Wednesday, 2015-08-05

*** SourabhP has quit IRC00:01
*** banix has joined #kolla00:03
*** athomas has quit IRC00:04
*** achanda_ has quit IRC00:11
*** shadower has quit IRC00:23
*** shadower has joined #kolla00:23
*** weiyu has quit IRC00:26
*** weiyu has joined #kolla00:28
*** vinkman has left #kolla00:37
SamYaplesdake_: ping00:39
*** weiyu has quit IRC00:40
*** weiyu has joined #kolla00:40
SamYapledaneyon_: how are you breaking this stuff?!?00:41
SamYapledaneyon_: you are coming up with things ive never seen!00:41
SamYapledaneyon_: are you running the liberty-2 images sdake pushed up into the Dockerhub or self-built images?00:41
daneyon_SamYaple last time i just pulled from kollaglue00:42
SamYapleok daneyon_, I havent' tested those images yet. I will tonight.00:43
*** weiyu has quit IRC00:44
SamYapledid you see kolla + big tent merged?00:44
SamYaplehttps://review.openstack.org/#/c/206789/00:44
daneyon_SamYaple ya... pretty awesome!!! Congratulations.00:45
SamYapleim excited. nice to have all our work validated!00:45
*** weiyu has joined #kolla00:48
rhalliseydaneyon_, build them dewd00:51
rhalliseyI have to build them a few times to get them to work00:51
rhalliseyhad some packaging issues00:51
*** banix has quit IRC00:57
*** weiyu has quit IRC00:58
*** echoingumesh has quit IRC01:04
*** echoingumesh has joined #kolla01:04
*** ashuk has quit IRC01:04
*** jasonsb_ has quit IRC01:08
*** echoingumesh has quit IRC01:08
*** diogogmt has joined #kolla01:19
*** erkules has joined #kolla01:20
*** ashuk has joined #kolla01:21
*** erkules_ has quit IRC01:22
*** weiyu_ has joined #kolla01:29
*** ashuk has quit IRC01:30
*** jruano has quit IRC01:30
*** jruano has joined #kolla01:44
*** unicell1 has quit IRC01:48
*** daneyon_ has quit IRC01:49
*** rhallisey has quit IRC01:55
*** Haomeng has joined #kolla02:01
*** sdake_ has quit IRC02:04
*** dims has quit IRC02:04
*** sdake has joined #kolla02:04
*** achanda has joined #kolla02:07
*** achanda has quit IRC02:14
*** achanda has joined #kolla02:28
sdakesamyaple yes I was at the tc meeting today to help get us approved ;-)02:30
sdakesamyaple daneyon is running with only 50gb / partition02:30
sdakewhich I don't think is sufficient02:30
*** achanda_ has joined #kolla02:54
*** achanda has quit IRC02:55
*** dolpher has joined #kolla02:56
openstackgerritSam Yaple proposed stackforge/kolla: Reduce dependencies on the dest nodes  https://review.openstack.org/20845102:57
*** achanda has joined #kolla03:03
*** achanda_ has quit IRC03:03
*** jasonsb has joined #kolla03:12
*** diogogmt has quit IRC03:21
openstackgerritYongfeng Du proposed stackforge/kolla: Install git for ubuntu source base  https://review.openstack.org/20933103:30
*** achanda has quit IRC03:38
sdakedims around?03:52
SamYaplesdake: would he be around in this channel?03:53
sdakesomeimtes03:54
sdakei am asking asalkeld in #heat03:54
*** coolsvap|away is now known as coolsvap03:54
sdakehe will know03:54
SamYaplemy stacklytics stats got merged so that should update as well03:54
sdakehe says it slurps data once a day03:55
sdakebut it has been more then a day with no change so ymmv03:55
*** jruano has quit IRC03:55
SamYapleso how does the repo switch happen?03:56
SamYapleto hte openstack namespace i mean03:56
coolsvapSamYaple, your change of affiliation is merged i think, the stackalytics refresh may take couple of days03:57
coolsvapdepending on its refresh cycle03:57
*** jruano has joined #kolla03:57
SamYaplecoolsvap: i see some updates in stats (Rackspace shows 36 commits) will it refresh past stats as well?03:57
SamYaple36 reviews*03:58
sdakethat is a different samyaple I think03:59
sdakeit was there already03:59
SamYapleno those 36 reviews are from me03:59
coolsvapSamYaple, it all depends on the date you have mentioned in the default_dat03:59
SamYaplebut it should be like 26003:59
sdakeoh today you mean?03:59
SamYapleyea04:00
coolsvapits showing 26604:00
SamYapleits updated a bit04:00
coolsvaphttp://stackalytics.com/?release=all&metric=marks&project_type=all&user_id=s8m04:00
sdake[20:59:52]  <sdake>asalkeld will stacklaytics update old reviews for the affiliation change from the past?04:00
sdake[21:00:01]  <sdake>and if not is there someone we can bug to fix it :)04:00
sdake[21:00:05]  <asalkeld>sdake: yeah04:00
sdake[21:00:10]  <asalkeld>(it does)04:00
SamYaplecoolsvap: but not on the kolla page04:00
SamYapleoh04:00
SamYaplenvm04:00
SamYapleit likterally just updated04:00
SamYaplepage refresh04:01
coolsvapSamYaple, it is same on kolla page as weel04:01
coolsvapwell04:01
SamYapleyes coolsvap. i did another refresh and its changed for me04:01
coolsvapstackalytics is all about when they refresh04:01
coolsvapSamYaple, regarding the ansible cinder review i had a couple of doubts04:02
sdakeour diversity looks even more fantastic now samyaple thanks for the change :)04:03
SamYaplecoolsvap: ok04:03
coolsvapregarding the cinder.conf template04:03
SamYapleyea04:03
coolsvapi picked up the minimum values available in a devstack env04:03
coolsvapi am not sure about the driver part04:04
SamYaplethe lvmdriver section should not be there, anything related to lvm should be removed (what if i use ceph?)04:04
coolsvaphmm04:04
SamYaplethe nova_catalog_admin_info is set to a default value04:04
sdakeor alterntaely lvm vs ceph shoudl be a config option04:04
