Friday, 2015-06-12

*** jasonsb has quit IRC00:51
*** jasonsb has joined #kolla00:51
sdakemy line magnetic tube amp is in the house00:56
sdakeyay ;)00:56
*** jasonsb has quit IRC00:56
*** bmace has quit IRC00:58
*** zhiwei has joined #kolla00:59
*** juggler has quit IRC01:00
*** juggler has joined #kolla01:00
sdaketime for surgery in my office :(01:09
* sdake hates moving stuff around01:09
lothsdake: memcache not mentioned?01:12
*** erkules_ has joined #kolla01:27
*** erkules has quit IRC01:29
*** dims_ has quit IRC01:35
*** jtriley has joined #kolla01:57
*** jtriley has quit IRC02:02
*** dims has joined #kolla02:05
sdakevti rack assembled02:09
sdakewow what a pita02:09
sdaketalk about frustration-event packaging02:09
sdakea million pieces of tape and bubble wrap02:09
sdakeloth memcache not mentioned where, in the ha spec?02:10
sdakeloth I thought memcache was a performance thing not a ha thing02:10
*** prad_ has quit IRC02:16
loth./shrug its listed in the offical HA guide http://docs.openstack.org/high-availability-guide/content/_memcached.html02:17
*** dims has quit IRC02:39
*** vinkman has joined #kolla03:01
*** ccrouch has joined #kolla03:04
*** jasonsb has joined #kolla03:12
*** ccrouch has left #kolla03:16
*** jasonsb has quit IRC03:30
*** jasonsb has joined #kolla03:31
*** jasonsb has quit IRC03:39
*** dims has joined #kolla03:40
*** dims has quit IRC03:45
sdakeloth please add a comment to the spec pointing that out plz03:59
sdakewe want to follow that ha guide as much as possible03:59
sdakeloth if the folks that put that guide thought it was important enough to go through the pain of putting in the docs, its probably something that hsould be on our radar04:01
*** vinkman has quit IRC04:02
*** tobe has joined #kolla04:04
*** sami__ has quit IRC04:29
*** sami__ has joined #kolla04:29
*** tobe has quit IRC05:06
*** bmace has joined #kolla05:10
*** jasonsb has joined #kolla05:19
*** tobe has joined #kolla05:28
*** gfidente has joined #kolla05:53
*** nihilifer has joined #kolla06:03
*** tobe has quit IRC06:05
SamYaplesdake: memcache is required for ha for some components06:07
SamYapleand undocumneted for some as well06:07
SamYaplelike consoleauth06:08
*** sdake has quit IRC06:08
SamYaplealso while there is a push to make that guide better, it has historically been outdated, so deviating from it isnt the worst decision06:10
*** inc0 has joined #kolla06:26
inc0good morning everyone06:26
SamYaplemorning inc006:28
nihilifergood morning06:29
inc0SamYaple, I'm working about keepalived and one thing got me concerned, and I see that you didn't seem to take account on that in yaodu as well06:30
inc0namely, when keepalived uses IP, it binds it to an interface06:31
inc0but with default mask of /3206:31
inc0and that might be problem06:31
SamYapleas it should06:31
SamYaplehow so06:31
inc0because if you have management network with /2406:31
inc0it won't route to it via /32 interface06:32
SamYaplethats not how things work06:32
inc0that might not be the case if we'd have another IP06:32
inc0for gateway to management06:32
inc0but I don't think we do in kolla06:32
SamYapleyou _have_ to have another ip in the network. how else are the nodes communicating06:32
inc0also I think I'll need to create bridge on host system06:33
SamYaplewhys that06:33
inc0because if I have just one iface connecting to management06:33
inc0it can't have 2 ips at the same time06:34
inc0so normal and floating06:34
SamYapleit most certainly can06:34
inc0if you create eth0:1 thingy you mean06:34
inc0?06:34
SamYapleno06:34
SamYapleyou can assign as many ips as you want to an interface06:34
inc0hmm, let me check if keepalived does that06:35
SamYapleit does06:35
SamYapleits not out there overwritting ip addresses06:35
inc0ah yes, ifconfig vs ip a06:40
inc0rookie error06:40
SamYapleifconfig can add multiple ips too06:40
inc0it doesn't show vrrp ip normally06:40
inc0ip a does06:40
SamYapleit doesnt show multiple ips in a readable way, correc06:41
SamYaplebut net-tools/ifconfig have been deprecated for many many years06:41
*** tobe has joined #kolla06:42
inc0brb06:43
*** Haomeng has joined #kolla07:05
inc0SamYaple, I'll also start new doc in documnetation - ha guide for kolla07:09
inc0We'll need to explicitly say that you need to change every occurence of public ip in openstack.env07:09
*** athomas has joined #kolla07:11
*** Haomeng|2 has joined #kolla07:14
*** Haomeng has quit IRC07:16
*** shardy has joined #kolla07:37
SamYapleinc0: sure07:40
*** jmccarthy has quit IRC07:49
*** jmccarthy has joined #kolla07:53
*** dims has joined #kolla08:02
*** dims has quit IRC08:07
openstackgerritMichal Jastrzebski (inc0) proposed stackforge/kolla: Keepalived container  https://review.openstack.org/18798108:21
