Tuesday, 2016-02-02

*** salmon_ has quit IRC00:20
*** angdraug has quit IRC01:39
*** dshulyak has joined #openstack-solar06:15
openstackgerritMerged openstack/solar: Removed deprecated ansible handler documentation  https://review.openstack.org/27467107:58
openstackgerritMerged openstack/solar: Fixed missing texts in docs  https://review.openstack.org/27461408:01
*** salmon_ has joined #openstack-solar08:10
openstackgerritLukasz Oles proposed openstack/solar: Use overlay images, 3 min test speedup  https://review.openstack.org/27506208:14
openstackgerritLukasz Oles proposed openstack/solar: Use overlay images, 3 min test speedup  https://review.openstack.org/27506208:21
pigmejok I will rewrite this ansible then salmon_ dshulyak  :)08:54
pigmejI wonder why this api is that weird08:54
pigmejdshulyak: https://github.com/openstack/solar/blob/master/solar/orchestration/workers/__init__.py#L15-L17 ?09:05
pigmejwhat are those imports doing there?09:05
dshulyakwo them you cant use modules like this, workers.tasks.Tasks09:06
pigmejso it's just for that alias ?09:06
pigmejI will add them as NOQA because I have no idea how to make it "ok" without NOQA09:07
dshulyakhm, what do you mean?09:08
pigmejthese imports are unsued in that file09:08
dshulyaktox didnt raise any errors09:09
pigmejbecause __init__ are in ignore09:09
pigmejI'm removing it09:09
pigmejI had it working yesterday, but then we merged your changes ;P09:09
openstackgerritJedrzej Nowak proposed openstack/solar: Removed all __init__ files from pep8 ignore  https://review.openstack.org/27463409:09
pigmejsalmon_, dshulyak rebased pep8 in __init__09:10
salmon_pigmej: what's h104?09:11
pigmej * [H104] Files with no code shouldn’t contain any license header nor comments, and must be left completely empty.09:11
pigmejdshulyak: can you please review https://github.com/Mirantis/solar-resources/pull/13 + https://github.com/Mirantis/solar-resources/pull/14 + https://review.openstack.org/#/c/273665/ ?09:55
salmon_please please09:55
pigmejsalmon_: now it's your time to try "rebase workflow"09:55
dshulyakyes, i will add last paraghraph to docs and do review09:56
salmon_everything is rebased09:56
pigmejsalmon_: for now yeah, but let's wait when we will have other changes; D09:56
salmon_ 2,500 commits in Solar!10:03
pigmejsalmon_: sadly it belongs to jenkins :(10:07
pigmejsalmon_: hmm, old tests execution 4m3s, new tests (after speedup) 3m25s10:09
salmon_you have faste disk :P10:09
dshulyakwhere we can host solar deb/rpm packages?10:14
pigmejfor rpm we could use fedorapeople10:15
salmon_for deb ppa?10:15
pigmejbut do you want to package deb/rpm now ?10:15
pigmejI would wait till we split properly extra stuff from main package10:15
pigmej(Computable Inputs, etc)10:15
pigmejsame for worker10:15
dshulyaki just checked fpm doc and it is quite easy to package python library + upstart/systemd10:16
dshulyakyou want to create separate repos for that stuff?10:17
*** openstackgerrit has quit IRC10:17
*** openstackgerrit has joined #openstack-solar10:17
pigmejdshulyak: I hope not, but having one big python pacakge with solar 'core' + solar worker + solar computable inputs it's kinda nonsense10:18
openstackgerritDmitry Shulyak proposed openstack/solar: Introduce documentation related to new worker  https://review.openstack.org/27464210:33
dshulyaki merged all patches related to ips10:42
pigmejon github too ? :)10:50
dshulyakpigmej: 2 on github and jenkins patch10:53
openstackgerritMerged openstack/solar: Allow to use any IPs with fuel-devops  https://review.openstack.org/27366510:59
*** tzn has joined #openstack-solar11:52
openstackgerritYuriy Taraday proposed openstack/solar: Fixes for 'tox -e docs' to run on my machine  https://review.openstack.org/27514411:59