SamYapleiscsi should be removed (what about ceph)04:04
SamYapleagreed, it could be in an if block04:05
SamYaplethats fine too, but not hardcoded like it is04:05
sdakeright if block is the way to go04:05
SamYaplethe logging stuff should just be straight up removed04:05
SamYaplevolume_clear should also be removed (unneeded and actually bad if you use ceph)04:05
SamYapleifblock even04:06
SamYaplejasonsb: did you hear back from your rabbitmq contact? I have some questions04:07
coolsvapSamYaple, one question04:12
coolsvapif we use if block for volume drivers is there a way we can import a block of key-value pairs04:13
coolsvapor do we need to hard code the values for both lvm and ceph?04:13
SamYaplewhat do yo umean coolsvap? like do a jinja2 include?04:14
coolsvaplike the options will be different for both lvm and rbd04:14
SamYapleright so well need if blocks04:15
SamYaplebut they don't really need to be configurable04:15
SamYaplesince the deployer can override everyhting in the augment files04:15
SamYaplejust an opinionated section of options04:15
openstackgerritSwapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add ansible support for Cinder  https://review.openstack.org/20886804:30
*** weiyu_ has quit IRC04:31
openstackgerritSwapnil Kulkarni (coolsvap) proposed stackforge/kolla: Fedora - Install from source for Designate  https://review.openstack.org/20933604:32
openstackgerritSwapnil Kulkarni (coolsvap) proposed stackforge/kolla: Adds push to build script  https://review.openstack.org/20826604:40
coolsvapSamYaple, removed pep8 issues in https://review.openstack.org/#/c/208266/04:40
*** achanda has joined #kolla04:41
*** jruano has quit IRC04:46
SamYaplecoolsvap: please dont patch over my commits04:49
SamYapleit messes with the stats04:49
openstackgerritSam Yaple proposed stackforge/kolla: Adds push to build script  https://review.openstack.org/20826604:51
sdakeit doesn't mess with the stats if you add a co-authored-by line04:52
SamYaplesdake: i added it04:52
sdakestackalytics keeps the stats in that case04:52
SamYapleand it does actually mess with the commiter stats04:52
SamYaplei updated the co-authored stats for coolsvap04:52
sdakedoes it?04:52
SamYapleit does04:52
sdakeI thought it tracked both04:52
sdakerather it treated the commit s coming from both people04:53
SamYaplewell either way i added the co-athoer which coolsvap was missing04:53
sdakei am pretty sure it treats the commit as coming from both04:53
SamYaplebut its best to cordinate two people working on the same patchset since i had local changes and wouldn't have b=picked up both04:53
sdakeagree04:53
sdakeits ok imo to take over a review for someone with permission04:54
SamYapleagreed04:54
sdakeor if they are out for a week or something and the work needs to be done04:54
SamYaplethats perfectly fine04:54
coolsvapSamYaple, i am sorry, but i dont believe it will mess up with your stats, the change set will still be authored by sam yaple04:55
SamYaplecoolsvap: yea but you added the fixes so you need to add co-authored-by04:55
coolsvapi didnt wanted to add co-author since i just helped up with a very minor thing04:55
SamYapleno no you made changes, you get that too04:55
coolsvapand it was up there for couple of days so thought might just fix that up04:55
SamYaplebut yea if you could just reach out the the author before pushing a new patchset so you dont run into a situation where a different patchset stomps your changes04:56
coolsvapsorry for that, i will take care of that in future04:56
SamYapleno problem :) I wouldn't have found those pep errors anytime soon. the gate logs werent very clear to me04:56
SamYaplethanks for the assist04:57
SamYapleand always add co-authorer-by, it makes kolla looks better and it makes you look better04:57
coolsvapSamYaple, if you do tox -e pep8, the logs are pretty straight forward04:58
SamYaplewil ltry that, thanks coolsvap04:58
coolsvapit gives you stats like this http://paste.openstack.org/show/407968/04:59
SamYapleah ok. anyway to include that in the gate?05:00
openstackgerritSwapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add ansible support for Cinder  https://review.openstack.org/20886805:04
coolsvapSamYaple, i think it should give similar results in gate, i actually did not see the gate results, for pep8 stuff i generally do it locally05:05
sdakesamyaple https://review.openstack.org/20934305:08
coolsvapsdake, :)05:09
sdakecoolsvap05:26
sdakecan you make a separate blueprint for cinder please05:26
SamYaplehello coolsvap, i added a full review of cinder. The conf file looks great!05:26
sdakei'd like all new blueprints for new services individuually05:26
sdakecall it ansible-cinder05:26
coolsvapsdake, sure will do05:26
sdakeif you want to really do me a solid make blueprints for the other services as well ;-)05:27
coolsvapsdake, sure will do05:27
sdakethe unfinished ones05:27
sdakeand we can jsut close out ansible-service05:27
coolsvapyes based on the etherpad05:27
coolsvapi will update the etherpad with details05:28
sdakemake sure to link the etherpad in each blueprint tho if possible05:28
SamYaplesdake: ansible-service was closed in l205:28
sdakeya thats wht i meant05:28
sdakesorry late here not enough sleep05:28
coolsvapsdake, will do the required05:30
