| *** openstack has joined #savanna | 17:19 | |
| *** IlyaE has joined #savanna | 17:20 | |
| *** dina_belova has quit IRC | 17:20 | |
| *** dina_belova has joined #savanna | 17:23 | |
| *** ruhe has quit IRC | 17:28 | |
| *** NikitaKonovalov has joined #savanna | 17:29 | |
| SergeyLukjanov | that's the first message in #savanna at eavesdrop | 17:32 |
|---|---|---|
| *** NikitaKonovalov has quit IRC | 17:34 | |
| *** ChanServ changes topic to "Savanna project (Hadoop on OpenStack) - http://savanna.rtfd.org; channel logs: http://eavesdrop.openstack.org/irclogs/%23savanna/; discussions in this (#savanna) channel; weekly meetings on Thursdays at 18:00 UTC in #openstack-meeting-alt channel" | 17:35 | |
| *** NikitaKonovalov has joined #savanna | 17:38 | |
| *** mattf is now known as _mattf | 17:49 | |
| *** NikitaKonovalov has quit IRC | 17:51 | |
| *** _mattf is now known as mattf | 17:53 | |
| *** rnirmal has joined #savanna | 18:07 | |
| *** dina_belova has quit IRC | 18:49 | |
| *** NikitaKonovalov has joined #savanna | 18:51 | |
| openstackgerrit | Jonathan Maron proposed a change to stackforge/savanna: Fix to convert parsing failure https://review.openstack.org/39553 | 18:51 |
| *** dina_belova has joined #savanna | 18:52 | |
| *** NikitaKonovalov has quit IRC | 18:55 | |
| *** dmitryme has quit IRC | 19:03 | |
| tmckayrh | Hey savanna folks, crobertsrh and I have been chatting about job source things and were looking at https://etherpad.openstack.org/savanna_API_draft_EDP_extensions | 19:15 |
| tmckayrh | On the example Job Source Object that is shown there, type is "Hive". I think that's not correct -- Hive is a type of job, not a type of Job Source (meaning a place where jobs are retrieved from) | 19:16 |
| SergeyLukjanov | tmckayrh, yes, I think it's correct | 19:17 |
| tmckayrh | Two thoughts: Maybe we should rename Job Source to something else (Job Depot, Job Store, Job Repository, etc) so that "Job Source" meaning actual runtime code for a job is not confused with "Job Source" meaning where it is stored | 19:17 |
| tmckayrh | hmm, forgot my second thought | 19:20 |
| tmckayrh | Okay, so job source type should be something like "git", "internal", "swift", "hdfs", "mercurial", etc -- something that makes it easier to map to a plugin, maybe? | 19:22 |
| tmckayrh | And I suppose for "SCM" type solutions like git, mercurial, svn, cvs, etc it makes sense to break them out as separate distinct types rather than just "SCM" because they have to be subclassed at some point anyway. | 19:23 |
| *** SergeyLukjanov has quit IRC | 19:24 | |
| crobertsrh | +1 for separating them out | 19:24 |
| tmckayrh | woohoo, +1! I need to save those up, maybe I can trade them in for something | 19:25 |
| *** SergeyLukjanov has joined #savanna | 19:27 | |
| openstackgerrit | Jonathan Maron proposed a change to stackforge/savanna: Fix to convert parsing failure https://review.openstack.org/39553 | 19:28 |
| SergeyLukjanov | tmckayrh, crobertsrh, I think that there are good ideas, let's discuss them on tomorrows meeting | 19:30 |
| tmckayrh | SergeyLukjanov, crobertrh, ack, I have updated https://etherpad.openstack.org/savanna_API_draft_EDP_extensions with notes based on ^^ | 19:32 |
| tmckayrh | I think maybe the terminology is not always clear. "Source" is overloaded | 19:32 |
| *** dina_belova has quit IRC | 19:33 | |
| openstackgerrit | Jonathan Maron proposed a change to stackforge/savanna: Fix to convert parsing failure https://review.openstack.org/39553 | 19:36 |
| tmckayrh | SergeyLukjanov, crobertsrh, also note that there maybe is a conflict in the sequence diagram vs the blueprint. On execution, when job executable code needs to be retrieved, does the job manager component synthesize information from the job object and job source object to do the code retrieval itself (what the sequence says) or does the job manager pass the id from the job object to the job source component and it does the retrieval f | 19:41 |
| tmckayrh | I think the second. I think job manager should call job source component and say "give me this job from this source" | 19:41 |
| tmckayrh | If that's true, I should update the diagram | 19:42 |
| *** IlyaE has quit IRC | 19:45 | |
| *** NikitaKonovalov has joined #savanna | 19:48 | |
| crobertsrh | tmckahrh: that might be the right thing to do | 19:49 |
| tmckayrh | crobertsrh, yes, because otherwise the job manager component will have to have knowledge of job retrieval plugins. Doesn't seem right. | 19:50 |
| crobertsrh | Yes, that would be poor | 19:50 |
| tmckayrh | So it's maybe an extra hop between components, but it's better design and encapsulation. I think it was still a little fuzzy when I made the diagram | 19:51 |
| *** NikitaKonovalov has quit IRC | 19:52 | |
| *** dina_belova has joined #savanna | 20:04 | |
| *** dina_belova has quit IRC | 20:09 | |
| *** SergeyLukjanov has quit IRC | 20:12 | |
| *** qwerty_nor has quit IRC | 20:15 | |
| *** dmitryme has joined #savanna | 20:18 | |
| *** dmitryme has quit IRC | 20:29 | |
| *** benl has joined #savanna | 20:31 | |
| *** dina_belova has joined #savanna | 20:44 | |
| *** NikitaKonovalov has joined #savanna | 20:48 | |
| *** dina_belova has quit IRC | 20:49 | |
| *** NikitaKo_ has joined #savanna | 20:51 | |
| *** NikitaKonovalov has quit IRC | 20:51 | |
| *** crobertsrh is now known as _crobertsrh | 20:54 | |
| *** NikitaKo_ has quit IRC | 20:56 | |
| *** IlyaE has joined #savanna | 21:11 | |
| *** IlyaE has quit IRC | 21:22 | |
| *** aignatov has quit IRC | 21:42 | |
| *** aignatov has joined #savanna | 21:43 | |
| *** dina_belova has joined #savanna | 21:46 | |
| *** dina_belova has quit IRC | 21:50 | |
| *** tstclair is now known as _tstclair | 22:00 | |
| *** mattf is now known as _mattf | 22:17 | |
| *** rnirmal has quit IRC | 22:33 | |
| *** dina_belova has joined #savanna | 22:47 | |
| *** lastidiot has quit IRC | 22:48 | |
| *** NikitaKonovalov has joined #savanna | 22:52 | |
| *** dina_belova has quit IRC | 22:52 | |
| *** NikitaKonovalov has quit IRC | 22:56 | |
| *** benl has quit IRC | 23:00 | |
| *** lastidiot has joined #savanna | 23:46 | |
| *** NikitaKonovalov has joined #savanna | 23:53 | |
| *** NikitaKonovalov has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!