| *** gokrokve has quit IRC | 00:21 | |
| *** gokrokve has joined #murano | 00:21 | |
| *** gokrokve has quit IRC | 00:25 | |
| *** ankurrr has quit IRC | 01:28 | |
| *** bharath has joined #murano | 02:33 | |
| *** sjmc7 has quit IRC | 02:40 | |
| bharath | serg_melikyan, ping | 02:48 |
|---|---|---|
| *** bharath has quit IRC | 03:06 | |
| *** drupalmonkey has joined #murano | 04:26 | |
| *** drupalmonkey has quit IRC | 05:17 | |
| *** chandankumar has joined #murano | 05:44 | |
| *** bharath has joined #murano | 06:15 | |
| bharath | How to pass IP address details of instances we created in the package to execution plan? | 06:17 |
| *** bharath has quit IRC | 06:21 | |
| *** bharath has joined #murano | 06:22 | |
| *** bharath has quit IRC | 06:31 | |
| *** bharath has joined #murano | 06:31 | |
| *** bharath has quit IRC | 06:34 | |
| *** bharath has joined #murano | 06:35 | |
| *** bharath has quit IRC | 06:35 | |
| *** bharath has joined #murano | 07:02 | |
| *** bharath has quit IRC | 07:07 | |
| *** bharath has joined #murano | 07:08 | |
| *** bharath has quit IRC | 07:14 | |
| *** bharath has joined #murano | 07:15 | |
| *** bharath has quit IRC | 07:16 | |
| *** bharath has joined #murano | 07:17 | |
| *** bharath has quit IRC | 07:18 | |
| *** bharath has joined #murano | 07:19 | |
| *** bharath has quit IRC | 07:19 | |
| *** bharath has joined #murano | 07:20 | |
| *** bharath has quit IRC | 07:24 | |
| *** sergmelikyan has quit IRC | 08:26 | |
| *** sergmelikyan has joined #murano | 08:26 | |
| *** tnurlygayanov has quit IRC | 09:14 | |
| *** tnurlygayanov has joined #murano | 09:15 | |
| *** openstackgerrit has quit IRC | 09:16 | |
| *** openstackgerrit has joined #murano | 09:17 | |
| *** sergmelikyan has quit IRC | 09:17 | |
| *** sergmelikyan has joined #murano | 09:22 | |
| *** stanlagun has joined #murano | 09:43 | |
| *** ativelkov has quit IRC | 09:44 | |
| *** ativelkov has joined #murano | 09:44 | |
| *** stan_lagun has quit IRC | 09:45 | |
| *** IgorYozhikov is now known as igoryozhikov | 10:21 | |
| *** igoryozhikov is now known as IgorYozhikov | 10:28 | |
| *** bharath has joined #murano | 11:00 | |
| *** asalkeld has quit IRC | 11:01 | |
| *** bharath has quit IRC | 11:04 | |
| *** bharath has joined #murano | 11:05 | |
| sergmelikyan | I think we need to consider updating our DB layer, since oslo.db is released | 11:12 |
| sergmelikyan | ruhe, what do you think? | 11:12 |
| ruhe | sergmelikyan: yeah, i have a local WIP branch with oslo.db | 11:35 |
| sergmelikyan | ruhe, cool! | 11:35 |
| ruhe | i just want to write more tests for db layer before we switch to oslo.db; also, i'm eager to refactor db layer, but i understand it's not the best time | 11:36 |
| ruhe | by refactor i mean just re-arranging of functions and files | 11:37 |
| smurashov | hi guys | 11:50 |
| smurashov | may be return to discussion of https://blueprints.launchpad.net/murano/+spec/engine-test-based-on-murano-pythonclient | 11:50 |
| smurashov | ? | 11:50 |
| smurashov | i want to start implementing it | 11:51 |
| smurashov | 1st variant is easier than 2nd, because we will not need change our ci settings | 11:53 |
| smurashov | but if in the future we will launch our engine tests against murano devstack installation, would be better implement a 2nd variant | 11:55 |
| ruhe | smurashov: in case if we start without tempest now | 11:56 |
| ruhe | smurashov: and would like to switch to tempest in future, would it cause a lot of changes? | 11:56 |
| smurashov | we will not have many changes in tests | 11:58 |
| ruhe | smurashov: in that case, i think that we can start without tempest | 11:58 |
| smurashov | most of changes would be in infrastructure of our ci | 11:59 |
| *** asalkeld has joined #murano | 11:59 | |
| smurashov | ok, i will implement a firsh variant | 11:59 |
| smurashov | and when the time comes | 12:00 |
| smurashov | i will upgrade it | 12:00 |
| ruhe | smurashov: ok. great | 12:01 |
| *** bharath has quit IRC | 12:29 | |
| *** lecalcot has joined #murano | 12:47 | |
