openstackgerrit | Merged openstack/ansible-role-tripleo-modify-image master: Fix exists check https://review.opendev.org/684134 | 00:02 |
---|---|---|
openstackgerrit | Merged openstack/tripleo-common master: Always close src/dst sessions and raise to retry https://review.opendev.org/684786 | 00:02 |
*** zaneb has quit IRC | 00:12 | |
*** zaneb has joined #tripleo | 00:12 | |
openstackgerrit | Merged openstack/puppet-tripleo master: Use correct paths to configure ovn dbs certs https://review.opendev.org/683916 | 00:20 |
*** toure is now known as toure|gone | 00:20 | |
*** zaneb has quit IRC | 00:34 | |
*** zaneb has joined #tripleo | 00:35 | |
*** zaneb has quit IRC | 00:36 | |
*** zaneb has joined #tripleo | 00:36 | |
*** gbarros has quit IRC | 00:45 | |
*** ricolin has joined #tripleo | 00:59 | |
*** mschuppert has quit IRC | 01:06 | |
openstackgerrit | Toure Dunnon proposed openstack/tripleo-ansible master: [WIP] Add service management support. https://review.opendev.org/681017 | 01:15 |
openstackgerrit | Toure Dunnon proposed openstack/tripleo-ansible master: [WIP] Add service management support. https://review.opendev.org/681017 | 01:19 |
*** ricolin has quit IRC | 01:34 | |
*** ricolin has joined #tripleo | 01:35 | |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: Add podman_container_facts https://review.opendev.org/686035 | 01:38 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: podman_image_facts: fix listing all images https://review.opendev.org/686080 | 01:40 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: Add podman_container_facts https://review.opendev.org/686035 | 01:45 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP: POC of podman_container https://review.opendev.org/682458 | 01:46 |
*** spsurya has joined #tripleo | 01:47 | |
*** cdearborn has quit IRC | 01:57 | |
*** ayoung has quit IRC | 02:08 | |
*** irclogbot_2 has quit IRC | 02:09 | |
*** irclogbot_1 has joined #tripleo | 02:13 | |
*** saneax has joined #tripleo | 03:18 | |
openstackgerrit | Merged openstack/tripleo-validations master: Constrain molecule https://review.opendev.org/685715 | 03:30 |
*** saneax has quit IRC | 03:37 | |
openstackgerrit | Merged openstack/tripleo-quickstart master: fix tq,cico setup https://review.opendev.org/685090 | 03:41 |
openstackgerrit | Merged openstack/tripleo-common master: Warn when looking up grandparent resource name https://review.opendev.org/682694 | 04:08 |
*** mmethot_ has joined #tripleo | 04:36 | |
*** ricolin_ has joined #tripleo | 04:36 | |
*** elfGamal has quit IRC | 04:37 | |
*** cgoncalves has quit IRC | 04:37 | |
*** numans has quit IRC | 04:37 | |
*** elGamal has joined #tripleo | 04:38 | |
*** d0ugal has quit IRC | 04:38 | |
*** ricolin has quit IRC | 04:40 | |
*** mmethot has quit IRC | 04:40 | |
*** dasp has quit IRC | 04:40 | |
*** jamesdenton has quit IRC | 04:41 | |
*** jamesdenton has joined #tripleo | 04:44 | |
*** cgoncalves has joined #tripleo | 04:46 | |
*** aakarsh has quit IRC | 04:48 | |
*** pcaruana has joined #tripleo | 04:52 | |
*** d0ugal has joined #tripleo | 04:52 | |
*** ykarel|away has joined #tripleo | 05:09 | |
*** marios has joined #tripleo | 05:18 | |
*** lmiccini has joined #tripleo | 05:27 | |
*** yprokule has joined #tripleo | 05:39 | |
*** hjensas has quit IRC | 05:44 | |
*** kopecmartin|off is now known as kopecmartin | 05:49 | |
*** jfrancoa has joined #tripleo | 05:54 | |
*** rcernin_ has joined #tripleo | 05:59 | |
openstackgerrit | Francesco Pantano proposed openstack/tripleo-heat-templates master: Add registry authentication parameters to pull ceph related containers https://review.opendev.org/677699 | 06:00 |
*** rcernin has quit IRC | 06:01 | |
*** mschuppert has joined #tripleo | 06:29 | |
*** dasp has joined #tripleo | 06:40 | |
*** rcernin_ has quit IRC | 06:45 | |
*** rcernin has joined #tripleo | 06:45 | |
*** xek has joined #tripleo | 06:55 | |
*** tesseract has joined #tripleo | 06:57 | |
*** jbadiapa has joined #tripleo | 07:00 | |
*** iurygregory has joined #tripleo | 07:00 | |
*** dciabrin has joined #tripleo | 07:01 | |
*** jaosorior has joined #tripleo | 07:03 | |
*** rcernin has quit IRC | 07:04 | |
*** slaweq has joined #tripleo | 07:10 | |
*** hjensas has joined #tripleo | 07:10 | |
*** jpich has joined #tripleo | 07:11 | |
*** jbadiapa has quit IRC | 07:13 | |
*** tosky has joined #tripleo | 07:17 | |
*** jbadiapa has joined #tripleo | 07:19 | |
openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: [wip] Add NetworkManager and dhcp-client for CentOS 8 https://review.opendev.org/686090 | 07:20 |
*** pkopec has joined #tripleo | 07:23 | |
*** pierreprinetti has joined #tripleo | 07:28 | |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 07:28 |
openstackgerrit | Merged openstack/tripleo-heat-templates stable/rocky: Add parameters for dateext in logrotate https://review.opendev.org/684938 | 07:29 |
*** amoralej|off is now known as amoralej | 07:30 | |
*** bogdando has joined #tripleo | 07:33 | |
openstackgerrit | Michele Baldessari proposed openstack/puppet-tripleo master: Add a configurable delay to Nova Evacuate calls https://review.opendev.org/684785 | 07:33 |
*** holser has joined #tripleo | 07:34 | |
*** jbadiapa has quit IRC | 07:34 | |
*** slaweq has quit IRC | 07:38 | |
openstackgerrit | Gael Chamoulaud proposed openstack/tripleo-validations stable/stein: New validation: detect failed containers https://review.opendev.org/686093 | 07:43 |
*** jpena|off is now known as jpena | 07:48 | |
openstackgerrit | Keigo Noha proposed openstack/tripleo-heat-templates stable/queens: Add parameters for dateext in logrotate https://review.opendev.org/686098 | 07:52 |
*** pkopec has quit IRC | 07:56 | |
*** florianf has joined #tripleo | 07:56 | |
*** rpittau|afk is now known as rpittau | 07:57 | |
*** slaweq has joined #tripleo | 07:57 | |
*** suuuper has joined #tripleo | 08:03 | |
openstackgerrit | Javier Peña proposed openstack/paunch master: Properly mock cpu_affinity on all unit tests https://review.opendev.org/686100 | 08:04 |
*** tkajinam has quit IRC | 08:04 | |
*** jpena is now known as jpena|brb | 08:10 | |
openstackgerrit | Natal Ngétal proposed openstack/python-tripleoclient master: Fix message of unused_params and invalid_role_specific_params. https://review.opendev.org/674570 | 08:11 |
*** jpich has quit IRC | 08:13 | |
*** paramite has joined #tripleo | 08:14 | |
*** weshay has quit IRC | 08:15 | |
*** weshay has joined #tripleo | 08:16 | |
Tengu | bogdando: all the patches merged! I'm pretty sure CI will be better from now on :) | 08:21 |
*** pkopec has joined #tripleo | 08:25 | |
openstackgerrit | Jose Luis Franco proposed openstack/tripleo-upgrade master: Sync up upgrades and updates workarounds mechanism. https://review.opendev.org/684419 | 08:26 |
*** ykarel|away has quit IRC | 08:28 | |
*** ykarel has joined #tripleo | 08:30 | |
*** ykarel is now known as ykarel|away | 08:30 | |
*** gfidente has joined #tripleo | 08:32 | |
*** jpena|brb is now known as jpena | 08:32 | |
*** alexmcleod has joined #tripleo | 08:33 | |
*** jaosorior has quit IRC | 08:37 | |
*** derekh has joined #tripleo | 08:38 | |
*** ricolin_ is now known as ricolin | 08:41 | |
*** panda has quit IRC | 08:48 | |
*** panda has joined #tripleo | 08:49 | |
*** holser has quit IRC | 09:02 | |
*** holser has joined #tripleo | 09:02 | |
*** holser has quit IRC | 09:03 | |
*** holser has joined #tripleo | 09:03 | |
*** jfrancoa has quit IRC | 09:03 | |
*** florianf has quit IRC | 09:11 | |
*** jpich has joined #tripleo | 09:17 | |
*** jfrancoa has joined #tripleo | 09:18 | |
*** ab-a has quit IRC | 09:30 | |
*** jbadiapa has joined #tripleo | 09:30 | |
*** ab-a has joined #tripleo | 09:31 | |
*** holser has quit IRC | 09:34 | |
*** holser has joined #tripleo | 09:36 | |
*** florianf has joined #tripleo | 09:39 | |
openstackgerrit | Merged openstack/tripleo-validations stable/stein: Add playbook creation for the automation https://review.opendev.org/685938 | 09:44 |
openstackgerrit | Gael Chamoulaud proposed openstack/tripleo-validations stable/stein: Constrain molecule https://review.opendev.org/686119 | 09:52 |
