| *** kebray has quit IRC | 01:29 | |
| *** kebray has joined #openstack-barbican | 01:47 | |
| *** dimtruck is now known as zz_dimtruck | 02:17 | |
| *** zz_dimtruck is now known as dimtruck | 04:08 | |
| *** crc32 has joined #openstack-barbican | 04:12 | |
| *** ajc_ has joined #openstack-barbican | 04:28 | |
| *** dimtruck is now known as zz_dimtruck | 05:21 | |
| *** jaosorior has joined #openstack-barbican | 06:39 | |
| *** crc32 has quit IRC | 06:56 | |
| *** Nirupama has joined #openstack-barbican | 07:41 | |
| *** chlong has quit IRC | 08:00 | |
| *** woodster_ has quit IRC | 08:06 | |
| *** kebray has quit IRC | 08:35 | |
| *** darrenmoffat1 has quit IRC | 08:39 | |
| *** darrenmoffat has joined #openstack-barbican | 08:40 | |
| *** jaosorior has quit IRC | 08:46 | |
| *** nkinder has joined #openstack-barbican | 08:48 | |
| *** ajc__ has joined #openstack-barbican | 12:14 | |
| *** ajc__ has quit IRC | 12:15 | |
| *** miqui_away has quit IRC | 12:21 | |
| *** woodster_ has joined #openstack-barbican | 12:25 | |
| *** Nirupama has quit IRC | 12:28 | |
| *** ajc_ has quit IRC | 13:16 | |
| *** jaosorior has joined #openstack-barbican | 13:29 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/barbican: Refactor exception handling in the app side https://review.openstack.org/152123 | 14:03 |
|---|---|---|
| jaosorior | I'm getting some pep8 issues on the commit above, can someone help me out figuring out what's up with those? | 14:04 |
| *** lisaclark1 has joined #openstack-barbican | 14:15 | |
| *** tkelsey has joined #openstack-barbican | 14:28 | |
| openstackgerrit | Tim Kelsey proposed openstack/barbican-specs: Adding spec for Barbican MKEK Model. https://review.openstack.org/148948 | 14:30 |
| *** lisaclark1 has quit IRC | 14:41 | |
| *** lisaclark1 has joined #openstack-barbican | 14:45 | |
| *** rm_work is now known as rm_work|away | 14:47 | |
| *** lisaclark1 has quit IRC | 14:51 | |
| *** openstackgerrit has quit IRC | 14:52 | |
| *** openstackgerrit has joined #openstack-barbican | 14:52 | |
| *** lisaclark1 has joined #openstack-barbican | 14:53 | |
| *** rm_work|away is now known as rm_work | 15:04 | |
| *** paul_glass has joined #openstack-barbican | 15:07 | |
| *** paul_glass has quit IRC | 15:10 | |
| *** zz_dimtruck is now known as dimtruck | 15:16 | |
| *** darrenmoffat has left #openstack-barbican | 15:17 | |
| *** lisaclark1 has quit IRC | 15:28 | |
| *** lisaclark1 has joined #openstack-barbican | 15:34 | |
| *** nkinder has quit IRC | 15:43 | |
| *** crc32 has joined #openstack-barbican | 15:48 | |
| *** crc32 has quit IRC | 15:52 | |
| *** kgriffs|afk is now known as kgriffs | 15:58 | |
| *** SheenaG1 has joined #openstack-barbican | 16:01 | |
| *** lisaclark1 has quit IRC | 16:14 | |
| *** lisaclark1 has joined #openstack-barbican | 16:17 | |
| *** david-lyle_afk is now known as david-lyle | 16:19 | |
| *** kebray has joined #openstack-barbican | 16:19 | |
| *** crc32 has joined #openstack-barbican | 16:24 | |
| alee | woodster_, rm_work , morganfainberg -responded to comments in https://review.openstack.org/#/c/127353/4 | 16:24 |
| *** crc32 has quit IRC | 16:25 | |
| rm_work | kk | 16:29 |
| woodster_ | alee will take a look after our standup | 16:30 |
| alee | woodster_, thanks | 16:30 |
| rm_work | alee: lookin' good :P | 16:33 |
| alee | thanks :) | 16:33 |
| *** kgriffs is now known as kgriffs|afk | 16:43 | |
| *** paul_glass has joined #openstack-barbican | 16:43 | |
| jaosorior | woodster_, alee, rm_work, jvrbanac are you guys acquainted with the i18n stuff? I'm getting some strange pep8 issues because of it that I haven't figured out in this CR https://review.openstack.org/148948 | 16:47 |
| jvrbanac | jaosorior, I think you linked the wrong CR? | 16:48 |
| jaosorior | Lol yes | 16:49 |
| jaosorior | https://review.openstack.org/152123 | 16:49 |
| jaosorior | That one | 16:50 |
| *** gyee has joined #openstack-barbican | 16:50 | |
| rm_work | jaosorior: they have weird rules for i18n | 16:50 |
| rm_work | it violates everything that you're used to doing | 16:50 |
| rm_work | you have to import _LE / _LW / etc directly | 16:50 |
| woodster_ | rm_work you don't because we import | 16:51 |
| rm_work | and use those (not just _()) | 16:51 |
| woodster_ | 'u' instead of _ | 16:51 |
| rm_work | right but | 16:51 |
| rm_work | you're not supposed to | 16:51 |
| rm_work | even though it violates typical pep8 stuff | 16:51 |
| rm_work | the openstack i18n stuff is different | 16:51 |
| woodster_ | well it passes their gate! the one we are using anyway :) | 16:52 |