*** pigmej has quit IRC12:36
*** pigmej has joined #openstack-solar12:38
openstackgerritYuriy Taraday proposed openstack/solar: Fixes for 'tox -e docs' in some dev environments  https://review.openstack.org/27514412:45
*** YorikSar has joined #openstack-solar12:57
pigmejYorikSar: hey :)12:57
YorikSarOh, this is the right channel.12:57
YorikSarI thought I created a new one again - there's no title here :)12:58
YorikSarSo I'm looking at Solar internals for "Python API" that I should consume to provide Config Service API.12:59
pigmejcool :)13:00
YorikSarIt looks like one can only create resources based on definitions in some Repository that is somewhere in fs, right?13:00
YorikSarSo I cannot provide an API call like "Create resource definition".13:01
pigmejOR you can set full path, but yeah basically it's like taht13:01
openstackgerritLukasz Oles proposed openstack/solar: Preinstall more stuff on an image  https://review.openstack.org/27518013:01
pigmejresource definition is resource repository13:01
pigmejYorikSar: but what is your question exactly ?13:02
openstackgerritMerged openstack/solar: Added info that our examples require nodes  https://review.openstack.org/27471313:02
salmon_pigmej: with this change hosts example is more than 2x faster from 8minutes to 3 minutes. With overlay patch, hosts example takes 1,5 minute :)13:02
pigmejsalmon_: but we should install less and less stuff on image:P13:03
YorikSarBut we can't make plugins install files into some repo, can we?13:03
pigmejYorikSar: sure you can13:03
salmon_pigmej: yup, but this change doesn't add anything. It just installs it earlier :P13:03
pigmejYorikSar: but also, you can manipulate on inputs later13:03
salmon_YorikSar: see solar repo commands13:04
pigmejso we could add generic "empty" resource for DR13:04
pigmejor for whatever13:04
salmon_YorikSar: what workflow do you need?13:04
pigmejand btw, you should ignore that your resource will have transports_id and location_id13:05
YorikSarOk, here's what I'm thinking about: user installed Fuel 9.0 with Config Service and Solar as two separate experimental features; now user shouldn't have to go through Solar Repo CLI to add some resource defs, Config Service API should allow user to add resource definitions w/o Solar CLI.13:06
YorikSarSo I'd like to have resource definitions stored in Solar DB itself somehow...13:07
pigmejok, so we have 2 choices :)13:08
pigmej#1 configdb adds some configdb resource repository which will have some "empty" reosurce definition13:09
pigmejthen you can easily add / remove / change inputs there13:09
YorikSarBut I'll need to change definitions, not resources themselves.13:09
pigmej#2 configdb provides resource definition in resoruce repository for any entity13:09
pigmejwhy do you want to change definitions ?13:10
YorikSarE.g. user wants to add new "template" - I need to somehow remember what inputs it had, with types and everything. And then use it for every hierarchy level in every env that uses this "template".13:12
YorikSar"template" in terms of Config Service (old one)13:12
pigmejso that template == resource definition, which is stored in resource repository, isn't it?13:12
pigmejit has some simple version parsing etc13:13
YorikSarRemember, we had "createComponent(ResDef[] definitions, Resource env)" on the board? So I want to be able to actually pass resource definition there.13:13
pigmejthen you should create resource repository content on demand probably13:13
pigmejI had an idea of storing repositories in DB at very first13:14
pigmejanyway, why ResDef[] can't be ResourceRepositoryNames[] ?13:14
