| *** paulfantom___ has joined #ara | 00:05 | |
| *** boxrick_ has joined #ara | 00:06 | |
| *** boxrick has quit IRC | 00:06 | |
| *** paulfantom has quit IRC | 00:06 | |
| *** paulfantom___ is now known as paulfantom | 00:06 | |
| *** ChanServ has quit IRC | 00:12 | |
| *** ChanServ has joined #ara | 00:33 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 00:33 | |
| *** jlozadad[m] has quit IRC | 00:38 | |
| *** harlowja has quit IRC | 01:00 | |
| *** jlozadad[m] has joined #ara | 01:05 | |
| *** themurph has joined #ara | 02:27 | |
| *** themurph has quit IRC | 02:38 | |
| *** themurph has joined #ara | 02:38 | |
| *** dcreno has joined #ara | 02:42 | |
| *** themurph has quit IRC | 02:43 | |
| *** bcoca has quit IRC | 03:07 | |
| *** dcreno has quit IRC | 03:13 | |
| *** harlowja has joined #ara | 03:29 | |
| *** harlowja has quit IRC | 04:14 | |
| *** dcreno has joined #ara | 05:10 | |
| *** dcreno has quit IRC | 05:42 | |
| *** sshnaidm is now known as sshnaidm_pto | 07:20 | |
| *** dcreno has joined #ara | 07:39 | |
| *** rvgate has quit IRC | 07:46 | |
| *** rvgate has joined #ara | 07:49 | |
| *** dcreno has quit IRC | 08:13 | |
| *** dcreno has joined #ara | 10:10 | |
| *** dcreno has quit IRC | 10:42 | |
| *** jlozadad has quit IRC | 11:30 | |
| *** jlozadad has joined #ara | 11:32 | |
| *** jlozadad has quit IRC | 11:44 | |
| *** jlozadad has joined #ara | 12:18 | |
| *** dcreno has joined #ara | 12:25 | |
| *** dcreno has quit IRC | 12:30 | |
| *** dcreno has joined #ara | 12:57 | |
| *** dcreno has quit IRC | 13:02 | |
| *** themurph has joined #ara | 13:37 | |
| *** Bakey has joined #ara | 13:38 | |
| *** dcreno has joined #ara | 13:57 | |
| ssbarnea | dmsimard: hi! are you here? a quick question on ara (btw really happy to see the junit export, works very nice with jenkins) | 14:01 |
|---|---|---|
| dmsimard | ssbarnea: I am | 14:02 |
| ssbarnea | dmsimard: is this normal https://ci.centos.org/artifacts/rdo/jenkins-tripleo-quickstart-gate-newton-delorean-quick-basic-5031/ara_oooq/ ? you see the .gz extension. | 14:02 |
| dmsimard | the junit support was contributed by RAX :) | 14:02 |
| ssbarnea | mainly making ara unbrowsable. not sure who's fault is, web server config or the archival part. | 14:03 |
| dmsimard | ssbarnea: tripleo-quickstart likely gzips the generated files by default, the vhost can then set decompression and mimetypes like here: https://github.com/openstack-infra/puppet-openstackci/blob/master/templates/logs.vhost.erb#L28-L61 | 14:04 |
| dmsimard | We don't control the web server on ci.centos.org so that's probably why it doesn't work there. | 14:04 |
| ssbarnea | dmsimard: and adding index.html.gz to DirectoryIndex i suppose, right? | 14:06 |
| dmsimard | ssbarnea: as far as openstack is concerned, we have rewrite rules that attempt to load the compressed version first and falls back to non-compressed | 14:06 |
| dmsimard | i.e, https://github.com/openstack-infra/puppet-openstackci/blob/821256b02d6bd7fdac99b055142f6593ca32cc96/templates/logs.vhost.erb#L111-L115 | 14:07 |
| ssbarnea | cool. time for me to raise a bug for ci.centos.org | 14:07 |
| ssbarnea | me not sure why people use apache instead of nginx, but that's a personal choice :D | 14:08 |
| ssbarnea | in this case this instance is an nginx server. i guess i will have to test the config items myself first, before raising a bug. | 14:10 |
| dmsimard | ssbarnea: FWIW this is likely something we can manage ourselves | 14:11 |
| dmsimard | ssbarnea: well, ourselves as in RDO/TripleO | 14:11 |
| dmsimard | ssbarnea: there's already a fair amount of jobs running from ci.centos.org that upload their logs to logs.rdoproject.org intead | 14:11 |
| dmsimard | i.e, https://centos.logs.rdoproject.org/weirdo-generic-packstack-scenario001/2604/ara/ | 14:12 |
| dmsimard | (that's gzipped) | 14:12 |
| ssbarnea | dmsimard: i have one hidden reason for documenting it for nginx, i want to use same config in other places, like downstream. | 14:12 |