*** jmccarthy has quit IRC08:28
*** jmccarthy has joined #kolla08:32
*** erkules_ is now known as erkules08:37
*** erkules has joined #kolla08:37
*** tobe has quit IRC08:47
*** tobe has joined #kolla08:48
*** diogogmt has quit IRC08:56
*** diogogmt has joined #kolla08:57
*** jmccarthy has quit IRC09:01
*** jmccarthy has joined #kolla09:06
*** dasm|afk is now known as dasm09:08
*** pdb has quit IRC09:40
*** pdb has joined #kolla09:41
*** dims has joined #kolla09:49
*** dims has quit IRC09:50
*** zhiwei has quit IRC09:55
*** dims has joined #kolla09:56
*** inc0_ has joined #kolla09:57
*** inc0 has quit IRC09:58
*** inc0_ has quit IRC10:02
*** inc0 has joined #kolla10:02
*** pdb has quit IRC10:12
*** pdb has joined #kolla10:18
*** athomas has quit IRC10:23
*** athomas has joined #kolla10:29
*** alisonholloway has quit IRC10:35
*** vbel has quit IRC10:38
*** vbel has joined #kolla10:38
*** inc0 has quit IRC10:53
*** inc0 has joined #kolla10:55
inc0guys, do we have tripleK? Kolla-kontainer in kolla?;)10:55
SamYaplei dont think we want to be known as KKK10:57
inc0yeah...KOK then?10:58
*** inc0_ has joined #kolla11:05
*** shardy_ has joined #kolla11:06
*** shardy has quit IRC11:07
*** inc0 has quit IRC11:08
*** tobe has quit IRC11:10
SamYaplehow about we all acknowledge openstack on openstack is a bad idea nad just drop it11:10
*** shardy_ has quit IRC11:11
SamYapleim about to pop out a whole lotta ansible patches11:12
*** shardy has joined #kolla11:12
SamYapleas soon as that spec gets approved11:12
openstackgerritPaul Bourke proposed stackforge/kolla: Add base image for oraclelinux  https://review.openstack.org/19101311:15
pdbSamYaple: +111:15
inc0_SamYaple, thing about ansible, it will be pain to write enough logic to take care about different networks and so on11:16
*** sami_ has joined #kolla11:17
*** sami__ has quit IRC11:20
SamYaplepdb: what is that +1 too?11:30
SamYapleinc0_: this may not have been discussed, but i dont believe that is in the per-view of a deployment tool11:30
SamYaplethats all admin stuff11:30
SamYaplei have no intention of writing in network creation logic11:31
*** shardy_ has joined #kolla11:33
*** shardy has quit IRC11:35
pdbSamYaple: openstack on openstack been a bad idea11:38
*** shardy_ has quit IRC11:39
*** shardy has joined #kolla11:39
SamYapleyup11:40
SamYapleit _could_ work out with ironic and baremetal11:40
SamYaplepast that is a bad thing11:40
SamYaplepdb: inc0_ updated the multi-host spec. i think the comment covers all use cases11:41
inc0_actually I don't think tripleo + kolla would be bad11:42
inc0_right now it is, no argument there11:42
inc0_but with kolla, tripleo will just deploy base images with docker onboard and then heat will start deploying kolla containers11:43
inc0_but, we'll see about that I guess11:44
inc0_if kolla lands in tripleo, it lands in rdo, and then it lands in number of users, and that's a good thing11:45
pdbSamYaple: your latest comment looks good to me12:58
*** dims is now known as dimsum__13:02
*** jtriley has joined #kolla13:06
*** sdake has joined #kolla13:21
*** nihilifer has quit IRC13:22
pdbwould any cores please be able to have a look at https://review.openstack.org/#/c/190117/13:24
SamYaplepdb it looks good. i like it. i coubt youll get a fight13:29
SamYaplebut the blueprint isnt approved yet13:29
SamYapleand i cant approve it (i dont think)13:29
pdbis it only sdake that does that?13:29
SamYaplei know he can13:29
*** bradjones has quit IRC13:35
sdakesamyaple you should be able to13:38
*** bradjones has joined #kolla13:38
*** bradjones has joined #kolla13:38
sdakebut maybe i didn't get the permissions correct13:38
sdakepdb morning13:38
sdakepdb sec le tme get sam yaple setup properly13:38
pdbsdake: hey13:38
SamYaplejust getting refreshed on all the config changes for kilo13:39
SamYaplethere are a bunch13:39
SamYaplemost notably the far superious auth_plugin13:39
sdakesamyaple try now13:39
SamYapleyup that got it thanks sdake13:40
openstackgerritPaul Bourke proposed stackforge/kolla: Create keystone user in start.sh if it doesn't exist  https://review.openstack.org/19107113:43
*** mstachow has joined #kolla13:43
SamYaplesdake: added concise language to the multi-host spec to discuss the choices. it appears to cover everyones usecase and wont add duplicate code anywhere13:43
sdakesounds good13:43
sdakei'll publish an update soon13:44
sdakebeen waiting for everyone to get all their thoughts into the spec13:44
SamYapleindeed13:44
sdakepdb did samyaple get you set13:47
sdakepdb i can add you to the drivers group (this is different from the core team) so you can edit launchpad on your own13:47
pdbsdake: yes thanks a lot13:47