sdakesamyaple if i'm not on the ball with that review over the next couple of days please  take it over05:30
sdakethe infra review05:30
*** bmace has quit IRC05:30
sdakei am on the road05:31
coolsvapSamYaple, one question regarding the logging options in containers05:31
coolsvapcan we keep it global05:31
coolsvapas for keystone i can see the logging is configurable05:32
SamYaplecoolsvap: im not sure i understand05:35
coolsvapSamYaple, https://github.com/stackforge/kolla/blob/master/ansible/roles/keystone/templates/keystone.conf.j2#L205:35
SamYaplecoolsvap: ho yea were updating all the logging options across Kolla during L305:36
SamYapledon't worry too much about it right now05:36
coolsvaphttps://github.com/stackforge/kolla/blob/master/ansible/roles/keystone/templates/keystone.conf.j2#L205:36
SamYapleitll be configurable05:36
coolsvapSamYaple, can i submit a patchset for that?05:36
coolsvapi have done a bit of work for that05:36
coolsvapjust wanted to check how we want to make it05:37
SamYaplewell im not sure if we want to keep it in the template like this05:37
coolsvapper service or global05:37
SamYapleor move it to the global.conf augment file05:37
SamYaplesince thats what the files are for05:37
coolsvapalright i will keep that aside05:37
openstackgerritSam Yaple proposed stackforge/kolla: Reduce dependencies on the dest nodes  https://review.openstack.org/20845105:38
sdakeboy i get so much cisco spam at cisco05:39
sdakeits like poeple keep adding me to mailing lists05:39
SamYaplei would never sdake05:39
SamYaplenever05:39
sdakeno i mean internally05:40
SamYaplei make jokes. its my thing05:40
sdakewe have a pe mailing list05:40
SamYaplei understnad05:40
sdakeand everyone spams the shit out of it05:40
sdakeall the directors + want all the pes at cisco to know every little detail of what is happening05:40
sdakethere are 70k people at cisco05:41
sdakethat is alot of details :)05:41
sdakethere is a de list too with 85 people on it -i'm sure thats worse05:41
sdakeall pe+ get added to every new major initiative like devops, openstack etc05:42
SamYapledoes anyone wish to take on fixing up the build-script?05:42
SamYaplei do not have alot of time for that05:42
sdakemandre may when he gets back in05:43
sdakejust get push working05:43
sdakewe can go from there05:43
coolsvapsdake, i am creating another bluepring for ansible-heat05:43
SamYapleyea ive got a review for push05:43
coolsvapis it fine?05:43
sdakeyes plase do05:43
SamYapleit works, but it gives no feedback05:43
sdakenice ;-)05:43
SamYapleits only single threaded too,but thats because my docker daemon freaked out multithreaded05:43
sdakeyou ca't push multithreaded05:44
sdakedocker implodes05:44
*** bmace has joined #kolla05:44
*** SourabhP has joined #kolla05:48
*** banix has joined #kolla05:53
coolsvapsdake, created all blueprints05:58
coolsvapupdated in https://etherpad.openstack.org/p/ansibalising-containers05:59
coolsvapSamYaple, regarding cinder data i believe it will need in case of lvm06:03
SamYaplecoolsvap: why?06:05
coolsvapthe lvm driver will use the data container to store the volumes created06:05
SamYaplewhat do you mean? it resides on a disk...?06:07
SamYapleor ar eyou talking about the current implementation of cinder with the loopback files?06:07
coolsvapi am talking about the current impl06:07
coolsvapalso regarding the defaults.yml, isn't cinder a core service06:08
SamYapleoh then that makes since06:08
SamYaplecoolsvap: no it shouldnt be deployed by default since it requires physical disks in any real deployment06:08
*** weiyu_ has joined #kolla06:09
coolsvapalright then moving it to globals06:09
sdakecoolsvap nice thanks - you should do more reviews so we can get you on the core reviewer team ;)06:09
SamYapleindeed06:09
sdakecoolsvap will you be in tokyo?06:09
coolsvapsdake, not sure at this point06:09
coolsvapSamYaple, just one question regarding the cinder review, do you think multinode need updates similar to allinone?06:13
SamYapleyea keep the files the same06:14
*** banix has quit IRC06:16
coolsvapsame as in similar to all in one or no updates?06:18
sdakehttps://launchpad.net/kolla/+milestone/liberty-306:18
sdake31 blueprints to eat through06:19
sdakefortunately many of them are 4-8 hour jobs ;)06:19
sdakebedtime night folks06:20
sdakeif folks are interested pick up blueprints ;-)06:20
sdakesamyaple if you could get the blueprints you want to tackle int othe tracker that would be great06:20
openstackgerritSwapnil Kulkarni (coolsvap) proposed stackforge/kolla: Add ansible support for Cinder  https://review.openstack.org/20886806:22
*** SourabhP has quit IRC06:26
*** achanda has quit IRC06:42
openstackgerritSam Yaple proposed stackforge/kolla: Reduce dependencies on the dest nodes  https://review.openstack.org/20845106:47
*** unicell has joined #kolla06:54
*** erkules has quit IRC07:13
*** erkules has joined #kolla07:13
*** pbourke has quit IRC07:33
*** pbourke has joined #kolla07:33
*** jmccarthy has quit IRC07:34
*** jmccarthy has joined #kolla07:35
*** tfukushima has joined #kolla07:51
*** athomas has joined #kolla07:57
*** diga has quit IRC08:02
*** diga has joined #kolla08:03
*** inc0 has joined #kolla08:14
inc0good morning08:14
SamYapleinc0:08:15
