14:00:26 #startmeeting sahara 14:00:27 Meeting started Thu Jan 24 14:00:26 2019 UTC and is due to finish in 60 minutes. The chair is tellesnobrega. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:31 The meeting name has been set to 'sahara' 14:00:47 show off hands 14:01:28 jeremyfreudberg, tosky ?? 14:02:01 o/ 14:02:32 waiting a couple more minutes 14:04:14 lets get started 14:04:19 #topic News/Updates 14:04:34 I'm still fighting with image creation on the gate 14:04:43 not sure why it is not loading the plugins 14:04:46 have to double check 14:05:05 also I got a ubuntu and rhel image working for spark using sahara-image-pack 14:05:49 the functionality for sahara-image-pack is basically the remaining issue to complete the patch which fixes the usage of a reduced number of plugins 14:06:02 it is 14:06:13 hopefully I will figure out what is going on soon 14:06:14 on my side: 14:06:35 I followed the packaging process of the new plugins for RDO 14:06:57 and pushed the adding of sahara to requirements 14:07:20 even if the requirement teams is requiring a sahara-lib in the near future 14:08:17 I finished the work (for real) on the S3 support for sahara-scenario 14:08:34 yes 14:08:37 and I realized that my pending patches for unversioned endpoint are in a better shape that what I though 14:08:38 thought* 14:08:50 great news :) 14:09:08 also thanks to Jeremy's fix which made tempest tests work with APIv1.1 on unversioned endpoint 14:09:20 about sahara-lib, it shouldn't be too complicated to get it done 14:10:15 (i'm here) 14:10:20 hey jeremyfreudberg 14:10:24 welcome 14:10:27 is it something that you still think should be done after Stein, or is there some time? 14:10:34 welcome back jeremyfreudberg 14:10:52 we could try in stein, finish it up properly 14:11:38 we need to ask governance if they are fine with an exception to the process 14:11:41 if requirements team is really forcing our hand, it should be more important than plugin updates and so on 14:12:37 it should be a group decision i guess 14:12:37 * jeremyfreudberg lacking a bit of context (don't have all the messages) 14:13:52 so, tosky pushed for adding sahara to requirements 14:14:07 but requirements team said that we should have a sahara-lib instead of sahara 14:14:14 so another split 14:14:37 containing basically the plugins folder of current sahara code 14:14:43 and updating imports 14:15:01 and probably some issues to fiz 14:15:02 fix 14:15:04 well, I think the other way: not just the plugins folder, but everything required for the plugins folder to work 14:16:01 tosky, probably yes 14:16:25 but anyway, at least the review which adds sahara to the blacklist.txt of requirements still allow us to proceed with untangling and fixing the dependencies 14:16:57 yes 14:17:04 we can continue moving forward 14:17:14 jeremyfreudberg, any updates? 14:17:22 (without that, if we re-introduce the jobs which checks the requirements to each plugin, they will fail) 14:17:26 * tosky shuts up 14:18:00 no need 14:18:15 if you have more to add please do :) 14:18:50 we need sahara installed with plugins to run unit tests on the plugins 14:19:49 that too 14:20:14 tellesnobrega, not much to report on my side at the moment, the semester is going again 14:21:11 got it 14:21:13 and as I reduced the number of dependencies on the RDO packages to the really needed, we need to reduce them in the plugins, so we need again the job which checks the requirements 14:21:42 right 14:21:46 yes 14:23:15 anything else on this? 14:23:52 moving on 14:24:13 #topic Missing for Stein 14:24:37 I would like to take some time to put together what is missing for us to finish for Stein release 14:24:53 on APIv2 we are still missing some unversioned patches 14:24:57 and flip the flag 14:25:00 anything else? 14:25:01 (see also: http://eavesdrop.openstack.org/meetings/sahara/2019/sahara.2019-01-17-14.03.log.html#l-38 ) 14:25:23 APIv2: pending changes for sahara-tests (tempest and scenario) 14:25:51 and maybe afterwards some fixes to sahara to run the v2 jobs too 14:26:20 I think that jeremyfreudberg had few open TODOs for the boot from volume 14:26:32 I think he got those in 14:26:34 already 14:26:35 more on the client, though, so not blockers for the API stabilization 14:26:57 oh, ok 14:27:09 apiv2 is in really good state 14:27:10 yeah, the new boot from volume stuff needs to be on client/osc/dashboard 14:27:25 and yesterday I found this old story (from the summer) about APIv2 which may be outdated: https://storyboard.openstack.org/#!/story/2003092 14:27:39 or superseded by other stories 14:27:41 but Jeremy may know better :) 14:28:02 it may be superseded by other stories for sake of organization, but the content is still true 14:28:12 various features are in fact missing from various places 14:28:20 which ones? 14:28:32 can you take some time to update the story with more information? 