Wednesday, 2015-06-17

stevellea little subtle and adding a second filter to filter on facts directly would be needed but could be based on this00:00
*** javeriak has quit IRC00:13
*** sdake has quit IRC00:24
*** sdake has joined #openstack-ansible00:48
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: changed container bind mounts to use abspath  https://review.openstack.org/18855800:53
*** sdake_ has joined #openstack-ansible00:58
*** sdake has quit IRC01:01
*** abitha has quit IRC01:03
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Added apt update tasks to everything using apt  https://review.openstack.org/19248401:20
*** Mudpuppy has quit IRC01:23
*** Mudpuppy_ has joined #openstack-ansible01:23
*** dkalleg has quit IRC01:25
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Allow protocol to be set per endpoint-type  https://review.openstack.org/19214801:26
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Split environment file into component parts  https://review.openstack.org/19210201:27
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Configurable lxc cache and rootfs path  https://review.openstack.org/19216501:27
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Fix spelling of runlevel in init scripts  https://review.openstack.org/19248901:28
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Backported tempest roles from Master  https://review.openstack.org/19244502:06
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Updated Juno for new dev work  https://review.openstack.org/18897502:06
openstackgerritMerged stackforge/os-ansible-deployment: remove '{{' from variable in loop expression  https://review.openstack.org/19100602:07
*** dkalleg has joined #openstack-ansible02:26
*** sdake_ is now known as sdake03:00
*** dkalleg has quit IRC03:08
openstackgerritMerged stackforge/os-ansible-deployment: Allow Horizon to access multiple regions  https://review.openstack.org/19030203:12
*** dkalleg has joined #openstack-ansible03:19
*** markvoelker has quit IRC03:31
*** JRobinson__ is now known as JRobinson__afk03:46
*** stevemar has quit IRC03:47
*** stevemar has joined #openstack-ansible03:48
*** dkalleg has quit IRC03:58
*** JRobinson__afk is now known as JRobinson__04:27
*** markvoelker has joined #openstack-ansible04:32
*** markvoelker has quit IRC04:37
*** fawadkhaliq has joined #openstack-ansible04:55
*** JRobinson__ has quit IRC05:06
*** javeriak has joined #openstack-ansible05:11
*** shausy has joined #openstack-ansible05:14
*** stevemar has quit IRC05:14
*** Mudpuppy_ has quit IRC05:41
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Updated Juno for new dev work  https://review.openstack.org/18897505:55
*** markvoelker has joined #openstack-ansible05:55
*** markvoelker has quit IRC06:00
*** shausy has quit IRC06:09
*** shausy has joined #openstack-ansible06:10
*** javeriak has quit IRC06:12
*** javeriak has joined #openstack-ansible06:12
*** JRobinson__ has joined #openstack-ansible06:13
*** radek_ has joined #openstack-ansible06:15
*** javeriak has quit IRC06:16
*** javeriak has joined #openstack-ansible06:17
*** fawadkhaliq has quit IRC06:17
*** daneyon has quit IRC07:20
*** sdake has quit IRC07:20
*** Mudpuppy has joined #openstack-ansible07:38
*** Mudpuppy has quit IRC07:42
*** amit213 has joined #openstack-ansible07:44
*** markvoelker has joined #openstack-ansible07:44
*** markvoelker has quit IRC07:49
*** amit213 has quit IRC07:51
*** javeriak has quit IRC08:02
*** JRobinson__ has quit IRC08:16
*** daneyon has joined #openstack-ansible09:09
*** daneyon has quit IRC09:14
openstackgerritHugh Saunders proposed stackforge/os-ansible-deployment: Backported tempest roles from Master  https://review.openstack.org/19244509:26
*** markvoelker has joined #openstack-ansible09:33
*** markvoelker has quit IRC09:37
openstackgerritHugh Saunders proposed stackforge/os-ansible-deployment: Updated Juno for new dev work  https://review.openstack.org/18897509:47
*** jmccrory has quit IRC10:10
*** jmccrory has joined #openstack-ansible10:12
*** daneyon has joined #openstack-ansible10:58
*** daneyon has quit IRC11:02
openstackgerritMerged stackforge/os-ansible-deployment: Backported tempest roles from Master  https://review.openstack.org/19244511:05