*** jaosorior has joined #tripleo | 09:52 | |
*** jpich has quit IRC | 09:55 | |
*** jpich has joined #tripleo | 09:55 | |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 09:56 |
*** iurygregory has quit IRC | 09:57 | |
*** hjensas has quit IRC | 10:06 | |
openstackgerrit | Carlos Goncalves proposed openstack/tripleo-heat-templates master: Fix placement of Octavia services in roles https://review.opendev.org/686125 | 10:08 |
*** gfidente has quit IRC | 10:09 | |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 10:12 |
*** iurygregory has joined #tripleo | 10:17 | |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 10:20 |
openstackgerrit | Martin Schuppert proposed openstack/tripleo-heat-templates stable/queens: [Queens] Run cellv2 host discovery via workflow or deploy_steps_tasks https://review.opendev.org/683069 | 10:22 |
*** rcernin has joined #tripleo | 10:25 | |
openstackgerrit | Merged openstack/python-tripleoclient stable/stein: Close some file descriptors that were left open https://review.opendev.org/686015 | 10:28 |
*** ykarel|away has quit IRC | 10:31 | |
*** slaweq has quit IRC | 10:38 | |
*** zbr|ruck is now known as zbr|lunch | 10:39 | |
*** slaweq has joined #tripleo | 10:42 | |
*** dsneddon has quit IRC | 10:44 | |
*** tosky_ has joined #tripleo | 10:49 | |
*** slaweq has quit IRC | 10:51 | |
*** tosky has quit IRC | 10:52 | |
*** slaweq has joined #tripleo | 10:57 | |
*** jbadiapa has quit IRC | 10:57 | |
*** tosky_ is now known as tosky | 10:58 | |
*** tosky_ has joined #tripleo | 11:02 | |
*** tosky is now known as Guest18639 | 11:02 | |
*** tosky_ is now known as tosky | 11:02 | |
*** jaosorior has quit IRC | 11:02 | |
*** rh-jelabarre has quit IRC | 11:04 | |
*** Guest18639 has quit IRC | 11:04 | |
*** rh-jelabarre has joined #tripleo | 11:05 | |
*** csmart has quit IRC | 11:07 | |
*** dsneddon has joined #tripleo | 11:09 | |
*** openstackstatus has quit IRC | 11:09 | |
*** rnoriega- has quit IRC | 11:09 | |
*** rnoriega_ has joined #tripleo | 11:13 | |
*** csmart has joined #tripleo | 11:18 | |
*** gfidente has joined #tripleo | 11:22 | |
*** jpena is now known as jpena|lunch | 11:24 | |
*** jaosorior has joined #tripleo | 11:28 | |
*** ansmith has quit IRC | 11:35 | |
openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: Add NetworkManager and dhcp-client for CentOS 8 https://review.opendev.org/686090 | 11:36 |
openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: CentOS 8 minimal testing and support https://review.opendev.org/684860 | 11:36 |
openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: [wip] use mirrors for functional test builds https://review.opendev.org/685885 | 11:36 |
*** dpeacock has quit IRC | 11:37 | |
*** raildo has joined #tripleo | 11:38 | |
*** rpittau is now known as rpittau|afk | 11:38 | |
*** dpeacock has joined #tripleo | 11:40 | |
*** artom has joined #tripleo | 11:47 | |
*** raildo has quit IRC | 11:52 | |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 11:52 |
*** raildo has joined #tripleo | 11:54 | |
*** dsneddon has quit IRC | 11:55 | |
*** ekultails has joined #tripleo | 11:58 | |
*** suuuper has quit IRC | 11:59 | |
*** dsneddon has joined #tripleo | 12:03 | |
*** jbadiapa has joined #tripleo | 12:06 | |
*** amoralej is now known as amoralej|lunch | 12:07 | |
openstackgerrit | Martin Schuppert proposed openstack/tripleo-heat-templates stable/queens: [Queens] Run cellv2 host discovery via workflow or deploy_steps_tasks https://review.opendev.org/683069 | 12:07 |
*** jpich has quit IRC | 12:12 | |
*** jpich has joined #tripleo | 12:13 | |
*** dsneddon has quit IRC | 12:13 | |
openstackgerrit | Tom Barron proposed openstack/tripleo-heat-templates master: Fix Manila when run without pacemaker https://review.opendev.org/682853 | 12:15 |
*** gbarros has joined #tripleo | 12:16 | |
*** dsneddon has joined #tripleo | 12:19 | |
*** dsneddon has quit IRC | 12:25 | |
*** derekh has quit IRC | 12:27 | |
*** ansmith has joined #tripleo | 12:27 | |
*** jpena|lunch is now known as jpena | 12:29 | |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 12:30 |
openstackgerrit | Sagi Shnaidman proposed openstack/tripleo-ansible master: Add podman_container_facts https://review.opendev.org/686035 | 12:32 |
*** gbarros has quit IRC | 12:32 | |
*** jbadiapa has quit IRC | 12:35 | |
slaweq | chandankumar: hi | 12:37 |
slaweq | chandankumar: I have a question according Your last comment in https://review.opendev.org/#/c/655423/ | 12:39 |
slaweq | chandankumar: how I can check deployment type and choose tobiko_cloud_name value based on that? | 12:39 |
slaweq | chandankumar: is there any variable which says that deployment is standalone or multinode? | 12:40 |
*** rlandy has joined #tripleo | 12:40 | |
openstackgerrit | Merged openstack/tripleo-validations master: Update documentation URL link for CPU & RAM Undercloud recommendations https://review.opendev.org/685971 | 12:43 |
openstackgerrit | Gael Chamoulaud proposed openstack/tripleo-validations stable/stein: Update documentation URL link for CPU & RAM Undercloud recommendations https://review.opendev.org/686145 | 12:44 |
*** hjensas has joined #tripleo | 12:48 | |
*** amoralej|lunch is now known as amoralej | 12:51 | |
*** aakarsh has joined #tripleo | 12:52 | |
*** Goneri has joined #tripleo | 12:53 | |
*** mcornea has joined #tripleo | 12:53 | |
chandankumar | slaweq: Hello | 12:56 |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 12:57 |
openstackgerrit | Martin Mágr proposed openstack/puppet-tripleo master: Configuration changes to support Qdr-mesh topology. https://review.opendev.org/647908 | 12:58 |
*** dsneddon has joined #tripleo | 12:58 | |
slaweq | chandankumar: I think I found some way to get it | 12:58 |
chandankumar | slaweq: cool | 12:58 |
chandankumar | slaweq: you can add a value in fs override and reuse that | 12:59 |
slaweq | chandankumar: I'm checking if standalone.sh script exists, if not, I assume cloud name will be "overcloud" | 12:59 |
*** derekh has joined #tripleo | 12:59 | |
chandankumar | slaweq: for standalone it is 'standalone' for overcloud it is overcloud and for undercloud it is undercloud | 12:59 |
slaweq | chandankumar: what value I can add in fs override exactly? | 12:59 |
*** gbarros has joined #tripleo | 13:00 | |
openstackgerrit | Chandan Kumar (raukadah) proposed openstack/tripleo-ci master: Use tempest_cloud_name in featureset_override https://review.opendev.org/686149 | 13:03 |
chandankumar | slaweq: ^6 use this patch as depends-on and just below featureset_override add tempest_cloud_name: 'standalone' and reuse the var in tobiko playbook | 13:04 |
*** jaosorior has quit IRC | 13:08 | |
*** jpich has quit IRC | 13:14 | |
slaweq | chandankumar: thx, I will do in that way | 13:14 |
*** jpich has joined #tripleo | 13:15 | |
slaweq | chandankumar: one more thing, can You take a look at https://review.opendev.org/#/c/685784/ - it seems that in last patch I somehow missed to change this one name also :/ | 13:15 |
slaweq | and because of that multinode job with fs 016 is failing still for me :/ | 13:16 |
chandankumar | slaweq: checking | 13:16 |
slaweq | chandankumar: thx | 13:16 |
weshay | fultonj, and tom :) I need a little time to checkout the rhel containers | 13:19 |
weshay | 2019-10-02 08:36:41 | tripleo_common.image.exception.ImageNotFoundException: Not found image: docker://trunk.registry.rdoproject.org/tripleomaster/rhel-binary-cron:fd46c80632bbd7abc8411277bfc4819f6b95a485_bf6d0acd | 13:19 |
weshay | http://logs.rdoproject.org/53/682853/10/openstack-check/tripleo-ci-rhel-8-scenario004-standalone-rdo/d9a9074/logs/undercloud/home/zuul/standalone_deploy.log.txt.gz | 13:19 |
weshay | fultonj, to test w/ rhel 8.. I *think* putting your patch as a depends-on to: https://review.opendev.org/#/c/682853/ would be the right thing? | 13:20 |
openstackgerrit | Slawek Kaplonski proposed openstack/tripleo-quickstart-extras master: Add playbook and role to run tobiko https://review.opendev.org/655423 | 13:20 |
fultonj | weshay: or i could make a tht patch which depends on it | 13:22 |
weshay | aye | 13:22 |
fultonj | one that has no content if we want to test them separately | 13:22 |
