Friday, 2018-08-03

openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Update repo_server role to remove pypi reverse proxy  https://review.openstack.org/58573500:05
*** vnogin has quit IRC00:05
*** hamzy has joined #openstack-ansible00:09
*** cshen has joined #openstack-ansible00:19
*** cshen has quit IRC00:24
*** gyee has quit IRC00:36
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_hosts master: Add route when required on the local table  https://review.openstack.org/58840602:19
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Update index retention policy to ensure its an int  https://review.openstack.org/58838902:21
*** dave-mccowan has quit IRC02:35
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_cinder master: Resolve cinder config deprecation warnings  https://review.openstack.org/58812702:58
*** dxiri has joined #openstack-ansible03:08
openstackgerritMerged openstack/openstack-ansible stable/queens: Update repo_server role to remove pypi reverse proxy  https://review.openstack.org/58573503:09
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible master: Update nspawn container create  https://review.openstack.org/58843003:23
*** udesale has joined #openstack-ansible03:31
*** dxiri has quit IRC03:41
openstackgerritMerged openstack/openstack-ansible-ops master: Update index retention policy to ensure its an int  https://review.openstack.org/58838903:48
*** hwoarang has quit IRC03:52
*** poopcat has quit IRC04:07
*** dxiri has joined #openstack-ansible04:21
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible master: Update nspawn container create  https://review.openstack.org/58843004:25
*** cshen has joined #openstack-ansible04:29
*** cshen has quit IRC04:34
*** cshen has joined #openstack-ansible04:35
*** udesale has quit IRC04:35
*** viks_ has joined #openstack-ansible04:37
*** cshen has quit IRC04:40
*** viks_ has quit IRC04:42
*** udesale has joined #openstack-ansible04:55
*** cshen has joined #openstack-ansible04:55
*** cshen has quit IRC05:00
*** czunker has joined #openstack-ansible05:11
*** pmannidi has joined #openstack-ansible05:13
*** dxiri has quit IRC05:14
*** shyambiradar has joined #openstack-ansible05:17
*** pradiprwt has quit IRC05:17
*** cshen has joined #openstack-ansible05:30
*** cshen has quit IRC05:34
*** shyambiradar has quit IRC05:41
*** shyambiradar has joined #openstack-ansible05:50
openstackgerritwangqi proposed openstack/openstack-ansible-lxc_hosts master: Add the project source code repository in README  https://review.openstack.org/58703006:28
*** chason has joined #openstack-ansible06:43
evrardjpgood morning06:52
*** shyambiradar has quit IRC06:54
*** shyambiradar has joined #openstack-ansible06:58
*** cshen has joined #openstack-ansible06:58
*** gillesMo has joined #openstack-ansible06:58
*** chason has quit IRC07:01
evrardjpodyssey4me: I would prefer smoke testing if we can -- or using ansible modules07:01
*** chason has joined #openstack-ansible07:02
openstackgerritMerged openstack/openstack-ansible-os_monasca-agent master: Add Ubuntu Bionic 18.04 support  https://review.openstack.org/58693307:02
kaiokmomnaser: bionic passing on https://review.openstack.org/#/c/586933/07:03
kaiokmomorning evrardjp07:04
*** chason has quit IRC07:08
*** pmannidi has quit IRC07:16
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Add support for CentOS 7  https://review.openstack.org/58845607:24
*** tosky has joined #openstack-ansible07:26
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Add support for openSUSE  https://review.openstack.org/58846207:37
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Add support for CentOS 7  https://review.openstack.org/58845607:40
*** chason has joined #openstack-ansible07:43
*** shyambiradar has quit IRC07:44
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Add support for openSUSE  https://review.openstack.org/58846207:45
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Switch to using project-templates  https://review.openstack.org/58846407:49
evrardjphello kaiokmo :)07:52
*** Bhujay has joined #openstack-ansible07:56
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] use original cache prep script on bionic  https://review.openstack.org/58846608:03
*** Bhujay has quit IRC08:05
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-os_glance master: Add bionic to galaxy_info metadata  https://review.openstack.org/58826408:06
*** cshen has quit IRC08:26
odyssey4meevrardjp: sure, but smoke testing isn't working either - and we have a ton of patches waiting to merge, some of which fix things (like logging, rabbitmq, etc)08:37
odyssey4meif you look carefully at the experimental check on that smoke testing patch, it failed tempest08:37
odyssey4meultimately, all I'm proposing is a temporary stop-gap to allow patches to continue moving while we figure out a better option08:37
*** shyambiradar has joined #openstack-ansible08:38
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Add support for CentOS 7  https://review.openstack.org/58845608:38
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] use original cache prep script on bionic  https://review.openstack.org/58846608:38
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Add support for openSUSE  https://review.openstack.org/58846208:41
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Add support for CentOS 7  https://review.openstack.org/58845608:42
*** electrofelix has joined #openstack-ansible08:49
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] Update to bionic 18.04.1 base image  https://review.openstack.org/58846608:57
*** DanyC has quit IRC09:04
odyssey4mejrosser evrardjp could I get another review for https://review.openstack.org/588362 please - I'd like to get that ported back ASAP09:04
*** DanyC has joined #openstack-ansible09:04
jrosserodyssey4me: done09:07
odyssey4metyvm09:07
*** DanyC has quit IRC09:09
*** chason has quit IRC09:10
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Update to bionic 18.04.1 base image  https://review.openstack.org/58846609:19
jrosserodyssey4me: based on a sample of 1 that might fix the lxc cache prep - can you keep an eye on it? ^09:19
* jrosser away09:19
odyssey4mejrosser: yep, will do09:26
*** chason has joined #openstack-ansible09:28
evrardjpodyssey4me: I am fine with the blacklist, I didn't know it was that easy to bring in.09:32
evrardjpI'd have put the cinder volume backup in it09:32
evrardjpjrosser: what's in the .1 image?09:33
evrardjpjust curious : )09:33
evrardjpodyssey4me: fine for me for the backport btw09:33
evrardjpodyssey4me: should I continue my healthcheck work?09:34
evrardjpI have the impression that those healthchecks could then be moved into the respective roles :)09:35
*** shyambiradar has quit IRC09:35
evrardjpor plays09:35
*** vnogin has joined #openstack-ansible09:35
odyssey4meevrardjp: the health check has operator value, so even if it's not used for gating, it's useful - we could also use it for gating with or without tempest - so yes, continue it!09:36
evrardjpok09:36
evrardjpWell09:36
evrardjpfunny story09:36
*** shyambiradar has joined #openstack-ansible09:36
evrardjpit failed today with timeout -- so I think it's not only tempest. It's a real thing we have.09:36
odyssey4mebut yes, I am a fan of the idea of putting some sort of health check into the roles09:36
odyssey4mewell yes, so there are two issues - one is the tempest test failing, perhaps due to load or some other issue09:37
odyssey4meanother is that cinder is failing, and while we haven't pinpointed why, we do know that the removal of all the rabbitmq things fixes it09:37
evrardjpodyssey4me: so you mean rollback to old rabbitmq things ?09:38