*** markvoelker has joined #openstack-ansible11:22
*** markvoelker has quit IRC11:27
*** sdake has joined #openstack-ansible11:41
*** Mudpuppy has joined #openstack-ansible11:54
*** sdake has quit IRC12:04
*** markvoelker has joined #openstack-ansible12:05
*** jlvillal has quit IRC12:11
*** shausy has quit IRC12:18
*** shausy has joined #openstack-ansible12:19
*** jlvillal has joined #openstack-ansible12:23
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Updated Juno for new dev work  https://review.openstack.org/18897512:44
*** daneyon has joined #openstack-ansible12:46
*** Mudpuppy has quit IRC12:50
*** daneyon has quit IRC12:51
*** tlian has joined #openstack-ansible13:03
openstackgerritKen Johnston proposed stackforge/os-ansible-deployment-specs: Spec for multiregion ansible deployment  https://review.openstack.org/19242113:27
*** stevemar has joined #openstack-ansible13:30
*** Mudpuppy has joined #openstack-ansible13:52
*** Mudpuppy has quit IRC13:57
*** Mudpuppy has joined #openstack-ansible13:57
*** sigmavirus24_awa is now known as sigmavirus2413:58
palendaeDoes https://review.openstack.org/#/c/188975/ potentially need to get the tempest venv treatment?14:00
palendaeI know cloudnull mentioned it14:00
*** Mudpuppy is now known as sigmavirus2514:01
*** sigmavirus25 is now known as Mudpuppy14:01
*** yaya_ has joined #openstack-ansible14:06
*** galstrom_zzz is now known as galstrom14:31
*** shausy has quit IRC14:31
*** daneyon has joined #openstack-ansible14:35
sigmavirus24odyssey4me: on https://review.openstack.org/#/c/189998/ how would delegating to keystone_all[0]'s host solve the problem for synchronize?14:36
sigmavirus24wouldn't we have to synchronize the keys to that host and then out to the other containers anyway?14:36
odyssey4mesigmavirus24 it should probably delegate to the container - my brain's stuck in Microsoft-land right now... essentially there should be a way to do a direct sync from the first keystone container to the others14:37
odyssey4mecloudnull would probably know off-hand14:38
* sigmavirus24 tried doing that14:39
sigmavirus24It didn't work14:39
sigmavirus24I probably did it wrong but I used delegate_to: keystone_all[0] and it tried but there was no way to ssh from that container into the others14:39
sigmavirus24And I agree if we could that'd be super elegant but in the absence of elegance, we have this =/14:40
*** daneyon has quit IRC14:40
odyssey4mesigmavirus24 yeah, I'm happy with leaving the push/pull mechanism - but I would like to see that the keys don't get recreated and forcibply moved around every time the role is executed :)14:44
sigmavirus24so cloudnull specifically asked for key rotation akin to nova's14:44
sigmavirus24And I copied that pattern although after Monday night's discussion with dolphm I'm pretty sure we don't need to destroy the directory to recreate since fernet_setup will rotate them for us14:45
openstackgerritHugh Saunders proposed stackforge/os-ansible-deployment: Handle proxies through environment variables  https://review.openstack.org/19271714:52
*** sdake_ has joined #openstack-ansible14:53
palendaehughsaunders: lol, I was gonna be looking at that, but thanks :p14:53
palendaeActually nm, different bug14:53
hughsaunders:)14:53
odyssey4mesigmavirus24 ok, I wasn't aware that you did it that way based on a request from cloudnull - alright, I suppose that effectively means that we need to wait for cloudnull to verify that the patch achieves what he was looking for...15:03
sigmavirus24it does, but it can probably be done in a simpler way15:04
sigmavirus24I'd want to verify that synchronize will effectively copy over the existing keys though on a regeneration15:04
odyssey4meI'll sigmavirus24 so I've pulled my -1 and will wait to see cloudnull's review before proceeding.15:04
*** openstackgerrit has quit IRC15:05
*** openstackgerrit has joined #openstack-ansible15:05
sigmavirus24odyssey4me: nah, your nit is valid15:05
sigmavirus24-1 can stay15:05
sigmavirus24I love the -1s15:05
sigmavirus24I live off -1s15:06
sigmavirus24-1s are my lifeblood15:06
odyssey4mesigmavirus24 haha :p15:06
sigmavirus24If I'm not getting -1's I'm doing something wrong =P15:06
*** Mudpuppy has quit IRC15:12
*** Mudpuppy_ has joined #openstack-ansible15:12
*** Mudpuppy_ is now known as Mudpuppy15:12
*** yaya_ has quit IRC15:29
*** Mudpuppy has quit IRC15:38
*** Mudpuppy has joined #openstack-ansible15:38
svghello folks15:47
*** stevemar has quit IRC15:50
*** javeriak has joined #openstack-ansible15:51
*** stevemar has joined #openstack-ansible15:52
*** daneyon has joined #openstack-ansible16:04
*** daneyon_ has joined #openstack-ansible16:04
*** daneyon has quit IRC16:08
svgWho's familiar with memcache configs?  We bumped into issues somehow related to https://bugs.launchpad.net/keystone/+bug/1332058, where we had some nodes down (metal, including memcache containers)16:11
