| openstackgerrit | Jim Baker proposed openstack/craton master: JSON Path-like querying for variables https://review.openstack.org/443941 | 01:40 |
|---|---|---|
| *** Guest4566 has joined #craton | 02:51 | |
| *** Guest4566 has quit IRC | 03:32 | |
| *** VW has joined #craton | 04:15 | |
| *** VW has quit IRC | 04:17 | |
| *** VW has joined #craton | 04:18 | |
| *** VW has quit IRC | 04:22 | |
| *** VW has joined #craton | 04:33 | |
| *** VW has quit IRC | 04:37 | |
| *** VW_ has joined #craton | 04:51 | |
| *** VW_ has quit IRC | 04:55 | |
| *** VW has joined #craton | 05:05 | |
| *** VW has quit IRC | 05:06 | |
| *** VW has joined #craton | 05:06 | |
| *** VW has quit IRC | 05:15 | |
| *** VW has joined #craton | 05:17 | |
| *** VW has quit IRC | 06:02 | |
| *** openstackgerrit has quit IRC | 08:18 | |
| *** openstackgerrit has joined #craton | 09:20 | |
| openstackgerrit | iswarya vakati proposed openstack/craton master: Replaces uuid.uuid4 with uuidutils.generate_uuid() https://review.openstack.org/444562 | 09:20 |
| openstackgerrit | iswarya vakati proposed openstack/craton master: Replaces uuid.uuid4 with uuidutils.generate_uuid() https://review.openstack.org/444562 | 10:39 |
| openstackgerrit | iswarya vakati proposed openstack/craton master: Replaces uuid.uuid4 with uuidutils.generate_uuid() https://review.openstack.org/444562 | 10:56 |
| sigmavirus | morning | 12:03 |
| tojuvone | Hi | 13:26 |
| thomasem | Hello! | 13:47 |
| thomasem | jimbaker: I think the descendant override problem is also present in your patch, fyi. | 14:01 |
| thomasem | https://gist.github.com/thomasem/ae6b65777af003d3d945ea47ec1e3dd9 | 14:02 |
| thomasem | The thing I was hung up on is how that was some how only an issue when I added the JSON path searching. | 14:03 |
| thomasem | I don't think it is. | 14:03 |
| thomasem | So, I think we ought to fix that in the parent (https://review.openstack.org/#/c/440929) | 14:04 |
| thomasem | And then make any appropriate modification in the JSON path patch, if needed. | 14:04 |
| fsaad | morning all | 14:08 |
| thomasem | o/ | 14:08 |
| fsaad | just fyi, I'm on PDT this week although I'm trying to keep up with central | 14:12 |
| fsaad | I really am on mountain time, but there hasn't been a daylight savings change here yet so puts me in PDT :/ | 14:13 |
| thomasem | Yikes... Feels like we could globally simplify this system quite a bit... | 14:13 |
| thomasem | :P | 14:15 |
| fsaad | lol | 14:19 |
| fsaad | yeah daylight savings happens here in like two weeks or something | 14:19 |
| fsaad | so picked quite a bad week to come visit family, what's one going to do... school districts chose this week for spring break :) | 14:20 |
| fsaad | things going allright ? | 14:20 |
| thomasem | Lol, c'est la vie. | 14:21 |
| thomasem | Oh, yeah. All's well. How was the weekend? | 14:21 |
| fsaad | thomasem: good, some driving on saturday and family time sunday, happily uneventful, how about yours ? | 14:25 |
| thomasem | Haha, that's always good. | 14:26 |
| thomasem | Mine was nice. Got a lot done around the house, caught up with some old friends yesterday. | 14:27 |
| fsaad | good stuff thomasem | 14:29 |
| fsaad | thomasem: I only have the tuesday invite copy, not part of the original, mind adding me? not sure if the time was also extended half hour early btw | 14:29 |
| jimbaker | thomasem, totally agreed on the cherry picking | 14:31 |
| jimbaker | and moving it back to the original patch | 14:31 |
| jimbaker | given how it is currently stacked, it seemed easier to work on the end patch however. at least until we have figured out how it all works together | 14:31 |
| jimbaker | especially if you found a counter case, as it seems you have. i believe the approach i worked out makes sense, but need to look at *all* corner cases to be sure | 14:32 |
| thomasem | jimbaker: I gotcha. I guess, to me, it felt easier layered than squashed. That's just how I think about things, though. | 14:33 |
| jimbaker | thomasem, well the json stuff introduces more nuance | 14:33 |