odyssey4meso I'm thinking we kill the failing tempest test, get the log and rabbitmq patches merged, then we have a working set again09:38
evrardjpthat's worth pursuing09:38
evrardjpprobably a wrong var09:38
odyssey4meevrardjp: nope, this patch fixes things - https://review.openstack.org/58550409:38
evrardjpwhat do you mean by get the log and rabbitmq patches merged?09:38
odyssey4meright now there's something in the rabbit bits broken - so we need to get https://review.openstack.org/585504 merged to fix it09:38
odyssey4mealso, the log gathering is broken - so we need to get https://review.openstack.org/587708 merged to fix that09:39
evrardjpI've seen for the log gathering09:39
evrardjpdo you want me to test the rabbitmq patch below mine to see it works better?09:39
odyssey4me'below mine' ?09:40
odyssey4meI've verified that it works on a test node, as has jrosser.09:40
evrardjpodyssey4me: below my healthchecks09:41
odyssey4meSo, the quickest way for us to get the integrated repo moving again is to get https://review.openstack.org/588360 merged.09:41
evrardjpto see how the healthchecks cinder patches behaves09:41
evrardjpok09:41
evrardjpI voted positively09:41
evrardjpthen we move the content of this patch into blacklist  things09:42
evrardjp(with the blacklist patch in)09:42
evrardjpwe backport the whole thing09:42
evrardjpwell09:42
evrardjpit's not mandatory, but I think it would be a good idea09:42
odyssey4meI'm not sure I understand the point of moving the tests we've remarked out into the blacklist. If they're not in the whitelist then that works just fine.09:43
evrardjpyeah that's true09:43
odyssey4meWhy do you want to backport those changes - queens and back seems fine to me?09:43
evrardjpwe never needed one09:43
evrardjpI think blacklist is just a nice feature09:43
evrardjpthe example is what was done for cinder volumes09:44
evrardjpwe should just follow devstack09:44
odyssey4meyeah, it does make it easier to remove tests if you need to09:44
evrardjpso that would be a blacklist09:44
odyssey4mewell, I find myself thinking that we should perhaps just use devstack's tempest instead09:44
evrardjpit is just out of luck that the blacklist removed ALL the tests from that scenario09:44
evrardjpodyssey4me: it's gonna move to ansible anyway09:45
evrardjp: )09:45
odyssey4meor make our role good for setting tempest up and all that, but not for running it - running it should not actually be done via ansible because the output is unreadable09:45
evrardjpodyssey4me: so.. in my healthchecks I consider two things09:45
evrardjpwhether install of tempest has been done (then skip the resource creation)09:45
odyssey4mebut I also like using the 'smoke' tests - it's really simple to maintain09:46
evrardjpor not09:46
*** shyambiradar has quit IRC09:46
evrardjpyeah I don't know the contents09:46
evrardjpit's probably better than using os modules09:46
evrardjpansible os modules*09:46
evrardjpbut I thought: hey09:46
evrardjpwhy not having ansible os modules in gates09:46
evrardjpand smoke in periodics09:46
odyssey4meevery project designates its own 'smoke' tests with a label, so basically they're saying this is a minimal canary test to verify the service09:47
odyssey4meso we could literally just use 'smoke' everywhere09:47
*** vnogin has quit IRC09:47
evrardjpok09:47
evrardjpI think it's silly to continue with current tests anyway09:47
kaiokmoevrardjp: about the 18.04.1 image: https://bugs.launchpad.net/ubuntu/+source/apt/+bug/177399209:48
openstackLaunchpad bug 1773992 in apt (Ubuntu) "apt-key adv should gpgconf --kill all the things after execution" [Medium,Fix released] - Assigned to Julian Andres Klode (juliank)09:48
evrardjpkaiokmo: thanks for the info09:48
kaiokmonp09:48
odyssey4meevrardjp: yeah, that's certainly a topic for the PTG09:48
evrardjpodyssey4me: let's start the etherpad for the PTG conversations09:51
evrardjphttps://etherpad.openstack.org/p/osa-denver-PTG-planning09:51
evrardjpwoops09:51
evrardjpwrong one09:51
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift master: Allow tags to be used for MQ tasks  https://review.openstack.org/58819609:54
odyssey4meevrardjp: ^ what do you think about that adjustment after your feedback?09:55
evrardjpodyssey4me: let me review that quick -- I will give you an etherpad later, sorry for the wrong link09:56
evrardjpodyssey4me: looks far more readable now09:57
evrardjplgtm09:58
odyssey4meok, we see if it passes tests - if it does, then I'll remove -w and we can get a move on with those patches09:58
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift master: Allow tags to be used for MQ tasks  https://review.openstack.org/58819609:59
odyssey4mesorry - one minor adjustment to prevent a patch storm later09:59
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder master: Allow tags to be used for MQ tasks  https://review.openstack.org/58819010:00
odyssey4mejrosser evrardjp if you guys have some time, I'd appreciate reviews for https://goo.gl/k9cCfF10:01
*** chason has quit IRC10:06
*** chason has joined #openstack-ansible10:07
*** cshen has joined #openstack-ansible10:08
*** chason has quit IRC10:12
odyssey4meevrardjp: shall we vote this through? https://review.openstack.org/58846610:12
jrosserodyssey4me: perhaps the release note should be referenced in the commit?10:15
jrosserhttps://wiki.ubuntu.com/BionicBeaver/ReleaseNotes/ChangeSummary/18.04.110:16
jrosserevrardjp: there are the changes in 18.04.1 ^10:16
evrardjpcores and anyone else: if you want to fill this PTG etherpad that would be great: https://etherpad.openstack.org/p/osa-stein-ptg10:16
odyssey4meNot essential for me, but it is usually good to include 'why' in the commit message.10:17
kaiokmoevrardjp: link the etherpad here for future reference https://wiki.openstack.org/wiki/PTG/Stein/Etherpads :)10:19
evrardjpI was updating all the refs to it, slowly :010:20
evrardjp:)10:20
evrardjpthanks for the link10:20
evrardjpit won't make me search in my emails hahah10:20
*** vnogin has joined #openstack-ansible10:20
evrardjpodyssey4me: the WHY is the only reason we have commit messages, IMO10:21
evrardjp:p10:21
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-os_keystone master: [TEST] bionic  https://review.openstack.org/58663510:21
odyssey4mejrosser: so yeah, if there's a specific reason why you've opted for .1 for the image, usually better to call iit out10:22
jrosserodyssey4me: lets see if that 18.04.1 image makes any difference https://review.openstack.org/#/c/586635/4 (if i've wired it right)10:22
odyssey4meif it's a specific bug fix, note that - otherwise you can just justify it by saying the old one doesn't work and this one does :p10:22
odyssey4medepends-on is supposed to be done without the /#/c/ ... so it should be https://review.openstack.org/58846610:23
odyssey4meit does work with the rest, but infra has asked that we try not to do it with the extra things10:23
odyssey4mecan't remember why10:24
jrosserah right - i'll try to remember that10:24
jrosserjust got into the habit of pasting url from browser10:24
kaiokmosame10:24
*** DanyC has joined #openstack-ansible10:24
*** DanyC has quit IRC10:29
*** DanyC has joined #openstack-ansible10:32
odyssey4mejrosser: FYI if you're the second core reviewer on a patch you can add +w too once you've approved a patch10:32
jrosserok - i was a little unsure of the etiquette there10:33
jrosserif that did/didnt require a 3rd name10:33
*** bradm_ has quit IRC10:33
odyssey4menah if there're two +2's then +w is appropriate10:33
openstackgerritMerged openstack/openstack-ansible-os_tempest master: Add tempest test blacklist capability  https://review.openstack.org/58836210:34