| rm_work | for example: https://github.com/stackforge/octavia/blob/master/octavia/certificates/manager/barbican.py#L24 | 16:52 |
| rm_work | they mandate the imports in that format | 16:52 |
| woodster_ | jaosorior could the issue be with the format(*args, **kwargs)?....I think the {slot} mode for format requires a key/value mapping maybe? | 16:53 |
| jaosorior | I have no clue :/ | 16:53 |
| rm_work | OH the just a string error | 16:53 |
| rm_work | yeah | 16:53 |
| woodster_ | rm_work I'd think this would be a valid import: from barbican import i18n as u | 16:53 |
| rm_work | sec | 16:53 |
| rm_work | this is an easy fix | 16:54 |
| rm_work | woodster_: normally yes but they require weird shit <_< | 16:54 |
| jaosorior | I was stomping my head for a fair bit | 16:54 |
| rm_work | yeah jaosorior | 16:54 |
| woodster_ | jaosorior, can you remove the *args part there? ARe you seeing this locally or just in the gate? | 16:54 |
| rm_work | jaosorior: so on line 80 | 16:54 |
| rm_work | jaosorior: you have u.("sometext".format(thing)) | 16:55 |
| woodster_ | rm_work I remember seeing an example that does the _LE stuff, but I didn't take that as prescriptive if we have a better way of doing it 8^) | 16:55 |
| rm_work | put an extra parens around the stuff | 16:55 |
| rm_work | u_(("sometext".format(thing))) | 16:55 |
| rm_work | it'll work | 16:55 |
| rm_work | i have not yet figured out WHY, but it's a peculiarity of how it handles the parsing of .format() | 16:56 |
| woodster_ | really? That doesn't make sense unless some sort of pre-parsing thing? | 16:56 |
| rm_work | at the point at which the i18n macros deal with stuff | 16:56 |
| rm_work | it hasn't figured it out yet | 16:56 |
| *** crc32 has joined #openstack-barbican | 16:56 | |
| rm_work | if you understand WHY, feel free to explain it to me, but it works | 16:56 |
| rm_work | :) | 16:56 |
| rm_work | anywhere you do a .format() inside a u._() | 16:57 |
| rm_work | just need extra parens :) | 16:57 |
| woodster_ | oh I see...checkout this: https://github.com/cloudkeep/barbican/blob/master/barbican/model/repositories.py#L183 | 16:57 |
| rm_work | ah | 16:57 |
| rm_work | you can do that too | 16:57 |
| woodster_ | jaosorior, you need to keep the slot in the I18n text for translation, but the .format is performed outside of that | 16:58 |
| rm_work | hmm, we may need to fix some stuff in our code in that case... lol | 16:58 |
| woodster_ | rm_work the double parens is not good...you'll never get a localization match that way | 16:58 |
| rm_work | yeah I may have to dig through and find where we did that | 16:58 |
| rm_work | it was ... a few places | 16:59 |
| *** lisaclark1 has quit IRC | 16:59 | |
| woodster_ | jaosorior ^^^^ Please see https://github.com/cloudkeep/barbican/blob/master/barbican/model/repositories.py#L183 | 16:59 |
| jaosorior | Let's see | 17:00 |
| jaosorior | Ooooh I see! | 17:02 |
| jaosorior | I'll try it when I get home | 17:02 |
| *** lisaclark1 has joined #openstack-barbican | 17:02 | |
| jaosorior | Thanks people :D | 17:03 |
| rm_work | do you guys remember if %s type string substitution was deprecated in favor of .format()? | 17:04 |
| woodster_ | I recall .format is more pythonic than % | 17:06 |
| rm_work | i guess https://www.python.org/dev/peps/pep-3101/ | 17:07 |
| rm_work | says "The new system does not collide with any of | 17:08 |
| rm_work | the method names of the existing string formatting techniques, so | 17:08 |
| rm_work | both systems can co-exist until it comes time to deprecate the | 17:08 |
| rm_work | older system." | 17:08 |
| rm_work | so I guess that *implies* that at some point the old system would be deprecated? | 17:08 |
| openstackgerrit | Thomas Dinkjian proposed openstack/python-barbicanclient: Adds base behaviors, secret behaviors and the secret smoke tests https://review.openstack.org/151777 | 17:20 |
| *** lisaclark2 has joined #openstack-barbican | 17:23 | |
| *** lisaclark1 has quit IRC | 17:26 | |
| jvrbanac | rm_work, they don't collide, and % isn't fully deprecated yet; however, .format() is the preferred method for normal strings as it has better error handling mechanisms and more functionality | 17:31 |
| rm_work | right | 17:31 |
| jvrbanac | rm_work, in Barbican. I believe our policy is .format() for everything other than log messages | 17:31 |
| *** xaeth has joined #openstack-barbican | 17:32 | |
| jvrbanac | rm_work, or unless you have a really good reason ;) | 17:32 |
| jvrbanac | however, those are quite rare | 17:34 |