| openstackgerrit | A change was merged to stackforge/murano: Move and rename functional tests https://review.openstack.org/109987 | 12:49 |
| *** lecalcot has quit IRC | 12:50 | |
| *** asalkeld has quit IRC | 13:03 | |
| openstackgerrit | A change was merged to stackforge/murano-dashboard: Move functional tests under muranodashboard/tests https://review.openstack.org/109954 | 13:37 |
| openstackgerrit | A change was merged to stackforge/murano-deployment: Update dasboard script due to test reorganization https://review.openstack.org/109990 | 13:44 |
| openstackgerrit | A change was merged to stackforge/murano-deployment: Update murano engine ci script https://review.openstack.org/110327 | 13:44 |
| ruhe | ok. now when we moved all the functional tests around we need to make sure that other patches can me merged and pass murano-ci | 13:46 |
| ruhe | for instance | 13:47 |
| ruhe | i can apply https://review.openstack.org/#/c/109151/ locally on master, but murano-ci says "This change was unable to be automatically merged with the current state of the repository" | 13:47 |
| ruhe | dteselkin: could you please take a look? ^^ | 13:47 |
| dteselkin | Yep | 13:48 |
| katyafervent | so we should restart murano-ci in all patches | 13:54 |
| katyafervent | BTW, why murano-ci on engine returns UNSTABLE instead of ERROR? | 13:54 |
| dteselkin | We've merged https://review.openstack.org/#/c/109987/ so there is no 'tests' subfolder. | 13:54 |
| dteselkin | Someone must reapply 109151 to the files in their new locations | 13:55 |
| ruhe | katyafervent: that's because of some jenkins plugin | 13:58 |
| ruhe | katyafervent: it requires a rewrite of murano-ci shell scripts to make it comment ERROR | 13:58 |
| *** drupalmonkey has joined #murano | 14:00 | |
| ruhe | dteselkin: but i can apply this change to murano master without any modifications | 14:02 |
| dteselkin | That's strange. I checked out latest master, create a patch based on commit you've provided, and when I tried to apply it it failed. | 14:06 |
| *** murali has joined #murano | 14:06 | |
| dteselkin | http://paste.openstack.org/show/eXKtiQYRyebzEAWqBtAI | 14:07 |
| dteselkin | http://paste.openstack.org/show/xdCIgJB4E80114zccCzB | 14:09 |
| ruhe | dteselkin: i'm just using checkout link from review | 14:14 |
| ruhe | git fetch https://review.openstack.org/stackforge/murano refs/changes/05/110305/1 && git checkout FETCH_HEAD | 14:15 |
| ruhe | git fetch https://review.openstack.org/stackforge/murano refs/changes/51/109151/3 && git checkout FETCH_HEAD | 14:15 |
| ruhe | the latter one is correct link | 14:15 |
| openstackgerrit | Dmitry Teselkin proposed a change to stackforge/murano-dashboard: Fix url_prefix in tests https://review.openstack.org/110654 | 14:15 |
| dteselkin | When you checkout the change, you are switching to the commit which parent is commit in master before the change. | 14:16 |
| ruhe | what about cherry-pick? it works for me too | 14:18 |
| ruhe | how does zuul apply the change? | 14:18 |
| dteselkin | Does zuul apply the change? AFAIK it only pushes the link to the git ref inside jenkins job. | 14:19 |
| dteselkin | I belive the best check is try to rebase the change on top of master | 14:20 |
| dteselkin | I've tried to cherry-pick the change. It failed. | 14:23 |
| ruhe | dteselkin: ok. thanks | 14:24 |
| dteselkin | Wait | 14:25 |
| dteselkin | Something went wrong, let me double-check :( | 14:25 |
| dteselkin | cherry-pick was successfull | 14:27 |
| dteselkin | However, the raw patch is unable to be applied | 14:32 |
| *** chandankumar_ has joined #murano | 14:51 | |
| *** chandankumar has quit IRC | 14:53 | |
| *** sjmc7 has joined #murano | 14:56 | |
| sjmc7 | morning all | 14:59 |
| sjmc7 | every time i set up one of our environments i get hit by the proxy issue described in https://bugs.launchpad.net/python-muranoclient/+bug/1326804 | 15:00 |
| sjmc7 | it looks like there is an HTTP client in the oslo incubator - are there any problems anyone knows about with moving to that? or it just hasn't been done yet? | 15:00 |
| *** bharath has joined #murano | 15:09 | |