sdakepdb what is your launchpad id or email13:47
pdbsdake: https://launchpad.net/~pauldbourke13:47
openstackgerritPaul Bourke proposed stackforge/kolla: Symlink all non Dockerfile resources in images  https://review.openstack.org/19011713:48
sdakepdb cool try not to break it :)13:48
sdakeie stay away from the details buttons :)13:48
sdakeand the create release/milestone/etc buttons :)13:48
pdbsdake: sure thing :)13:50
sdakepdb do you see a delete button availble to you?13:51
sdakelooks like a trashcan13:51
sdakedont press it if you do13:51
pdbhmm im not sure I do13:52
sdakeclick overview13:52
pdbwhat is the drivers group in a sentence?13:52
sdakeon the upper right hand it says "Delete" with a trashcan next to it13:52
sdakeit is people that have admin control over the launchpad bug tracker and blueprint process13:52
sdakebut I need to make sure you don't have access to delete13:53
sdakeotherwise I will have to rescind my offer :(13:53
pdbhttps://i.imgur.com/3C1AJlz.png13:53
pdbthats what I see in overview ^13:54
sdakelooks good i guess only admins can delete13:57
sdakeyay :)13:57
sdakespread the word folks, if you want driver access, pls contact me in channel13:57
SamYaplealright guys. I have painstakingly gone through all of the kilo code and determined the _bare_ _minimum_ non-deprecated options needed to run multihost13:59
SamYaplefor all the main services at least, keystone glance cinder nova neutron14:00
*** walters has quit IRC14:00
SamYaplewe are ready to start laying down templates14:00
pdbSamYaple: can we help with this somehow?14:02
SamYapleeh the hard parts over. im just waiting for the spec to be approved to finalize the config gen and remove the docker-compose logic14:03
*** walters has joined #kolla14:03
pdbkk14:06
*** shardy_ has joined #kolla14:07
pdbstill wondering about that question on buildconfs from yesterday...14:08
pdbpdb question: in order to build alternate images I currently need to place an empty .buildconf in the root of the repo to prevent the default centos-base- being used. This seems fine, but they're in the gitignore so can't be part of the repo14:08
SamYaplepdb:AHH i still havent submited the prefix patch14:08
SamYaplethatll be incoming14:08
pdbwhat does it do?14:08
SamYaplecommandline --prefix patch14:08
SamYaple--prefix centos-base-14:08
pdbhmm I also have one of those sitting in my repo :)14:09
pdbI somehow thought you said that wasn't a good idea14:09
*** shardy has quit IRC14:09
SamYapleme? never!14:10
SamYaplei actually included that in the original patch but removed it for patch length sake14:10
pdbhttps://gist.github.com/brk3/c1df5d9f2e17144765cc14:11
SamYaplepdb: submit it if you would like, but you are missing the override pieces14:12
pdbin some ways I think I prefer committing default buildconfs though. as its automatic then14:12
pdbif a user tries to build say a fedora image and doesn't know about that flag, they'll be confused as to why they're getting centos based images14:12
sdakejpeeler around?14:13
*** shardy_ has quit IRC14:13
SamYaplefrankly the whole build method is convoluted. ive been thinking about a cleaner way of doing it14:13
jpeelerhey14:13
*** shardy has joined #kolla14:13
SamYaplei dont think it should require so much opaque code14:14
sdakejpeeler mind submitting a  voting  gate change14:14
sdakei'd do the work, but i don't know whaty ou reverted14:14
pdbyeah fair enough14:15
pdbquick different topic, the current keystone is not set up to run behind apache, right?14:15
SamYaplepdb: correct14:15
sdake./build isn't super opaque imo14:15
SamYaplesdake: thats for a single container14:15
SamYaplethat code is fine14:15
sdakeyou dislike build all containers?14:15
SamYapleits a bit confusing esspecially with all the structural changes we are doing14:16
sdakeagree mandre did a fantastic job of solving the problem but the shell is hard to understand14:16
sdakewe should probably rework both scripts into python in the future - be easier to comprehend14:16
SamYapleagree on all accounts as well14:16
SamYaplepersonally, ansible solves this great. very very clean14:17
SamYaplebut im not sure i want that as a build dep14:17
jpeelersdake: just for the image building test, right?14:17
sdakejpeeler we don't have naything else do we ?14:17
jpeeleri wasn't sure if the bashate stuff was ready yet, you got it close14:17
sdakewhen we get to functional testing we can add a second gate14:17
sdakeand just incude the content of the first14:18
sdakeand later merge them once the functional gating is solid14:18
jpeelerSamYaple: were we in agreement to remove docker-compose? or did i misread what you said earlier14:18
sdakeif you want to rename the gate functional-build that would rock ;)14:18
