| *** tellesnobrega has quit IRC | 00:26 | |
| *** tellesnobrega has joined #openstack-sahara | 00:32 | |
| *** masber has joined #openstack-sahara | 01:13 | |
| *** tmckay has quit IRC | 02:32 | |
| *** Bhujay has joined #openstack-sahara | 06:00 | |
| *** openstackgerrit has quit IRC | 06:49 | |
| *** rcernin has quit IRC | 06:55 | |
| *** tesseract has joined #openstack-sahara | 07:06 | |
| *** tosky has joined #openstack-sahara | 08:20 | |
| *** Bhujay has quit IRC | 08:44 | |
| *** Bhujay has joined #openstack-sahara | 09:30 | |
| *** Bhujay has quit IRC | 11:08 | |
| *** Bhujay has joined #openstack-sahara | 11:32 | |
| *** tmckay has joined #openstack-sahara | 11:59 | |
| *** jeremyfreudberg has joined #openstack-sahara | 12:53 | |
| *** Bhujay has quit IRC | 13:37 | |
| tellesnobrega | tosky, about the ambari version | 13:39 |
|---|---|---|
| tellesnobrega | what I'm thinking on doing is, remove the ambari_version argument | 13:39 |
| tellesnobrega | and use the plugin_version passed to select the ambari, if it is 2.3 we install 2.4.3.0 if not 2.3 (anything above it) we deploy 2.6.2.0 | 13:40 |
| tellesnobrega | does that sound good? | 13:40 |
| tosky | fine by me; as long as we produce a working setup, it's fine | 13:40 |
| tellesnobrega | cool | 13:41 |
| tellesnobrega | I think it is better than what we have now | 13:41 |
| tellesnobrega | which we allow the user to screw up | 13:41 |
| tosky | did you see my comment about my try to instantiate the image? | 13:41 |
| tosky | or better, try to create a cluster | 13:41 |
| tellesnobrega | yes | 13:51 |
| tellesnobrega | not sure what the issue is | 13:51 |
| jeremyfreudberg | tosky, how do you suggest i manage these backwards incompatible changes to saharaclient in relation to sahara-tests? | 14:20 |
| jeremyfreudberg | https://review.openstack.org/#/c/579680/ | 14:20 |
| tosky | jeremyfreudberg: we switch to openstacksdk /o\ | 14:22 |
| tosky | :D | 14:22 |
| tosky | joking aside, when I read the proposal, I though "if we need to do one breaking change, and another one later, why not unify them?" | 14:23 |
| jeremyfreudberg | tosky: what is the right workflow in terms of requirements/releasing here | 14:27 |
| jeremyfreudberg | change sahara-tests so client patch can be merged, release new client, bump requirement? | 14:27 |
| jeremyfreudberg | or something else | 14:27 |
| jeremyfreudberg | it would work, but i don't like the part where we have to temporarily break sahara-tests to do it | 14:28 |
| tosky | probably change sahara-tests to consider the future changes (which should be just some wrapper with try...catch), apply the change, update the requirements | 14:28 |
| tosky | we don't break sahara-tests, or we lose scenario tests | 14:28 |
| tosky | I don't remember how other projects managed this (and how) | 14:29 |
| tosky | probably microversioning | 14:29 |
| tosky | couldn't those changes wait and be applied only for v2? | 14:29 |
| jeremyfreudberg | tosky: oh, you mean to make these changes only in DataSourceManagerV2 | 14:30 |
| jeremyfreudberg | so if the user instantiates a v1.1 client they get the old args | 14:30 |
| jeremyfreudberg | we could do that | 14:30 |
| *** jeremyfreudberg has left #openstack-sahara | 16:17 | |
| *** openstackgerrit has joined #openstack-sahara | 16:19 | |
| openstackgerrit | Merged openstack/sahara-image-elements master: Totally rewrite s3_hadoop https://review.openstack.org/578193 | 16:19 |
| *** tesseract has quit IRC | 16:20 | |
| *** jeremyfreudberg has joined #openstack-sahara | 17:21 | |