| stanlagun | are there any objections for me taking https://bugs.launchpad.net/murano/+bug/1334352 ? | 15:10 |
| stanlagun | sjmc7: ^^ | 15:10 |
| sjmc7 | no, please do :) | 15:10 |
| sjmc7 | while you're working on it, i noticed yesterday that the engine is under very high CPU load during the repeated downloads as well | 15:11 |
| stanlagun | thats strange. Downloading is pure I/O, it should not affect CPU usage | 15:12 |
| sjmc7 | yeah, i was surprised too. it was taking a long time between each request, so something was going on | 15:12 |
| sjmc7 | i didn't get a chance to look into it much, and it may not matter once that is fixed | 15:12 |
| *** bharath has joined #murano | 15:12 | |
| bharath | serg_melikyan, Hi | 15:13 |
| stanlagun | I know | 15:13 |
| stanlagun | I guess it parses all YAMLs and then all YAQL expressions in it | 15:13 |
| sjmc7 | yeah. a lot of times for the core package | 15:13 |
| stanlagun | parsing is CPU-constraint thing | 15:13 |
| sjmc7 | out of interest, what was the reason for having the API communicate with the DB rather than the engine (as in heat)? | 15:13 |
| stanlagun | this is unavoidable but with proper caching this will happen only once per package | 15:14 |
| sergmelikyan | bharath, hi! | 15:14 |
| sergmelikyan | Had chance to take a look on examples that I had shared a little bit closer? | 15:15 |
| bharath | serg_melikyan, I was trying to develop a small package which create two instances and to store IP address of every instance in a file in each instance | 15:15 |
| bharath | sergmelikyan, yeah | 15:15 |
| sergmelikyan | bharath, you also can use this example: https://github.com/murano-project/murano-app-incubator/blob/master/io.murano.apps.java.HelloWorld/Classes/HelloWorld.yaml#L57 | 15:15 |
| bharath | sergmelikyan, https://github.com/bharaththiruveedula/io.murano.apps.linux.TwoInstances | 15:16 |
| sergmelikyan | this one is most suitable for your needs. This example shows how to get IP of instance and send it to VM | 15:16 |
| bharath | sergmelikyan, Okay I will have a look at it | 15:16 |
| stanlagun | sjmc7: because API can tell current state of the world so it must access DB. So the decision was do we want 2 components to access DB or just one API. If engine relies on API it will be easier to migrate it to Glance plugin | 15:16 |
| sjmc7 | tnurlygayanov - in your comment to https://bugs.launchpad.net/python-muranoclient/+bug/1326804 you mention rewriting with the requests library. was that instead of the client in oslo? or was the oslo client new? | 15:16 |
| sjmc7 | aaah, ok | 15:16 |
| sjmc7 | makes sense stanlagun | 15:17 |
| tnurlygayanov | sjmc7 it is about python client - we want to rewrite client on requests library | 15:17 |
| stanlagun | sending rpc request from API on each call es very inefficient. Especially when there is a lot of polling happening from dashboard. I believe that is bad design in Heat | 15:17 |
| sjmc7 | yeah, it's less efficient in that case stanlagun | 15:18 |
| sjmc7 | tnurlygayanov - yeah - but there's an apiclient module in oslo - https://github.com/openstack/oslo-incubator/blob/master/openstack/common/apiclient/client.py | 15:19 |
| sjmc7 | which uses requests | 15:19 |
| sjmc7 | that might also get us keystone v3 for free | 15:19 |
| *** gokrokve has joined #murano | 15:20 | |
| sjmc7 | seems crazy everyone's got their own HTTPClient | 15:24 |
| sjmc7 | .. but nobody is apparently using the oslo one | 15:25 |
| bharath | sergmelikyan, so here http://goo.gl/DU0mj1 I can change that line to ipaddress1 => $.instances[0].ipAddress | 15:28 |
| bharath | ? | 15:28 |
| ruhe | sjmc7: we need proxy support too. and i guess we'll need to use requests instead of code from oslo (just because everybody else uses requests) | 15:38 |
| sjmc7 | the oslo one uses requests | 15:38 |
| sjmc7 | https://blueprints.launchpad.net/oslo/+spec/common-client-library | 15:39 |
| ruhe | there is some db related code in oslo which i wanted to use. but then it turned out that nobody's using it and it is not being tested | 15:39 |