fultonj | weshay: i'll make a tht review like that | 13:26 |
fultonj | if tbarron wants to add a depends on too, then we'd consider that an extra test | 13:26 |
*** toure|gone is now known as toure | 13:26 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/tripleo-common master: Implement Ansible fact cache for Mistral executor https://review.opendev.org/682855 | 13:27 |
*** mcornea has quit IRC | 13:27 | |
*** hjensas has quit IRC | 13:36 | |
*** mcornea has joined #tripleo | 13:42 | |
*** ykarel has joined #tripleo | 13:45 | |
*** zbr|lunch is now known as zbr|ruck | 13:48 | |
openstackgerrit | Martin Mágr proposed openstack/puppet-tripleo master: Configuration changes to support Qdr-mesh topology. https://review.opendev.org/647908 | 13:49 |
*** bdodd has joined #tripleo | 13:54 | |
*** jaosorior has joined #tripleo | 13:54 | |
*** jhakimra has joined #tripleo | 13:59 | |
bogdando | owalsh, Tengu: hi, some testing shows the lazy activation cannot be disabled just like that https://pastebin.com/tV87RfRC (tested for py27). Therefore no memory leaks was discovered as well. I wonder how to disable it properly and then repeat testing but for py3 perhaps?.. | 14:06 |
bogdando | wrt https://review.opendev.org/#/c/685698 | 14:06 |
fultonj | weshay: looks like https://review.opendev.org/#/c/682853 already depends on https://review.opendev.org/#/c/685449 | 14:06 |
*** panda has quit IRC | 14:06 | |
Tengu | bogdando: well, apparently that option will be removed. That's why I dropped the change. | 14:07 |
bogdando | would be nice to know more about other alternatives | 14:07 |
Tengu | bogdando: owalsh pointed one, should be good. | 14:08 |
*** panda has joined #tripleo | 14:08 | |
bogdando | as now we have a testing snippet to easily evaluate it against possible mem leaks, owalsh ^^ | 14:08 |
bogdando | WSGIImportScript ? | 14:08 |
owalsh | bogdando: yea, that is the only documented solution | 14:08 |
owalsh | I don't see how disabling lazy loading solves the problem, wsgi app stil isn't imported till the first requests IIUC | 14:10 |
bogdando | will add that to my testing then | 14:10 |
bogdando | owalsh: that's what I was observing right | 14:11 |
bogdando | :) | 14:11 |
openstackgerrit | Merged openstack/puppet-tripleo master: Add a configurable delay to Nova Evacuate calls https://review.opendev.org/684785 | 14:11 |
owalsh | also the mod_wsgi release notes say to never use it | 14:11 |
*** gkadam has joined #tripleo | 14:12 | |
*** gkadam has quit IRC | 14:12 | |
tbarron | weshay: I dunno why we are getting that image not found in the rhel8 job: 2019-10-02 08:36:41 | tripleo_common.image.exception.ImageNotFoundException: Not found image: docker://trunk.registry.rdoproject.org/tripleomaster/rhel-binary-cron:fd46c80632bbd7abc8411277bfc4819f6b95a485_bf6d0acd | 14:12 |
weshay | tbarron, aye.. that's on me and arxcruz|ruck right now | 14:12 |
tbarron | weshay: I think other rhel8 scenario jobs are getting it too | 14:13 |
weshay | one fire at a time.. we'll get through this | 14:13 |
tbarron | weshay: k, just checking :) | 14:13 |
weshay | tbarron, aye.. it's not job specific | 14:13 |
weshay | it's the busy season :) | 14:13 |
tbarron | :) | 14:13 |
*** jpich has quit IRC | 14:18 | |
*** jaosorior has quit IRC | 14:18 | |
*** jpich has joined #tripleo | 14:18 | |
bogdando | owalsh, Tengu: so I've tried WSGIImportScript and it changed nothing | 14:21 |
bogdando | nothing loaded | 14:21 |
bogdando | the only option left is redesign health checks to be active and send a wake-up requests | 14:21 |
bogdando | to the API port | 14:22 |
owalsh | bogdando: really? What did you set WSGIImportScript to? | 14:22 |
bogdando | WSGIImportScript /var/www/cgi-bin/heat/heat_api | 14:22 |
bogdando | where I put my mem-alloc tester | 14:23 |
owalsh | bogdando: needs more than that | 14:23 |
openstackgerrit | Sorin Sbarnea proposed openstack/ansible-role-container-registry master: Disable docker iptables support https://review.opendev.org/686173 | 14:23 |
bogdando | mhm | 14:23 |
openstackgerrit | Brent Eagles proposed openstack/tripleo-heat-templates master: Add Octavia driver agent service https://review.opendev.org/658118 | 14:23 |
* bogdando back to rtfms | 14:23 | |
*** openstackstatus has joined #tripleo | 14:29 | |
*** ChanServ sets mode: +v openstackstatus | 14:29 | |
bogdando | owalsh: tried https://pastebin.com/PcLx516S but nothing changes | 14:30 |
*** rcernin has quit IRC | 14:30 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/tripleo-common master: Implement Ansible fact cache for Mistral executor https://review.opendev.org/682855 | 14:35 |
openstackgerrit | Sorin Sbarnea proposed openstack/ansible-role-container-registry master: Disable docker iptables support https://review.opendev.org/686173 | 14:35 |
owalsh | bogdando: could be that heat wsgi module isn't doing the right thing? | 14:35 |
bogdando | owalsh: it allocates memory by a request, so my substituted code works | 14:36 |
bogdando | it also returns the expected output (the name of process group) | 14:36 |
owalsh | bogdando: not sure what you mean, what substituted code? | 14:37 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/tripleo-common master: Implement Ansible fact cache for Mistral executor https://review.opendev.org/682855 | 14:37 |
bogdando | did you look through the testing results? | 14:37 |
bogdando | owalsh: I'm using a mocked WSGI app | 14:37 |
*** yprokule has quit IRC | 14:37 | |
owalsh | bogdando: what testing results? | 14:38 |
bogdando | owalsh: https://gist.githubusercontent.com/bogdando/43a6c5938c964fa4ee94f714e0638ee5/raw/6065e9ce1b146ea9cf3bf7de064f5217efaa3a10/gistfile1.txt and results are here: | 14:38 |
bogdando | 16:06:19 - bogdando: owalsh, Tengu: hi, some testing shows the lazy activation cannot be disabled just like that https://pastebin.com/tV87RfRC (tested for py27). […] | 14:38 |
*** jpich has quit IRC | 14:38 | |
owalsh | bogdando: you're wsgi app does nothing on import | 14:39 |
owalsh | s/you're/your | 14:39 |
*** jpich has joined #tripleo | 14:39 | |
bogdando | owalsh: dunno, it works | 14:39 |
bogdando | could you look through results first :) ? | 14:39 |
owalsh | bogdando: yea, but when WSGIImportScript imports it it needs to do the init | 14:39 |
bogdando | it keeps memory allocated | 14:39 |
bogdando | ugh | 14:40 |
openstackgerrit | John Fulton proposed openstack/tripleo-quickstart master: Use custom ceph-ansible until c8 storage sig ready https://review.opendev.org/685449 | 14:40 |
bogdando | that's getting complicated to keep testing it right... | 14:40 |
openstackgerrit | Sagi Shnaidman proposed openstack/tripleo-ansible master: Add podman_container_info https://review.opendev.org/686035 | 14:40 |
bogdando | I give up :) | 14:40 |
zbr|ruck | mwhahaha: if fixed iptrables order, but i have a question: do we really need to do this for all installs? i personally don't really like that customization. | 14:41 |
mwhahaha | zbr|ruck: no we don't. this is happening because you're doing this outside of tripleo. this is a tripleo ci specific thing | 14:41 |
mwhahaha | zbr|ruck: this should be handled elsewhere or configurable in the install | 14:42 |
mwhahaha | We actually already do this in that role | 14:42 |
mwhahaha | https://opendev.org/openstack/ansible-role-container-registry/src/branch/master/tasks/docker.yml#L22-L28 | 14:43 |
* mwhahaha doesn't exactly have the correct context for this entire issue | 14:43 | |
fultonj | tbarron: now that the depends-on of your patch has been udpated, i think you need to rebase https://review.opendev.org/#/c/682853/ | 14:44 |
mwhahaha | i'd just move it and make it an option. eg. disable_iptables_management | 14:44 |
mwhahaha | or something | 14:44 |
openstackgerrit | Tom Barron proposed openstack/tripleo-heat-templates master: Fix Manila when run without pacemaker https://review.opendev.org/682853 | 14:44 |
tbarron | fultonj: done | 14:45 |
owalsh | bogdando: if you look at the regular /var/www/cgi-bin/heat/heat_api it does it | 14:46 |
owalsh | f application is None: | 14:46 |
owalsh | application = init_application() | 14:46 |