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/queens: Add tempest test blacklist capability  https://review.openstack.org/58850510:34
*** cshen has quit IRC10:36
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/pike: Add tempest test blacklist capability  https://review.openstack.org/58850810:37
*** bradm has joined #openstack-ansible10:37
jrosserodyssey4me: does fixing typos in commit messages force stuff to be reckecked or is it smart?10:38
odyssey4mechanging the commit message will cause a recheck10:39
jrosserwell this is very odd - the lxc cache prep works here https://review.openstack.org/#/c/588466/ but i just watched the console here and it fails https://review.openstack.org/#/c/586635/10:44
odyssey4meis it actually pulling that dependent patch in?10:46
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Do no use tempest  https://review.openstack.org/58754410:49
*** cshen has joined #openstack-ansible10:50
*** shyambiradar has joined #openstack-ansible10:55
*** madorn has quit IRC10:57
*** madorn has joined #openstack-ansible10:58
*** vnogin has quit IRC10:59
*** vnogin has joined #openstack-ansible11:05
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Do no use tempest  https://review.openstack.org/58754411:07
*** czunker has quit IRC11:07
openstackgerritMerged openstack/openstack-ansible-os_tempest master: Disable tempest.scenario.test_minimum_basic temporarily  https://review.openstack.org/58836011:07
evrardjpjrosser: so much green: https://review.openstack.org/#/c/588466/411:10
evrardjphahah11:10
evrardjpjrosser: might want to move lxc_hosts bionic to voting?11:10
*** savvas has joined #openstack-ansible11:11
*** DanyC has quit IRC11:12
*** DanyC has joined #openstack-ansible11:13
*** shyambiradar has quit IRC11:15
*** shyambiradar has joined #openstack-ansible11:15
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Cleanup duplicate 'when'  https://review.openstack.org/58835411:17
*** DanyC has quit IRC11:17
openstackgerritMerged openstack/openstack-ansible-pip_install master: Allow the role to only configure pip, not install it  https://review.openstack.org/58784011:24
jrosserevrardjp: i'm still not convinced that the role tests work11:25
jrosserevrardjp: see here - https://review.openstack.org/#/c/586635/ that pattern exhibits in many many places11:26
openstackgerritMerged openstack/openstack-ansible-lxc_hosts master: Update to bionic 18.04.1 base image  https://review.openstack.org/58846611:27
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-os_keystone master: [TEST] bionic  https://review.openstack.org/58663511:34
*** shyambiradar has quit IRC11:34
*** shyambiradar has joined #openstack-ansible11:35
*** ansmith has quit IRC11:45
jrosserodyssey4me: i'm not sure i see anything in the logs about depends-on being applied, even when i'm pretty sure its worked11:46
*** savvas has quit IRC11:59
*** zul has joined #openstack-ansible12:01
*** savvas has joined #openstack-ansible12:08
*** shyambiradar has quit IRC12:10
*** rmart04 has joined #openstack-ansible12:14
*** savvas has quit IRC12:21
odyssey4mejrosser: well, you might find it by SHA - but you're more likely to see it via resulting task changes or whatever - in this case a different image being downloaded12:23
odyssey4meanyway, it's merged now12:23
odyssey4meevrardjp can we merge https://review.openstack.org/588191 now so that I can get on with doing all the role changes again12:28
evrardjpodyssey4me: ok12:32
evrardjpbypassing things12:32
odyssey4metyvm12:32
*** czunker has joined #openstack-ansible12:35
openstackgerritJames Denton proposed openstack/openstack-ansible master: [WIP] Adds haproxy endpoints for networking-ovn support  https://review.openstack.org/58406912:43
odyssey4memorning jamesdenton - FYI the integrated built tests for master should no longer have the tempest failures12:45
jamesdentonoh good. i was beating my head trying to figure out why12:45
openstackgerritFlorian Haas proposed openstack/openstack-ansible stable/queens: Drop --tags=haproxy_server-config from run-upgrade.sh  https://review.openstack.org/58853112:45
evrardjphaha12:45
jamesdentonbut, i did discover i was missing a few things for tests, so maybe it was better that way :D12:46
evrardjpjrosser: it didn't12:48
evrardjpcheck the current run start of the log12:48
evrardjpvs the time the other patch was merged:12:48
evrardjphttps://review.openstack.org/#/c/588466/ merged at 13:27. Current log is http://logs.openstack.org/35/586635/5/check/openstack-ansible-functional-ubuntu-bionic/66087d4/job-output.txt.gz12:49
evrardjpjrosser: little trick: abandon and reopen. Use the depends-on too : )12:49
evrardjpor be a little patient : D12:50
evrardjpI mean12:50
evrardjpmeh I can't properly speak12:50
evrardjpIf you edit the commit-message with a depends-on it will force things, and re-trigger gating, which should do the trick12:50
odyssey4melosing your ability to speak is a skill learned when working from home12:51
openstackgerritJames Denton proposed openstack/openstack-ansible-os_neutron master: [WIP] Adds networking-ovn support  https://review.openstack.org/58406712:54
odyssey4meevrardjp: if only people would listen to you the first time ;)12:56
odyssey4mebummer, there was a brilliant cartoon showing the evolution of a person working from home - I thought it was xkcd, but I can't seem to find it12:59
jamesdentoni feel like i'm letting myself go in some regards12:59
jamesdentonbeen WFH for a year now12:59
*** nicolasbock has joined #openstack-ansible13:00
odyssey4medid find this gem though: https://xkcd.com/1075/13:01
odyssey4mehaha, this is spot on: http://theoatmeal.com/comics/working_home13:02
odyssey4me'The horrible degradation of social skills' is the thing I was looking for13:02
odyssey4me"English is speak becoming hardness." :D13:03
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca-agent master: Switch to using project-templates  https://review.openstack.org/58846413:04
jamesdentonpretty spot on13:04
evrardjpaha13:05
* odyssey4me just discovered https://www.explainxkcd.com/13:05
evrardjpjamesdenton: you're one year in you said? :D13:06
jamesdentonYes, one year in. I'm still holding on to my humanity.. still try to shower semi-regularly13:07
evrardjpodyssey4me: it's a good friday then13:07
odyssey4methe comments here, so good: https://www.reddit.com/r/xkcd/comments/3tan6w/xkcd_1605_dna/13:09
odyssey4mejamesdenton: if you haven't grown a beard yet, you're still grooming13:09
evrardjpjamesdenton: when you said 1 year, I was precisely on the 1 year image. It was priceless.13:09
jamesdentonoh i've been working the beard for years now. It's no rip-van-winkle beard, though13:10
*** vnogin has quit IRC13:10
odyssey4mejamesdenton: When we next meet, I expect to find you looking akin to Hagrid.13:10
jamesdentonI do have to deal with commoners in this town semi-regularly13:10
jamesdentonhah13:10
evrardjpI need sleep I read "I do have to deal with coroners in this town"13:11
jamesdentonhmm13:12
evrardjpthat was super weird.13:12
jamesdentonget some rest :D13:12
*** ansmith has joined #openstack-ansible13:12
odyssey4meevrardjp: perhaps you're living in a cowboy western fantasy13:12
evrardjpodyssey4me: in space?13:13
evrardjpodyssey4me: like cowboy bebop?13:13
odyssey4meevrardjp: no coroners in that tale13:14
evrardjpor Firefly13:14
evrardjpodyssey4me: well, there is IIRC.13:15
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca master: Add support for CentOS 7  https://review.openstack.org/58854513:16
evrardjpthat was a good show.13:16
odyssey4meyeah13:17
odyssey4meBlack Lagoon is similar and pretty good.13:17