| *** SheenaG1 has quit IRC | 17:36 | |
| *** jkf has joined #openstack-barbican | 17:39 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/barbican: Updated from global requirements https://review.openstack.org/152206 | 17:39 |
| *** atiwari1 has quit IRC | 17:46 | |
| *** kebray has quit IRC | 17:49 | |
| *** kebray has joined #openstack-barbican | 17:59 | |
| *** lisaclark2 has quit IRC | 18:00 | |
| *** SheenaG1 has joined #openstack-barbican | 18:08 | |
| *** atiwari has joined #openstack-barbican | 18:08 | |
| *** dimtruck is now known as zz_dimtruck | 18:12 | |
| rm_work | redrobot: you there? | 18:14 |
| rm_work | redrobot: redrobot~~ | 18:18 |
| rm_work | oh, probably at lunch <_< | 18:18 |
| *** zz_dimtruck is now known as dimtruck | 18:21 | |
| *** bdpayne has joined #openstack-barbican | 18:28 | |
| openstackgerrit | Thomas Dinkjian proposed openstack/python-barbicanclient: Adds base behaviors, secret behaviors and the secret smoke tests https://review.openstack.org/151777 | 18:40 |
| openstackgerrit | Merged openstack/barbican: Updated from global requirements https://review.openstack.org/152206 | 18:50 |
| *** lisaclark1 has joined #openstack-barbican | 18:55 | |
| openstackgerrit | Thomas Dinkjian proposed openstack/python-barbicanclient: Adds base behaviors, secret behaviors and the secret smoke tests https://review.openstack.org/151777 | 19:02 |
| *** lisaclark1 has quit IRC | 19:02 | |
| *** kgriffs|afk is now known as kgriffs | 19:19 | |
| *** kgriffs is now known as kgriffs|afk | 19:29 | |
| jaosorior | jvrbanac: You're on fire reviewing today :D | 19:48 |
| jvrbanac | jaosorior, ? | 19:49 |
| jaosorior | just read a couple of particularly good reviews | 19:50 |
| jvrbanac | jaosorior, ahh lol | 19:50 |
| rm_work | redrobot: around? | 19:51 |
| *** kfarr has joined #openstack-barbican | 19:53 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/barbican: Refactor exception handling in the app side https://review.openstack.org/152123 | 19:54 |
| jaosorior | rm_work, jvrbanac, woodster_: It worked :D | 19:54 |
| jaosorior | http://www.blog.equals6.com/wp-content/uploads/2014/03/mind-blown.gif | 19:55 |
| *** lisaclark1 has joined #openstack-barbican | 19:57 | |
| rm_work | heh | 19:57 |
| rm_work | yeah, I'm glad we had that talk because I didn't realize there was a technical difference between double-parens and moving the .format() outside | 19:58 |
| rm_work | so, now I know it matters :P | 19:58 |
| woodster_ | alee, rm_work, I added comments to the rbac bp...basically agreeing with you guys\ | 19:59 |
| redrobot | the barbican weekly meeting is starting now on #openstack-meeting-alt | 20:00 |
| rm_work | redrobot: I need to talk to you after :P | 20:01 |
| * redrobot hides | 20:01 | |
| jaosorior | lol | 20:01 |
| redrobot | alee you around for the meeting? | 20:02 |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/barbican: Drop old namespace for some oslo libraries https://review.openstack.org/150372 | 20:12 |
| openstackgerrit | Merged openstack/barbican: Fix error in "tenants to projects" migration script https://review.openstack.org/151145 | 20:13 |
| openstackgerrit | Merged openstack/barbican: Fix downgrade for revision 254495565185 https://review.openstack.org/151172 | 20:14 |
| alee | SheenaG1, ping | 20:46 |
| SheenaG1 | what's up alee? | 20:46 |
| alee | SheenaG1, were you going to send me an abstract? | 20:46 |
| alee | SheenaG1, or did you send it already and I missed it .. | 20:46 |
| SheenaG1 | alee: we're working on it - I have Chelsea taking a look right now, we'll try to have it over to you in the next day or so | 20:47 |
| *** lisaclark1 has quit IRC | 20:47 | |
| alee | SheenaG1, OK great thanks | 20:47 |
| SheenaG1 | alee: no problem, we haven't forgotten - it's just not in a state that's worth reviewing yet, IMO | 20:47 |
| alee | SheenaG1, ok --- deadline's in a week though | 20:48 |
| SheenaG1 | alee: understood. We'll definitely want your input before we submit | 20:49 |
| *** paul_glass has quit IRC | 20:54 | |
| *** paul_glass has joined #openstack-barbican | 20:57 | |
| *** lisaclark1 has joined #openstack-barbican | 21:00 | |
| morganfainberg | jaosorior, so the reason we started collapsing is the migrations are hard to maintain as things like oslo.db update | 21:00 |
| *** lisaclark1 has quit IRC | 21:01 | |
| jaosorior | öhö | 21:01 |
| jaosorior | can you elaborate? | 21:01 |
| morganfainberg | jaosorior, sure, the original migrations (go look at essex and such) are very simplistic | 21:01 |