| jimbaker | because now the matching is done at the db level. i suppose we could use http://jsonpath-rw.readthedocs.io/en/latest/ to do json matching in python as well | 14:34 |
| thomasem | It does. The problem is now when I find a bug, I have to consider whether it's something to do with JSON path, or a problem with the original algorithm. It adds potential culprits. If we make the parent airtight, it means I can have more confidence that it's the JSON path addition causing the problem. | 14:34 |
| jimbaker | so to generalize, we have to assume some predicate matcher is done at the db level | 14:35 |
| jimbaker | anyway, i don't think we are only converging on this, is there a bug?, at this point. before the algorithm was flawed in the two implementations we had tried | 14:36 |
| thomasem | fsaad: you're talking about the Craton meeting invites? | 14:37 |
| fsaad | the tuesday one we talked about using for sprint planning yeah | 14:38 |
| fsaad | thomasem: ^ | 14:38 |
| fsaad | also, looks like we got access to a real customer installation to start trying craton out on ;) | 14:39 |
| thomasem | Oh nice | 14:39 |
| thomasem | fsaad: Ahhh, gotcha. There isn't an original. :) This was just my own I created when I started. | 14:39 |
| thomasem | unless jimbaker or someone else has an original invite that keeps track of the Craton meetings for our calendars? | 14:40 |
| thomasem | Then again, we do have that Craton Google calendar? | 14:40 |
| jimbaker | thomasem, perfectly happy to put that on the craton google cal instead | 14:41 |
| thomasem | fsaad: thoughts? I don't mind maintaining an invite for the Rackspace calendars, of course, but the Google Calendar will suit the public forum. | 14:42 |
| thomasem | possibly better? | 14:42 |
| jimbaker | btw, speaking of calendars - i will be taking vacation this thurs/fri | 14:44 |
| thomasem | Got it. I'm also going to be in training tomorrow afternoon. | 14:45 |
| jimbaker | thomasem, the key thing presumably is to publish the .ics publicly. how we do it, i don't think it matters | 14:45 |
| thomasem | jimbaker: this might fix the parent patch https://gist.github.com/thomasem/ae6b65777af003d3d945ea47ec1e3dd9#gistcomment-2025750 | 14:46 |
| thomasem | jimbaker: correct | 14:47 |
| jimbaker | thomasem, cool, will take a look | 14:48 |
| thomasem | It was adapting your fix to the parent :) | 14:48 |
| thomasem | So, hope I understood it correctly | 14:48 |
| jimbaker | also: due to double bass transport for an orchestra festival, i might be running a few minutes late to today's meeting in just over 10 min | 14:48 |
| jimbaker | biab | 14:49 |
| *** VW has joined #craton | 14:50 | |
| fsaad | google calendar works fine for me too | 14:50 |
| *** VW has quit IRC | 14:50 | |
| thomasem | cool | 14:51 |
| *** VW has joined #craton | 14:51 | |
| fsaad | no worries jimbaker | 14:51 |
| *** VW has quit IRC | 14:52 | |
| *** VW has joined #craton | 14:53 | |
| *** VW has quit IRC | 14:55 | |
| *** thomasem has quit IRC | 14:58 | |
| *** thomasem has joined #craton | 14:59 | |
| thomasem | meeting in #openstack-meeting-4 | 15:00 |
| thomasem | sigmavirus: git-harry: ^^ | 15:02 |
| thomasem | tojuvone: ^^ | 15:03 |
| thomasem | o/ | 15:39 |
| antonym | fsaad: hacked together a playbook that'll check if a server exists by mac in craton, and if not, create an entry: https://github.com/antonym/ecopoiesis-utility/blob/master/roles/discover/tasks/main.yml | 15:42 |
| antonym | any plans to have craton take care of assigning out ip addresses if a cidr is set? would be nice to create an entry and grab the next available ip | 15:43 |
| jimbaker | antonym, it was discussed last year, but set aside. but we could help here | 15:48 |
| antonym | yeah, i know that it slowly turns it into neutron, heh | 15:48 |
| fsaad | yeah was thinking about neutron being already there for that purpose right | 15:48 |
| antonym | but don't really have a way to utilize that field if i can't auto increment | 15:49 |
| antonym | otherwise it requires setting all of the configurations somewhere else before injecting them into craton | 15:49 |