SamYaplejpeeler: for ansible things14:18
sdakedocker-compose is staying for the dev workflow experience and to support the tirpeo deployment model14:18
sdakewish slagle would chime in on that review and let us know which way he wants to roll with tripleo14:18
jpeelerwhich review14:19
SamYaplethe docker-compose python stuff doesnt work well when wrapped in other python code since they keep changing methods and things14:19
sdakethe ansible-multi14:19
mstachowo/14:19
sdakehey mstachow14:19
sdakesamyaple its just unnecessary overhead imo - assuming we can get container recreation to work properply with raw docker14:19
SamYaplesdake: well its got some cool abstraction things that would be _nice_ but it also means maintaining and ansible module that i dont want to do14:20
SamYapleand its less flexible for sure14:20
sdakemsstachow do you want to be ont he drivers team to edit blueprints and launchpad tracker in general - this is different from the core team14:22
sdakeif so need your email or launchpad id14:23
inc0_yeah, it would be nice to have someone from EMEA timezones14:23
inc0_mstachow, ^^14:24
mstachowsdake - sure14:24
mstachowmy launchpad id is mstachow :)14:25
SamYaplethats suprising mstachow14:25
inc0_coincidences happen14:26
*** inc0_ is now known as inc014:26
*** mstachow is now known as secret_mstachow14:27
secret_mstachowSamYaple now it's more suprising ;)14:28
*** inc0 is now known as not_mstachow14:28
not_mstachow;)14:28
*** not_mstachow is now known as inc014:29
*** SamYaple is now known as mstachow14:29
mstachowinteresting...14:29
*** pdb is now known as SamYaple14:29
mstachowyou should probably register this name14:29
SamYaplexD14:29
*** inc0 is now known as pdb14:29
pdbdoes that work?14:29
secret_mstachowii is registered :O14:29
SamYapleI thought Id registered mine14:29
mstachowme too....14:29
*** mstachow is now known as SamYaple_14:29
pdbnickserv did show error14:29
pdbbut that's it14:30
*** pdb is now known as inc014:30
*** SamYaple is now known as pdb_14:30
*** pdb_ is now known as pdb14:30
SamYaple_ah i have those messages blocked14:30
*** secret_mstachow is now known as mstachow14:30
inc0but I guess if it *just* show error and nothing else14:30
inc0that still kinda sux14:30
SamYaple_we should probably turn that stuff on14:30
*** SamYaple_ is now known as SamYaple14:30
pdbto add to the confusion my client just notified me of someone in another channel writing pdb.set_trace()14:31
SamYapleits your fault for naming yourself that14:31
SamYaplelets be honest14:31
inc0that's price of having nick in python stdlib14:31
SamYaple /nick openstack14:31
pdbim no good at coming up with aliases14:32
*** pdb is now known as pbourke14:32
SamYapleI got super creative with mine14:32
*** vinkman has joined #kolla14:33
pbourkeso, for source based images. do you see the ones we probably wont have source (e.g. rabbit, or base images), being symlinked under the source directory? or is there any point in that14:35
SamYaplepbourke: source-based was only for Openstack services in my plan14:35
pbourkeright14:36
SamYapleoh i see. yea i would think linking would be the correct path14:36
pbourkeis there any need though? I mean, the build-all script can just chop and change14:37
*** nihilifer has joined #kolla14:37
SamYapleit cant currently14:37
SamYaple?14:37
SamYaplesource vs binary?14:37
pbourkeah I see14:40
*** sdake_ has joined #kolla15:05
*** sdake has quit IRC15:09
*** sdake has joined #kolla15:10
*** sdake_ has quit IRC15:14
*** inc0 has quit IRC15:17
*** blahRus has joined #kolla15:20
*** pradk has joined #kolla15:30
*** jasonsb has quit IRC15:30
*** jmccarthy has quit IRC15:37
*** absubram has joined #kolla15:41
*** absubram has quit IRC15:42
openstackgerritPaul Bourke proposed stackforge/kolla: Add base image for oraclelinux  https://review.openstack.org/19101315:42
*** absubram has joined #kolla15:43
openstackgerritPaul Bourke proposed stackforge/kolla: Create keystone user in start.sh if it doesn't exist  https://review.openstack.org/19107115:43
jpeelersdake: would it be frowned upon to wrap all Dockerfile yum install lines with a retry loop?15:47
sdakegroan15:48
jpeelerif you look at the above review, two image failures happened due to yum15:48
sdakeseems like we should trust yum to do the job ;)15:48
sdakelet me look at the gate log15:49
* jpeeler starts to get scared about making it voting15:49
sdakejpeeler recheck ftw15:49
sdakeits a bit inconvinent15:49
sdakebut gating always is :)15:50
sdakepborke around?15:50
sdakepbourke around?15:50
pbourkesdake: yup15:50
jpeeleri wonder if it's networking weirdness and not really yum's fault15:50
sdakepbourke you said your going to install oraclelinux from source15:50