*** weiyu_ has quit IRC08:21
*** akwasnie has joined #kolla08:21
*** athomas has quit IRC08:21
*** athomas has joined #kolla08:22
*** weiyu_ has joined #kolla08:22
*** barra204 has quit IRC08:25
*** macjack has joined #kolla08:35
SamYaplepbourke: the containerized deps work well08:51
pbourkeSamYaple: I'll take a look08:52
SamYaplei rekicked my server to a clean one and it worked08:52
openstackgerritMerged stackforge/kolla: Install git for ubuntu source base  https://review.openstack.org/20933108:58
pbourkedoes pypi take a while to make packages available?09:00
pbourkecant seem to fetch sdake's new openstack-crux package09:00
coolsvapi can search it though09:01
SamYapleidk i think that was larsks wasnt it?09:01
SamYapleor did sdake do it09:01
coolsvapCould not find a version that satisfies the requirement openstack-crux (from versions: )09:02
coolsvapNo matching distribution found for openstack-crux09:02
pbourkeyup09:03
pbourkeSamYaple: no it was sdake09:03
SamYapleok09:03
vbelpbourke: we use mirror, right? :)09:27
pbourkewe do but others dont09:27
pbourkeand we can't mirror whats not there!09:27
pbourkeIm sure it just takes some time09:28
vbelyes, it is not there09:28
openstackgerritPaul Bourke proposed stackforge/kolla: Fix missing config in horizon source image  https://review.openstack.org/20941509:32
*** gfidente has joined #kolla09:33
*** gfidente has joined #kolla09:33
openstackgerritPaul Bourke proposed stackforge/kolla: Install correct crux in centos-source-base  https://review.openstack.org/20497709:33
openstackgerritPaul Bourke proposed stackforge/kolla: Fix missing config in horizon source image  https://review.openstack.org/20941509:34
*** dolpher has quit IRC09:37
openstackgerritPaul Bourke proposed stackforge/kolla: Set curl as the default source install method  https://review.openstack.org/20942109:37
*** vbel has quit IRC09:37
openstackgerritPaul Bourke proposed stackforge/kolla: Add horizon source for oraclelinux  https://review.openstack.org/20943009:53
openstackgerritPaul Bourke proposed stackforge/kolla: Add horizon source for oraclelinux  https://review.openstack.org/20943009:53
*** dims has joined #kolla10:00
*** dims has quit IRC10:00
*** dims has joined #kolla10:00
*** weiyu_ has quit IRC10:02
*** athomas has quit IRC10:09
*** athomas has joined #kolla10:15
*** weiyu_ has joined #kolla10:18
*** coolsvap is now known as coolsvap|away10:33
*** jruano has joined #kolla10:50
pbourkeSamYaple: is there a reason you're not installing ansible using pip in the new kolla-ansible image10:53
SamYaplepbourke: yes im cloning devel10:54
pbourkeI dont have an issue with it per say, just cant help but think of the same situation with crux10:56
SamYapleidk about crux but i need upstream ansible stuffs10:57
pbourkedo you also require ansible dev on the deploy node?11:02
SamYapleno11:05
pbourkeFailed to template {{ (database.stdout.split('localhost | SUCCESS => ')[1]|from_json).changed }}: an unexpected type error occurred. Error was expected string or buffer11:05
SamYaplein fact this reduces the requirement down to ansible 1.8.4 I believe11:05
SamYaplepbourke: youre going to have to figure out how i can reproduce that because it works fine for me11:06
pbourkeok, just wondering if you'd seen it. will try a clean install11:07
SamYapleyea whats thats saying is its not valid json out11:08
SamYaplebut that line is parsing the output from the ansible command11:08
SamYaplethe one in the container11:08
openstackgerritSam Yaple proposed stackforge/kolla: Add check_script for haproxy in keepalived  https://review.openstack.org/20946311:13
SamYaplepbourke: what step did that fail on, keystone glance nova?11:17
pbourkekeystone11:17
*** rhallisey has joined #kolla11:17
SamYapleyea i dont know seems to work for me11:18
SamYaplei just did a fresh kick with no deps to test this patch11:18
SamYapleah im sorry pbourke11:24
SamYapleyoure right11:24
SamYaplemissing quote11:25
SamYapleworks with dns name, but ip needs to be quoted11:26
*** dims_ has joined #kolla11:26
*** dims has quit IRC11:28
openstackgerritSam Yaple proposed stackforge/kolla: Reduce dependencies on the dest nodes  https://review.openstack.org/20845111:30
SamYaplepbourke: give that a go11:31
SamYapleshould work with ip addresses as well11:31
openstackgerritweiyu proposed stackforge/kolla: Ceilometer config-ceilometer.sh file without backslash  https://review.openstack.org/20947311:40
openstackgerritSam Yaple proposed stackforge/kolla: Update the documentation for Ansible  https://review.openstack.org/20947411:43
SamYaplepbourke: can you confirm the latest patch fixes your issue?11:44
*** coolsvap|away is now known as coolsvap11:44
pbourkeSamYaple: in process11:44
SamYaplethanks11:44
pbourkeSamYaple: I tore down my env11:44
*** weiyu_ has quit IRC11:44
*** dims_ has quit IRC11:44
SamYapleare you reasonably happy with this solution (should it ever work for you)?11:45
pbourkeyes11:48
*** dims has joined #kolla11:51
SamYaplewell thats good because i can think of no other way to do this11:54
coolsvapSamYaple, i am trying to run the ansible allinone but it gets stuck at starting mariadb12:00
SamYaplecoolsvap: ill need a little more info than that12:00
coolsvapSamYaple, :) sorry i am trying to run the ansible all in one, i configured the ansible-thing according to instructions and running ansible-playbook command for all-in-one, it starts haproxy and keepalived and then it gets stuck at mariadb12:02