YorikSarOh, right. Config Service can manage its own repo and allow user to CRUD definitions there.13:14
openstackgerritMerged openstack/solar: Fixes for 'tox -e docs' in some dev environments  https://review.openstack.org/27514413:14
YorikSarBecause I don't want to force client to be near server.13:15
pigmejthe only problem is that you will need to have this repo installed "everywhere" but it's general problem of current solar + solar repository design13:15
YorikSarYay! I've contributet to Solar!13:15
pigmejYorikSar: congrats :)13:15
pigmejyou're 15th contributor :D13:15
pigmejanyway, back to the topic ;)13:15
pigmejYorikSar: could you for now assume that you manage your own resource repository ?13:16
pigmejeven ignoring fact that it's some directory ?13:16
YorikSarYeah. And then I'll provide API to create resdef in it, for example.13:16
YorikSarI'll repeat some of Solar CLI then.13:17
pigmejsounds pretty reasonable to me, at least for now13:17
YorikSarHm... But for "library" part we'll have to pass just bare strings (specs) as ResDef[].13:18
pigmejYorikSar: also, I'm pretty sure that you will find some bugs in repository module :D13:21
pigmejfeel free to ask / fix / request for help then :)13:22
pigmejalso, if you prefer to not use repositories at all (you can find it then) you can use https://github.com/openstack/solar/blob/master/solar/dblayer/solar_models.py#L807 which is dblayer part of Resource13:23
pigmejit's not aware of *any* repository part13:23
YorikSarpigmej: Yeah, but I'd like to be able to let Config Service work in the same context as Solar later, so probably I'll have to use (fix ;)) Repository.13:25
pigmejrepos are for sure workirng, but there are some places that assume that everything is fine13:26
pigmejso in case of fuckups you may need rm -fr /some/path13:26
YorikSarOh, "_REPOS_LOCATION = '/var/lib/solar/repositories'"... It's hardcoded there.13:27
YorikSarSo I'll have to just always make sure I use a separate repo in it, right?13:27
pigmejlike configdb13:28
pigmejgenerally repository is a directory with proper structure13:28
pigmejyou can add stuff there without using repository module at all13:28
pigmejnot that I recommend it or so, but .. :)13:28
YorikSarBTW, why didn't that idea about storing repos in DB itself didn't fly?13:30
pigmejbecause it would mean that we will somehow store files inside database13:31
pigmejlike puppet manifests etc, it could be complicated to maintain etc13:32
YorikSarWell... Just add a FUSE on top of DB - and it becomes simple! :) (j/k)13:33
pigmejyeah I know13:34
pigmejI even wanted to do it :P13:37
pigmejYorikSar: just ignore this fact for now, maybe we will figure out something smarter in future :)13:37
pigmejdshulyak: https://review.openstack.org/#/c/274634/ review it please :)13:39
YorikSarMy main consern here is that we need to provide some API that clients should be able to use later.13:39
pigmejwell, for now soalr is a library13:40
pigmejwe will have some HTTP API13:40
pigmejbut yeah, that is my concern too13:40
*** tzn has quit IRC13:44
*** tzn has joined #openstack-solar13:47
openstackgerritLukasz Oles proposed openstack/solar: Set StrictHostKeyChecking=no during jenkins tests  https://review.openstack.org/27521813:53
openstackgerritLukasz Oles proposed openstack/solar: Set StrictHostKeyChecking=no during jenkins tests  https://review.openstack.org/27521813:55
openstackgerritLukasz Oles proposed openstack/solar: Preinstall more stuff on an image  https://review.openstack.org/27518013:56
openstackgerritLukasz Oles proposed openstack/solar: Use overlay images, it speedups tests  https://review.openstack.org/27506213:56