sdakeshouldn't it be oraclelinux-source-base?15:50
*** shardy_ has joined #kolla15:51
pbourkesdake: Well, it's like SamYaple was saying earlier, the source based installs are only for the openstack images15:51
pbourkesdake: I have a keystone source one ready that symlinks binary/oraclelinux-binary-base to source/oraclelinux-source-base15:52
sdakelet me check out that review15:52
*** shardy has quit IRC15:52
sdakedoes oel have packaging for openstack?15:53
pbourkeI haven't submitted the keystone one yet as my patch dependency chain is starting to get confusing :/15:53
sdakegit rebase -i ftw ;)15:53
sdakesee my twitter feed for a video on how to use git rebase15:53
pbourkeI consider myself pretty good with that kind of thing :)15:53
sdakeany objection to calling it oel-binary-base15:53
bmaceassuming you don't use linux sdake? :)15:53
sdakebmace ack :)15:53
pbourkebut sometimes it still throws me15:53
sdakethe problem you will run into with oraclelinux as a name is docker has a 31 character limit on image names15:54
sdakedumb I know15:54
pbourkeno objection here15:54
pbourkebmace: ?15:54
sdakewe are right at that limit on some of our image names15:54
sdakeor if there is some other short acronym you would prefer15:54
bmacehey pbourke.  was just talking about that vid / link sdake was mentioning.  had issues seeing it on chrome on linux.. it is some webex thing and they don't always play nice on the penguin platform15:55
SamYaplei vote badlinux15:55
pbourkeooooo15:55
sdakebadlinux lol15:55
SamYaple:D15:55
pbourkewanna step outside? ;)15:55
sdakehey sam likes ubuntu15:55
sdakeso that alone is enough reason to step outside :)15:55
pbourkei actually use ubuntu myself15:55
pbourkeas does bmace I believe15:56
sdakedoes oel have openstack packaging15:56
SamYapletruthfully i dont like ubuntu, i only use it since it has the widest userbase15:56
SamYaplei use arch/gentoo/LFS15:56
pbourkeit does but its old (icehouse)15:56
sdakeI dont think it makes sense to go in binary then15:56
sdakeunless you want to install rdo on it, which would make sense15:56
pbourkeim not sure what the exact plans are there yet, right now we'd prefer to go source based15:56
*** shardy_ has quit IRC15:57
sdakerdo i suspect would work on oel15:57
pbourkeour icehouse pkgs are essentially rdo15:57
*** shardy has joined #kolla15:57
bmaceyup, for a desktop i would prefer mint, but it is pretty behind on its library set and i don't want to sort of hand craft it using the latest ubuntu.. so i am sucking it up w/ unity and 15.04.  still a little bloaty but better than the first unity release15:57
sdakeok well your in for some serious work to implement source based installs :)15:57
pbourkesdake: haven't found any major issues so far?15:57
SamYapleits really really easy to do source installs actually15:57
sdakeif its source only then it should go in the source directory i think rather then the binary directory15:58
SamYaplei prefer them15:58
pbourkehmm15:58
sdakebinary is for packaged versions15:58
sdakeeg, binary packaging15:58
sdakeRPM or DEB specifically15:58
sdakeif it goes in binary it needs to either support rdo installs or oel needs to update from icehouse :)15:59
jpeelerSamYaple: you may find this interesting (maybe) https://www.nylas.com/blog/graduating-past-playbooks15:59
pbourkeit makes sense to me to have both binary and source directories. the non openstack services can go under binary, and source for the rest. that leaves the option to fill in the missing OS services under binary at a later date15:59
sdakeintereting hadn't thought of that problem16:00
SamYaplejpeeler: we use roles16:00
sdakepbourke as long as build-all-containers works i guess I dont care ;)16:00
sdakeor whatever that script is called16:00
SamYaplejust-do-it.sh16:00
jpeelercool, i hadn't checked16:01
SamYaplejpeeler: yay we do all those things16:01
sdakepbourke waiting on gate to +2 that change16:02
sdakebut it looks gtm16:02
pbourkesdake: great :)16:02
sdakeas we add more support for other distros and install methods, wht to do about our compose directory?16:02
sdakepbourke actually I am -1' because of  the name, needs to be shorter16:03
sdakei forgot about that objection16:04
sdakebut good thing I rmeembered - or it would have been seirous pain later ;)16:04
*** jtriley has quit IRC16:05
* vinkman reading back...16:05
vinkmanPlease do not call it OEL, but OL or something along that line...16:05
sdakewhatever you want to call it16:05
vinkmanOEL was the previous name and will make for major confusion...16:05
sdakeit justneeds to be shorter16:05
sdake6 chars or less16:06
pbourkevinkman: ol is fine16:06
sdakesince I know 6 chars works :)16:06