| morganfainberg | and needed a lot of work to carry forward when exceptions changed etc | 21:01 |
| morganfainberg | so, we collapsed and make the first migration just create the schema expected at n-2 | 21:02 |
| morganfainberg | that way you have work to maintain the collapsed start | 21:02 |
| morganfainberg | but it's a small scope of migrations | 21:02 |
| morganfainberg | *and* it eliminates all the migration test maintenance | 21:02 |
| jaosorior | that sounds good! | 21:02 |
| morganfainberg | if we move to alembic we would either need to forever maintain SQL-A migrate or convert everything to alembic | 21:02 |
| morganfainberg | both are unfun | 21:02 |
| morganfainberg | so this way we can say when we move to alembic, we only have 2 releases where SQL_A migrate is used [if we don | 21:03 |
| morganfainberg | 't convert everything] | 21:03 |
| woodster_ | morganfainberg, we did our alembic stuff before oslo db I believe. Do you think we should move away from alembic? | 21:03 |
| morganfainberg | no | 21:03 |
| morganfainberg | stay with it | 21:03 |
| morganfainberg | we're planning on moving to it | 21:03 |
| morganfainberg | but we have to wait for some fixes to land in oslo.db | 21:04 |
| *** lisaclark1 has joined #openstack-barbican | 21:04 | |
| morganfainberg | specific to how we do migrations | 21:04 |
| woodster_ | morganfainberg, oh I see. So moving *from* oslo.db eventually then? I wasn't expecting that :) | 21:04 |
| morganfainberg | well no | 21:04 |
| morganfainberg | alembic is migrations | 21:04 |
| morganfainberg | oslo.db is the layer over SQL-Alchemy | 21:05 |
| *** kfarr has quit IRC | 21:05 | |
| morganfainberg | (not to be confused with sql-alchemy-migrate which alembic replaces) | 21:05 |
| *** tkelsey has quit IRC | 21:05 | |
| morganfainberg | oslo.db works fine for us w/ sql-a-migrate, but alembic not yet. | 21:05 |
| morganfainberg | oslo.db we're happy with [for the most part] | 21:05 |
| jaosorior | morgainfainberg: We are using straight SQL-Alchemy, how good has oslo.db for you? Should we start migrating to oslo.db soonish? | 21:05 |
| jaosorior | ohooh | 21:05 |
| morganfainberg | oslo.db is just a way to do config over sql-a, nice way to be consistent | 21:06 |
| morganfainberg | some openstack-specific unification | 21:06 |
| morganfainberg | i can't tell you if barbican would benefit from it. but it would make you gjuys more consistent with the rest of OpenStack | 21:07 |
| morganfainberg | might also reduce some code you need to carry | 21:07 |
| jaosorior | I'll take a look into it (later this week) | 21:07 |
| jaosorior | probably will see how it's used in keystone anyway | 21:07 |
| *** kgriffs|afk is now known as kgriffs | 21:07 | |
| jaosorior | morgainfainberg: Thanks for the info Mr. :D | 21:08 |
| jaosorior | redrobot, woodster_: So... I'm guessing the conclusion is that I should fix this https://bugs.launchpad.net/barbican/+bug/1415869 ? | 21:08 |
| redrobot | jaosorior what is rev 4070806f6972? | 21:09 |
| jaosorior | revision | 21:09 |
| jaosorior | the script's hash | 21:09 |
| redrobot | jaosorior right, but what point in time is it? Juno-final? | 21:10 |
| jaosorior | ....no idea | 21:10 |
| redrobot | jaosorior so, I think that being able to upgrade from Juno->Kilo is a must-have for us. I consider downgrades a "nice-to-have" | 21:11 |
| jaosorior | 4070806f6972 dsa in container type modelbase_to | 21:11 |
| jaosorior | upgrade works | 21:11 |
| jaosorior | well | 21:11 |
| jaosorior | for all the other ones above it | 21:11 |
| jaosorior | I have no idea if the upgrade for that specific script works | 21:12 |
| jaosorior | because I can't downgrade to it... | 21:12 |
| openstackgerrit | Steve Heyman proposed openstack/barbican: Add the ability to use either identity v2 or v3 API https://review.openstack.org/152277 | 21:12 |
| jaosorior | it's actually this script 2ab3f5371bde_dsa_in_container_type_modelbase_to.py | 21:13 |
| jaosorior | got confused with the hash, sorry | 21:13 |
| jaosorior | atiwari: are you around? | 21:13 |
| atiwari | jaosorior, yes | 21:14 |
| jaosorior | can you help me out? | 21:14 |
| jaosorior | 2ab3f5371bde_dsa_in_container_type_modelbase_to.py I can't figure out that script... Sorry :/ | 21:14 |
| atiwari | sure | 21:14 |
| jaosorior | the downgrade doesn't seem to work for me (I use Postgres,) | 21:14 |
| jaosorior | and I'm trying to make it database independant | 21:14 |
| jaosorior | (or independent... however the actual word in english is..) | 21:15 |