14:29:16 yeah, i need to take some time especially to look at the client/osc code, i forget a lot of what was already implemented there / can be accomplished more generically 14:29:24 so i will update the story after looking 14:29:28 thanks 14:29:42 none of it will be a major amount of work, i will definitely tackle it soon 14:29:54 awesome 14:30:03 especially because we need to know the potential blocker for APIv2 vs possible internal-only changes 14:30:08 the split work is looking better each time 14:30:28 we are still missing the work on the deployment tools 14:30:52 that is all right tosky? 14:31:08 and maybe we need to do another split for a sahara-lib, but lets see 14:31:21 yes, correct: we need to have sahara in requirements (as discussed earlier, underway) 14:31:34 so that we can readd most of the missing jobs to the plugins, and avoid regressions 14:31:41 perfect 14:32:15 on the deployment side, we need fixes on the ansible deployer, the puppet deployer, and maybe tripleo to override the values in puppet 14:32:41 do you think those will be hard changes? 14:33:28 or should be straight forward changes? 14:33:41 again there is a bit of chicken-and-egg at least in ansible, because we are missing the plugins on pip, which require a release but that depends on the requirements 14:34:04 the changes may not be too complicated; they only require some thought on how to arrange the code 14:34:20 probably we need an overridable variable with the list of plugins 14:34:35 so that the number of plugins installed can be controlled through the deployment tool 14:34:42 the requirements should be in soon, I see a +2 already 14:35:09 I see 14:36:21 and finally we have python3 14:36:37 up to this point we are stuck on getting sahara to connect with instances using python3 14:36:42 all before that is working 14:36:47 we are not sure after 14:37:11 but it seems that it shouldn't be too complicated to get python3 compatibility 14:37:23 the "after" is mostly "run something on the remote node", I think it should mostly work 14:37:30 yes 14:37:36 that is my assumption as well 14:37:48 also, there is an issue when trying to create image using python3 14:37:57 we have to check libvirt with python3 14:38:40 the error with python3 is something related with pickle and the way it's used in sahara/utils/ssh_remote.py, which relies on sahara/utils/procutils.py (and sahara/cli/sahara_subprocess.py) 14:38:40 other than that, sahara should be fine for migrating to python3 14:38:46 something weird happens there 14:39:07 which issue with images on python3? Issues with libvirt/python3? 14:39:21 at least for me 14:39:33 it fails to create image when I run on python3 env 14:39:50 it throws that guestfs is not installed 14:40:07 maybe the libguestfs python3 binding is missing 14:40:15 maybe I installed the wrong libguestfs 14:42:06 I see there is a python3-libguestfs 14:42:12 I will install it and see if it works 14:42:16 if it does, perfect :) 14:42:52 \o 14:42:54 \o/ 14:43:18 I will try after the meeting 14:43:29 and that is basically what we have to finish by M3 14:43:45 for some reason i'm thinking that there is no python3 guestfs in centos, but it was a long time since i looked 14:44:41 I think that tellesnobrega is testing it on Fedora 14:44:46 I'm 14:44:58 python3-libguestfs is available 14:45:07 yeah 14:45:56 ok, so it worked :) 14:46:02 only missing the bindings 14:46:04 thanks tosky 14:46:10 nice 14:46:56 sooo, 14:47:09 about the sahara-lib, tosky do you think this is a must? 14:48:06 tellesnobrega: the priority is the patch which make sahara works regardless of the plugins 14:48:13 of the amount of plugins installed 14:48:27 and the deployment with various deployment systems 14:49:00 ok 14:49:04 the split of sahara-lib depends also on whether the government team is willing to accept it now, after the freeze (apparently) 14:49:05 so, lets focus on those 14:49:18 I'd say that asking the TC wouldn't hurt 14:49:18 and if time allows it, we can get to it 14:49:36 but better ask first, because they may say "sure, but before M3" :) 14:50:20 I will ask soon, but not commit to it 14:50:26 since we have priorities 14:50:44 sure, I mentioned it because I think it's better to have the full picture 14:50:55 of course 14:51:46 related to APIv2, there is the pending patch which fixes the policies ; jeremyfreudberg, as you raised some points, do you think it's fixed? 14:51:51 (when you can) 14:52:07 oh, right, it might be 14:52:26 i'll need to look again, but that patch is definitely a priority 14:52:51 jeremyfreudberg, please take a look and let me know what I need to improve there 14:53:43 and everyone, (especially jeremyfreudberg again, sorry :) please recheck the sahara-tests patches floating around, I'd like to cut a new release which could be used also for rocky too 14:55:40 yup 14:56:19 we are 4 minutes away 14:56:27 anything else for the day? 14:57:54 not from me 14:58:35 before we close, I just want to again say thank you both for the great work done up to now 14:58:54 sahara will be in a better place once we finish all these big features 14:58:56 we all worked hard :) 14:59:04 so, thanks all :) 14:59:11 we sure did 15:00:02 see you all next week 15:00:07 #endmeeting