openstackLaunchpad bug 1332058 in Mirantis OpenStack "keystone behavior when one memcache backend is down" [Critical,Fix committed] - Assigned to Yuriy Taraday (yorik-sar)16:11
openstackgerritJesse Pretorius proposed stackforge/os-ansible-deployment: Implement service_region user_group_var  https://review.openstack.org/18966416:11
svgseems osad configs rely on the defaults, which gives long timeouts when one or more memcaches is down16:12
*** Mudpuppy has quit IRC16:22
*** yaya has joined #openstack-ansible16:24
*** sdake_ is now known as sdake16:25
*** yaya has quit IRC16:29
*** openstackgerrit has quit IRC16:33
*** openstackgerrit has joined #openstack-ansible16:34
*** rrrobbb has joined #openstack-ansible16:42
*** a| has joined #openstack-ansible16:46
*** a| has quit IRC16:46
*** mfisch has quit IRC16:47
*** logan2 has joined #openstack-ansible16:47
*** javeriak has quit IRC16:48
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Updated Juno for new dev work  https://review.openstack.org/18897516:48
*** mfisch has joined #openstack-ansible16:49
*** mfisch has quit IRC16:50
*** mfisch has joined #openstack-ansible16:50
*** abitha has joined #openstack-ansible16:58
*** abitha has quit IRC17:00
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Updated Juno for new dev work  https://review.openstack.org/18897517:00
*** abitha has joined #openstack-ansible17:07
*** javeriak has joined #openstack-ansible17:08
*** javeriak has quit IRC17:08
*** javeriak has joined #openstack-ansible17:10
*** sdake_ has joined #openstack-ansible17:30
*** sdake has quit IRC17:34
*** Mudpuppy has joined #openstack-ansible17:49
openstackgerritMerged stackforge/os-ansible-deployment: Added apt update tasks to everything using apt  https://review.openstack.org/19248417:57
*** sdake has joined #openstack-ansible17:58
*** sdake_ has quit IRC18:01
jmccroryhi all, is there are current option to not deploy utility containers?18:02
cloudnullcores https://review.openstack.org/#/c/188975/13 - please make this go for juno18:06
cloudnulljmccrory: you'd have to modify the environment to not deploy utility18:06
cloudnulljmccrory: https://github.com/stackforge/os-ansible-deployment/tree/master/etc/openstack_deploy/env.d18:07
cloudnullin master kilo you can simply remove the utility.yml from your environment setup18:07
*** annashen has joined #openstack-ansible18:07
cloudnullin juno/icehouse you'd have to remove the the same bits from the monolithic environment.yml file.18:08
*** sdake_ has joined #openstack-ansible18:10
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Fix spelling of runlevel in init scripts  https://review.openstack.org/19248918:13
*** annashen has quit IRC18:13
*** sdake has quit IRC18:14
*** yaya has joined #openstack-ansible18:14
*** annashen has joined #openstack-ansible18:15
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Configurable lxc cache and rootfs path  https://review.openstack.org/19216518:18
jmccrorycloudnull: ah ok, thanks18:18
*** markvoelker has quit IRC18:25
*** markvoelker has joined #openstack-ansible18:25
svghello folks18:37
cloudnullhey svg18:45
*** daneyon_ has quit IRC18:45
*** daneyon has joined #openstack-ansible18:46
svghi cloudnull18:47
svgdid you notice my messages? ^^ 2 hours up18:47
*** markvoelker_ has joined #openstack-ansible18:47
*** markvoelker has quit IRC18:47
svgwas one of the reasons things got real sluggish at certain times18:48
*** daneyon_ has joined #openstack-ansible18:49
cloudnullno... been intermittently connected. im in training all this weelk.18:50
* cloudnull looking back through scroll back.18:50
*** daneyon has quit IRC18:52
cloudnullsvg: the "memcache configs" things?18:53
svgyes18:53
svgall os daemons get a list of all the memcached servers18:54
svgbut it takes some time to time out when some are offline18:54
svgand the result is, all commands (horizon or cli) get slow - sometimes 10-50s18:54
svgSo I'm wondering if that might be a config problem18:55
svgthe bug I linked to, shows some options to tell the client driver to remember down connections etc18:55
svgbut those ar not used in osad, so I'm wondering if this might be an optimisation opportunity18:56
* cloudnull looking.18:57
svgone of the first thing I think of is ordering: the first server should point to the container on the same metal host18:58