SamYaplecoolsvap: i believe the docs may be outdated in that regard. did you configure kolla_internal_address to an unused ip on your network to actas a VIP?12:03
SamYapleupdated docs patch is here: https://review.openstack.org/20947412:03
rhalliseycoolsvap, just set kolla internal and external ip to something random that's un used12:04
rhalliseykolla_external_address: "192.168.1.148"12:04
rhalliseykolla_internal_address: "192.168.1.148"12:04
rhalliseywhat I use12:04
coolsvapletme do that one sec12:05
SamYapleyou could also disable the haproxy role with "enable_haproxy: no" in hte globals.yml12:05
SamYaple(and remove the containers)12:05
*** dwalsh has joined #kolla12:08
*** weiyu_ has joined #kolla12:14
coolsvapSamYaple, do i need to install ansible from source?12:15
SamYapleno12:16
coolsvapI am getting KeyError: 'Memory'12:16
SamYaplecoolsvap: you will need a newer version than that12:17
SamYaplepip install ansible12:17
SamYapleshould pull in 1.9.212:17
SamYaplethats if your repo doesnt have a new enough version12:18
coolsvapi have ansible-1.9.2-112:18
SamYaplei dont know then rhallisey i think fixed it by tweaking source code12:19
coolsvaphmm i got that github issue12:19
coolsvapi will see and add that fix12:19
rhalliseyI used what he did12:19
rhalliseyit worked for me12:20
rhalliseywe can carry a patch in the short term if we need to12:20
SamYaplerhallisey: how so?12:20
rhalliseywe can just get a diff of the patch and apply it during setup12:21
SamYapleright but that can't be in the kolla repo12:21
rhalliseyjust for a temp fix?  how long until they release the new fix?12:22
SamYaplegplv3 issues12:22
rhalliseyohh12:22
rhalliseyyour right12:22
rhalliseyboo12:22
SamYaplea solution is the same thing I am doing for the other modules12:22
SamYaplemove docker-py into the kolla-ansible container12:22
SamYapleits a rather quick change due ot our start.yml abstraction12:23
SamYapleidk having docker-py in a container would be helpful, but its more complicated than i like12:31
SamYapleit would be the quickest way to ensure that we are able to fix upstream issues quickly12:31
rhalliseyI mean this is the only one we've encountered and it isn't that bad12:32
SamYapleits not the only one12:32
SamYapleits the only breaking one12:32
SamYaplewe have work arounds for the others12:32
rhalliseyoh I don't know of the other12:32
rhalliseyothers12:32
rhallisey:)12:33
SamYapleyea one option is broken (so we dont use it) and it doesnt block until the container exits so we have another task to wait on it12:33
SamYaplealso the pid issue12:33
rhalliseygotcha12:33
*** diogogmt has joined #kolla12:40
*** inc0_ has joined #kolla12:51
*** inc0 has quit IRC12:54
*** dims_ has joined #kolla12:57
*** diogogmt has quit IRC12:57
*** dims has quit IRC12:59
*** thumpba has joined #kolla13:01
*** inc0_ has quit IRC13:04
*** thumpba has quit IRC13:04
sdakemorning13:05
SamYaplemorning13:05
SamYaplesdake: check out the containerized deps13:08
SamYaplethey work pretty solid13:08
SamYapledid a fresh system, only required docker and docker-py :D13:08
sdakegetting ready to head to airport13:13
sdakegotta jet13:14
sdakettyl13:14
sdakemake sure to announce big tent today rhalliey13:14
sdakein case people missed it13:14
sdaketry to get people to take on blueprints13:14
sdakei may or may not be able to join the meeting - depends on where i am in the air13:14
rhalliseyall over it13:19
*** sdake has quit IRC13:21
pbourkesdake: is there any additional steps you need to make the new openstack-crux available? Can see it via the website but pip can't seem to find it13:21
*** weiyu_ has quit IRC13:27
*** coolsvap is now known as coolsvap|away13:30
*** dwalsh has quit IRC13:31
*** dims has joined #kolla13:36
*** dims_ has quit IRC13:38
*** ashuk has joined #kolla13:39
larskspbourke: I suspect he didn't actually upload a package.  I'd like to take over the PyPi distribution in any case, since I'm the author.  I'll talke with sdake about that later today.13:39
pbourkelarsks: appreciate it thanks13:39
*** inc0_ has joined #kolla13:40
*** ashuk has quit IRC13:40
*** dims_ has joined #kolla13:42
*** dims has quit IRC13:44
*** jtriley has joined #kolla13:45
*** dwalsh has joined #kolla13:47
*** dims has joined #kolla13:55
*** dims_ has quit IRC13:57
*** thumpba has joined #kolla14:07
SamYaplerhallisey: what time is meeting today?14:08
rhallisey2 hours14:08
rhallisey160014:08
SamYapleok no sleep for sam14:08
rhalliseyno worries were just going through bps14:09
rhalliseyyou already have a few14:09
SamYaplecool14:09
SamYaplewell ill probably join from the laptop in bed so ill be there in any case14:09
rhalliseySamYaple, I have a question when using until in ansbile14:21
rhalliseythis docs are not helpful at all14:21
SamYaplei have a docs patch14:22
SamYapleits in review14:22
rhalliseyno not kolla docs14:22
SamYapleoh ok14:22
rhalliseycloud.keystone_client.services.list()14:22
rhalliseywill fail in glance when keystone isn't ready14:22
SamYapleyea14:22
SamYapleright14:22
rhalliseyI want to retry14:22
SamYapleyup i have that in a patch already14:22
rhalliseyehh where14:22
SamYaplecheck the "containerized deps" patch14:23
SamYaplei believe its there...14:23
SamYapleits a simple "until successful" block14:23
