19:00:13 <catherineD|2> #startmeeting refstack 19:00:14 <openstack> Meeting started Tue Jul 5 19:00:13 2016 UTC and is due to finish in 60 minutes. The chair is catherineD|2. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:15 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:00:17 <openstack> The meeting name has been set to 'refstack' 19:00:36 <sslypushenko> o/ 19:00:56 <catherineD|2> sslypushenko: hello 19:00:58 <pvaneck> o/ 19:01:11 <catherineD|2> #link meeting agenda and notes, please feel free to add items https://etherpad.openstack.org/p/refstack-meeting-16-07-05 19:01:59 <catherineD|2> let's give the other a couple more mins 19:02:09 <andrey-mp> o/ 19:02:15 <sslypushenko> before we start - here the link on RefStaclClient application for AppCatalog https://review.openstack.org/#/c/336656/ 19:02:40 <sslypushenko> Client can be runned 19:02:50 <sslypushenko> in cloud in 2 clicks) 19:03:24 <andrey-mp> in which cloud? 19:03:41 <andrey-mp> inside/outside? 19:03:50 <sslypushenko> any OpenStack cloud with Murano support 19:04:32 <catherineD|2> sslypushenko: great I will add the link to our agenda so the refstack team can follow 19:04:34 <pvaneck> cool. will try it out 19:05:01 <sslypushenko> hmmm... do you have Openstack with Murano? O_o 19:05:53 <andrey-mp> i'm sorry - i don't know much about Murano... it will run vm with refstack client to test this cloud? 19:05:58 <pvaneck> not currently, but can add murano support 19:06:05 <sslypushenko> anyway, we can move on as soon as we are ready... 19:06:32 <sslypushenko> just a small piece of information to share 19:07:00 <catherineD|2> sslypushenko: will take a look thx 19:07:26 <catherineD|2> I am not very familiar with Murano either ... 19:10:01 <catherineD|2> Let's discuss it next week when we each learn about what it does .. we will ask sslypushenko: for more info if needed next week 19:10:26 <pvaneck> yea, i believe that is the gist of it, a vm will be spun up with refstack-client automatically installed 19:10:29 <sslypushenko> catherineD|2: Lucky you)) 19:10:53 <sslypushenko> pvaneck: yeap, it is just vm with refstack-client 19:11:22 <catherineD|2> today's agenda https://etherpad.openstack.org/p/refstack-meeting-16-07-05 19:11:31 <catherineD|2> ready to move on ? 19:12:18 <pvaneck> yep 19:12:28 <catherineD|2> alright 19:12:34 <catherineD|2> #topic Test results ownership 19:13:06 <andrey-mp> catherineD|2 19:13:10 <andrey-mp> catherineD|2 19:13:14 <andrey-mp> sorry 19:14:04 <andrey-mp> catherineD|2: Alex told me that his question still doesn't have answers 19:14:15 <catherineD|2> so we had a meeting last week but we ran out of time ... I will schedule additional meeting this week .. I plan on Thurs 18:00 UTC .. does that work for everyone 19:14:27 <catherineD|2> andrey-mp: that is why we need additional meeting schedule 19:15:10 <sslypushenko> will work for me 19:15:15 <andrey-mp> catherineD|2: for us is better 17:00 UTC... 19:15:53 <andrey-mp> or 20:00 UTC 19:16:50 <catherineD|2> sslypushenko: pvaneck: andrey-mp: I need your help to answers Alex's questions too ... so far the conversation is between Alex and I .., I would appreciate your inputs .. 19:17:10 <catherineD|2> andrey-mp: noted will try 17:00 UTC or 20:00 UTC 19:17:41 <sslypushenko> catherineD|2: 20:00 is too late... 17:00 much better 19:18:12 <catherineD|2> sslypushenko: ok 17:00 UTC on Thurs 19:18:20 <sslypushenko> yeap +! 19:19:25 <catherineD|2> meanwhile I really would like additonal inputs on the conversion between Alex and I ... I am sorry for the lengthy conversation but your inputs are important to unlock the deadlock 19:20:20 <andrey-mp> as I understood Chris promised to get some answers 19:20:35 <catherineD|2> andrey-mp: I am not sure what I can answer to Alex ... my anser will be the same like what I had answer and to him it is not good enough ... no point to go on .. at this point we need others' input 19:21:16 <catherineD|2> andrey-mp: that is why I will schedule addtional meeting 19:21:37 <andrey-mp> Does Chris can attend it? 19:22:54 <catherineD|2> I don't know the point is the have Chris and Mark to attend if Thursday is not good for them we will have to schedulte a different time ... their present to the meeting is necessary 19:23:02 <catherineD|2> ok moving on ? 