| dmsimard | ssbarnea: nothing wrong with nginx -- did you know we had a role to set up ARA in different flavors ? nginx is one of them :D | 14:13 |
| ssbarnea | i should be able to find a solution easily, i being using nginx for 10+ years. | 14:13 |
| dmsimard | https://github.com/openstack/ansible-role-ara | 14:13 |
| dmsimard | ssbarnea: a fair warning, I'm not 100% positive ara 1.0 will have HTML generation | 14:14 |
| ssbarnea | apparently you underestimate the power of *static* content, there is nothing easier to scale and with better uptime than static ;) | 14:15 |
| ssbarnea | i am trying to make ara more popular here, so better if you keep html alive a bit more :D | 14:18 |
| dmsimard | ssbarnea: html generation is simple and straightforward but it scales very horribly | 14:20 |
| dmsimard | ssbarnea: I explain why it doesn't scale well here: https://ara.readthedocs.io/en/latest/advanced.html | 14:20 |
| dmsimard | tl;dr, if you run a 1000 task playbook against 50 hosts, that's 50 000 task results and 50 000 small html files (on top of whatever else there may be) | 14:20 |
| dmsimard | it takes time to compress those, transfer those (rsync) and they take up a significant amount of disk inodes versus what is a tiny sqlite database | 14:21 |
| ssbarnea | dmsimard: hmm, yes I agree but how about other approaches like JSON? i wonder how much JSON can a browser load until it chokes. | 14:23 |
| dmsimard | The plan with 1.0 is to have an optional API server with a stateless client interface | 14:24 |
| dmsimard | only example of that implementation that comes to mind right now is zuul -- it's what makes them available to gate their client interface against real data | 14:25 |
| * dmsimard tries to find an example | 14:26 | |
| *** bcoca has joined #ara | 14:51 | |
| *** bcoca has joined #ara | 14:51 | |
| *** themurph has quit IRC | 14:53 | |
| *** themurph has joined #ara | 15:02 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/ara master: Documented nginx config for static serving https://review.openstack.org/571220 | 15:07 |
| *** jlozadad_ has joined #ara | 15:52 | |
| *** jlozadad has quit IRC | 15:52 | |
| ssbarnea | dmsimard: yep, sqlite seems like the optimum way to store a report. what I do not know is why JS cannot download the database. why the middleware is really needed? | 16:14 |
| dmsimard | ssbarnea: the frontend relies on the database model since there is no API | 16:15 |
| dmsimard | ssbarnea: so the frontend needs actual access to the database | 16:15 |
| ssbarnea | if I remember well, JS can work with sqlite files directly, am I wrong? | 16:16 |
| dmsimard | I don't know, the backend is really sqlalchemy, sqlite happens to be the default driver so it wouldn't work with mysql/postgresql/etc | 16:17 |
| dmsimard | anyway, that's why things are properly decoupled in 1.0 | 16:17 |
| *** jlozadad_ has quit IRC | 16:18 | |
| ssbarnea | i have no time to check now but when I have time I will check https://github.com/kripken/sql.js/issues/230 | 16:25 |
| ssbarnea | (there are other alternative, too many maybe so I am not sure which one is future proof) | 16:25 |
| ssbarnea | HTML5 world is changing very fast | 16:26 |
| dmsimard | ssbarnea: doing this in 0.x would not be very productive, maybe we can sync with gvincent_ -- we've been slowly setting up the pieces for the new UI and it'd be great to have your input there | 16:28 |
| ssbarnea | good idea. i would be happy to provide feedback, especially on UI part. | 16:28 |
| dmsimard | ++ | 16:30 |
| ssbarnea | regarding index.html.gz, i contacted centos, raised a bug at https://bugs.centos.org/view.php?id=14890 -- and it is likely to be fixed soon. | 16:31 |
| ssbarnea | i am not sure why some ara checks failed on my patch, seems totally unrelated. | 16:32 |
| ssbarnea | ERROR! Unexpected Exception, this is probably a bug: No module named ara | 16:33 |
| dmsimard | yup, seems like a new regression | 16:35 |