*** DanyC has joined #openstack-ansible13:18
*** suggestable has joined #openstack-ansible13:18
odyssey4meevrardjp jrosser shall we? https://review.openstack.org/58550413:20
openstackgerritMerged openstack/openstack-ansible-tests master: Prevent common MQ task set from synchronizing  https://review.openstack.org/58819113:20
odyssey4mejrosser: also a good patch needing another review: https://review.openstack.org/58537813:20
evrardjpodyssey4me: yup watched it too.13:21
evrardjpI think the coroner is in Black Lagoon though.13:21
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Replace tempest with healtchecks  https://review.openstack.org/58754413:22
odyssey4meevrardjp: yes, I think that's right - the style of Cowboy Bebop is much more focused on the humour, whereas in Black Lagoon it's more gritty13:23
* odyssey4me likes dem healt checks13:23
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Replace tempest with healtchecks  https://review.openstack.org/58754413:25
evrardjpodyssey4me: yeah I am thinking healthchecks should be simple to read, and easy to execute/teardown13:26
*** lbragstad has quit IRC13:26
evrardjpalso I think it's closer to our users. I would expect more users running things with ansible rather that run tempest :p13:26
odyssey4meevrardjp: mind upgrading your vote here? https://review.openstack.org/#/c/587870/113:27
evrardjptempest is a great tool for devs13:27
odyssey4meevrardjp: perhaps you're missing my intent with the comment13:27
evrardjp?13:27
odyssey4mehealt checks are very healty13:28
evrardjphaha13:28
* odyssey4me wonders what being healty is all about13:28
evrardjpI se13:28
odyssey4me:p13:28
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Replace tempest with healthchecks  https://review.openstack.org/58754413:28
odyssey4meSorry, it's friday. It's been a long week.13:28
evrardjpFriday?13:29
odyssey4me:p13:30
evrardjpjamesdenton: you go to the ptg by any chance?13:31
odyssey4meevrardjp: jamesdenton doesn't leave his room, except to get supplies13:32
evrardjpa room full of supplies odyssey4me ?13:32
odyssey4methat'd be supplies-ing13:33
evrardjphttp://gph.is/29cBCaY13:33
evrardjphaha13:33
* odyssey4me blames mhayden13:33
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca master: Add support for CentOS 7  https://review.openstack.org/58854513:33
evrardjpplecisely113:34
evrardjps/1/!/13:34
odyssey4meomg laughing so hard right now13:34
evrardjp:)13:35
jamesdentonevrardjp no i did not go to the PTG. I'm not l33t enough13:38
*** ansmith has quit IRC13:40
odyssey4mejamesdenton: UR l337 d4wG13:40
* odyssey4me puts himself to bed13:40
jamesdentonheh13:41
jamesdentonI hope to be in Berlin, though13:41
*** ansmith has joined #openstack-ansible13:44
cloudnullmornings13:52
jamesdenton++13:53
openstackgerritMerged openstack/openstack-ansible master: Fix gate log collection  https://review.openstack.org/58770813:56
*** czunker has quit IRC13:59
evrardjpjamesdenton: I am very sad.14:01
evrardjplet's drink some beverages in Berlin then!14:02
jamesdentonlet's!14:02
jamesdentonfirst round is on me14:02
* odyssey4me orders a double of Dalmore 62 Single Highland Malt Scotch (Matheson 1942)14:09
odyssey4metotally kidding14:09
jamesdentonDalmore 62 Single Highland Malt Scotch (Matheson 2018)14:11
odyssey4melet's see - one Dalmore 62 Single Highland Malt Scotch (Matheson 1942) @ $58,000 - is 16 one-and-a-half ounce shots, making it $3,625 per (american) shot14:13
odyssey4mein case you're all wondering: https://moneyinc.com/most-expensive-whiskey-bottles/14:15
* odyssey4me continues muttering to himself14:16
jamesdentonYou're getting Cutty Sark and you're gonna like it14:16
odyssey4mehahahaha14:16
*** rpittau has quit IRC14:18
odyssey4meAlright, I think it's time to call it a day for me. It's friday, and the weather's lovely. Someone chase evrardjp into the w/end.14:20
odyssey4meCheers folks!14:20
cloudnulltake care odyssey4me14:21
jamesdentonenjoy the weekend14:21
*** DanyC has quit IRC14:22
*** DanyC has joined #openstack-ansible14:22
openstackgerritKaio Kassiano Moura Oliveira proposed openstack/openstack-ansible-os_monasca master: Add support for openSUSE  https://review.openstack.org/58858414:29
*** dxiri has joined #openstack-ansible14:38
evrardjpgood weekend odyssey4me14:44
d34dh0r53evrardjp: go away :)14:51
d34dh0r53evrardjp: enjoy your weekend!14:51
*** FrankZhang has joined #openstack-ansible14:52
evrardjpd34dh0r53: I still have meetings and all14:52
d34dh0r53evrardjp: ohh, bummer man14:53
d34dh0r53I'm pretty sure it's against the Geneva Convention to schedule meetings on Friday afternoon14:53
evrardjpthat's fine my last one starts in 6 minutes and should only last 1h if it doesn't exceed its planned time (unlikely though)14:53
evrardjphaha14:54
evrardjpI smiled d34dh0r53 !14:54
d34dh0r53evrardjp: good :)14:54
guilhermesptake care odyssey4me !!14:56
*** cshen has quit IRC14:58
*** dxiri has quit IRC14:58
eumel8o/15:01
eumel8is there a way for a list of ip addresses in the same network sort by the last octed?15:03
*** spatel has joined #openstack-ansible15:08
*** ullbeking has quit IRC15:21
*** ullbeking has joined #openstack-ansible15:22
*** ullbeking has quit IRC15:22
*** ullbeking has joined #openstack-ansible15:23
*** ullbeking has quit IRC15:23
*** ullbeking has joined #openstack-ansible15:23
*** ullbeking has quit IRC15:24
openstackgerritMerged openstack/openstack-ansible master: Remove the unnecessary space  https://review.openstack.org/58281415:25
*** ullbeking has joined #openstack-ansible15:25
*** ullbeking has quit IRC15:25
openstackgerritMerged openstack/openstack-ansible master: Remove all MQ vhost/user and DB create tasks  https://review.openstack.org/58550415:25
*** ullbeking has joined #openstack-ansible15:25
*** ullbeking has quit IRC15:25
jamesdentonTFW the gate fails on something other than your patch15:28
bgmccollumM-m-m-m-m-m-m-erged...!15:29
openstackgerritMerged openstack/openstack-ansible master: Change the osa wrapper from a heredoc to a file  https://review.openstack.org/58263315:29
cloudnulleumel8 in ansible, probably not. there is the ipaddr filter which can do some of that but probably not specifically what you want.15:37
*** vnogin has joined #openstack-ansible15:37
cloudnullin python you could do that given the list of ips and simply calling sorted15:37
*** vnogin has quit IRC15:38
eumel8cloudnull: yeah, I saw it. also a shell task could do this with sort -n -t . -k 4,4 but nothing in jinja15:40
cloudnull++15:41
devxwhats the best way to update the /etc/hosts file?15:42
cloudnulldevx probably lineinfile15:42
devxsounds good15:43
mnasermorning everyone15:43
mnaserkaiokmo: nice job on the merge :)15:43
kaiokmothanks mnaser15:44
guilhermespyeah indeed! kaiokmo is doing a great job while I'm finishing some stuff out of monasca scope15:44
guilhermespI will come back porting grafana asap15:44
kaiokmothere's still work to do on monasca role and monasca-agent too15:44
openstackgerritMerged openstack/openstack-ansible master: Do not install linux-image-extra by default  https://review.openstack.org/58537815:45
kaiokmofor now I'm focusing on add support for centos and opensuse on both monasca and monasca-agent15:45
jamesdentonAny suggestions on pinning down the issue for timeouts on openstack-ansible-deploy-aio_lxc-ubuntu-xenial check?15:52
mnaserjamesdenton: i think there was a proposed check that odyssey4me pushed up to disable the check that has been timing out frequently15:54
jamesdentonthanks mnaser. Would this be it? https://review.openstack.org/#/c/588360/15:58
mnaseryes sir15:58
bgmccollumjamesdenton yeah thats the one, it merged a few hours ago. when was your last recheck?15:58
mnaserthats the one15:58
jamesdenton2-3 hrs ago15:58
