jangutter | (facepalm - rh-python35-python-shade -> rh-python35-python-openstacksdk ) | 15:16 |
| *** chandankumar is now known as raukadah | 15:32 | |
| dmsimard | jangutter: shade was eventually replaced by openstacksdk | 17:45 |
| dmsimard | it was merged and integrated in openstacksdk | 17:45 |
| jangutter | dmsimard: I have to doublecheck if ansible picks it up though.... | 17:48 |
| dmsimard | jangutter: which ansible ? the one used to deploy SF ? | 17:49 |
| jangutter | dmsimard: the one running inside SF: I had a nice docker set up with sf-3.1's ansible scl and used the os_server module for example. | 17:50 |
| jangutter | dmsimard: it's not critical - the one shipped by centos (non-scl version) also works. | 17:51 |
| dmsimard | jangutter: the ansible modules import openstacksdk iirc -- it should work | 17:51 |
| jangutter | dmsimard: I'll run some checks later when I get more time to verify - there's a .80 chance it's PEBCAK on my part. | 17:54 |
| dmsimard | jangutter: sure, keep us updated :D | 17:56 |
| jangutter | weird question: I'm looking at your nodepool.yaml config and I see: ELEMENTS_PATH: /etc/opt/rh/rh-python35/nodepool/elements:/usr/share/project-config/nodepool/elements | 18:10 |
| jangutter | is /usr/share/project-config/nodepool/elements something specific to your setup? | 18:11 |
| dmsimard | jangutter: yeah. | 18:12 |
| dmsimard | jangutter: Some of our images rely on nodepool elements which are provided by the upstream openstack-infra/project-config repository | 18:13 |
| jangutter | dmsimard: aah, so if I clone that to my builder....... | 18:13 |
| dmsimard | If you're not interested in reproducing images that are fairly similar to the openstack gate ones, you can disregard that | 18:13 |
| dmsimard | yes, we have an adhoc process that syncs project-config to the builder node (read: a cron) | 18:14 |
| *** jpena is now known as jpena|off | 18:27 | |
| *** sbadia has quit IRC | 18:35 | |
| *** sbadia has joined #softwarefactory | 19:37 | |
| *** ChanServ sets mode: +o sbadia | 19:37 | |
| *** sbadia has quit IRC | 20:34 | |
| *** sbadia has joined #softwarefactory | 20:39 | |
| *** ChanServ sets mode: +o sbadia | 20:39 | |
| *** shanemcd has quit IRC | 21:53 | |
| *** shanemcd has joined #softwarefactory | 21:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!