| jaosorior | first, what did you mean by dsa in container type modelbase_to ? | 21:15 |
| jaosorior | "dsa in container type modelbase_to" | 21:15 |
| atiwari | wait, let me see | 21:15 |
| atiwari | I am off for sometime | 21:15 |
| jaosorior | Alright | 21:17 |
| jaosorior | well, it's quite late here (Finnish time) | 21:17 |
| jaosorior | so i'm gonna go soon | 21:17 |
| jaosorior | but when you have time | 21:17 |
| jaosorior | can you help me out figuring out what it was meant for initially in this bug https://bugs.launchpad.net/barbican/+bug/1415869 ? | 21:17 |
| *** kgriffs is now known as kgriffs|afk | 21:17 | |
| atiwari | we have added rya and dsa type in the container table | 21:17 |
| atiwari | before it was just generic and certificate | 21:18 |
| jaosorior | and also | 21:19 |
| jaosorior | in the upgrade and downgrade functions | 21:19 |
| jaosorior | the alter table operations are actually the same | 21:19 |
| jaosorior | is that the way it was intended? | 21:20 |
| jaosorior | oh no | 21:20 |
| jaosorior | wait up | 21:20 |
| jaosorior | sorry, I mis-read that part | 21:20 |
| jaosorior | they do change | 21:20 |
| jaosorior | anyway | 21:20 |
| jaosorior | now I get it | 21:20 |
| atiwari | no it is not same | 21:20 |
| jaosorior | will fix it tomorrow | 21:20 |
| atiwari | grwat | 21:20 |
| jaosorior | thanks atiwari | 21:20 |
| atiwari | great | 21:20 |
| atiwari | yew | 21:20 |
| atiwari | yrw | 21:21 |
| rm_work | redrobot: so, you around/ | 21:25 |
| xaeth | alee, if i wanted to have a conversation around the rpm spec file you provided, who would i have that w/ | 21:26 |
| alee | xaeth, me most likely :) | 21:27 |
| xaeth | ;) | 21:27 |
| xaeth | so i'm going through and doing some general cross release bits, like el6 vs fedora (since it has upstart bits, and we'd be using el6 internally) | 21:27 |
| xaeth | and as i was doing this i noticed lots of overlap on file ownership | 21:28 |
| xaeth | i've tried to clean some of that up, but i'm wondering if it wouldn't be better to go back to their original common package with a sub for api and worker ? the model in yours is that the common is the api | 21:28 |
| alee | xaeth, just a sec -- let me pull up the rpm | 21:30 |
| xaeth | np | 21:30 |
| *** chipmanc has joined #openstack-barbican | 21:30 | |
| xaeth | alee, this might be easier: https://gist.github.com/gregswift/78081dca9adabd14ad63 | 21:31 |
| alee | xaeth, possibly -- I think this was done this way to b more in line with what had been done for openstack-cinder and openstack-glance | 21:35 |
| xaeth | makes sense, which is why i wasn't going to move away from that at first. Do either of those have the concept of worker nodes? | 21:35 |
| alee | xaeth, not to say that it couldn't be done as you suggest -- we should just choose to do what makes sense. | 21:36 |
| *** kebray has quit IRC | 21:36 | |
| xaeth | or a parrallel | 21:36 |
| alee | xaeth, let me look at the other two .. | 21:36 |
| xaeth | i didn't see one in glance | 21:36 |
| xaeth | hadn't looked at cinder | 21:36 |
| alee | xaeth, in openstack-cinder, there is no worker mode | 21:38 |
| xaeth | kk | 21:38 |
| alee | but what is common in the fedora packaging is the idea that the common library code is in python-* | 21:38 |
| xaeth | problem is that its configs not code | 21:39 |
| alee | so for example, in openstack-cinder, you have openstack-cinder which is the server, and python-cinder which is the underlying code | 21:39 |
| xaeth | right, and in this python-barbican is the underlying code | 21:40 |
| xaeth | i added another file to that gist. this one includes a -qlp dump on the main and worker rpms | 21:40 |
| alee | xaeth, I'm checking the client to see what it depends on | 21:40 |
| alee | xaeth, I was checking the python-cinderclient to see if I depended on python-cinder .. | 21:43 |
| alee | with the idea being that the "common code" would be the code that was common to client and server | 21:43 |
| alee | but it doesn't | 21:43 |
| alee | let me look at another -- glance say | 21:43 |
| xaeth | k | 21:44 |
| xaeth | woodster_, redrobot : does work log to barbican-api.log ? | 21:44 |
| woodster_ | xaeth, worker logs to barbican-worker.log | 21:45 |
| xaeth | hrmm... kk :) your missing a logrotate entry then | 21:45 |
| *** lisaclark1 has quit IRC | 21:45 | |
| woodster_ | xaeth, ha nice! | 21:46 |
| *** kebray has joined #openstack-barbican | 21:47 | |
| *** lisaclark1 has joined #openstack-barbican | 21:47 | |
| xaeth | woodster_, i would also imagine that the api nodes shouldn't own the db manage script? | 21:47 |