eumel8mnaser: as the upcoming PTL I have some proposals for you  to review: https://review.openstack.org/#/c/586867/  https://review.openstack.org/#/c/586866/ https://review.openstack.org/#/c/586774/15:59
bgmccollumjamesdenton doesn't look like that check even made it to the tempest bits...so probably unrelated.16:00
eumel8don't know if you have a list on etherpad. won't pass the queue16:00
bgmccollumhas anyone though about task time regression testing?16:00
bgmccollumseems like we have the data16:00
-openstackstatus- NOTICE: The infra team is renaming projects in Gerrit. There will be a short ~10 minute Gerrit downtime in a few minutes as a result.16:01
mnasereumel8: that's interesting, i think we'll have to add something to release time to update the branch there in `_horizon_translations_pull`16:01
eumel8ok16:03
mnasereumel8: do you happen to have an extra few cycles to add a change into https://docs.openstack.org/openstack-ansible/latest/contributor/periodic-work.html#development-cycle-checklist under "After creating a new stable branch:" ?16:03
mnaseras a note under "For all the repos, update the eventual static files refering to master/previous branch name"16:03
mnaserif you don't, i can do it but it'd be a ton of help if you had a few spare cycles :)16:04
mnaserthat way wee don't end up deploying from master translation and break things16:04
bgmccollumjamesdenton under `setup-infrastructure.yml` the task `repo_build : Clone git repositories` took 43 minutes...which isn't normal...16:04
devxcloudnull have you tried to run macvlan in conjuction with netplan? i know it doesn't supported but I was wondering if there is a better way than using networkd-dispatch scripts :(16:04
*** gyee has joined #openstack-ansible16:09
jamesdentonthanks bgmccollum. have to step away but will check that out again16:10
cloudnulldevx everything i've been doing with networkd is in this role https://github.com/openstack/ansible-role-systemd_networkd16:12
*** ivve has joined #openstack-ansible16:12
cloudnullthe issue i had with macvlan yesterday is resolved. the route add trick worked a treat  :)16:12
cloudnullthe basic network setup i've been testing with is here http://paste.openstack.org/show/72725016:14
cloudnull^ look ma, no bridges!16:14
cloudnulland here's a fully configured node http://paste.openstack.org/show/727251 which is a compute/infra host16:15
cloudnullonly the brq that neutron creates.16:15
cloudnullhttp://paste.openstack.org/show/727252/16:15
*** suggestable has quit IRC16:20
*** udesale has quit IRC16:20
cloudnulland the contianers / vms http://paste.openstack.org/show/727254/16:20
devx@clou16:20
*** cshen has joined #openstack-ansible16:20
cloudnulldogs and cats, living together :)16:20
devxcloudnull i will check it out thanks!16:21
* mnaser twitches16:21
mnaserno gerrit16:21
mnaserwhat do16:21
*** FrankZhang has quit IRC16:22
cloudnullWAT?!16:23
cloudnullgerrit is down ?16:23
kaiokmoback again16:23
cloudnullwooooo, crisis averted16:23
cloudnullhahaha16:23
kaiokmoinfra team is updating things in gerrit16:23
cloudnullmakes sense.16:25
*** rmart04 has left #openstack-ansible16:26
*** cshen has quit IRC16:33
cloudnullanyone seen this error in the neutron logs http://paste.openstack.org/show/727256/  ?16:39
cloudnulldoesn't seem to be causing any failures. but its spamming the logs16:39
evrardjpmeetings over! I am done for today and the week-end. I will try to work on the release toolkit this week-end if I got 5 minutes.16:48
jrossercloudnull: my rocky/bionic AIO was totally full of neutron log spam but i didnt cat chance to check it out yet16:48
evrardjpeveryone: have a nice week-end.16:48
jrossero/ have a good one evrardjp16:49
cloudnulltake care evrardjp16:49
*** openstackgerrit has quit IRC16:49
*** openstackgerrit has joined #openstack-ansible16:51
openstackgerritMerged openstack/openstack-ansible-os_octavia master: Execute service setup against a delegated host using Ansible built-in modules  https://review.openstack.org/58232416:51
openstackgerritMerged openstack/openstack-ansible-os_octavia master: Execute image setup against octavia_service_setup_host  https://review.openstack.org/58816516:51
openstackgerritMerged openstack/openstack-ansible-os_octavia master: Execute flavor setup against octavia_service_setup_host  https://review.openstack.org/58816616:51
openstackgerritMerged openstack/openstack-ansible-os_horizon master: fix execution of horizon_translations_update.yml  https://review.openstack.org/58686616:53
*** cshen has joined #openstack-ansible16:55
*** gillesMo has quit IRC16:59
*** cshen has quit IRC17:01
openstackgerritMerged openstack/openstack-ansible-os_octavia master: Execute network setup against octavia_service_setup_host  https://review.openstack.org/58817417:16
openstackgerritMerged openstack/openstack-ansible-os_octavia master: Make the project home link in next line  https://review.openstack.org/58458717:16
jamesdentonbgmccollum any hints on where you found that task duration?17:21
bgmccollumjamesdenton it was this change, right? -- https://review.openstack.org/#/c/584069/617:22
jamesdentonyes17:22
bgmccollumhttp://logs.openstack.org/69/584069/6/check/openstack-ansible-deploy-aio_lxc-ubuntu-xenial/c9d95e3/logs/ara-report/17:22
bgmccollumI saw setup-infrastructure.yml took 72 minutes, so clicked into that playbook, and sorted the tasks by duration...17:23
jamesdentonahh ok, i was looking at plays not tasks.17:23
bgmccollumyeah, the clones should only be about 5-10 minutes in my experience, so something was definitely wrong. i kicked off a recheck already...17:24
jamesdentonoh thanbk you17:24
bgmccollum👍17:24
bgmccollumwould be nice to see some task time averaging / regressing testing17:25
-openstackstatus- NOTICE: Project renames and review.openstack.org downtime are complete without any major issue.17:27
jrosserbgmccollum: if you fancy this could be applied to the repo git clone https://review.openstack.org/#/c/588372/17:29
jrosserthat was my plan to give that a try but not had time to wire it in yet17:29
* bgmccollum clicks17:30
jrosserit works locally but i've messed up installing the pip packages in the gate i think17:30
jrosserand it gives 10x or more speed increase on git cloning, so long as the repos and underlying network are up to it17:31
bgmccollumreally should have called it git_fireball_mode17:31
jrosserthats only used during bootstrap-ansible as it stands, but theres no reason why the same thing can't fit into the repo build as well17:32
*** ullbeking has joined #openstack-ansible17:33
*** ullbeking has quit IRC17:33
*** ullbeking has joined #openstack-ansible17:35
bgmccollummight be able to get some parallelism out of using delegate_to?17:36
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] check if gpg agent is running, and can be made to run  https://review.openstack.org/58862917:43
openstackgerritByron McCollum proposed openstack/openstack-ansible master: [TEST] Parallelize Git Clones  https://review.openstack.org/58863017:47
spatelguys can you help me to pint point my Queens horizon issue17:47
spatelits extremely slow17:47
jrosserbgmccollum: that whole play targets localhost17:48
spatelI have test basic stuff17:48
spatelmemcache etc.. they all looks good17:48
bgmccollumjrosser so, the delegate_to should parallelize each with_item...if ansible still behaves that way...17:49
spatelDid you guys experience slowness in your Queens deployment?17:49
openstackgerritMerged openstack/openstack-ansible master: Allow flexible tag for Rabbit monitoring user  https://review.openstack.org/57953617:51
spatelmy "nova list" taking 3 to 4 second17:54
spateldoes that means its not horizon issue?17:54