SamYaplerhallisey: https://review.openstack.org/#/c/208451/9/ansible/roles/glance/tasks/register.yml14:25
SamYapleretired every 5 seconds up to 10 times14:25
SamYapleretries*14:25
*** mordred has quit IRC14:25
rhalliseykk14:26
rhalliseyya that's totally different than my approach14:26
SamYaplewhat is your suggestion?14:26
SamYapleI am happy to change it for better ideas14:26
SamYaplelets wieght the procs and cons14:26
SamYaplepros*14:26
rhalliseyI don't really see a con with either14:27
SamYaplethis is literally all the "first thing that came to my mind" stuff14:27
rhalliseyit's just differnt14:27
SamYapleok can you expand on that? we may need to use it14:27
rhalliseyI was trying to do the until on kolla_keystone_service:14:27
rhalliseyin roles/glance/tasks/register.yml14:27
*** dims_ has joined #kolla14:27
rhalliseyhttp://fpaste.org/251796/87848821/14:28
SamYapleoh thats still whats happening14:28
rhalliseyya exactly14:28
SamYapleyea exact same thing is happening14:28
rhalliseyyou're just doing it in a diff way14:28
SamYaplethe big difference is this is running with the containerized deps14:28
SamYaplethats really the only different14:28
SamYaplethat had to happen anyway14:28
rhalliseyya I figured14:29
rhalliseyI didn't notice it I'm glad I asked14:29
SamYaplecool. glad to see we think alike14:29
SamYaplehow do you like the playbooks so far? i see youre getting into them14:29
SamYapleany major complaints?14:29
*** dims has quit IRC14:29
rhalliseyhttps://bugs.launchpad.net/kolla/+bug/148149514:30
openstackLaunchpad bug 1481495 in kolla "Race condition between keystone and glance using ansible deployment" [High,Confirmed] - Assigned to Ryan Hallisey (rthall14)14:30
rhalliseyyou can reference that14:30
SamYapleok ill throw the tag in there i have to update another tag as well 'DocImpact'14:30
rhalliseySamYaple, ya I'm getting into it.  It's going ok14:30
*** dims has joined #kolla14:31
*** dims_ has quit IRC14:33
*** blahRus has joined #kolla14:33
SamYaplecool. well with this patch it should make things much better14:34
*** dims_ has joined #kolla14:34
SamYapleonly requiring docker and docker-py on the host!14:35
SamYaplethats an accomplishment in itself!14:35
openstackgerritSam Yaple proposed stackforge/kolla: Reduce dependencies on the dest nodes  https://review.openstack.org/20845114:36
rhalliseythat's great!14:36
SamYapleupdated the commit message for you14:37
*** dims has quit IRC14:37
*** dims has joined #kolla14:38
SamYaplepbourke: for the record you can run with -v for verbose and see the failed output14:39
SamYaplejust fyi14:39
SamYaple-vvvv will show you ssh debug info too14:39
pbourkeI know thanks14:39
*** inc0_ has quit IRC14:40
SamYapleim not sure how to achieve better output while still validating input14:40
*** dims_ has quit IRC14:40
SamYaplethis may be something we have to live with/document well to acheive the goal of containerized dependencies14:40
SamYapleill keep searching for an answer though14:41
SamYaplerhallisey: fyi on that pastebin you posed14:41
SamYapleposeted*14:41
SamYaplewhatever14:41
SamYapleyou can do "until: result|changed" for the same affect14:42
SamYapleit will use ansibles internal changed filter rather than greping the output14:42
pbourkeas far as I can see -vvv only gives the command ran, not the output14:42
SamYapleah it probably wont since we wrap it in until14:43
SamYapleif you didnt wrap it in until it would return the failed output14:43
SamYapleI bet I can adjust tha though14:44
SamYaplelet me try something14:44
pbourkeinstead of saying split('localhost | SUCCESS')14:44
pbourkecan we not do something like split('localhost | ')14:44
pbourketo account for both success and failure outcomes14:44
rhalliseySamYaple, ah ok14:45
rhalliseyI couldn't figure out what to use14:45
SamYaplepbourke: no we can't since we need to valited json14:45
SamYaplerhallisey: yea its alot of learn by example. i like to learn by studying other good ansible projects14:45
rhalliseythe docs had one example14:46
rhalliseyso I had no idea what else could be done14:46
SamYaplebasically if we split on that pbourke, well be left with something we cant validate14:46
*** banix has joined #kolla14:47
pbourkeI see14:47
pbourkehate to hold up the change, I just have a real problem with this error as is14:48
pbourkeits guaranteed confusion for the end user14:48
pbourkeand looks like a bug on our part14:48
*** jruano has quit IRC14:49
SamYaplecan you think of a way around it?14:50
pbourkeim trying :)14:50
SamYapleim trying to validate the variable _before_ pasing it through the filtter right now14:51
pbourkeneed vbel to work some more magic14:51
SamYapleyea im not a jinja2 guy in that big a way14:51
SamYaple  changed_when: "{{ database.stdout.find('localhost | SUCCESS => ') != -1 and (database.stdout.split('localhost | SUCCESS => ')[1]|from_json).changed }}"14:53
SamYaplebam14:53
SamYapleteamwork14:53
SamYaplethat spits out the error14:53
SamYapleand not a json validation issue14:53
SamYaplepbourke ^14:53
vincent_vdkSamYaple: what are you trying to do?14:54
SamYaplevincent_vdk: lots and lots. basically containerize all the dependencies for ansible14:55
pbourkevery nice14:55
SamYaplewhich works14:55
SamYaplebut getting the output back is tough14:55
pbourkeI think it should work nicely14:55