openstackgerrit | Daniel King proposed openstack/python-tripleoclient master: Removed references to bm-deploy images https://review.opendev.org/686064 | 14:46 |
bogdando | owalsh: I'd appreciate if you could fix it for me :) | 14:46 |
*** Vorrtex has joined #tripleo | 14:47 | |
bogdando | owalsh: I'll try WSGIScriptAlias | 14:48 |
bogdando | just reading https://stackoverflow.com/questions/20790077/django-startup-code-with-wsgi-lazy-loading | 14:48 |
ade_lee | mwhahaha, hey - when trying to debug by running container-puppet.py, its failing with message "Missing CNI default network" | 14:48 |
ade_lee | mwhahaha, any idea how to get past that? | 14:48 |
mwhahaha | --net=host | 14:48 |
zbr|ruck | mwhahaha: wow, i discovered that we already had this in container_registry_docker_options in default, which should have worked anyway. | 14:49 |
zbr|ruck | weshay: ^ mainly we are talking about a change that is already in the role. where do we call the role from? | 14:50 |
mwhahaha | so docker.yaml is for the registry setup | 14:50 |
mwhahaha | which should be before the docker install | 14:50 |
mwhahaha | at least it is in tripleo deploys | 14:51 |
ade_lee | mwhahaha, so according to doc, we are running "SHOW_DIFF=True PROCESS_COUNT=1 CONFIG=glance_api.json ./container-puppet.py" , we need to add --net=host ? | 14:52 |
mwhahaha | no the container needs a --net=host when it's launched | 14:52 |
openstackgerrit | John Fulton proposed openstack/tripleo-quickstart master: Use custom ceph-ansible until c8 storage sig ready https://review.opendev.org/685449 | 14:52 |
mwhahaha | in THT that's usually net: host in the definition | 14:52 |
zbr|ruck | if i discover that this was caused by use of tasks_from... i am going to the first pub and not return till tomorrow. | 14:52 |
mwhahaha | zbr|ruck: enjoy your beer then | 14:52 |
zbr|ruck | not sure what is your oppinion re use of tasks_from with roles because I see it like using "jmp" in C code.... | 14:53 |
zbr|ruck | no way to control the entry points | 14:53 |
owalsh | bogdando: see the last comment here https://stackoverflow.com/questions/20958553/wsgi-startup-why-two-identical-processes | 14:53 |
openstackgerrit | Tom Barron proposed openstack/tripleo-heat-templates master: Fix Manila when run without pacemaker https://review.opendev.org/682853 | 14:55 |
hrybacki | weshay: do you happen to recall where in oooq/oooq-e the undercloud gets enrolled to the IPA server? I'm grooming through the repos but I can't find an `ipa-client-install` anywhere | 14:56 |
bogdando | owalsh: thanks, interesting | 14:59 |
owalsh | bogdando: there are some caveats when setting application_group though | 14:59 |
bogdando | but it seems it needs to be non-empty | 15:00 |
owalsh | bogdando: really, it's set to %{GLOBAL} on my deployment | 15:01 |
owalsh | which means '' | 15:01 |
bogdando | right | 15:01 |
bogdando | now I'm editing it directly in /etc/httpd/conf.d/10-heat_api_wsgi.conf | 15:01 |
bogdando | but it has yet worked for me | 15:01 |
openstackgerrit | Tom Barron proposed openstack/tripleo-heat-templates master: Fix Manila when run without pacemaker https://review.opendev.org/682853 | 15:02 |
mwhahaha | weshay: i figured out how to speed up container processing even more btw. | 15:03 |
weshay | mwhahaha++ how's that? | 15:03 |
mwhahaha | magic | 15:03 |
mwhahaha | rm -rf /* and leaving | 15:03 |
weshay | mwhahaha, and cloudnull I think we need to figure out how to make the case w/ infra about the proxies | 15:03 |
weshay | mwhahaha, :) | 15:03 |
mwhahaha | the proxies are likely fine | 15:04 |
weshay | hrm.. why the huge diff per provider then? | 15:04 |
weshay | location? | 15:04 |
owalsh | bogdando: what does worked mean? | 15:04 |
mwhahaha | few different reasons. let me post a patch and we can talk | 15:04 |
bogdando | it it does not start | 15:04 |
bogdando | although I have no ideas for the proper initialization in the wsgi app , what you've mentioned earlier | 15:05 |
owalsh | bogdando: don't see how you can tell if that wsgi app was imported at process start or now. If you add a call to application() at the end maybe then you'll see the memory being allocated on import | 15:05 |
bogdando | just dumb print works for some folks | 15:05 |
weshay | zbr|ruck, how can I help? | 15:07 |
zbr|ruck | https://bluejeans.com/2655417928 | 15:07 |
openstackgerrit | Alex Schultz proposed openstack/tripleo-common master: Switch to ProcessPoolExecutor https://review.opendev.org/686181 | 15:08 |
mwhahaha | weshay, EmilienM, bogdando: -^ | 15:08 |
mwhahaha | cut my processing down to 10 mins | 15:08 |
mwhahaha | we started with 32 in my env | 15:08 |
mwhahaha | 22 after randomizing | 15:08 |
mwhahaha | 10 after ProcessPoolExecutor | 15:08 |
EmilienM | nicely done | 15:08 |
mwhahaha | next we can tweak the workers for that and we might get more | 15:08 |
Tengu | wow. | 15:10 |
*** Vorrtex has quit IRC | 15:10 | |
*** jpich has quit IRC | 15:11 | |
*** jpich has joined #tripleo | 15:12 | |
ykarel | mwhahaha, re. ^^ is the issue only with py2? or both py2/py3 | 15:14 |
ykarel | i noticed in rhel8 container prep was faster | 15:14 |
EmilienM | i think it's the same for both py2/py3 | 15:15 |
EmilienM | (on multiprocess vs thread) | 15:15 |
ykarel | in rhel8 i saw container prep was taking around 5 minutes | 15:15 |
EmilienM | in which infra? | 15:15 |
ykarel | so i thought the performance was related to py2 only | 15:15 |
ykarel | rdo | 15:15 |
ykarel | i also thought it can be related to subprocess perf on py2, but can be other reasons as well | 15:17 |
mwhahaha | might be mostly py2 | 15:18 |
mwhahaha | i haven't been testing under py3 | 15:18 |
mwhahaha | i think we'll see improvements in both | 15:18 |
openstackgerrit | Merged openstack/tripleo-validations stable/stein: New validation: detect failed containers https://review.opendev.org/686093 | 15:18 |
mwhahaha | since CI is currently py2, it's the primary target :D | 15:19 |
mwhahaha | py3 likely has better stream processing | 15:19 |
*** gfidente has quit IRC | 15:20 | |
*** mcornea has quit IRC | 15:21 | |
*** dtrainor has quit IRC | 15:21 | |
ykarel | mwhahaha, okk | 15:22 |
ykarel | mwhahaha, EmilienM just for reference:- https://logs.rdoproject.org/openstack-periodic-master/opendev.org/openstack/tripleo-ci/master/periodic-tripleo-ci-rhel-8-scenario002-standalone-master/164c330/logs/undercloud/home/zuul/standalone_deploy.log.txt.gz#_2019-10-01_19_29_13 | 15:22 |
ykarel | https://logs.rdoproject.org/openstack-periodic-master/opendev.org/openstack/tripleo-ci/master/periodic-tripleo-ci-rhel-8-scenario003-standalone-master/a289723/logs/undercloud/home/zuul/standalone_deploy.log.txt.gz#_2019-10-01_19_28_23 | 15:23 |
mwhahaha | standalone tends to be faster than undercloud i think | 15:24 |
mwhahaha | i think it's less containers | 15:24 |
mwhahaha | but yea that's definately faster | 15:24 |
openstackgerrit | Bogdan Dobrelya (bogdando) proposed openstack/tripleo-quickstart master: Fix secondary (overcloud) subnode indexes https://review.opendev.org/682346 | 15:25 |
openstackgerrit | Bogdan Dobrelya (bogdando) proposed openstack/tripleo-ci master: Fallback to public IPv4 when there is no priv. IP https://review.opendev.org/682092 | 15:25 |
mwhahaha | buildah is probably also better on rhel8 because of the correct kernel | 15:25 |
mwhahaha | so in general, it's going to have better perf for a lot of things | 15:25 |
bogdando | owalsh: I've tried wsgi for a few hours only and already hate it by all my soul :D | 15:26 |
owalsh | owalsh: lets go back to cgi scripts :-) | 15:26 |
ykarel | mwhahaha, same data for undercloud/overcloud as well:- https://logs.rdoproject.org/openstack-periodic-master/opendev.org/openstack/tripleo-ci/master/periodic-tripleo-ci-rhel-8-ovb-3ctlr_1comp-featureset001-master/a000b39/logs/undercloud/home/zuul/undercloud_install.log.txt.gz#_2019-09-15_09_58_31 | 15:26 |
ykarel | around 5 minutes | 15:26 |
bogdando | yea | 15:26 |
mwhahaha | ykarel: yea buildah tasks 10-15 seconds instead for like 30-60 on centos. though i'm not seeing any actual yum updates occuring | 15:26 |
mwhahaha | who knows :D | 15:27 |