dshulyakpigmej: are you sure that this is not used anywhere https://review.openstack.org/#/c/274634/3/solar/cli/__init__.py ?13:57
pigmejdshulyak: in this file it's not13:57
pigmejyou mean that session ?13:58
pigmejit's not, it's certainly not, it's old, because some time ago just import was enough13:58
pigmejthen you need still to make_all()13:58
pigmejor something like that13:58
dshulyakpigmej: yes, session, previously it was called on import?13:58
pigmejlike very previously13:58
pigmejthen we moved some stuff, refactored but this line was still there13:59
pigmejI think salmon_ discovered it when he saw that update is not working on sqlite13:59
pigmejbecasue of session;)13:59
pigmejdshulyak: https://review.openstack.org/#/c/274642/ this is ready ?14:00
salmon_yeah, it was nice bug :D14:00
salmon_like just before release14:01
pigmejI'm not sure if 'nice' is correct word there :P14:01
dshulyakyes, docs patch is ready, maybe i should add ‘How to write extension’ but it is not easy job, so i would prefer to do it a later :)14:02
salmon_I will review it today14:02
pigmejdshulyak: lol :D14:03
pigmejbtw guys I think before 0.2 we should somehow cover more examples or explain inputs connections14:03
pigmejbecause I suppose it's magic for any newcomer14:03
openstackgerritMerged openstack/solar: Removed all __init__ files from pep8 ignore  https://review.openstack.org/27463414:03
salmon_for sure we need cli example14:03
pigmejsalmon_: you can do it! I believe in you :)14:04
salmon_after jenkins :P14:04
pigmejfor you everything is 'after jenkins' :P14:04
salmon_"It takes too long" ;)14:04
pigmejthen speed++ :P14:04
salmon_waiting for gerrit now14:05
pigmejsalmon_: http://imgs.xkcd.com/comics/compiling.png14:06
pigmejwe literaly did it when we were releasing 0.1 ;D14:06
*** tzn has quit IRC14:07
*** tzn has joined #openstack-solar14:07
salmon_dshulyak: can you workflow +1 here https://review.openstack.org/#/c/275180/2 ?14:27
openstackgerritMerged openstack/solar: Set StrictHostKeyChecking=no during jenkins tests  https://review.openstack.org/27521814:28
dshulyaksalmon_: added14:29
openstackgerritMerged openstack/solar: Preinstall more stuff on an image  https://review.openstack.org/27518014:32
YorikSarNow we have dshulyak's approval on typo in comments :)14:33
salmon_dshulyak: thx14:36
pigmejwhat typo ?:D14:38
YorikSar> # pre install solare requirements14:44
YorikSarSolare - sounds french :)14:45
pigmejit's not even first typo that we commited ;P14:45
pigmejnot that I'm proud of it...14:45
openstackgerritDmitry Shulyak proposed openstack/solar: [WIP] Use oslo.config  https://review.openstack.org/27524114:48
pigmejdshulyak: does oslo.config support ENV ?14:49
dshulyaki will check source a bit later, but i wasnt able to find it in documentation14:51
pigmejbut it should be possible to add logic for it14:51
dshulyaki tried to play around sys.argv :) in order to allow nested parsers and i dont like it completetly14:51
dshulyakit looks like total s..14:52
pigmejhehe ;D14:53
dshulyakmaybe openstack clients solved it somehow, i need to take a look14:53
pigmejbtw do we need command line config ?14:54
dshulyakit would be good to expose atleast basic options, like config-file14:55
pigmejthis one yup14:56
*** salmon_ has quit IRC14:56
*** salmon_ has joined #openstack-solar14:59
*** angdraug has joined #openstack-solar14:59
openstackgerritLukasz Oles proposed openstack/solar: Do not use archive mode. It copies owner and perms  https://review.openstack.org/27527315:36
openstackgerritLukasz Oles proposed openstack/solar: Do not use archive mode. It copies owner and perms  https://review.openstack.org/27527715:38