| jeremyfreudberg | apparently kolla has a [trustee] patch too: https://review.openstack.org/#/c/573763/ | 17:23 |
| tosky | which looks sane | 17:24 |
| tosky | I think I found the issue with openstack-ansible-os_sahara | 17:25 |
| tosky | we don't support service_token_roles_required=True | 17:25 |
| tosky | or at least not in the configuration deployed by openstack-ansible-os_sahara | 17:25 |
| tosky | I didn't check that | 17:26 |
| jeremyfreudberg | yeah, in our own (devstack) jobs none of the services have that config option as True (or present) | 17:28 |
| jeremyfreudberg | i'm not sure if i really understand the option either | 17:30 |
| tellesnobrega | tosky, jeremyfreudberg I might drop for a while this afternoon, having an ISP change today | 17:36 |
| jeremyfreudberg | tellesnobrega: a better ISP? | 17:36 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_sahara master: Disable keystone_authtoken/service_token_roles_required for now https://review.openstack.org/579932 | 17:36 |
| tellesnobrega | hopefully I'll be able to do more testing (cdh and ambari) after today | 17:37 |
| tellesnobrega | jeremyfreudberg, hopefully yes, I'm changing from 25mb (never really is) to 100mb | 17:37 |
| tellesnobrega | even if its not 100mb, if it gets to 50mb I'll be at more than double what I got now | 17:37 |
| jeremyfreudberg | sounds good | 17:39 |
| openstackgerrit | Luigi Toscano proposed openstack/sahara master: DNM Test with service_token_roles_required=True https://review.openstack.org/579933 | 17:39 |
| jeremyfreudberg | tosky: i'm not sure if that patch will be all that interesting... i would expect issues if we set that to True in the services sahara talks to, not in sahara itself | 17:40 |
| tosky | jeremyfreudberg: interesting too | 17:41 |
| jeremyfreudberg | yes, we may as well be sure (since the old behavior is deprecated) | 17:42 |
| tosky | if it works, we may even flip it | 17:42 |
| tosky | one warning less in the logs | 17:42 |
| jeremyfreudberg | that would be nice | 17:43 |
| jeremyfreudberg | someone should prepare another patch, with service_token_roles_required=True in cinder,glance,heat,neutron,nova (and I suppose manila, if we had testing) | 17:43 |
| jeremyfreudberg | as we talk to all those services directly, sometimes | 17:44 |
| *** tellesnobrega has quit IRC | 17:52 | |
| jeremyfreudberg | also, tosky, going back to our discussion early today, i cannot make the saharaclient change on v2 only, because s3 datasouces will be on v1 too | 17:53 |
| tosky | jeremyfreudberg: uhm, but is the change needed to have it working? Can't the extra parameter be ignored? | 17:56 |
| jeremyfreudberg | i guess i could write it in a totally different way for v1 | 17:57 |
| jeremyfreudberg | with lots of optional args | 17:57 |
| jeremyfreudberg | seems a bit throwaaway | 17:57 |
| jeremyfreudberg | throwaway | 17:57 |
| jeremyfreudberg | i will think about it more | 17:57 |
| *** tosky has quit IRC | 18:14 | |
| *** jeremyfreudberg has quit IRC | 18:32 | |
| *** tellesnobrega has joined #openstack-sahara | 19:22 | |
| tellesnobrega | hey | 21:07 |
| tellesnobrega | I'm back | 21:07 |
| tellesnobrega | with new internet | 21:07 |
| tellesnobrega | apparently a lot better than before | 21:07 |
| *** rcernin has joined #openstack-sahara | 22:09 | |
| *** rcernin has quit IRC | 22:49 | |
| *** rcernin has joined #openstack-sahara | 23:02 | |
| *** rcernin_ has joined #openstack-sahara | 23:23 | |
| *** rcernin has quit IRC | 23:24 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!