vinkmanheh… cool… thanks...16:07
*** sdake_ has joined #kolla16:12
pbourkedo we want to build these images as part of the gate?16:16
*** sdake has quit IRC16:16
*** jasonsb has joined #kolla16:16
sdake_they will be built automatically16:20
sdake_so yes16:20
sdake_i htink we need a separate gate for the various distro types16:21
sdake_because we will run into time limits if we do it in one go16:21
pbourkeagree on separate gate. but do we want build-all to build *everything*16:22
pbourke?16:22
pbourkemight be nicer just to have it build the default with an optional override16:23
sdake_that is controlled by PREFIX16:23
sdake_only the PREFIX distro will be built16:23
sdake_I dont know buid all everything makes sense16:23
sdake_atm prefix defaults to centos-rdo16:23
sdake_if you want to build ol-source, change PREFIX to ol-source16:23
sdake_make a nonvoting gate for it similar t owhat jeff has done for the current centos-rdo gate16:24
pbourkegotcha16:24
pbourkeseparate commit for the gate?16:25
pbourkeor same16:25
sdake_ya try to make each discrete unit of work separate16:34
sdake_the gate work goes to openstack-infra/project-config repo16:34
pbourkeoh didn't know that16:36
pbourkeok thanks for help all. will pick up the rename and other bits and pieces monday. have a good weekend16:36
*** athomas has quit IRC16:52
*** vinkman1 has joined #kolla16:57
*** vinkman has quit IRC16:57
*** shardy has quit IRC17:07
*** inc0 has joined #kolla17:21
sdake_got the amp on the rack17:28
sdake_boy that was a pita17:28
inc0hmf, keepalived seems to be workign17:31
inc0I have cluster of 4 nodes, and vrrp moves floating like it should17:32
SamYapleinc0: did you do that multicast or unicast?17:36
SamYaplein my experince with keepalived in the container, unicast works best. but i wont reject it otherwise if its working17:36
inc0I didn't specify this, so probably multicast17:37
inc0configuring unicast would be more painful, but I'd be happy if someone else would test it out17:40
inc0btw have you seen that latest ha guide don't even mention keepalived?17:42
SamYapleagain i wouldnt put to much worry on sticking to that guide17:42
SamYapleits a reference at best17:42
SamYaplethey have cleaned it up so there is no _wrong_ information, but it certainly doesnt cover everything17:43
inc0I know, but I also know what network partition is;)17:43
SamYapleyea frankly i feel the keepalived and haproxy components will turn into option ones as the future goes on. admin can setup external LB17:44
SamYaplesame with mariadb/rabbitmq/ceph as well. but for now it just needs to work17:44
SamYaplepartitioning should be prevented if at all possible but we cant account for every case17:44
SamYaplebtw, if you havent already gotten your commits to qualify as ATC, the HA docs are a great place to send some commits17:45
inc0I have, thanks;)17:45
SamYaplewell dont let that stop you from contributing17:46
SamYapleslaker17:46
inc0right now I need to write 2 specs, 1 bug and quite a bit feature to heat17:46
inc0so it's not lack of problems to solve that limits me17:46
SamYaplei have new baby and job that doesnt involve contributing17:47
SamYapletime limits me17:47
inc0you sir still amaze me by hours you're present here;)17:48
inc0by the way, where exactly are you from?17:50
mstachowGreat news! Galera is tested on fedora and centos. I've got working replication on every single method, including rsync and percona stuff ^^17:53
inc0nice one Michal:)17:54
mstachowPatchset should be uploaded in next monday. I must address last review =)17:54
*** gfidente has quit IRC17:58
inc0anyway 8pm friday, going for a weekend17:58
inc0cya all, have a good one17:58
mstachowhave a nice weekend Michal ;)17:58
*** inc0 has quit IRC17:59
sdake_note you wont get atc status by committing to kolla18:08
sdake_as it is a stackforge project18:08
mstachowatc status? what do you mean?18:14
nihiliferactive technical contributor18:16
nihiliferof openstack18:16
nihiliferin general, atc status == openstack summit for free18:17
nihilifer;)18:17
mstachowmeh that isn't my intention :)18:20
mstachowbut good to know :)18:20
mstachowthat information was mostly FYI18:20
mstachowmoreover I strongly belive that one day kolla will be out of stackforge into openstack ;)18:22
mstachowThis is only my enthusiasm :)18:22
mstachowok It's 8PM and this is Friday - after couple of exams and hard working here I need to start my weekend now18:23
mstachowhave a nice time guys at weekend.18:24
mstachowbye o/ :)18:24
nihiliferbye18:24
*** mstachow has quit IRC18:25
sdake_agree re openstack namespace18:25
sdake_almost almost dropped my tube amp getting it in the stereo rack18:26
* sdake_ was sweating bullets18:26