jrosserspatel: horizon makes a ton of calls to the underlying APIs17:54
bgmccollumspatel -- try "openstack --timing server list" and see which calls are slow...17:55
jrosserso if you see the cli being slow too then it's probably not something within horizon17:55
bgmccollumspatel these are the results from my queens AIO from the utility container -- http://paste.openstack.org/show/727271/17:58
bgmccollumranges from 2.25s to 0.9s17:58
spatelbgmccollum: here is the output http://paste.openstack.org/show/727272/17:58
bgmccollumand that AIO is a VM itself...17:59
bgmccollumlet me compare to my pike...17:59
spatelI have metal and with 40 core and 32G mem18:00
spatelmy pike was extremely fast compare to current Queens so something is wrong18:00
bgmccollummy pike is faster, but its running on metal, compared to queens in vm -- http://paste.openstack.org/show/727273/18:03
spateljrosser: I understand but this is extremely slow.. hard to watch this slowness18:03
bgmccollumnot sure the cause of the differences you are seeing...18:03
spatelanybody using queen who can verify18:04
spateli was very happy with pike but then i upgrade and now i am here stuck18:04
openstackgerritMerged openstack/openstack-ansible master: Simplify bootstrap-ansible.sh  https://review.openstack.org/58390518:05
openstackgerritMerged openstack/openstack-ansible stable/queens: Drop --tags=haproxy_server-config from run-upgrade.sh  https://review.openstack.org/58853118:05
spatelWhat i should look for slowness ?18:05
spatelCan you give me some tip or hints i should look18:05
*** electrofelix has quit IRC18:06
bgmccollumProbably some MySQL tuning that could be done, but I'm far from an expert on that18:07
spatelany thing i should look in mysql ? like any query etc..?18:08
bgmccollumlike i said, not an expert.18:08
spatelREDHAT Document saying: When you click on admin tab, in backgroud it access/list all configured API and if there is any delay in API response then you will observe delay in horizon admin tab. Check how much time it take from OpenStack CLI to fetch the data. For ex: nova list , cinder list etc.18:08
cloudnullwelp that was a nice rabbit hole. https://github.com/openstack/neutron/blob/master/neutron/services/trunk/drivers/linuxbridge/agent/trunk_plumber.py#L123-L156 - seems neutron is working as expected. RE: http://paste.openstack.org/show/72725618:11
openstackgerritByron McCollum proposed openstack/openstack-ansible master: [TEST] Parallelize Git Clones  https://review.openstack.org/58863018:12
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] check if gpg agent is running, and can be made to run  https://review.openstack.org/58862918:18
*** openstackgerrit has quit IRC18:19
spatelbgmccollum: what kind of LB you have ?18:19
spatelF5 or Haproxy?18:19
bgmccollumhaproxy18:19
spatelwhat is the box configuration?18:19
bgmccollumQueens? Or Pike?18:19
spatelin Production metal18:20
bgmccollumQueens in a VM18:20
bgmccollumPike is a Ryzen 7 1700 / 32GB Memory18:20
spatelpossible my haproxy is slow18:20
spatelbecause all API call flying over LB18:21
* mhayden rips threads18:24
jamesdentonduuuuuuude this DPDK config just keeps getting better and better....18:26
jamesdenton</s>18:26
* bgmccollum has that sinking feeling its going to timeout18:41
bgmccollumphoto finish18:41
*** coreywright has quit IRC18:42
*** openstackgerrit has joined #openstack-ansible18:45
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] check if gpg agent is running, and can be made to run  https://review.openstack.org/58862918:45
spatelbeam.smp eating lots of CPU on infra* nodes18:47
spatelis that rabittmq ?18:47
spatelis that normal for idle cluster?18:48
*** coreywright has joined #openstack-ansible18:54
spatelbgmccollum: does it normal ?18:57
*** cshen has joined #openstack-ansible18:57
*** cshen has quit IRC19:02
*** cshen has joined #openstack-ansible19:09
openstackgerritMerged openstack/openstack-ansible-os_octavia master: Move role setup into service setup tasks  https://review.openstack.org/58817519:09
openstackgerritMerged openstack/openstack-ansible-os_octavia master: Execute security group setup against octavia_service_setup_host  https://review.openstack.org/58817619:09
bgmccollumspatel -- A Pike Lab with 40 Cores has a load average of 0.15, with rabbit bouncing around 35% of a single core19:17
spatel31.3  0.8   3:20.88 beam.smp19:17
spatel31%19:17
spatelDo you feel like your horizon is slow?19:18
bgmccollumno19:18
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] check if gpg agent is running, and can be made to run  https://review.openstack.org/58862919:19
spatelbgmccollum: can you tell me is this looks good to you? http://paste.openstack.org/show/727277/19:20
spateli have these setting in my use_variable19:20
bgmccollumi really don't mess with any of those, i just use the defaults19:22
bgmccollumcant really speak on when or why to change any of them19:23
spatelHmm!19:24
spateli think i should remove all those setting and try19:25
spatelto see how it works19:25
spatelbgmccollum: in short you have user_variable empty?19:26
bgmccollumcan i ask why you changed these from the defaults?19:28
spatelto control CPU resources19:29
spateli want to to increase them on demand19:29
spatelbut now i am going to comment them out and going to run playbook to see what it does19:29
*** mhayden has quit IRC19:32
*** mhayden has joined #openstack-ansible19:33
*** cshen has quit IRC19:33
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] check if gpg agent is running, and can be made to run  https://review.openstack.org/58862919:39
spatelAlso i am planning to move my haproxy to F519:40
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Cleanup files and templates using smart sources  https://review.openstack.org/58864719:40
cloudnullmnaser ^ if you're around mind having a loot at that ?19:41
mnasercloudnull: lookin19:41
cloudnullthanks !19:41
cloudnullwas trying to copy a lot of what you had started elsewhere.19:42
jrossercloudnull: i am closing in on why bionic lxc cache prep fails - i might need your help later19:43
cloudnulli'm around for a while longer. let me know.19:44
cloudnullhappy to help make that go19:44
jrosserok, cool, thanks19:44
logan-cloudnull: suggestion in there for user friendliness19:51
mnaserhaha wow19:51
cloudnull++ logan- will do19:51
mnasercloudnull: same comment as logan- alongside a few other things19:51
mnasercloudnull: please ping when you push a revision, i can have a look at it right away19:52
cloudnull++19:52
* cloudnull revising 19:52
mnaserlogan-: mind rebasing ?https://review.openstack.org/#/c/582272/19:52
logan-ah yep, one sec. thanks mnaser19:53
jrossercloudnull: so this lxc-cache prep business seems to be failing for gpg related reasons, and i've thrown in a bunch of stuff here to try to figure it out https://review.openstack.org/#/c/588629/19:54
openstackgerritLogan V proposed openstack/openstack-ansible master: Move get-ansible-role-requirements to scripts  https://review.openstack.org/58227119:54
openstackgerritLogan V proposed openstack/openstack-ansible master: Use an absolute path for the role requirements file  https://review.openstack.org/58227219:54
jrossercloudnull: this is useful to see http://logs.openstack.org/29/588629/3/check/openstack-ansible-functional-ubuntu-bionic/389dde6/logs/host/lxc-cache-prep-commands.log.txt.gz19:55
jrosserand i have a hunch it is caused by this http://logs.openstack.org/29/588629/5/check/openstack-ansible-functional-ubuntu-bionic/d911d08/logs/host/lxc-cache-prep-commands.log.txt.gz19:55
openstackgerritLogan V proposed openstack/openstack-ansible master: Use ZUUL_SRC_PATH to pass the path to Zuul git sources  https://review.openstack.org/58227319:56