19:23:20 <catherineD|2> #topic Pending reviews 19:23:39 <catherineD|2> #topic Disable the Catalog tabs on the UI. ( https://review.openstack.org/#/c/336739/ ) 19:23:47 <catherineD|2> #link Disable the Catalog tabs on the UI. ( https://review.openstack.org/#/c/336739/ ) 19:24:31 <catherineD|2> andrey-mp: did you see pvaneck: 's latest comments ... ? 19:24:40 <andrey-mp> yeah 19:24:52 <catherineD|2> it was my mistake 19:25:42 <andrey-mp> we want to introduce these changes when it will ready? 19:26:22 <catherineD|2> yea 19:26:58 <catherineD|2> as soon as we finish what we need to do in feature/vendor branch ... we should merge it before the next summit 19:27:25 <andrey-mp> ok. do you have a plan what we must implement? 19:28:20 <pvaneck> product + test association is the last big piece remaining I believe 19:29:03 <andrey-mp> what about guidelines? what about switch for UI - original version/extended version ? 19:29:13 <catherineD|2> pvaneck: yea .. and small thing here and there to improve the website (like enable editing , cancel ... _ 19:29:39 <catherineD|2> andrey-mp: after test association we will go to guideline ... 19:29:49 <andrey-mp> last change from Chris is about registering guidelines (by Foundation and by users)... 19:29:51 <catherineD|2> with guideline we will need validation 19:30:06 <andrey-mp> prototype already implemets guidelines 19:30:20 <catherineD|2> andrey-mp: CHris already made the change https://review.openstack.org/#/c/336704/ 19:30:24 <andrey-mp> but we still can't agree about test association 19:30:31 <catherineD|2> from Foundation point of vie 19:31:08 <catherineD|2> that is why we schedule meeting on Thursday .. 19:31:33 <andrey-mp> catherineD|2: yeah, I saw it. I would ask - is it suitable way in prototype for adding new guidelines? 19:31:43 <catherineD|2> vendor/product registration is useless without test association 19:31:55 <catherineD|2> that would come after test association 19:32:27 <andrey-mp> catherineD|2: but we can do these task in parallel... 19:32:35 <catherineD|2> Chris would like to link data from the Marketplace to RefStack .. that is our goal for Newton cycle 19:33:01 <catherineD|2> we can if we have time 19:33:01 <andrey-mp> catherineD|2: what data? 19:33:25 <catherineD|2> the certfication on marketplace to refstack result data 19:34:38 <andrey-mp> ok. next my question for this theme will be later when we will talk about my reviews ) 19:35:39 <catherineD|2> sure let's discuss about your reviews 19:35:55 <catherineD|2> #topic Using refstacl-client to update product table product_ref_id 19:36:23 <catherineD|2> basically I do not have technical issues with both review ... 19:37:26 <catherineD|2> andrey-mp: I want to make sure that we all understand how the CPID is created 19:37:43 <andrey-mp> I have only one - refstack-client will support feature that is not present in RefStack API... but we don't have release model for refstack-client 19:38:15 <andrey-mp> as I know - CPID now is a ID of public endpoint of keystone 19:38:16 <catherineD|2> andrey-mp: agree 19:38:51 <catherineD|2> andrey-mp: yea most of the time CPID is the hash of the public endpoint of keyston ... 19:39:28 <andrey-mp> For me - it's ok for cloud identification 19:39:33 <catherineD|2> There are a lot of code in refstack-client trying to get the service id either of v2 or v3 keystone API 19:40:17 <andrey-mp> (and it didn't work on my cloud :) ) 19:40:38 <andrey-mp> because tempest.conf changed 19:40:42 <catherineD|2> but most of the time these algorithms failed because of the way we code and the user tempest config file 19:40:51 <catherineD|2> andrey-mp: exactly 19:41:53 <catherineD|2> pvaneck: sslypushenko: so I wander whether this is the time we should review this CPID code so we do not give user the false idea that the cpids can be service id 19:42:57 <sslypushenko> yeap, it is good time for review CPID getting code 19:43:07 <andrey-mp> catherineD|2: you mean that CPID should not be public keystone endpoint ID? 19:43:32 <catherineD|2> andrey-mp: that is what it is now most of the time .. 