| alee | xaeth, ok - so it looks like python-X is the python code, openstack-X is the init scripts/config files/scripts | 21:48 |
| alee | xaeth, are we not following that model .. looking? | 21:48 |
| xaeth | the model we are following is that both packages have both | 21:48 |
| xaeth | err | 21:49 |
| xaeth | bad statement | 21:49 |
| xaeth | python-X is the python code | 21:49 |
| xaeth | openstack-X and openstack-X-worker have a lot of the same stuph | 21:49 |
| rm_work | redrobot >_> | 21:49 |
| redrobot | rm_work what up? | 21:49 |
| alee | xaeth, right -- but I would think they would not be deployed on the same machine , right? | 21:50 |
| rm_work | redrobot: PM | 21:50 |
| redrobot | xaeth barbican-worker is just a python script. It logs to wherever you send the std-out and std-err. | 21:51 |
| redrobot | xaeth looks like we have it set up to log to barbican_worker.log https://github.com/openstack/barbican/blob/master/etc/init/barbican-worker.conf#L9 | 21:51 |
| alee | redrobot, xaeth , woodster_ I think we need some clarification as to how these workers are deployed | 21:52 |
| xaeth | i guess i would too | 21:53 |
| alee | that is do you expect to see a barbican worker on the same server as a barbican server? | 21:53 |
| xaeth | i'm under the impression that while they can be deployed together in a small environment, they are designed to be deployed separately for scale | 21:53 |
| alee | xaeth, and you could well be right :) | 21:53 |
| alee | xaeth, in which case I'd be curious as to which files it would need to share with barbican-server | 21:54 |
| alee | can you have multiple workers on the same machine? | 21:54 |
| redrobot | alee xaeth the intention is for the worker to run in it's own machine | 21:54 |
| alee | redrobot, and barbican-server is not on the same machine as a barbican-worker? | 21:55 |
| *** chipmanc has quit IRC | 21:56 | |
| redrobot | alee correct, see the first diagram here https://github.com/cloudkeep/barbican/wiki/Architecture | 21:56 |
| woodster_ | xeath so the worker process and keystone listener are optional on a worker node, the database manage script is optional on any of the nodes really. Rackspace probably run that on a dedicated node/process, with a different set of db user credentials for example. | 21:56 |
| redrobot | alee each box is its own machine | 21:56 |
| alee | redrobot, look at https://gist.github.com/gregswift/78081dca9adabd14ad63 lines 173-187 -- are these the files that correspond to a worker? | 21:57 |
| alee | woodster_, ^^ | 21:57 |
| woodster_ | the nuance is things lke the worker and keystone listener and (eventually) a retry/scheduler service. They could all be deployed on one worker node, or else optionally installed if desired. I'm not sure if keeping them separate is worth the rpm hassle though | 21:58 |
| woodster_ | alee, no not all of those files are relevant | 21:58 |
| xaeth | question... why separate out worker ? | 21:59 |
| xaeth | sounds like it shares enough with the main that its all just configuration at that point | 21:59 |
| woodster_ | alee in fact most are irrelevant. Only the /etc/barbican/barbican-api.conf and associated rotate file are relevant | 21:59 |
| xaeth | the actual binary being the main difference | 21:59 |
| alee | redrobot, woodster_ the key thing is to separate out what consitutes a server and what constitutes a worker | 21:59 |
| woodster_ | not sure what /etc/init/barbican-api.conf is... | 22:00 |
| xaeth | thats the upstart config | 22:00 |
| woodster_ | ...where server is defined as an api node? | 22:00 |
| xaeth | yes | 22:00 |
| woodster_ | xaeth, oh yes, thanks | 22:00 |
| alee | once we know what is in each of those are we can determine how to separate them out | 22:00 |
| redrobot | xaeth the reason we separate them is that we want to separate the concerns | 22:01 |
| woodster_ | the /etc/barbican/barbican-api.conf files is the only one shared between the api and worker node types | 22:01 |
| alee | woodster_, redrobot - in this spec file, the line 155-166 are in an api node, 173-187 are in the worker node | 22:01 |
| woodster_ | the remainder are for the api only | 22:01 |
| xaeth | redrobot, ya i understand the general concept. | 22:02 |
| woodster_ | alee, can you link to that file? | 22:02 |
| alee | https://gist.github.com/gregswift/78081dca9adabd14ad63 | 22:03 |
| woodster_ | alee, sorry my client just render the top part...going direclty to link to see the rest of the stuff there.... | 22:03 |