bogdando | mwhahaha, ykarel: weshay said periodic jobs never update it | 15:27 |
ykarel | mwhahaha, yes yum update not happening as dnf strict fails if repo doesn't exist | 15:27 |
bogdando | so there is always an empty run | 15:27 |
mwhahaha | yea so it's apples to ornages on that then | 15:27 |
ykarel | mwhahaha, to rule out yum_update, i tried disabling yum_update, but no improvement with centos | 15:28 |
mwhahaha | all my testing is without yum update | 15:28 |
mwhahaha | so just the base fetching | 15:28 |
bogdando | owalsh: my best state of testing it now auto starts but fails with bad python code it seems I have no idea how to debug | 15:28 |
mwhahaha | yum update adds a bunch of time. in my testing it was adding back the 10 mins i gained from shuffling the list | 15:28 |
bogdando | 500 Internal Server Error :/ | 15:28 |
openstackgerrit | Carlos Goncalves proposed openstack/tripleo-common stable/stein: Fixup octavia-api healthcheck https://review.opendev.org/686186 | 15:29 |
bogdando | mwhahaha: my profiling shows it's always like 11-23 min for 20-50 packages updated per container | 15:29 |
*** aakarsh has quit IRC | 15:29 | |
ykarel | bogdando, mwhahaha so to rule out yum_update.sh i tried with https://review.rdoproject.org/r/#/c/22366/ | 15:30 |
bogdando | but the most weird was there is always a delay of a 45m -1h before yum updating the last two containers o.O | 15:30 |
ykarel | but not much improvement with centos | 15:30 |
*** gbarros has quit IRC | 15:30 | |
ykarel | and rhel8 was always faster w or without it | 15:30 |
* mwhahaha waits for centos8 to get on with it | 15:31 | |
Tengu | an official image for centos-8 is in the QA pipe afaik. | 15:31 |
openstackgerrit | Carlos Goncalves proposed openstack/tripleo-common stable/rocky: Fixup octavia-api healthcheck https://review.opendev.org/686189 | 15:32 |
openstackgerrit | Carlos Goncalves proposed openstack/tripleo-common stable/queens: Fixup octavia-api healthcheck https://review.opendev.org/686190 | 15:32 |
*** marios has quit IRC | 15:35 | |
*** mcornea has joined #tripleo | 15:36 | |
*** gbarros has joined #tripleo | 15:36 | |
*** iurygregory has quit IRC | 15:37 | |
openstackgerrit | Merged openstack/tripleo-quickstart master: Switching default centos url for image promotion https://review.opendev.org/685709 | 15:38 |
openstackgerrit | Merged openstack/paunch master: Properly mock cpu_affinity on all unit tests https://review.opendev.org/686100 | 15:38 |
*** tosky_ has joined #tripleo | 15:41 | |
*** tosky is now known as Guest83058 | 15:41 | |
*** tosky_ is now known as tosky | 15:41 | |
*** Guest83058 has quit IRC | 15:43 | |
openstackgerrit | Merged openstack/ansible-role-tripleo-modify-image master: Enable saving of rpms https://review.opendev.org/685806 | 15:47 |
openstackgerrit | Sorin Sbarnea proposed openstack/tripleo-quickstart-extras master: Disable docker iptables https://review.opendev.org/686192 | 15:48 |
*** mdbooth has quit IRC | 15:50 | |
*** aakarsh has joined #tripleo | 15:50 | |
*** dtrainor has joined #tripleo | 15:51 | |
openstackgerrit | Alex Schultz proposed openstack/tripleo-common master: Use tenancity to backoff when layer is locked https://review.opendev.org/686193 | 15:51 |
*** mdbooth has joined #tripleo | 15:51 | |
*** ekultails has quit IRC | 15:54 | |
*** ekultails has joined #tripleo | 15:58 | |
*** ykarel is now known as ykarel|away | 16:01 | |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP - Introduce tripleo-container-manage role https://review.opendev.org/686196 | 16:02 |
*** florianf has quit IRC | 16:03 | |
*** jfrancoa has quit IRC | 16:04 | |
openstackgerrit | Carlos Camacho proposed openstack/tripleo-upgrade master: Introduce molecule tests in the tripleo-upgrade role https://review.opendev.org/685041 | 16:07 |
openstackgerrit | Jeremy Stanley proposed openstack/diskimage-builder master: Add security mirror override for debian-minimal https://review.opendev.org/686199 | 16:09 |
*** gfidente has joined #tripleo | 16:09 | |
openstackgerrit | Luke Short proposed openstack/python-tripleoclient master: [WIP] Handle a deployment that was manually cancelled. https://review.opendev.org/686202 | 16:10 |
*** pcaruana has quit IRC | 16:10 | |
*** pcaruana has joined #tripleo | 16:11 | |
bogdando | owalsh: I ended up just patching the original heat wsgi app | 16:12 |
bogdando | with that allocator | 16:12 |
bogdando | so now it works as usual but just allocates 1G per worker | 16:13 |
bogdando | and whatever options I played, it never auto loads it after container restart ... | 16:13 |
bogdando | owalsh: here is the end configs state http://paste.openstack.org/show/kycalFEyFNAsgVyFJjLN/ | 16:14 |
Tengu | bogdando: "For some cases, such as when WSGIScriptAlias is being used, it is actually possible to preload the application code when the processes first starts, rather than when the first request arrives. To preload an application see the WSGIImportScript directive." | 16:15 |
Tengu | bogdando: the "SGIScriptAlias" is a typo right? | 16:16 |
openstackgerrit | Merged openstack/tripleo-validations stable/stein: Constrain molecule https://review.opendev.org/686119 | 16:16 |
bogdando | yeah, fixed that | 16:17 |
bogdando | wow, it worked if moved to the main 10-* script | 16:17 |
bogdando | never works if placed in the modules config | 16:17 |
Tengu | order does matter | 16:17 |
*** aakarsh has quit IRC | 16:17 | |
bogdando | Tengu, owalsh: so I re-tested it with autoload, no memory leaks | 16:18 |
bogdando | I think we can use that approach then | 16:19 |
bogdando | Tengu: just amend 10-heat_api_wsgi.conf instead | 16:19 |
Tengu | bogdando: do you have a way to inject that autoload in t-h-t or puppet? not 100% sure hiera is the place to be | 16:19 |
Tengu | ah? wait. | 16:19 |
bogdando | yea | 16:19 |
bogdando | just that file | 16:19 |
bogdando | gtg see you tomorrow | 16:20 |
Tengu | where/how is it generated? that's the question. in puppet, yes. but does it take things out of t-h-t? | 16:20 |
bogdando | that's prolly puppet-foo for each | 16:20 |
Tengu | hmmmmm | 16:20 |
*** jfrancoa has joined #tripleo | 16:20 | |
Tengu | MAYBE heat::wsgi::apache_api:: can be the entry within heat. | 16:21 |
openstackgerrit | Jeremy Stanley proposed openstack/diskimage-builder master: Add security mirror override for debian-minimal https://review.opendev.org/686199 | 16:21 |
Tengu | need to dig a bit in the puppet foo. | 16:21 |
mwhahaha | it comes from puppet-apache | 16:21 |
mwhahaha | puppet-heat just wraps the apache stuff | 16:21 |
*** rlandy is now known as rlandy|brb | 16:22 | |
Tengu | mwhahaha: meaning we need to edit every single puppet module then ? no way to inject that from the mighty puppet-tripleo? | 16:22 |
mwhahaha | likely | 16:22 |
Tengu | here we're talking about heat, but nova and friends will (probably) need that change as well. | 16:22 |
mwhahaha | https://github.com/openstack/puppet-heat/blob/master/manifests/wsgi/apache.pp#L140 | 16:23 |
mwhahaha | might be able to patch openstacklib | 16:23 |
Tengu | pfrrttt | 16:23 |
Tengu | ah, would be best. | 16:23 |
mwhahaha | https://github.com/openstack/puppet-openstacklib/blob/master/manifests/wsgi/apache.pp#L288 | 16:24 |
Tengu | yep. | 16:24 |
Tengu | I'm in the file. | 16:24 |
mwhahaha | basically just an apache::vhost, so if that exposes what you are looking for then you can put it in there | 16:24 |
mwhahaha | or do something else | 16:24 |
Tengu | yeah. we might want to make that optional though. | 16:24 |
bogdando | it should be alrady duable via custom_wsgi_process_options | 16:25 |
Tengu | so it might be good to add a new param. Need to talk with owalsh when he's connected (not sure about his TZ). | 16:25 |
bogdando | doable* | 16:25 |
Tengu | bogdando: lemme check. | 16:25 |
bogdando | or rather vhost_custom_fragment | 16:25 |
Tengu | apache::vhost::WSGIImportScript | 16:26 |
bogdando | nice | 16:26 |
Tengu | we might get a shortcut in the module. | 16:26 |
bogdando | handing that over to you folks :) | 16:26 |
bogdando | cu | 16:26 |
Tengu | just need to ensure it's present in our version. | 16:26 |
Tengu | yeah, won't do that today ;). | 16:26 |