jamesdentonbgmccollum no dice. bummer.19:56
spatelI just compare my queens with my LAB machine which is running on VMware that is super fast compare to my production :(20:03
bgmccollumjamesdenton look like it was in the middle of setting up tempest...almost made it. same thing happened to mine.20:03
spatelwhat a painful20:03
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Cleanup files and templates using smart sources  https://review.openstack.org/58864720:03
jamesdentonalright, well. thanks for the assist. I have another one failing on galera repo build. The mirror is missing the package. oy vey20:04
cloudnullspatel: queens in vmware is faster or slower?20:05
*** cshen has joined #openstack-ansible20:05
cloudnulllogan- mnaser updated that pr20:05
cloudnulljrosser looking at that log20:05
spatelqueens in vmware running faster20:05
spatelcloudnull: ^^20:05
cloudnullis queens in prod now too ?20:05
spatelcloudnull: but on my production where we have lots of core running super slow20:05
cloudnullor still pike20:05
cloudnullah is that a hardware issue?20:06
spatelI am all queens (I have upgraded pike)20:06
cloudnullnice!20:06
spatelI have 3 physical infra nodes20:06
spatel40 core 32G memory SSD20:06
openstackgerritLogan V proposed openstack/openstack-ansible master: Use ZUUL_SRC_PATH to pass the path to Zuul git sources  https://review.openstack.org/58227320:07
*** savvas has joined #openstack-ansible20:07
cloudnulljrosser I see - E: Can not write log (Is /dev/pts mounted?) - posix_openpt (19: No such device)20:09
cloudnullmaybe that's related?20:09
savvashey guys, to deploy Magnum or Octavia, how would I make sure the containers get created and the files downloaded to the repo server?20:09
jrossercloudnull: i figure that somethings changed very recently that now means extra stuff needs to be mounted in the chroot to make the cache prep script work20:10
jrosserif you diff the xenial/bionic prep scripts they are not quite the same, gpg related20:10
cloudnulljrosser is so we can `mknod -m 0666 /dev/pts/{0..5} c 136 $INT`20:11
spatelwhen i click on "instance" tab my browser wheel just spinning even i can see all contents on page20:11
spatelcloudnull: ^^20:11
cloudnullspatel do the CLI clients return fast?20:12
mnaserlogan-: what does 3pci stand for?20:12
spatelcloudnull: is there anyway or plugin i can install on browser to see where its getting stuck20:12
logan-mnaser: 3rd party ci20:12
mnaserdamn i was trying to figure that out for the longest time.20:12
spatelcloudnull: CLI is fast20:12
cloudnullspatel chrome as the developer panel20:12
cloudnullif the CLI is fast I'd suspect something in path making it slow20:12
mnasercloudnull: regarding that change, my concern is that, what if the user ran ansible once with overrides, then ran it again without them20:13
mnaserthe overrides won't be removed20:13
spatelyes.. something is wrong...20:13
cloudnullmnaser i see20:13
mnaserso to remove the overrides would require actually just deleting the lines in the file20:13
jrossercloudnull: i think that gpg-agent wants to make stuff in /run/user/0/gnupg20:14
mnasercloudnull: maybe we can always download them but we download them from the virtualenv rather than /etc.. im not sure :X20:14
cloudnullyea I think that's a side effect of not carrying the templates.20:15
openstackgerritGerman Eichberger proposed openstack/openstack-ansible-os_octavia master: ssh-key should land in $HOME  https://review.openstack.org/58865220:15
cloudnullI guess I could case those files and make a remote backup before mucking with them20:15
xgerman_^^ might fix Octavia gates20:16
cloudnullwhen no override is present, copy remote from the backup file20:16
spatelcloudnull: do you think it could be DNS ?20:17
mnasercloudnull: i think that might make sense20:17
spatelwhy horizon depend on DNS ?20:17
cloudnullwhat's tracepath tell you ?20:17
cloudnulltoo many hops ?20:18
cloudnulljrosser That would make some sense20:18
cloudnullwe can try adding the new devices in the chroot to https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/templates/prep-scripts/_container_sys_setup.sh.j220:18
openstackgerritJonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: [TEST] check if gpg agent is running, and can be made to run  https://review.openstack.org/58862920:19
jrossercloudnull: ^ that may be bogus but lets see20:19
jrosserjust mkdir -p /run/users/020:20
cloudnullhttp://paste.openstack.org/show/727295/20:20
cloudnullsomething like that would create those devices.20:20
openstackgerritLogan V proposed openstack/openstack-ansible master: Move get-ansible-role-requirements to scripts  https://review.openstack.org/58227120:21
openstackgerritLogan V proposed openstack/openstack-ansible master: Use an absolute path for the role requirements file  https://review.openstack.org/58227220:21
openstackgerritLogan V proposed openstack/openstack-ansible master: Use ZUUL_SRC_PATH to pass the path to Zuul git sources  https://review.openstack.org/58227320:21
spatelcloudnull: check this out http://paste.openstack.org/show/727296/20:21
spatelcloudnull: do you think 3 second is slow?20:22
cloudnullspatel that seems normal20:31
spatelI have shutdown infra-02 to see what is happening20:32
spateli hate browser spinning wheel :(20:34
*** ansmith has quit IRC20:34
spateli got 504 gateway timeout error20:34
spatellook like my HAproxy still sending request to down node20:34
spatelthat is not good20:35
spatelhow do i check haproxy status for down node?20:35
jamesdentonhatop?20:36
jamesdentonor check the logs20:36
openstackgerritGerman Eichberger proposed openstack/openstack-ansible-os_octavia master: ssh-key should land in $HOME  https://review.openstack.org/58865220:36
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Cleanup files and templates using smart sources  https://review.openstack.org/58864720:37
cloudnullmnaser how about that?20:37
spatel /var/log/haproxy/ is empty20:37
jamesdenton /var/log/haproxy.log20:39
spatelcrap!! this is very odd20:45
spateli should down infra-02 and my mysql is down20:45
spatel01 & 03 showing mysql is dead20:45
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Cleanup files and templates using smart sources  https://review.openstack.org/58864720:45
spateloh! wait..20:46
spateli think i am on wrong node20:46
spatelphew!!20:46
cloudnulllogan- mnaser updated with a couple comments. logan- good catch on the recursive option .20:46
spateljamesdenton: there is no log file /var/log/haproxy.log20:47
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Cleanup files and templates using smart sources  https://review.openstack.org/58864720:51
cloudnulland that should resolve the last two idempotentcy issues.20:51
jamesdentonspatel maybe syslog would see it, then. not sure20:52
spateli did check in syslog and its not there20:52
openstackgerritMerged openstack/openstack-ansible master: Normalise containers-lxc-* playbook structure  https://review.openstack.org/58622620:53
openstackgerritMerged openstack/openstack-ansible master: Add openstack client to Ansible venv  https://review.openstack.org/58249620:55
spatelif this thing won't get fix then i have to go back to pike :(20:56
spateli can't just keep fixing stuff everyday, just tired of this thing20:57
spatelmy BOSS going to kick me out :)20:57
spatelI am going to try tonight otherwise i am planning to re-install everything20:58
cloudnullspatel what's down?20:58
jamesdentonI did a Pike to Queens upgrade last week and it wasn't super bad. Had some galera wonkiness but thats about it. Probably my fault20:58
cloudnullis the galera cluster up?20:58
spatelI am just trying to find out why Horizon is slow20:58