sdake_most stressful thing i've experienced in the last 9 months18:26
*** sdake_ is now known as sdake18:34
sdakejpeeler aorund18:39
jpeelerhey18:40
openstackgerritMichal Rostecki proposed stackforge/kolla: [WIP] Add designate-sink service  https://review.openstack.org/18939318:44
sdakeyay more improvements :)18:53
openstackgerritMichal Rostecki proposed stackforge/kolla: [WIP] Add designate-sink service  https://review.openstack.org/18939318:57
openstackgerritMichal Rostecki proposed stackforge/kolla: [WIP] Add designate-sink service  https://review.openstack.org/18939319:05
*** pradk has quit IRC19:10
openstackgerritDaneyon Hansen proposed stackforge/kolla: Spec to Add Support for High Availability  https://review.openstack.org/18198319:13
sdakemy computer is beeping at me19:16
sdakeand i dont know where its coming from19:16
*** sdake_ has joined #kolla19:17
sdake_samyaple any luck with the isngle nic script thing?19:20
*** sdake has quit IRC19:21
*** sdake_ is now known as sdake19:26
*** absubram has quit IRC19:37
harmwsink, again19:49
harmwlets see! :)19:49
nihiliferlet's see how many comments I'll get from you this time :)20:00
harmwyea well, sorry for beeing a PITA :P20:01
harmwyou're taking it well though, +1 for that :)20:01
harmwI'm almost done btw20:01
harmwthere20:04
harmwnihilifer: think you can fix all that somewhre tonight?20:21
nihiliferharmw: yes20:21
harmwcool20:21
nihiliferone more question, more by curiosity20:22
nihiliferdoes really20:22
nihilifertest -n $FOO20:22
nihiliferdiffer much from20:22
nihilifertest -n "$FOO"20:22
nihilifertechnically?20:22
harmwjust test it :)20:22
nihiliferbecause for me they doesn't ;)20:23
nihiliferbut maybe I don't know about something20:23
harmwill paste something, gimme a sec20:24
harmwnihilifer: http://paste.fedoraproject.org/231612/14341407/20:26
*** pradk has joined #kolla20:27
nihiliferaww, ok, my fail20:28
nihiliferthanks20:28
harmw:)20:29
sdakehey folks20:37
sdakehow far off is that sink container?20:38
*** daneyon has joined #kolla20:51
sdakedaneyon did you catch my email20:55
sdakeor still recovering form travel ;)20:55
openstackgerritMichal Rostecki proposed stackforge/kolla: [WIP] Add designate-sink service  https://review.openstack.org/18939320:56
daneyonsdake It's been 2 hours since i looked at email. I've been working with Dane to dig into Magnum networking20:57
sdakenice20:57
sdakein the office then?20:58
sdakeif you could take a look and yay or nay today i'd appreciate it20:58
sdakeso I can plan for the next charlie foxtrot to hit our plates :)20:58
daneyonyup... just wanted to disconnect so I can focus. had to jump on irc to chat with Kyle20:58
sdakenice - ya i am distracting I know :)20:59
sdakehey good news, got my amp20:59
daneyonSoiunds like the Neutron community does not want to touch magnum/container networking until Tokyo20:59
sdakehaven't powered it on yet, afraid something will blow up20:59
daneyonnot u, just generally20:59
sdakeack to that20:59
daneyoni do much better when i can unplug and try to focus20:59
sdaketha tis good feedback, can we getsomeone from the neutron community to respond on adrian's thread21:00
sdakeso he will know its off radar for liberty21:00
sdakeand we can focus on other things21:00
sdakeperhaps kyle ;)21:00
daneyonlet me ask21:00
sdakeit would be best on the ml to inform the whole community rather then back channeled on irc21:01
sdakesince i'm not really part of the neutron community that sort of communication coming from me might look like cisco speak rather than neutron speak :)21:02
sdakemy opinion on the matter is its an optimization that can wait for later ;)21:04
daneyonsdake I agree21:04
sdakei uspect that is how the neutron community perceives it as well21:04
daneyonDane or I will ask Kyle to respond to the ML thread21:04
sdakeappreciate it, I know Kyle but we dont have a long history21:05
sdakeboy its hot in arizona21:06
sdakei wish my parents and the airport lived in flagstaff21:07
sdakealong with good schools and whatnot :)21:07
daneyonsdake is this the ML thread you're referring to: http://osdir.com/ml/openstack-dev/2015-05/msg01433.html21:09
sdakeno it hit the ml today21:09
daneyonsdake i've been to AZ a few times when it's 110 degrees plus... that's some crazy stuff21:10
sdake[openstack-dev] [Magnum] Docker Native Networking21:10
sdakeonly 100f today21:11
sdakeHOT21:11
sdakerelief in november ;)21:12
harmwsdake: that Sink container is nearly done21:27
harmwjust +1'd it21:27
sdakenice21:28
sdakei'm old and had a rough week, so going to take a nap :)21:28
sdakebbl :)21:28
harmwhehe21:28
harmwyea me to, the napping part that is :p21:28
*** pradk has quit IRC21:28
sdakeplane tire had a blowout21:28
harmwouch21:28