Tengu | still here due to a late mtg, but that's it :D | 16:26 |
*** tesseract has quit IRC | 16:29 | |
*** bogdando has quit IRC | 16:30 | |
openstackgerrit | Merged openstack/tripleo-heat-templates stable/stein: nova-libvirt: set 'cpuset_cpus' to 'all' https://review.opendev.org/686025 | 16:37 |
*** rlandy|brb is now known as rlandy | 16:37 | |
*** jpich has quit IRC | 16:37 | |
*** tosky_ has joined #tripleo | 16:40 | |
*** aakarsh has joined #tripleo | 16:40 | |
*** tosky has quit IRC | 16:43 | |
*** derekh has quit IRC | 16:43 | |
*** rfolco is now known as rfolco|dentist | 16:44 | |
*** tosky_ is now known as tosky | 16:45 | |
owalsh | Tengu: I'm GMT+! | 16:48 |
owalsh | err GMT+1 | 16:48 |
*** lmiccini has quit IRC | 16:58 | |
*** dsneddon has quit IRC | 16:58 | |
openstackgerrit | Kamil Sambor proposed openstack/tripleo-heat-templates master: Support connecting OVN DB over SSL https://review.opendev.org/680345 | 17:01 |
*** holser has quit IRC | 17:02 | |
openstackgerrit | Kamil Sambor proposed openstack/tripleo-heat-templates master: Support connecting OVN DB over SSL https://review.opendev.org/680345 | 17:06 |
*** dtrainor has quit IRC | 17:07 | |
Tengu | owalsh: ah, +2 here (for now) :D | 17:11 |
*** dtrainor has joined #tripleo | 17:11 | |
Tengu | owalsh: can we talk a bit more about that wsgi thingy and how to implement it within puppet-openstacklib tomorrow? | 17:11 |
owalsh | Tengu: sure | 17:11 |
Tengu | owalsh: good! signing off for today. | 17:12 |
Tengu | have a great evening/night ;). | 17:12 |
owalsh | you too | 17:12 |
*** gfidente has quit IRC | 17:22 | |
*** ykarel|away has quit IRC | 17:23 | |
openstackgerrit | Merged openstack/tripleo-validations master: Add new healthcheck-service-status validation https://review.opendev.org/682377 | 17:24 |
*** jpena is now known as jpena|off | 17:27 | |
*** dsneddon has joined #tripleo | 17:29 | |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP - Introduce tripleo-container-manage role https://review.opendev.org/686196 | 17:31 |
*** jhakimra has quit IRC | 17:33 | |
*** dsneddon has quit IRC | 17:37 | |
*** amoralej is now known as amoralej|off | 17:40 | |
*** dsneddon has joined #tripleo | 17:41 | |
*** pcaruana has quit IRC | 17:45 | |
openstackgerrit | Luke Short proposed openstack/python-tripleoclient master: Handle a deployment that was manually cancelled. https://review.opendev.org/686202 | 17:52 |
*** dsneddon has quit IRC | 17:53 | |
openstackgerrit | Merged openstack/tripleo-heat-templates master: Remove leftovers after NeutronExternalNetworkBridge option https://review.opendev.org/683380 | 17:53 |
openstackgerrit | Luke Short proposed openstack/python-tripleoclient master: Handle a deployment that was manually cancelled. https://review.opendev.org/686202 | 17:54 |
*** ykarel|away has joined #tripleo | 17:57 | |
openstackgerrit | Sagi Shnaidman proposed openstack/tripleo-ansible master: Add podman_volume_info ansible module https://review.opendev.org/686233 | 18:13 |
*** aakarsh has quit IRC | 18:13 | |
openstackgerrit | Sagi Shnaidman proposed openstack/tripleo-ansible master: Add podman_volume_info ansible module https://review.opendev.org/686233 | 18:13 |
*** dsneddon has joined #tripleo | 18:15 | |
*** ykarel|away has quit IRC | 18:16 | |
openstackgerrit | Merged openstack/tripleo-common stable/queens: [Queens] Add workflow to do cellv2 host discovery https://review.opendev.org/683067 | 18:16 |
dking_desktop | I'm attempting to deploy an overcloud. The last message, just under an hour ago was [overcloud.Controller.0.DeploymentActions]: CREATE_COMPLETE state changed. How can I check on the status? | 18:18 |
EmilienM | openstack overcloud status | 18:20 |
*** iurygregory has joined #tripleo | 18:21 | |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP - Introduce tripleo-container-manage role https://review.opendev.org/686196 | 18:23 |
openstackgerrit | Raildo Mascena proposed openstack/tripleo-quickstart-extras master: Add keycloak-setup role https://review.opendev.org/671262 | 18:31 |
*** cdearborn has joined #tripleo | 18:33 | |
*** alexmcleod has quit IRC | 18:33 | |
dking_desktop | EmilienM: Thank you. It says DEPLOYING. I supplse that means that I should just wait for it to finish, and that it's just taking an abnormally long time? | 18:33 |
weshay | cloudnull, hey.. looks like tripleo-bootstrap network config does not work w/ rhel 8 | 18:35 |
weshay | http://logs.rdoproject.org/93/22793/1/check/periodic-tripleo-ci-rhel-8-ovb-3ctlr_1comp-featureset001-master/6b8af1e/logs/undercloud/home/zuul/overcloud_deploy.log.txt.gz | 18:35 |
weshay | network.service is not a native service, | 18:35 |
ekultails | weshay: He's on a business trip. I can have a look. | 18:37 |
weshay | ekultails, hey cool.. which is weird.. maybe the ansible module is busted or not updated enough for rhel8 | 18:37 |
weshay | network.service loaded inactive dead LSB: Bring up/down networking | 18:37 |
weshay | NetworkManager.service loaded active running Network Manager | 18:38 |
weshay | http://logs.rdoproject.org/93/22793/1/check/periodic-tripleo-ci-rhel-8-ovb-3ctlr_1comp-featureset001-master/6b8af1e/logs/undercloud/var/log/extra/services.txt.gz | 18:38 |
weshay | ekultails, we can probably just loop on network vs. networkManager | 18:39 |
weshay | like we do https://opendev.org/openstack/tripleo-ansible/src/branch/master/tripleo_ansible/roles/tripleo-bootstrap/tasks/main.yml#L74 | 18:39 |
ekultails | I imagine we have it hardcoded in the role to use the network service instead of NetworkManager. We can have a check for the distribution. If EL7 then use network, elseif EL8||Fedora then use NetworkManager. | 18:40 |
weshay | aye | 18:41 |
openstackgerrit | Brent Eagles proposed openstack/tripleo-heat-templates stable/stein: Add NeutronPermittedEthertypes on OVS agent https://review.opendev.org/686245 | 18:42 |
weshay | wrote it up https://bugs.launchpad.net/tripleo/+bug/1846391 | 18:42 |
openstack | Launchpad bug 1846391 in tripleo "tripleo-bootstrap network service fails on rhel 8" [Critical,Triaged] | 18:42 |
ekultails | Our roles have functionality to include distro specific variables, too. Looks like they finally get to be useful. :-) | 18:43 |
ekultails | https://github.com/openstack/tripleo-ansible/blob/master/tripleo_ansible/roles/tripleo-bootstrap/tasks/main.yml#L26-31 | 18:43 |
ekultails | weshay: Cool, thanks for writing that bug report up. I'll get a patch out soon. | 18:44 |
openstackgerrit | Brent Eagles proposed openstack/tripleo-heat-templates stable/rocky: Add NeutronPermittedEthertypes on OVS agent https://review.opendev.org/686246 | 18:49 |
openstackgerrit | Brent Eagles proposed openstack/tripleo-heat-templates stable/queens: Add NeutronPermittedEthertypes on OVS agent https://review.opendev.org/686247 | 18:52 |
openstackgerrit | Luke Short proposed openstack/tripleo-ansible master: Only use the network service for EL 7. https://review.opendev.org/686248 | 18:55 |
ekultails | weshay: ^ That should do the trick. | 18:56 |
weshay | ekultails, k.. testing here https://review.rdoproject.org/r/#/c/22793/ | 18:57 |
*** holser has joined #tripleo | 19:02 | |
EmilienM | dking_desktop: very likely | 19:05 |
*** kopecmartin is now known as kopecmartin|off | 19:05 | |
dking_desktop | Well, now, it's almost been an hour and a half. The baremetal nodes I'm using are powered off. Is that normal during that deploy step? | 19:08 |
mwhahaha | dking_desktop: i seem to have lost that patch you posted the other day, can you send that to me again so i can review? | 19:08 |
*** spsurya has quit IRC | 19:09 | |
dking_desktop | mwhahaha: Sure: https://review.opendev.org/686064 | 19:10 |
mwhahaha | thanks | 19:10 |
mwhahaha | seems like the tests are fine | 19:10 |
dking_desktop | Unfortunately, even though it passed the tox unit tests and pep8 now (I had to shorten a couple lines), it's now failing a CI. | 19:10 |
dking_desktop | Oh? What about tripleo-ci-centos-7-scenario000-multinode-oooq-container-updates ? | 19:11 |
mwhahaha | likely unrelated but i'll look | 19:11 |
mwhahaha | we don't actually exercise this code in ci (clearly) | 19:11 |
mwhahaha | yea un related | 19:13 |