| ssbarnea | i wonder if is not the same issue with cmd2 that git me earlier. | 16:35 |
| ssbarnea | sadly the logs were not collected | 16:35 |
| dmsimard | looking at a job, the python -m ara commands worked but it looks like 'ara' might not be in $PATH, I'll see if I can reproduce locally. | 16:37 |
| ssbarnea | http://logs.openstack.org/20/571220/1/check/ara-integration-opensuse-py27-2.5.2/f918075/job-output.txt.gz | 16:41 |
| ssbarnea | this states that develop-install failed. | 16:41 |
| ssbarnea | i was right, same issue as mine. | 16:42 |
| ssbarnea | cmd2 broke most openstack project. | 16:43 |
| ssbarnea | search for cmd2 in logs. i suspect, cmd2 dropped support for py2, and this prevented cliff from being installed. Here is the fix: https://review.openstack.org/#/c/571243/1/test-requirements.txt | 16:43 |
| ssbarnea | mainly a pinning to last version that works. | 16:44 |
| ssbarnea | i copied it from official openstack reqs file. | 16:44 |
| *** themurph has quit IRC | 16:44 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/ara master: Freeze cmd2 library on python2 https://review.openstack.org/571253 | 16:47 |
| *** jlozadad_ has joined #ara | 16:57 | |
| *** jlozadad_ has quit IRC | 17:10 | |
| *** jlozadad_ has joined #ara | 17:13 | |
| dmsimard | ssbarnea: ah thanks for figuring that out | 17:22 |
| openstackgerrit | David Moreau Simard proposed openstack/ara master: Documented nginx config for static serving https://review.openstack.org/571220 | 17:25 |
| dmsimard | ssbarnea: added a comment on your doc patch and rebased it on top of your fix | 17:25 |
| ssbarnea | thanks | 17:27 |
| openstackgerrit | Merged openstack/ara master: Freeze cmd2 library on python2 https://review.openstack.org/571253 | 17:33 |
| *** themurph has joined #ara | 17:33 | |
| *** Guest9512 has quit IRC | 18:17 | |
| *** spiette has joined #ara | 18:20 | |
| *** themurph has quit IRC | 18:55 | |
| *** themurph has joined #ara | 19:05 | |
| *** themurph has quit IRC | 19:07 | |
| *** themurph has joined #ara | 19:12 | |
| dcreno | :dmsimard: I'm getting the following when running a lot of ansible jobs at once: OperationalError: (pymysql.err.OperationalError) (1040, u'Too many connections') | 19:34 |
| Bakey | you should try running less jobs at once | 19:37 |
| dmsimard | dcreno: what is your max connections ? | 19:38 |
| dcreno | dmsimard: probably the default 100, looks like about 72 jobs got stuck. Looking at the same thing right now, guess I need to double-check and kick that up to something higher. | 19:39 |
| *** jlozadad_ has quit IRC | 19:55 | |
| *** themurph has quit IRC | 19:56 | |
| dcreno | dmsimard: interesting, the DB reports that the Max_used_connections is only 108 while max_connections is the default of 151. The DB claims to have never hit 151. | 19:59 |
| *** themurph has joined #ara | 20:03 | |
| *** themurph has quit IRC | 20:03 | |
| *** themurph has joined #ara | 20:12 | |
| *** tbielawa has joined #ara | 20:18 | |
| *** jlozadad_ has joined #ara | 20:46 | |
| ssbarnea | dmsimard: is there a way to mark some failures as expected when generating junit? | 21:00 |
| ssbarnea | i enabled ara to generate junit for infrare junit, which gave me https://rhos-ci-staging-jenkins.rhev-ci-vms.eng.rdu2.redhat.com/job/gate-infrared-tox/8/ | 21:01 |
| ssbarnea | bug this execution was sucessful, as part of testing was to generate same ansible task failures. | 21:02 |
| dmsimard | ssbarnea: I don't believe there is a way to do that right now. | 21:25 |
| *** tbielawa is now known as tbielawa|g0n3 | 21:27 | |
| *** themurph has quit IRC | 21:30 | |
| *** tbielawa|g0n3 has quit IRC | 21:40 | |
| *** harlowja has joined #ara | 21:52 | |
| *** dcreno_ has joined #ara | 21:59 | |
| *** dcreno__ has joined #ara | 22:01 | |
| *** dcreno_ has quit IRC | 22:01 | |
| *** dcreno has quit IRC | 22:03 | |
| *** dcreno__ has quit IRC | 22:06 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!