| sjmc7 | that's an example usage - it's a just a basic client to handle auth and make requests | 15:39 |
| sjmc7 | ok. i'll spend an hour seeing if it works as a drop-in replacement | 15:39 |
| ruhe | i'll ping assignee of https://blueprints.launchpad.net/oslo/+spec/common-client-library-2 to see why it stopped | 15:40 |
| sjmc7 | looks like some projects are starting to use it (at the bottom of that list) | 15:41 |
| ruhe | i found references to apiclient from python-glanceclient | 15:43 |
| sjmc7 | glanceclient is using its own | 15:44 |
| sjmc7 | lots of projects have the code imported, but they're not using it | 15:44 |
| ruhe | also, according to our folks who were related to this BP, this code is no longer maintained | 15:45 |
| ruhe | there is ongoing work on common client sdk | 15:45 |
| *** chandankumar_ has quit IRC | 15:45 | |
| sjmc7 | they just abandoned this and started again? wonderful | 15:46 |
| ruhe | well, it seems to be complicated. the original BP was blocked by some decisions in keystone. and now there is a new group of people working on that sdk | 15:47 |
| sjmc7 | ok. i won't use it then. they really need to remove stuff if they're abandoning it and nobody's using it | 15:48 |
| *** gokrokve_ has joined #murano | 15:51 | |
| *** gokrokve has quit IRC | 15:55 | |
| *** gokrokve_ has quit IRC | 15:55 | |
| bharath | sergmelikyan, can we access one instances like instances[0]? | 16:18 |
| bharath | *one instance | 16:18 |
| stanlagun | yes | 16:39 |
| *** gokrokve has joined #murano | 16:40 | |
| bharath | stanlagun, so can I use this statement https://github.com/bharaththiruveedula/io.murano.apps.linux.TwoInstances/blob/master/Classes/twoInstances.yaml#L37 ? | 16:55 |
| bharath | stanlagun, I dont have img right now which has murano-agent installed to test? | 16:59 |
| stanlagun | you can, but it will fail if there will be less then 2 instances or no ipAddress in any of them. Also all of this is useless if you don't have Murano Agent | 17:01 |
| bharath | stanlagun, I thought we can create n number of instances using $.instances.pselect($.deploy()) | 17:04 |
| bharath | stanlagun, ipAddress is member in instance right? | 17:04 |
| stanlagun | $.instances.pselect($.deploy()) just deploys all the instances in $.instances list. There can be any number of them (including 0) | 17:06 |
| stanlagun | the same goes for ipAddresses | 17:06 |
| bharath | stanlagun, oh okay | 17:10 |
| bharath | stanlagun, can we have repo from where we can download images? | 17:14 |
| bharath | stanlagun, I mean murano-agent installed images?. | 17:14 |
| *** murali has quit IRC | 17:17 | |
| openstackgerrit | Georgy Okrokvertskhov proposed a change to stackforge/murano: Add caller information to Begin execute log entry https://review.openstack.org/110317 | 17:18 |
| gokrokve | Hi | 17:21 |
| gokrokve | I propose to change log level for entries like "DEBUG murano.dsl.executor [-] ef24f00d295a496da5fa8107454f0bfb: Begin execution: io.murano.system.InstanceNotifier.trackCloudInstance" to Info level | 17:21 |
| gokrokve | Debug level is to verbose as it dumps the whole HTTP traffic | 17:23 |
| openstackgerrit | A change was merged to stackforge/murano-dashboard: Add keypair '+' button https://review.openstack.org/110172 | 17:26 |
| openstackgerrit | Georgy Okrokvertskhov proposed a change to stackforge/murano: Add caller information to Begin execute log entry https://review.openstack.org/110317 | 17:28 |
| stanlagun | gokrokve: those "Begin execute..." messages are for debugging only. Thats why they have DEBUG level. They are useless in production | 17:28 |
| stanlagun | as for HTTP traffic that can be configured. You can have custom levels per individual libraries | 17:29 |
| gokrokve | They are useful in production if someone wants to trace execution | 17:29 |
| stanlagun | That is the same if someone would like to debug in production. That can happen | 17:30 |
| gokrokve | Per library change is bad for operations | 17:30 |
| stanlagun | But most of the time cloud admin that looks at those logs has no idea what those methods are doing and what is the correct order | 17:30 |