| alee | xaeth, if there is enough of an overlap, you could consider a openstack-barbican-common subpackage I suppose -- but I suspect it won't come to that. | 22:05 |
| redrobot | xaeth sorry, I'm not sure I understood your question then... I suppose the reason we originally separated the worker is so that you could "yum install barbican-worker" on the worker nodes... no real technical reason beyond that. | 22:06 |
| woodster_ | alee, xaeth, redrobot, I think line 160 (the db manage script) is the only one I could see separating into another package, but IMHO not worth it. #181 and #182 (assuming they are recurve grabbing vassal stuff too) are probably too broad, when only the barbican-api.conf file is needed. | 22:07 |
| xaeth | So as long as we drop it down so that only the api.conf is shared, then i think its pretty reasonable | 22:07 |
| woodster_ | alee, xaeth, redrobot, that all said...I've been bugged that the worker's conf file is called 'barbican-api.conf' :) | 22:07 |
| xaeth | and yes, i'd love to see it split so that it doesn't share that | 22:08 |
| woodster_ | ...or just call it barbican.conf, though I'd hate to accidently swap an api one for a worker one | 22:09 |
| xaeth | redrobot: my question was geared around the fact that they seemed to both include the same files, aside from the final binary. so if there isn't much of a split, why bother splitting? but if thats not the actual case and they should be fully split maybe we should get there? | 22:09 |
| xaeth | woodster_, i'd definitely lean towards separate because of that | 22:09 |
| woodster_ | xeath, redrobot, alee yeah the api and worker share much code in common so no point in splitting that. The division is really more of configuration files, and in the case of worker nodes some specific processes that can be run on them (but also using that same shared code base). | 22:14 |
| redrobot | xaeth, I should also point out that uwsgi is not a requirement | 22:16 |
| xaeth | thanks redrobot, i'll blame alee for that :) | 22:16 |
| redrobot | xaeth Rackspace deployment uses uwsgi, but barbican can be deployed using any server you like | 22:16 |
| redrobot | xaeth recently I've been running my dev environment server under Gunicorn | 22:17 |
| xaeth | actually i think its in the main spec file too so *shrug* | 22:17 |
| xaeth | but ya understood | 22:17 |
| redrobot | xaeth I wrote the original spec file, so it's probably my fault >_< | 22:17 |
| alee | redrobot, xaeth - well you need something .. | 22:17 |
| redrobot | alee xaeth it would be worthwhile looking into what openstack-keystone uses to run their server | 22:18 |
| xaeth | alee: yes, but it doesn't look like glance or keystone have one listed | 22:19 |
| xaeth | http://pkgs.fedoraproject.org/cgit/openstack-keystone.git/tree/openstack-keystone.spec?h=el6-icehouse | 22:19 |
| redrobot | alee it's possible to run it using the built-in wsgi host... | 22:19 |
| xaeth | redrobot, most of those commonalities are far as code goes are in python-barbican | 22:19 |
| alee | xaeth, redrobot ok | 22:19 |
| redrobot | which I think is what keystone is doing? | 22:19 |
| redrobot | xaeth so by extension, the vassals files are not necessary | 22:20 |
| xaeth | kewl | 22:20 |
| xaeth | hrm... so that is uwsgi specific? | 22:21 |
| redrobot | yes, uwsgi loads the vassal files, which point to the paste files, where the actual config is | 22:21 |
| xaeth | we could include them in the %docs or %datadir as a reference ... i saw that in one of hte other projects for something | 22:21 |
| xaeth | we dont have too. just could | 22:21 |
| redrobot | I'd rather not put any server specific stuff in, to avoid having to answer questions about it later :) | 22:22 |
| xaeth | ok | 22:23 |
| redrobot | xaeth I would think that we need to write a barbican-api.py script, which loads the barbican wsgi app and runs it in the built in wsgi host | 22:26 |
| xaeth | makes sense | 22:27 |
| *** kebray has quit IRC | 22:31 | |
| *** kebray has joined #openstack-barbican | 22:37 | |
| xaeth | got the shared list dropped down to just hte config and pushed the keystone listener to a separate package | 22:54 |
| xaeth | (which is what is in the spec file in barbican proper) | 22:54 |
| xaeth | update spec and output in the gist | 22:56 |
| *** kgriffs|afk is now known as kgriffs | 22:56 | |
| *** xaeth is now known as xaeth_afk | 23:05 | |
| *** kgriffs is now known as kgriffs|afk | 23:06 | |
| rm_work | hey tdink_ / hockeynut / jvrbanac: the Barbican functionaltests | 23:07 |