spatelno galera is happy20:58
spatelthat was my mistake20:59
spateli was looking into wrong box20:59
cloudnullok20:59
jamesdentonSlow is relative. Is everything functional at the moment?20:59
cloudnull^20:59
spatelDo you want to take a look i can get you on shared screen20:59
spatelEverything is working fine... i can create VM etc..20:59
cloudnullim not in a place to do that atm20:59
jamesdentonme either, about to hop on a call21:00
spatelbut its just painful slow..21:00
jamesdentonlike.. 10 secs? 1 min?21:00
cloudnullare you seeing anything in the error log?21:00
cloudnulltimeouts ?21:00
spatelI am going to look into logs now to see...21:00
spatelgoing to look into horizon log first21:00
cloudnullwas the cluster responding at roughly the same speed as it was on pike?21:00
bgmccollumCan I get a +W ? -- https://review.openstack.org/#/c/587229/21:00
* cloudnull looking 21:01
cloudnulldone.21:01
bgmccollumFinally got through the gauntlet21:01
bgmccollumCheers!21:01
spatelcloudnull: in my LAB queen is fast enough compare to production one..21:01
spateli am surprised why...??21:02
spatelLAB machine is running on VMWare21:02
cloudnullin the same datacenter?21:02
spatelyes21:02
spatelsame datacenter21:02
spateleven before pike was running very fast on same hardware21:02
spatelI am thinking i should bootstrap and run all playbook again21:03
cloudnullshouldn't hurt .21:03
spatelbut first let me find logs21:03
cloudnullmaybe something is  in a funky state post upgrade21:03
cloudnullI'd pore through the logs and see if there's anything obvious21:04
jamesdentonand just to confirm... CLI clients operate as expected?21:04
spatelCLI is working fine21:04
cloudnullalso -- https://github.com/openstack/openstack-ansible-ops/tree/master/elk_metrics_6x :)21:04
cloudnullgood for log processing :P21:04
jamesdentonhave you tried restarting horizon containers?21:05
spatelcloudnull: thanks  but at present i need to fix this issue, i may later add this in production21:05
spateljamesdenton: not yet21:05
jamesdentonmay be worth a shot21:05
cloudnullmemcached can cause issues like this21:05
jamesdenton^^21:05
cloudnullmaybe you need to up the memory limit21:06
spateli am running all those setting default21:06
cloudnulltrying upping them21:07
cloudnullhttps://github.com/openstack/openstack-ansible-memcached_server/blob/master/defaults/main.yml#L4621:07
cloudnullhttps://github.com/openstack/openstack-ansible-memcached_server/blob/master/defaults/main.yml#L5121:07
spatellet me try couple of stuff and see21:08
spateli will report here whatever i find or not find...21:08
cloudnullspatel what's the memory on your infra nodes?21:09
spatel32G21:09
spatelhttp://paste.openstack.org/show/727308/21:10
cloudnullif you do want to up the memory usage, `openstack-ansible memcached-install.yml -e memcached_connections=8192 -e memcached_memory=10240`21:11
cloudnullthat would double the connections and ste the cached memory to 10G21:12
cloudnullup from 8G in your case.21:12
spatelhmm! but i have 32G so is that enough to set that limit?21:12
spatelI am surprised i have 4G memory on LAB machine and they are running fast :)21:13
cloudnullbetter: `openstack-ansible memcached-install.yml -e memcached_connections=8192 -e memcached_memory=10240 -e memcached_threads=8`21:13
spatelsomething else going on here21:13
cloudnulllab has fewer nodes?21:13
spatelin LAB i have same setup like production 3 infra node 3 compute 3 ceph21:13
spatelall running on VMware21:13
cloudnullif you dont want to up the memory limit you can omit that setting21:14
spateli am gona try everything to see what is wrong21:14
mnasercloudnull: any reason you did a copy using command but then later with remote_src after?21:15
cloudnullif it's fast in one environment, and slow in another I'd suspect either something is wrong in the environment or something in the path to the environment is slow (bad routes, misconfigured switches, bad cable, etc)21:15
cloudnullmnaser I only wanted that copy to run once.21:15
*** savvas has quit IRC21:16
cloudnullwhich is why I added the creates arg21:16
cloudnullhowever if there's a way to do the same with the copy module I'd gladly do that instead21:16
* cloudnull might need to tell that command to skip lint 21:17
spatelhmm!21:17
*** cshen has quit IRC21:17
spateltailing horizon logs and other browser accessing horizon - nothing in logs21:19
spatel[pid 10589] (70007)The timeout specified has expired: [client 172.28.0.8:42494] AH01581: Timeout while writing data for URI /static/dashboard/css/9cd30430f096.css to the client, referer: https://10.30.2.9/project/instances/21:20
spateljust notice this error21:21
spatel10.30.2.9 is my external LB ip21:21
spatelRestarting horizon service21:23
cloudnullwhat about the api from outside the environment21:23
cloudnullis that slow ?21:23
spatel systemctl restart httpd  just hung on Infra-0321:24
spatelall other node quickly restart httpd21:24
spatelit took 2 minute to restart httpd on that node21:26
spatellet me dig into that21:27
spatelon other nodes its took 3 second21:27
cloudnullmaybe infra3 was responding but had a hung process21:27
spateli am thinking i should shutdown that node and see how horizon behaving21:28
spatelnow restart happened quick21:30
*** pcaruana has quit IRC21:38
spatelI am running developer tool and its showing  TTFB - took 15 second, means i got reply back after 15 second and then it took 16 second to download contents21:50
*** cshen has joined #openstack-ansible21:51
spatelI am planning to shutdown 2 nodes so its to troubleshoot with single node21:51
*** KeithMnemonic has quit IRC21:53
*** cshen has quit IRC21:55
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Cleanup files and templates using smart sources  https://review.openstack.org/58864722:05
cloudnull ^ sorry mnaser logan- needed another pr, we still offer deployments with dragonflow and vpnaas so i still need those rootwrap files22:06
cloudnullbut all the rest are still gone :)22:06
*** greatgatsby has joined #openstack-ansible22:14
spatelcloudnull: what is the best way to re-install everything here ?22:14
spateljust destroy all container and re-run all playbook?22:14
*** spatel has quit IRC22:22
cloudnullyes.22:26
cloudnullopenstack-ansible contianers-lxc-destroy.yml22:26
cloudnullthen rerun the playbooks22:27
*** nicolasbock has quit IRC22:43
*** tosky has quit IRC23:04
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Cleanup files and templates using smart sources  https://review.openstack.org/58864723:05
openstackgerritMerged openstack/openstack-ansible master: Cleanup duplicate 'when'  https://review.openstack.org/58835423:06
openstackgerritMerged openstack/openstack-ansible master: Fix Bootstrap AIO Passing Multiple BOOTSTRAP_OPTS  https://review.openstack.org/58722923:06
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts master: [TEST] check if gpg agent is running, and can be made to run  https://review.openstack.org/58862923:09
openstackgerritTony Fontenay proposed openstack/openstack-ansible-lxc_container_create master: CentOS : Fix TASK lxc_container_create : Execute first script  https://review.openstack.org/58868023:13
openstackgerritByron McCollum proposed openstack/openstack-ansible stable/queens: Fix Bootstrap AIO Passing Multiple BOOTSTRAP_OPTS  https://review.openstack.org/58883923:24
openstackgerritByron McCollum proposed openstack/openstack-ansible stable/pike: Fix Bootstrap AIO Passing Multiple BOOTSTRAP_OPTS  https://review.openstack.org/58884023:25
*** ullbeking has quit IRC23:41
*** ullbeking has joined #openstack-ansible23:41
*** ullbeking has joined #openstack-ansible23:41

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!