| openstackgerrit | A change was merged to stackforge/murano: Use with_variant method for dialects db types https://review.openstack.org/110305 | 17:31 |
| gokrokve | Hm. Probably we need to think how to pass this infor to the user | 17:31 |
| stanlagun | Actually begin/end messages are what is called "trace". They need to be of TRACE level :) | 17:31 |
| gokrokve | True | 17:32 |
| gokrokve | But this is trace for MuranoPL so for engine it is just info what is going on | 17:32 |
| stanlagun | I believe we need a separate debugging back door :) Like DebugView utility in Windows | 17:32 |
| gokrokve | That always a problem when there are multiple layers of execution | 17:32 |
| gokrokve | By the way do we add task# in logs? | 17:33 |
| gokrokve | If we have this number we can easily grep fro particular trace | 17:33 |
| stanlagun | actually because we have so many garbage messages we often don't see important warnings or errors if they don't come with stack trace | 17:33 |
| stanlagun | what is task#? | 17:34 |
| gokrokve | Yes. This is exactly my motivation to bump MuranoPL trace to higher level | 17:34 |
| stanlagun | probably thats environment ID | 17:34 |
| gokrokve | Task number | 17:34 |
| gokrokve | Or env ID | 17:34 |
| gokrokve | We need something which will help to distinguish traces from different deployments | 17:35 |
| gokrokve | and envs | 17:35 |
| stanlagun | Begin/End messages that are printed by hundreds for each deployment are also fall to garbage category | 17:35 |
| gokrokve | as engine can simultaneously execute different deployments | 17:35 |
| stanlagun | yes | 17:35 |
| gokrokve | No. It is not hundreds. For single VM it is about 10-15 | 17:36 |
| stanlagun | we need MuranoPL debugger :) Thats what you want for real | 17:36 |
| gokrokve | So it is quite possible to trace it even by eye | 17:36 |
| gokrokve | Now I want to have at least poor's man debugger | 17:36 |
| stanlagun | real environments are tens of apps | 17:37 |
| gokrokve | logs with traces to understand what was going wrong | 17:37 |
| stanlagun | and on production server that handles all deployments from all users there will be thousands of them | 17:38 |
| * stanlagun need to go | 17:38 | |
| gokrokve | ok | 17:38 |
| gokrokve | lets talk about that later | 17:38 |
| gokrokve | I think it still can be valuable to end-users | 17:39 |
| *** chandankumar_ has joined #murano | 17:39 | |
| *** chandankumar_ has quit IRC | 18:33 | |
| sjmc7 | tsufiev - looks like there were some changes in horizon/openstack_dashboard's css that are causing us CSS problems | 18:43 |
| sjmc7 | https://bugs.launchpad.net/murano/+bug/1349894 being one of them, but much of the dashboard looks messed up | 18:44 |
| serg_melikyan | My proposals for summit are published :) | 19:37 |
| serg_melikyan | [23:36:07] Serg Melikyan: https://www.openstack.org/vote-paris/Presentation/application-catalog-three-facets-of-managing-openstack-workloads-2 | 19:37 |
| serg_melikyan | https://www.openstack.org/vote-paris/Presentation/bring-your-own-application-to-the-cloud | 19:38 |
| *** bharath has quit IRC | 19:59 | |
| openstackgerrit | Georgy Okrokvertskhov proposed a change to stackforge/python-muranoclient: Provide method for calling action in an environment https://review.openstack.org/99502 | 20:35 |
| *** btully has joined #murano | 20:39 | |
| openstackgerrit | Steve McLellan proposed a change to stackforge/python-muranoclient: Convert muranoclient to 'requests' https://review.openstack.org/110781 | 20:42 |
| serg_melikyan | https://www.openstack.org/vote-paris/Presentation/astrologer - spammers are everywhere %) | 20:43 |
| serg_melikyan | https://www.openstack.org/vote-paris/Presentation/app-provisioning-with-head-and-puppet - head is definitely needed when you do provisioning %) | 20:47 |
| *** asalkeld has joined #murano | 20:55 | |
| *** peoplemerge has joined #murano | 21:34 | |
| *** btully has quit IRC | 22:13 | |
| *** drupalmonkey has quit IRC | 22:24 | |
| *** sjmc7 has quit IRC | 23:05 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!