pigmejsalmon_: you did it again15:39
pigmejhttps://review.openstack.org/275277 + https://review.openstack.org/27527315:39
salmon_pigmej: use this one https://review.openstack.org/#/c/275277/15:40
pigmejyou're the master of abanddoned patches15:40
pigmejah, so you abbandoned the one with + from me?15:40
pigmejsalmon_: .... :P15:41
openstackgerritLukasz Oles proposed openstack/solar: Do not use archive mode, it copies owner and perms  https://review.openstack.org/27527716:24
pigmejdshulyak: hmmm16:43
pigmejTotal Delta: -2908862385.5616:43
pigmejTotal Time: 8.1607761383116:43
dshulyakpigmej: restart?16:43
pigmejloooooooooong time between process and run-once16:44
dshulyakit doesnt affect16:45
pigmejafter all tasks were finished I have had16:45
pigmejTotal Delta: 20.382378816616:45
pigmejTotal Time: 11.294923067116:45
pigmejbut in the middle of report I had this weird thingy16:45
dshulyakit looks corect16:45
dshulyakcan you create a bug :) ?16:45
dshulyaki will take a look when it will be clear whats with that hcf16:46
pigmejI can :)16:46
pigmejdshulyak: ok16:46
salmon_jenkins is dead?17:00
pigmejsalmon_: jenkins is lazy17:02
pigmejcheck zuul status17:03
pigmejsome solar tasks are pending for more than 2h17:03
openstackgerritJedrzej Nowak proposed openstack/solar: Ansible 2.0 support in ansible handlers  https://review.openstack.org/27532517:03
pigmejit works BUT after we merge this we need https://bugs.launchpad.net/solar/+bug/154101617:04
openstackLaunchpad bug 1541016 in Solar "Share more code between ansible handlers" [Undecided,New]17:04
pigmejhmm salmon_ should I remove ansible 1.9 and replace it with 2.0 there ?17:05
openstackgerritJedrzej Nowak proposed openstack/solar: Ansible 2.0 support in ansible handlers  https://review.openstack.org/27532517:06
pigmejI removed it17:06
pigmejI would show you, but gerrit returns 50217:11
pigmejso you need to believe me17:11
pigmejbut the url should be like: https://review.openstack.org/#/c/275325/2/requirements.txt17:11
pigmejand https://review.openstack.org/#/c/275325/2/bootstrap/playbooks/files/ubuntu-ansible.sh17:11
salmon_ah, it's ok now17:16
openstackgerritMerged openstack/solar: Do not use archive mode, it copies owner and perms  https://review.openstack.org/27527717:30
pigmejok you jenkins... :)17:30
pigmejwhat a surprise at the very end of a workday :D17:30
pigmejsalmon_: ^17:30
*** openstackgerrit has quit IRC17:32
*** openstackgerrit has joined #openstack-solar17:32
*** salmon_ has quit IRC17:43
*** salmon_ has joined #openstack-solar17:44
*** tzn has quit IRC18:06
*** angdraug has quit IRC18:31
*** tzn has joined #openstack-solar18:47
*** tzn has quit IRC18:48
*** angdraug has joined #openstack-solar18:50
*** YorikSar_ has joined #openstack-solar19:07
*** YorikSar has quit IRC19:12
openstackgerritLukasz Oles proposed openstack/solar: Set correct user for /vagrant/solar-resources dir  https://review.openstack.org/27539620:13
salmon_pigmej: dshulyak ^ ;)20:15
*** dshulyak has quit IRC20:37
openstackgerritMerged openstack/solar: Set correct user for /vagrant/solar-resources dir  https://review.openstack.org/27539620:45
*** salmon_ has quit IRC20:55
*** salmon_ has joined #openstack-solar20:55
openstackgerritLukasz Oles proposed openstack/solar: Disable host key checking for whole net  https://review.openstack.org/27541021:11
*** tzn has joined #openstack-solar21:16
openstackgerritMerged openstack/solar: Disable host key checking for whole net  https://review.openstack.org/27541021:31
*** tzn has quit IRC23:18
*** salmon_ has quit IRC23:42
*** tzn has joined #openstack-solar23:44
*** tzn has quit IRC23:45

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