svgand possibl eusing those extra options from the bug link, I need to test that a bit though18:59
cloudnulllooks like an optimization we can make. in master/kilo we have a deterministic sorting process we can put the list of servers through.19:04
cloudnullExample: https://github.com/stackforge/os-ansible-deployment/blob/master/playbooks/os-cinder-install.yml#L80-L8819:04
cloudnullthat would make it so that were not using the same server as the first server all the time19:05
*** yaya has quit IRC19:07
dstaneki was looking at https://bugs.launchpad.net/openstack-ansible/+bug/1399383 and the review for it - is there any reason why the internal endpoints are not going over SSL?19:08
openstackLaunchpad bug 1399383 in openstack-ansible juno "General SSL support for all public endpoints including spice" [High,Confirmed] - Assigned to Andy McCrae (andrew-mccrae)19:08
cloudnullsomething like https://gist.github.com/cloudnull/bf7311fb93bca16c9c3b would help19:08
cloudnulldstanek: ssl w/ out a signed cert is next to impossible with most OS services. but it certainly could be set to https if they wanted to.19:10
dstanekcloudnull: why no signed cert? it's usually best practice to use SSL everywhere even on a "secure, private" network19:11
svgcloudnull: that sounds right - ho wis the sorrtic logic in that module?19:12
*** yaya has joined #openstack-ansible19:13
cloudnulldstanek: w/ a singed cert it "should" work. but not every service supports that and supports it the same way.19:15
svgreading the module code...19:15
cloudnulllast i tried, and im sure a few others went down this path too but, ssl all the way down w/ the enventlet wsgis is next to impossible19:16
*** openstackgerrit has quit IRC19:16
dstanekcloudnull: ah, so some of the services need some love19:16
*** openstackgerrit has joined #openstack-ansible19:17
cloudnulleventlet + monkey patching makes using a real web server impossible in some cases.19:17
cloudnullbut, rest assured, keystone works great with ssl :)19:17
dstanekcloudnull: what about with a terminator? i'm guessing that some don't respect the HTTPS header to rewrite links19:17
cloudnulllast i looked ssl terminated at the lb works fine.19:18
*** yaya has quit IRC19:49
*** stevemar has quit IRC20:00
*** javeriak has quit IRC20:03
*** Mudpuppy has quit IRC20:09
*** Mudpuppy has joined #openstack-ansible20:10
*** toddnni has quit IRC20:11
*** toddnni has joined #openstack-ansible20:13
*** Mudpuppy has quit IRC20:22
*** toddnni_ has joined #openstack-ansible20:23
*** toddnni has quit IRC20:26
*** toddnni_ is now known as toddnni20:26
*** galstrom is now known as galstrom_zzz20:30
*** Mudpuppy has joined #openstack-ansible20:48
*** Mudpuppy_ has joined #openstack-ansible20:50
*** Mudpuppy has quit IRC20:54
*** yaya has joined #openstack-ansible20:55
*** dkalleg has joined #openstack-ansible20:57
*** annashen has quit IRC21:05
*** javeriak has joined #openstack-ansible21:06
*** Mudpuppy_ is now known as Mudpuppy21:10
*** tlian has quit IRC21:13
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Ceph/RBD support  https://review.openstack.org/18195721:16
openstackgerritKevin Carter proposed stackforge/os-ansible-deployment: Updated to the latest stable rabbitmq release  https://review.openstack.org/19289921:18
*** openstackgerrit has quit IRC21:24
*** openstackgerrit has joined #openstack-ansible21:24
*** sdake_ is now known as sdake21:25
*** Mudpuppy has quit IRC21:28
*** radek_ has quit IRC21:38
*** toddnni has quit IRC21:38
*** toddnni has joined #openstack-ansible21:41
*** annashen has joined #openstack-ansible21:42
*** Mudpuppy has joined #openstack-ansible21:59
*** Mudpuppy has quit IRC22:02
*** sigmavirus24 is now known as sigmavirus24_awa22:08
*** yaya has quit IRC22:09
*** JRobinson__ has joined #openstack-ansible22:29
*** rrrobbb has quit IRC22:31
*** sigmavirus24_awa is now known as sigmavirus2422:55
*** sigmavirus24 is now known as sigmavirus24_awa22:56
*** JRobinson__ is now known as JRobinson__afk23:05
openstackgerritMerged stackforge/os-ansible-deployment: Fix spelling of runlevel in init scripts  https://review.openstack.org/19248923:05
*** JRobinson__afk is now known as JRobinson__23:14
openstackgerritMerged stackforge/os-ansible-deployment: Generate consistent length passwords in AIO  https://review.openstack.org/19222023:26
*** sdake_ has joined #openstack-ansible23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!