19:43:43 <sslypushenko> andrey-mp: we need something else for private clouds 19:44:01 <andrey-mp> sslypushenko: why and what? 19:44:55 <catherineD|2> so for public cloud this endpoint_id will be unique ... but for private cloud they are not necessary to be unique 19:45:01 <sslypushenko> because 127.0.0.1:5000/v2.0 is not unique url for many clouds) 19:45:04 <andrey-mp> btw - we can implement small tempest plugin inside of refstack client. it can be based on tempest infrastructure and will write somewhere detected information about the cloud. 19:45:47 <andrey-mp> sslypushenko - it is not the URL as I know - it is an ID of endpoint. or am I wrong? 19:46:09 <hogepodge> you can't rely on results always being generated in refstack-client 19:46:27 <andrey-mp> catherineD|2: with another CPID it will be another cloud. user can register/delete many private cloud as he want. 19:46:33 <hogepodge> sometimes users take output from their ci runs and then use the client to upload the data, so tempest isn't being run from the client 19:46:37 <sslypushenko> andrey-mp: for public clouds it is hash of url 19:46:56 <catherineD|2> andrey-mp: it is the URL in the case that refstack-client can not communicate to the cloud due to the way we code 19:47:30 <andrey-mp> hm. then my review for refstack-client is incorrect 19:47:59 <catherineD|2> andrey-mp: as hogepodge: indicates ... we can no assume that refstack-client can connect to the cloud .. 19:48:05 <sslypushenko> catherineD|2: btw, I don't like this idea, though 19:48:07 <andrey-mp> this is the question - how we can identify any cloud? 19:48:30 <catherineD|2> andrey-mp: this is why I agree with your review technically but not practically 19:48:38 <andrey-mp> or how we can associate uploaded results? 19:49:02 <catherineD|2> sslypushenko: pls let us know your thoughts 19:49:45 <catherineD|2> andrey-mp: this is the exact reason why I think result should be assocciate to product_id and not product_ref_id 19:50:09 <andrey-mp> with such thoughts - association is possible only in RefStack UI 19:50:26 <catherineD|2> andrey-mp: user upload their result and determine the product they want the result to associate to 19:50:35 <andrey-mp> in this case we should re-think use cases about Cloud ID detection 19:51:24 <sslypushenko> idea to upload raw results using refstack-client looks like a working temporary solution but it breaks the idea of RefStack as an origin of truth 19:51:49 <andrey-mp> sslypushenko: +100 19:51:55 <catherineD|2> Cloud ( identified by CloudID) is an instance of a product .. product is unique 19:52:26 <sslypushenko> subunit test results does not contain any kind on cloud identity - and it is a main issue here 19:52:31 <andrey-mp> catherineD|2: cloud can be built from several products 19:53:11 <andrey-mp> but it doesn't describe what is CloudID 19:53:13 <catherineD|2> sslypushenko: that is a request from hogepodge: to allow vendor to use the results from their CI run 19:53:47 <sslypushenko> catherineD|2: I know... unfortunately it is also a workaround not a soulution 19:53:55 <sslypushenko> *solution 19:54:34 <catherineD|2> sslypushenko: and since that is our mission (as we describe in our mission statement) we have to honor that .. 19:54:44 <catherineD|2> sslypushenko: I understand .. 19:55:50 <sslypushenko> we need to provide kind of tool for signing test results. This tool should be small and safe to run as a part of internal ci pipeline 19:56:01 <catherineD|2> sslypushenko: ++ 19:56:24 <andrey-mp> how it can work/ 19:56:44 <catherineD|2> we thought that we can rely on honor system ... but this thought is being challenge right now 19:56:48 <sslypushenko> Than we can assume that these test results was produced by cloud with specified CPID 19:57:20 <andrey-mp> question about what is the CloudID still present ) 19:57:50 <sslypushenko> catherineD|2: I'd like to rely on some peace of code besides honor) 19:58:23 <catherineD|2> andrey-mp: good question ... at this point cloudid is just a number 19:58:41 <catherineD|2> even that we can get it from a cloud it is not perfect ... 19:59:01 * catherineD|2 2 mins left 19:59:46 <catherineD|2> let's think about this some more 20:00:19 <catherineD|2> meanwhile please review Replace run_tempest.sh script with ostestr command ( https://review.openstack.org/#/c/329691/ ) 20:00:25 <catherineD|2> need to end the meeting ... 20:00:29 <catherineD|2> thank you all 20:00:35 <catherineD|2> #endmeeting