| rm_work | are they "tempest" tests? | 23:08 |
| rm_work | like, I see they pull in stuff from tempest, but | 23:08 |
| rm_work | they are SO different from the normal tempest tests I see, that I am not sure if they actually count :P | 23:08 |
| jvrbanac | rm_work, they are not | 23:09 |
| rm_work | ok | 23:09 |
| rm_work | so they don't count for the purposes of openstack's testing requirements? | 23:09 |
| rm_work | does Barbican have any real tempest tests anywhere? | 23:09 |
| jvrbanac | rm_work, they are just barbican functional tests | 23:09 |
| openstackgerrit | Venkat Sundaram proposed openstack/barbican-specs: Add Quota support for Barbican resources https://review.openstack.org/132091 | 23:10 |
| jvrbanac | rm_work, not that I know of. Tempest tests would go in the tempest repo. But if I recall correctly, that is aimed for integration between projects | 23:10 |
| *** chlong has joined #openstack-barbican | 23:12 | |
| rm_work | jvrbanac: err, I thought tempest tests were in the project-tree now | 23:12 |
| jvrbanac | rm_work, I don't know. I haven't kept up with that stuff. hockeynut should be able to trace that down though | 23:14 |
| hockeynut | rm_work the tests in tempest repo are indeed more integration-type | 23:14 |
| rm_work | hockeynut: so there ARE barbican tests in the tempest repo? | 23:14 |
| hockeynut | rm_work the tests in our repo are our funtional tests | 23:14 |
| rm_work | I thought tempest tests were supposed to be in the project repo now | 23:14 |
| *** paul_glass has quit IRC | 23:15 | |
| hockeynut | rm_work none of our tests are in tempest repo | 23:15 |
| rm_work | err | 23:15 |
| *** crc32 has quit IRC | 23:15 | |
| rm_work | ok so | 23:15 |
| rm_work | doesn't openstack require "tempest tests"? | 23:16 |
| hockeynut | rm_work you are correct - tempest tests go in our repo. the id is that our curent functional tests are (or will become) the "tempest tests" | 23:16 |
| rm_work | hockeynut: err, so | 23:16 |
| *** SheenaG1 has quit IRC | 23:16 | |
| rm_work | hockeynut: we're trying to figure out if we can do our required "tempest tests" in the way you guys did | 23:16 |
| rm_work | or whether we have to stick closer to the style of the ones in the main tempest repo (which kinda suck) | 23:17 |
| rm_work | who actually decides if the tests we write fulfill for the openstack testing requirements? | 23:17 |
| rm_work | s/for// | 23:17 |
| hockeynut | rm_work I believe that is the TC so Sean Dague I think is the tempest guy | 23:18 |
| hockeynut | rm_work I' | 23:18 |
| rm_work | k | 23:18 |
| redrobot | rm_work soooooooooo all that stuff is in flux | 23:18 |
| hockeynut | cast in jello | 23:18 |
| rm_work | whelp | 23:18 |
| rm_work | who can we ask for some direction? >_> | 23:18 |
| rm_work | sdague? | 23:18 |
| redrobot | rm_work does your project have a QA liasion? | 23:19 |
| rm_work | no? | 23:19 |
| rm_work | I don't know? | 23:19 |
| rm_work | I don't think so? maybe? | 23:19 |
| redrobot | rm_work https://wiki.openstack.org/wiki/CrossProjectLiaisons | 23:20 |
| rm_work | sooo | 23:20 |
| hockeynut | rm_work redrobot I've spoken a bit with Matt Trenish about this. He had some takes (mostly "more tests, please") but nothing clear | 23:20 |
| rm_work | I guess Salvatore | 23:20 |
| rm_work | we're neutron but kinda not | 23:20 |
| rm_work | we aren't on that project list yet :P | 23:21 |
| redrobot | so, with the "big tent" changes coming down the pipeline, it's not clear what anything means anymore | 23:21 |
| rm_work | thanks redrobot, makes me very comfortable :P | 23:21 |
| hockeynut | rm_work please hook me up with your QA/QE person and we can discuss | 23:21 |
| rm_work | hockeynut: alright I guess that'd be fnaval (Franklin) | 23:21 |
| redrobot | eventually they'll define some sort of testing-related "tags" and at that point the requirements for the tags will clear up what it is that needs to be done | 23:21 |
| rm_work | I'll get you guys talking tomorrow | 23:21 |
| redrobot | rm_work that's my hope anyway... all that stuff is super confusing now. >_< | 23:22 |
| hockeynut | rm_work sounds good! | 23:22 |
| rm_work | kk thanks | 23:22 |
| *** lisaclark1 has quit IRC | 23:23 | |
| *** lisaclark1 has joined #openstack-barbican | 23:36 | |
| *** dimtruck is now known as zz_dimtruck | 23:39 | |
| *** lisaclark1 has quit IRC | 23:46 | |
| *** jaosorior has quit IRC | 23:46 | |
| *** lisaclark1 has joined #openstack-barbican | 23:49 | |
| *** kgriffs|afk is now known as kgriffs | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