*** ykarel has joined #tripleo | 19:14 | |
dking_desktop | Okay. That makes me feel a little better. I wasn't quite sure how to do that test locally. It looks like it's an Ansible playbook run, but I hadn't been able to find the details in the logs yet. | 19:15 |
dking_desktop | I'm busy trying to get this overcloud deploy to work, but I can't tell if it's stuck. There's been no output in over an hour. The status is still DEPLOYING, and I see that nova processes are getting some commands, but I don't know enough yet to troubleshoot. | 19:16 |
mwhahaha | what the last thing outputed from the deploy? | 19:17 |
mwhahaha | if it's in the ansible section, you can also check the ansible log in /var/lib/mistral/overcloud/ansible.log (or something like that) | 19:17 |
dking_desktop | [overcloud.Controller.0.DeploymentActions]: CREATE_COMPLETE state changed | 19:18 |
mwhahaha | check openstack server list | 19:18 |
mwhahaha | are all the servers provisioned? | 19:18 |
dking_desktop | http://paste.openstack.org/show/780766/ | 19:20 |
dking_desktop | I just noticed that they both have the baremetal flavor, where I would have thought they would have had the compute and control flavors. | 19:21 |
openstackgerrit | Arx Cruz proposed openstack/tripleo-ansible master: Only use the network service for EL 7. https://review.opendev.org/686255 | 19:24 |
*** ccamacho has joined #tripleo | 19:24 | |
mwhahaha | weshay, arxcruz|ruck: is scenario010 broken? https://review.opendev.org/#/c/683735/ | 19:24 |
weshay | containers-multinode | 19:25 |
* weshay looks | 19:25 | |
mwhahaha | weshay: not featureset 10 | 19:25 |
mwhahaha | scenario10 | 19:25 |
weshay | oh gotcha | 19:25 |
weshay | mwhahaha, yup.. looks like to broke yesterday | 19:26 |
weshay | tripleo-ci-centos-7-scenario010-standaloneopenstack/tripleo-commonmastercheck678206,767112019-10-01T20:47:51 | 19:26 |
weshay | http://zuul.openstack.org/builds?job_name=tripleo-ci-centos-7-scenario010-standalone | 19:26 |
weshay | dang it | 19:26 |
mwhahaha | what is 10 anyway? octavia? | 19:26 |
mwhahaha | or is that 12 | 19:26 |
mwhahaha | TASK [octavia-controller-config : get management port net mask broke | 19:27 |
*** ccamacho has quit IRC | 19:28 | |
ykarel | omg that scenario010 job is voting and not in promotion criteria | 19:29 |
ykarel | things started happening just after promotion | 19:29 |
mwhahaha | plz switch to nonvoting | 19:29 |
ykarel | and it's not even running in promotion pipeline | 19:29 |
weshay | 2019-10-02 07:21:22 | TASK [Configure octavia on overcloud] ****************************************** | 19:30 |
weshay | awesomeness | 19:30 |
* weshay gets on it | 19:30 | |
ykarel | ack | 19:30 |
*** ykarel is now known as ykarel|away | 19:30 | |
openstackgerrit | wes hayutin proposed openstack/tripleo-ci master: move standalone-scenario010 to non-voting https://review.opendev.org/686258 | 19:33 |
weshay | https://review.opendev.org/#/c/652089/ | 19:34 |
ykarel|away | weshay, good to change at job definition level | 19:35 |
ykarel|away | line 521 | 19:35 |
ykarel|away | as i see that job is used in tripleo-ansible as well | 19:35 |
openstackgerrit | Jaganathan Palanisamy proposed openstack/tripleo-common stable/queens: NeutronPhysnetNUMANodesMapping parameter issue https://review.opendev.org/686259 | 19:36 |
dking_desktop | mwhahaha: Would you have any tips on how I can troubleshoot this deploy? | 19:38 |
mwhahaha | dking_desktop: check nova/ironic logs | 19:38 |
mwhahaha | dking_desktop: looks like it's not provisioning for whatever reason | 19:38 |
dking_desktop | Should the nodes be powered on? | 19:39 |
mwhahaha | it should do that itself unless there are ipmi erorr or something | 19:39 |
openstackgerrit | James Slagle proposed openstack/tripleo-common master: Raise exception when no IPs found https://review.opendev.org/686261 | 19:40 |
openstackgerrit | wes hayutin proposed openstack/tripleo-ansible master: scenario010 nonvoting https://review.opendev.org/686262 | 19:40 |
*** holser has quit IRC | 19:40 | |
openstackgerrit | James Slagle proposed openstack/tripleo-validations master: Fix Python 3 incomaptibility in tripleo-ansible-inventory https://review.opendev.org/686263 | 19:42 |
openstackgerrit | James Slagle proposed openstack/tripleo-validations master: Fix Python 3 incompatibility in tripleo-ansible-inventory https://review.opendev.org/686263 | 19:42 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: Remove podman_image* https://review.opendev.org/686265 | 19:45 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP: POC of podman_container https://review.opendev.org/682458 | 19:45 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP - Introduce tripleo-container-manage role https://review.opendev.org/686196 | 19:45 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: Add podman_container_info https://review.opendev.org/686035 | 19:46 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP: POC of podman_container https://review.opendev.org/682458 | 19:46 |
openstackgerrit | Emilien Macchi proposed openstack/tripleo-ansible master: WIP - Introduce tripleo-container-manage role https://review.opendev.org/686196 | 19:46 |
*** ykarel|away has quit IRC | 19:46 | |
openstackgerrit | wes hayutin proposed openstack/tripleo-ansible master: Handle empty cert related hostvar info https://review.opendev.org/684826 | 19:48 |
openstackgerrit | Merged openstack/tripleo-quickstart-extras master: update centos-release openstack to rocky https://review.opendev.org/685177 | 19:56 |
openstackgerrit | wes hayutin proposed openstack/tripleo-ci master: move standalone-scenario010 to non-voting https://review.opendev.org/686258 | 19:56 |
openstackgerrit | wes hayutin proposed openstack/tripleo-ci master: move standalone-scenario010 to non-voting https://review.opendev.org/686258 | 19:57 |
weshay | arxcruz|ruck, https://review.opendev.org/#/c/686262/ | 19:57 |
*** ansmith_ has joined #tripleo | 19:58 | |
mwhahaha | d0ugal: for tomorrow, can you tell me why https://review.opendev.org/#/c/686181/ breaks when running under mistral? Seems like it's not compatibile with being run under the mistral engine (see failures in containers-multinode) | 19:58 |
arxcruz|ruck | weshay: https://review.rdoproject.org/r/#/c/22801/ | 19:58 |
*** ansmith has quit IRC | 20:00 | |
*** holser has joined #tripleo | 20:00 | |
weshay | arxcruz|ruck, https://trunk-primary.rdoproject.org/api-centos-queens/api/civotes_detail.html?commit_hash=e2208117a31eea79c50330333501b6581f9faafe&distro_hash=15a9b671b405f92b4fc5f2f8b12ae92dfe136fd8 | 20:01 |
arxcruz|ruck | weshay: https://trunk-primary.rdoproject.org/api-centos-stein/api/civotes_detail.html?commit_hash=a361c5edd5734320f781e10358a602f4ba57fc04&distro_hash=647b08e47a72f7533142c09074f227628f08f9fa | 20:03 |
triple-oh-noob | Hi. So, trying to add an additional (compute) node. When I perform introspection, I get: "Can not transition from state 'starting' on event 'start' (no defined transition) | 20:07 |
triple-oh-noob | I have defined the node's "instack" information succefully. | 20:07 |
openstackgerrit | Merged openstack/tripleo-quickstart-extras master: ensure periodic job requirements are visible https://review.opendev.org/685184 | 20:09 |
openstackgerrit | Merged openstack/tripleo-quickstart master: Added release file for train https://review.opendev.org/683932 | 20:09 |
openstackgerrit | Merged openstack/os-refresh-config stable/stein: Update UPPER_CONSTRAINTS_FILE for stable/stein https://review.opendev.org/653758 | 20:09 |
*** aakarsh has joined #tripleo | 20:09 | |
*** cylopez has quit IRC | 20:12 | |
arxcruz|ruck | weshay: https://zuul.openstack.org/builds?job_name=tripleo-ci-centos-7-standalone-upgrade-stein | 20:14 |
*** ekultails has quit IRC | 20:16 | |
openstackgerrit | Martin Mágr proposed openstack/tripleo-heat-templates master: [WIP] Enable external public endpoint for MetricsQdr https://review.opendev.org/686282 | 20:24 |
openstackgerrit | Martin Mágr proposed openstack/puppet-tripleo master: [WIP] Enable metrics qdr to use public VIP https://review.opendev.org/686287 | 20:30 |