vincent_vdknot sure if it might help but there is a debug module14:55
pbourkelet me try14:55
SamYaplevincent_vdk: yea but thats for debug :P14:55
vincent_vdkyeah, not sure what you're trying to do14:56
SamYaplethe error for the enduser was awful vincent_vdk14:56
SamYaplenot helpful at all14:56
SamYaplethi sallows us to do our thigns and give the enduser a decent error14:56
vincent_vdki see14:56
*** banix has quit IRC14:57
vincent_vdkAnsible error handling is not optimal imho14:57
vincent_vdkas are the docs14:57
SamYapleespecially when you are wrapping ansible in ansible!14:57
SamYaplewhich we are doing here14:57
vincent_vdkthe mailinglist is full of hidden gems14:57
SamYapleill post a link to teh review in a second14:57
rhalliseyso I had no idea what else could be done14:58
SamYaple?14:59
rhalliseywas cleaning my keyboard lol15:00
*** banix has joined #kolla15:00
SamYapleah15:00
rhalliseysplit out the last thing I typed15:00
pbourkeSamYaple: works 100%15:00
SamYaplew00t15:00
pbourkeSamYaple: I get the output about not being able to talk to the database back the deploy console15:01
pbourkeSamYaple: nice job15:01
SamYaplethanks for the push15:01
SamYapledont ever do it again15:01
SamYaple:P15:01
pbourkeha15:01
openstackgerritSam Yaple proposed stackforge/kolla: Reduce dependencies on the dest nodes  https://review.openstack.org/20845115:01
*** banix has quit IRC15:06
pbourkeSamYaple: there is "-o oneline"15:10
SamYapleyea it seems to be b0rked in 2.x15:10
pbourkemaybe not in 2.015:10
pbourkeif it works without then all the better15:10
SamYaplewell its not working at the moment since i changed ansible versions it seems. i dont know15:11
SamYapleI think we need to pin on ansible <2.0 in the container15:12
SamYaplewe can backport moduels as needed15:12
*** diogogmt has joined #kolla15:13
pbourkewould it be hard to host a 2.0 rpm somewhere similar sdake's crux?15:16
SamYapleno because the output is different which breaks this15:17
SamYaplethe fact is without a licensing issue, we can backport moudles safely since its in a container15:17
pbourkeim really just referring to the fact we're pulling in a github repo15:17
SamYaplei dont want to rush into 2.0, it was just we needed modules15:17
pbourkeas he specifically said yesterday he wanted to avoid doing that15:17
SamYapleyea we arent anymore15:17
pbourkeoh15:17
openstackgerritSam Yaple proposed stackforge/kolla: Reduce dependencies on the dest nodes  https://review.openstack.org/20845115:17
SamYaplebam15:17
pbourkesweet15:18
pbourkeoh wow yeah the output is different15:18
SamYapleawell its kinda different15:19
SamYaplethe -o makes it the same15:19
SamYaplesometimes15:19
SamYaple2.0 is weird15:19
SamYaplesince we can backport modules, I am going to take that approach15:20
SamYaplemore stable15:20
SamYaplethere is nothing we require from 2.0 at this time15:20
SamYapleas long as the host playbooks remain compatible with 2.0 i dont care what version we run in the container15:20
SamYapledamn these typos15:21
pbourkecould've sworn I remember you saying you needed 2.015:21
SamYaple*for the modules15:21
SamYaple_but_ since we dont have a licensing issue i can fork and backport these modules all day long now15:21
SamYaplein that case we dont need 2.015:22
pbourkeyeah but you/we have to maintain the fork15:23
SamYaplehow so? if it works once it will always work ;)15:23
SamYaplebut in this case the modules i am refering too still are in even the devel branch so we would have to anyway we slice it15:23
pbourkeok personally I dont think the use case requires backporting15:25
*** dwalsh has quit IRC15:28
pbourkebut I dont feel strongly about it just throwing it out there15:31
SamYaplewhat do you mean? ansible 2.0 isnt even out15:31
SamYapleand these modules arent even in it15:31
pbourkeIt just seems easier to clone their devel branch and check out a ref that suits us rather than forking, backporting and rebasing etc.15:32
SamYaplethe devel branch doesnt have the modules yet...15:32
SamYaple15:24 < SamYaple> but in this case the modules i am refering too still *aren't* in even the devel branch so we would have to anyway we slice it15:32
SamYaplefixed ^15:32
pbourkeoh... so where are they located currently?15:33
pbourkepull requests?15:33
*** daneyon has joined #kolla15:33
SamYapleno they sit in mordreds git branch waiting on some shade stuff15:34
SamYapleno big deal, ive got al lthis covered itll all fall into place15:34
jpeelerharmw: you around? i'd really like to get this vagrantfile working for libvirt guests, but the networking works slightly differently15:34
SamYaplehey jpeeler !15:35
jpeelerhiiiiiii15:35
SamYaplehows married life15:35
jpeelerit's great!15:35
SamYaplegood good15:35
jpeeleri'm 31, so i've waited a long time heh15:35
*** shakamunyi has joined #kolla15:35
SamYaplewell its not like rushing into it is a good idea either15:35
jpeelervery true15:36
SamYaplegot any kids?15:36
jpeelernope15:36
*** daneyon has quit IRC15:36
SamYapledont rush into that either15:36
jpeelerdo you?15:36
SamYaple215:36
jpeelercool, didn't know15:37
SamYapleyup yup, they are 3 months and 13 months15:38
jpeelerSamYaple: i may finally get around to finishing the build.py changes. anything much change there aside from the push functionality?15:39
SamYaplenope15:39
SamYapleand push is still in review i think :P15:39