sdakesoda can exploded in my face at 1am21:28
sdakepresentation in big loud room with lots of people21:29
sdakealmost dropped my tube amp21:29
harmwawful21:29
sdakeya not much positive happened this week outside of work :)21:29
sdakewe did get our mission statement ocmpleted tho21:29
sdakeyay for rhallisey :)21:29
sdakethat makes up for all the pain :)21:30
sdakehey they are 42 users in our channel :)21:30
sdakemagic #s ftw21:30
*** sdake_ has joined #kolla21:36
lothsdake what is the difference between magnum/heat/nova-docker driver?21:38
lothseems there is some redundancy there21:38
*** sdake has quit IRC21:40
*** rhallisey|PTO has quit IRC21:40
-openstackstatus- NOTICE: Gerrit will be offline for project renames between 22:00 and 22:30 UTC21:41
*** ChanServ changes topic to "Gerrit will be offline for project renames between 22:00 and 22:30 UTC"21:41
*** diogogmt_ has joined #kolla21:41
*** diogogmt has quit IRC21:43
*** diogogmt_ is now known as diogogmt21:43
*** sdake_ has quit IRC21:44
*** pbourke has quit IRC21:52
*** pbourke has joined #kolla21:53
*** pbourke has quit IRC21:54
*** pdb has joined #kolla21:54
*** pdb has quit IRC21:55
*** pdb has joined #kolla21:56
-openstackstatus- NOTICE: Gerrit is offline for project renames. ETA 20:3022:02
*** ChanServ changes topic to "Gerrit is offline for project renames. ETA 20:30"22:02
-openstackstatus- NOTICE: Gerrit is offline for project renames. ETA 22:4022:06
*** ChanServ changes topic to "Gerrit is offline for project renames. ETA 22:40"22:06
*** pdb has quit IRC22:14
lothSamYaple: Is the keystone module here https://github.com/SamYaple/yaodu/blob/master/ansible/library/keystone still needed if http://docs.ansible.com/keystone_user_module.html is avalible?22:27
*** jasonsb has quit IRC22:28
*** bmace has quit IRC22:36
*** ChanServ changes topic to "Weekly Kolla meetings will be held every Wednesday 1600 UTC for even weeks of the month and 2200 UTC for odd weeks in #openstack-meeting-4 | https://wiki.openstack.org/wiki/Meetings/Kolla"22:40
-openstackstatus- NOTICE: Gerrit is back online. Zuul reconfiguration for renamed projects is still in progress, ETA 23:30.22:40
*** pdb has joined #kolla22:42
*** daneyon_ has joined #kolla22:43
*** sdake has joined #kolla22:44
*** daneyon has quit IRC22:46
sdakesweet amp works ;)22:47
sdakedidn't blow up on the boat from china22:47
sdakeloth magnum and nova-docker are 100% orthogonal to heat22:48
*** bmace has joined #kolla22:48
sdakenova-docker requires nova to run single containers just like they were vms22:48
sdakemagnum allows you to run multiple container orchestration systems such as swarm, kubernetes, or mesos (in the future) in a way that is highly compatible with openstack22:49
daneyon_SamYaple ping23:01
*** openstackgerrit has quit IRC23:22
*** openstackgerrit has joined #kolla23:22
nihilifersdake: here?23:24
sdakeyup shoot23:25
nihiliferI see spec about ubuntu in launchpad23:25
sdakeyou mean blueprint ?23:25
nihiliferyes, s/spec/blueprint23:25
sdakewhich one and what is the subject of your question about it :)23:26
nihiliferhow do we exactly plan to support non-RH distros? we want to write everything from scratch for Ubuntu, or move the most of common stuff in scripts to some "base" directory?23:27
sdakedo you mean for the container os or the host os23:27
nihilifercontainer os23:27
sdakei think via binary packaging or via source23:27
sdake(eg install debs in a dockerfile, or install packages via source)23:28
nihiliferbinary23:28
sdakewith same start.sh for all23:28
nihiliferok, so we plan to keep the same start.sh scripts for all distros? :)23:28
sdakei would like to do so yes23:28
nihiliferthat was my main doubt23:28
sdakenot sure if its feasable or not23:29
sdakei sure hope it is23:29
sdakebecause making dokerfiles is easy, making start.sh is hard23:29
sdakethere will be symlinks to a base directory where all the start.sh will be stored23:29
nihilifersounds good23:30
sdakebase/keystone for example would have a start.sh23:30
sdakebase/mariadb woudl hae a different start.sh23:30
sdakeand the binary/mariadb would link to start.sh23:30
sdakeor somethign to that effect ;-)23:30
nihiliferso we'll have to make a change with moving and symlinking all scripts before working with any distro (maybe with oracle linux soon)23:33
sdakethere is a dependency change in the repo for that already23:33
sdakehttps://review.openstack.org/#/c/190117/23:33
sdakerather int he review queue23:33
nihilifer+1'd ;)23:37
sdakeso my wife posted pics of our amp on facebook23:37
sdakeand one of her friends asked "what are you powering with this"23:38
sdakeand she said "The economy I think."23:38

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