dking_desktop | mwhahaha: I see that the process is progress for Heat resources Controller and Compute, and nova had sent them to the scheduler. There are log entries in nova logs for "Unknown column 'resources' in 'field list'" | 20:32 |
* mwhahaha shrugs | 20:32 | |
*** gbarros has quit IRC | 20:32 | |
dking_desktop | ...and there, it just finally failed for timeout creating the ResourceGroup. | 20:33 |
mwhahaha | check the ironic logs | 20:34 |
openstackgerrit | Francesco Pantano proposed openstack/tripleo-heat-templates master: Add prometheus port parameter https://review.opendev.org/685465 | 20:40 |
triple-oh-noob | Anyone any ideas? I'm just trying to launch introspect on a new server to add. Nothing much in the ironic logs other than this. | 20:41 |
*** paramite has quit IRC | 20:42 | |
weshay | mwhahaha, https://review.opendev.org/#/c/685777/ is that role run w/ a playbook w/ hosts: all | 20:43 |
weshay | ? | 20:43 |
mwhahaha | likely | 20:44 |
mwhahaha | it's in the output | 20:44 |
mwhahaha | it's run in the container tho | 20:44 |
mwhahaha | er no it's not | 20:45 |
weshay | error's on | 20:45 |
weshay | remove /var/lib/containers/storage/overlay/9cdd774a37289bdfa7e58dfba15dd357cfbf17f4e630ba5e45fe34bdbbe53a2a/merged: device or resource busy | 20:45 |
weshay | warning on: | 20:45 |
mwhahaha | yea it doesn't force it | 20:45 |
weshay | WARNING]: provided hosts list is empty, only localhost is available. Note\\nthat the implicit localhost does not match 'all | 20:45 |
mwhahaha | hosts: localhost | 20:45 |
weshay | "tripleo_common.image.exception.ImageUploaderException: Modifying image 192.168.24.1:8787/tripleomaster/rhel-binary-swift-proxy-server:968c9526ff2eeafaa13207a791f61173db604fff_9e598812 failed" | 20:45 |
mwhahaha | that's something else? | 20:46 |
mwhahaha | link to logs? | 20:46 |
dking_desktop | triple-oh-noob: I'm still new, myself, but I imagine that it would be helpful for you to provide your introspect rules, and/or maybe your instackenv.json? Maybe the specific commands you are running, as you can run introspection through several commands. | 20:46 |
openstackgerrit | James Slagle proposed openstack/tripleo-heat-templates master: Add missing role name to Host prep steps task https://review.opendev.org/686293 | 20:46 |
weshay | mwhahaha, http://logs.rdoproject.org/93/22793/3/check/periodic-tripleo-ci-rhel-8-ovb-3ctlr_1comp-featureset001-master/0ce4bae/logs/undercloud/home/zuul/install-undercloud.log.txt.gz | 20:47 |
weshay | mwhahaha, rhel-8 | 20:47 |
weshay | undercloud install | 20:47 |
mwhahaha | of course, we can't have consistent interactions between versions | 20:47 |
mwhahaha | we can slap an ignore errors on that tasks | 20:48 |
mwhahaha | we want to try and remove the buildah container, those kinda odd that it's conflicting there | 20:49 |
triple-oh-noob | dking_desktop: I have used the same format and detail types for the other servers successfully, I think there's something a bit "weird". | 20:49 |
weshay | arxcruz|ruck, check the tmux | 20:50 |
weshay | had to cherrypick the change onto it.. | 20:50 |
weshay | arxcruz|ruck, updates are turned off | 20:50 |
weshay | but the service is running | 20:50 |
mwhahaha | wonder if we need to retry the container remove because of parallel processing | 20:51 |
triple-oh-noob | .. or there's something over ipmi that's not working... Wondering if anyone can shed light on what the error is suggesting... | 20:51 |
weshay | mwhahaha, if you want to test a change.. just drop a depends on https://review.rdoproject.org/r/#/c/22793/ | 20:55 |
mwhahaha | do you mean update that change to dependon a patch? | 20:55 |
weshay | aye | 20:56 |
weshay | arxcruz|ruck, we need a bug on that probably | 20:56 |
arxcruz|ruck | weshay: sorry, on which one? | 20:57 |
weshay | arxcruz|ruck, https://review.opendev.org/#/c/685777/2/tasks/yum_update_buildah.yml | 20:57 |
*** paramite has joined #tripleo | 20:57 | |
weshay | arxcruz|ruck, http://logs.rdoproject.org/93/22793/3/check/periodic-tripleo-ci-rhel-8-ovb-3ctlr_1comp-featureset001-master/0ce4bae/logs/undercloud/home/zuul/install-undercloud.log.txt.gz | 20:57 |
arxcruz|ruck | weshay: ack | 20:58 |
weshay | arxcruz|ruck, cherry picked https://review.rdoproject.org/r/#/c/22803/ | 20:59 |
openstackgerrit | Alex Schultz proposed openstack/ansible-role-tripleo-modify-image master: Add retires/ignore errors for removing buildah containers https://review.opendev.org/686294 | 20:59 |
mwhahaha | weshay: -^ | 20:59 |
weshay | arxcruz|ruck, so we should see rhel8 master and stein centos7 promote | 20:59 |
weshay | mwhahaha, cool.. thanks | 20:59 |
arxcruz|ruck | weshay: https://bugs.launchpad.net/tripleo/+bug/1846413 | 21:00 |
openstack | Launchpad bug 1846413 in tripleo "RHEL-8 ovb job failing in tripleo-modify-image : Cleanup working container" [High,Triaged] | 21:00 |
weshay | mwhahaha, throws tables | 21:00 |
weshay | I kid | 21:00 |
weshay | now ur just causing timeouts ;) | 21:01 |
* weshay channels mwhahaha | 21:01 | |
mwhahaha | pffft rhel8 who cares about that | 21:01 |
weshay | :) | 21:01 |
weshay | f.. rhel8 I'm all about the stream | 21:01 |
mwhahaha | i'm kinda over this whole thing | 21:01 |
mwhahaha | i was over containers before, now i'm really over them | 21:01 |
mwhahaha | worst idea ever | 21:01 |
mwhahaha | this only cements it more and more | 21:02 |
*** Goneri has quit IRC | 21:02 | |
weshay | arxcruz|ruck, promoter Promoting the container images for dlrn hash 968c9526ff2eeafaa13207a791f61173db604fff on master to current-tripleo | 21:02 |
arxcruz|ruck | weshay: i saw :) | 21:02 |
arxcruz|ruck | \o/ | 21:02 |
weshay | arxcruz|ruck, I'll watch it from here.. thanks for the help.. tomorrow make sure stein and rhel8 master have gone through | 21:03 |
weshay | arxcruz|ruck, i'll continue on queens | 21:03 |
*** aakarsh has quit IRC | 21:03 | |
weshay | arxcruz|ruck, oh.. just the bug.. so mwhahaha can note it in the commit message | 21:04 |
weshay | arxcruz|ruck, mwhahaha get's paid by the patch / bug | 21:04 |
arxcruz|ruck | weshay: https://bugs.launchpad.net/tripleo/+bug/1846413 | 21:04 |
openstack | Launchpad bug 1846413 in tripleo "RHEL-8 ovb job failing in tripleo-modify-image : Cleanup working container" [High,Triaged] | 21:04 |
weshay | rock on king kong | 21:04 |
*** tosky_ has joined #tripleo | 21:05 | |
openstackgerrit | Alex Schultz proposed openstack/ansible-role-tripleo-modify-image master: Add retires/ignore errors for removing buildah containers https://review.opendev.org/686294 | 21:05 |
*** tosky has quit IRC | 21:06 | |
*** xek has quit IRC | 21:08 | |
*** tosky_ is now known as tosky | 21:08 | |
*** ansmith_ has quit IRC | 21:09 | |
*** rakhmerov has quit IRC | 21:09 | |
openstackgerrit | James Slagle proposed openstack/tripleo-heat-templates master: Remove outdated FIXME's https://review.opendev.org/686299 | 21:20 |
*** aakarsh has joined #tripleo | 21:22 | |
*** jfrancoa has quit IRC | 21:27 | |
openstackgerrit | James Slagle proposed openstack/tripleo-heat-templates master: Use single quotes for --start-at-task debug tasks https://review.opendev.org/686309 | 21:47 |
*** raildo has quit IRC | 21:47 | |
*** ansmith_ has joined #tripleo | 21:48 | |
openstackgerrit | Merged openstack/tripleo-ansible master: scenario010 nonvoting https://review.opendev.org/686262 | 21:48 |
*** holser has quit IRC | 21:56 | |
*** slagle_ is now known as slagle | 21:58 | |
*** rfolco|dentist is now known as rfolco | 22:17 | |
*** pkopec has quit IRC | 22:17 | |
*** paramite has quit IRC | 22:18 | |
*** mcornea has quit IRC | 22:35 | |
*** slaweq has quit IRC | 22:35 | |
weshay | mwhahaha, that resolved the issue on rhel8 | 22:47 |
mwhahaha | K | 22:47 |
*** rcernin has joined #tripleo | 22:48 | |
*** tkajinam has joined #tripleo | 22:54 | |
weshay | hrm.. container-prep time is way down :)) http://dashboard-ci.tripleo.org/d/si1tipHZk/jobs-exploration?orgId=1&from=now-6h&to=now | 23:07 |
*** tosky has quit IRC | 23:16 | |
*** cdearborn has quit IRC | 23:23 | |
*** panda is now known as panda|off | 23:39 | |
*** gbarros has joined #tripleo | 23:42 | |
*** gbarros has quit IRC | 23:55 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!