jpeelerit is15:39
SamYaplei havent had a chance to touch any of it15:39
SamYapleplease take over15:40
SamYapleim in over my head15:40
SamYapleso much stuff15:40
jpeeleri know the feeling15:40
jpeeleri'd be happy just to get it switched over in the gate15:40
SamYapleyea i need that to build our functional gating too15:40
*** inc0 has joined #kolla15:48
inc0afternoon guys15:49
*** sdake has joined #kolla15:50
SamYaplehey inc015:52
rhalliseyhey15:52
*** dwalsh has joined #kolla15:54
sdakeyo ddude15:57
sdakedudes15:57
*** daneyon has joined #kolla15:57
jasonsbSamYaple: i spoke with rabbit guy15:57
rhalliseyheyo15:57
jasonsbSamYaple: i can update you15:57
sdakeso rhallisey pleae run the meeting15:59
sdakebut i'll be in attendance15:59
rhalliseykk15:59
sdakei am in the air so i dont know how my internet wil faire15:59
rhalliseylol really15:59
SamYaplejasonsb: awesome!15:59
sdakeilll do the announcements tho15:59
sdakeif yu dont mind15:59
sdakeya gogo internet15:59
*** dims_ has joined #kolla16:00
rhalliseyya I'll start stuff you can just add announcements16:00
rhalliseynice16:00
*** echoingumesh has joined #kolla16:00
*** diogogmt has quit IRC16:00
rhalliseymandre, around?16:01
*** diogogmt has joined #kolla16:01
SamYaplemeeting in #openstack-meeting-4 everyone16:02
*** dims__ has joined #kolla16:02
*** dims has quit IRC16:02
*** dims_ has quit IRC16:04
*** dims has joined #kolla16:11
*** dims__ has quit IRC16:13
*** unicell has quit IRC16:19
*** vincent_vdk has quit IRC16:24
*** openstackgerrit_ has joined #kolla16:30
*** vincent_vdk has joined #kolla16:45
*** unicell has joined #kolla16:51
SamYaplefirst16:53
*** sdake has quit IRC16:58
*** inc0 has quit IRC16:59
*** dwalsh has quit IRC17:04
*** openstackgerrit_ has quit IRC17:05
*** jasonsb has quit IRC17:10
*** dims has quit IRC17:11
*** dims has joined #kolla17:12
*** dwalsh has joined #kolla17:20
*** openstackgerrit_ has joined #kolla17:24
*** bmace has quit IRC17:24
*** dwalsh has quit IRC17:25
*** achanda has joined #kolla17:25
*** dwalsh has joined #kolla17:27
*** openstackgerrit_ has quit IRC17:34
*** erkules has quit IRC17:35
*** bmace has joined #kolla17:37
*** erkules has joined #kolla17:39
*** inc0 has joined #kolla17:47
harmwjpeeler: I've got like 1 minute :)17:50
jpeelerharmw: is the bridge interface you used your wireless nic?17:52
harmwyep17:52
harmwif it doesn't exist, Vagrant will simply ask you which adapter to use instead17:52
jpeeleri think i'm getting closer, though it sounds like doing docker in docker will eventually be the way to go. will keep looking at it17:54
*** inc0 has quit IRC17:55
*** SourabhP has joined #kolla18:00
*** dwalsh has quit IRC18:01
*** dwalsh has joined #kolla18:09
*** diga has quit IRC18:09
*** diga has joined #kolla18:11
*** jasonsb has joined #kolla18:17
*** tfukushima has quit IRC18:19
*** athomas has quit IRC18:27
*** diga has quit IRC18:31
*** sdake has joined #kolla18:32
*** diga has joined #kolla18:33
*** dwalsh has quit IRC18:34
*** openstackgerrit has quit IRC18:46
*** openstackgerrit has joined #kolla18:46
*** dwalsh has joined #kolla18:47
*** akwasnie has quit IRC19:13
*** achanda has quit IRC19:16
*** thumpba has quit IRC19:58
*** tfukushima has joined #kolla20:30
*** tfukushima has quit IRC20:34
*** tfukushima has joined #kolla20:35
*** tfukushima has quit IRC20:40
-openstackstatus- NOTICE: Zuul has been restarted to resolve a reconfiguration failure: previously running jobs have been reenqueued but change events between 19:50-20:54 UTC have been lost and will need to be rechecked or their approvals reapplied to trigger testing.21:04
-openstackstatus- NOTICE: Correction: change events between 20:50-20:54 UTC (during the restart only) have been lost and will need to be rechecked or their approvals reapplied to trigger testing.21:09
*** jtriley has quit IRC21:54
*** bmace has quit IRC21:54
*** bmace has joined #kolla21:57
*** gfidente has quit IRC21:58
*** SourabhP has quit IRC22:22
*** echoingumesh has quit IRC22:26
*** echoingumesh has joined #kolla22:26
*** achanda has joined #kolla22:28
jpeelerok ansible experts: http://paste.fedoraproject.org/252018/88136511/ - i can't find a ansible_eth0 anywhere... but i have to go for today. maybe somebody knows the fix for me later22:28
SamYaplejpeeler: unless you have a node named 'network01' i think you need otupdate your inventory file to match your envrionemtn22:30
*** echoingumesh has quit IRC22:30
SamYapleadditionally are you sure eth0 exists? its not em0 or something?22:31
*** SourabhP has joined #kolla22:32
*** rhallisey has quit IRC22:33
*** dwalsh has quit IRC22:35
*** jtriley has joined #kolla22:39
*** diogogmt has quit IRC22:45
openstackgerritSam Yaple proposed stackforge/kolla: Update the documentation for Ansible  https://review.openstack.org/20947422:48
*** mordred has joined #kolla22:56
sdakejpeeler fedora doesn't use eth0 for its first interface22:58
*** achanda has quit IRC23:04
*** echoingumesh has joined #kolla23:19
*** echoingu_ has joined #kolla23:20
*** echoingumesh has quit IRC23:23
*** jasonsb has quit IRC23:25
*** weiyu_ has joined #kolla23:26
*** blahRus has quit IRC23:31
*** weiyu_ has quit IRC23:37
*** jtriley has quit IRC23:51
*** sdake has quit IRC23:56

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