| antonym | unless it was all static assigned dhcp, but bleh | 15:50 |
| jimbaker | to me, this is more of a generalization of what can be done with variables | 15:50 |
| jimbaker | what antonym then uses it for... craton wouldn't have to care | 15:50 |
| jimbaker | similar functionality in mongodb iirc, but i don't know if it's smart about the specific data element, such as an ip addr and what it means to inc | 15:51 |
| fsaad | jimbaker: you have the inventory data for the demo ? I think sulo used both aio and the osic one, don't need that right now, just real data in and show the filters again, the only thing pending from last meeting was real data | 15:51 |
| jimbaker | fsaad, i do not have the real data | 15:52 |
| fsaad | sulo: is the box you used available to the team ? | 15:52 |
| jimbaker | i know we have been circulating some gists around, etc, but i haven't been tracking | 15:52 |
| jimbaker | just need to get in a common repo, and document | 15:52 |
| sulo | fsaad: the data is the data you provided in that email | 15:52 |
| sulo | jimbaker: ^ | 15:52 |
| fsaad | yeah that I have, but asked for the box itself to re-use | 15:52 |
| antonym | yeah, just thinking if i could just say, cell1 requires cidr 10.0.1.0/24, then it could assign out until full | 15:53 |
| antonym | could also check if ip was already taken, but that might be more involved | 15:53 |
| sulo | fsaad: ah, no i use it for active dev, so it redone many times laready | 15:53 |
| jimbaker | antonym, so we have been discussing how we could use a namespace to assign such custom functionality | 15:53 |
| fsaad | sulo: ah cool, and the import script you used was tim's, plus the playbook right | 15:54 |
| jimbaker | this could be a good way of proving out this idea | 15:54 |
| sulo | fsaad: yes, correct | 15:54 |
| fsaad | sulo: good deal thanks | 15:54 |
| sulo | fsaad: i simply spin that docker and injest data into it | 15:54 |
| fsaad | jimbaker: not sure if you have a box to run the demo off of, I'll get one going we can share | 15:55 |
| jimbaker | antonym, i think it should be easy to experiment with something like that | 15:55 |
| jimbaker | fsaad, thanks, exactly what i need | 15:55 |
| fsaad | right on | 15:55 |
| antonym | jimbaker: cool | 15:56 |
| *** jovon has joined #craton | 16:49 | |
| thomasem | Anyone got a trick for getting our functional tests to only run the specified test(s)? | 16:58 |
| thomasem | Ope, might have gotten it | 17:02 |
| thomasem | Yep, got it. Nevermind! | 17:04 |
| thomasem | I was specifying the path incorrectly | 17:04 |
| thomasem | tox -e functional craton.tests.functional.test_host_calls:APIV1HostTest.test_host_get_all_vars_filter_resolved_region | 17:05 |
| thomasem | works | 17:05 |
| *** openstackgerrit has quit IRC | 17:18 | |
| fsaad | https://bugs.launchpad.net/python-cratonclient/+bug/1667109 | 17:20 |
| openstack | Launchpad bug 1667109 in Craton's Python Client "Allow environment export from client" [High,New] | 17:20 |
| fsaad | that one is still for this milestone, has there been any discussion about it ? ^ | 17:20 |
| fsaad | sigmavirus: looks like you had it briefly | 17:20 |
| sigmavirus | fsaad: not that I'm aware of | 17:21 |
| sigmavirus | With the devices endpoint in the client, I'm not sure what's needed there, to be honest | 17:21 |
| jimbaker | fsaad, sigmavirus, right - this is something we mentioned at the last demo with support | 17:25 |
| fsaad | jimbaker: is it? | 17:26 |
| fsaad | sigmavirus: with the devices endpoint we can get a list of all devices/facts for a given tenant/cloud or whatever right | 17:26 |
| jimbaker | i think we do want eventually a better import/export mechanism, possibly based on graphql | 17:26 |
| fsaad | so an ansible inventory plugin can call that and get what is needed ? | 17:26 |
| jimbaker | fsaad, correct | 17:26 |
| jimbaker | and it can every single device across every single project | 17:27 |
| jimbaker | can *get* | 17:27 |
| jimbaker | so it truly is bulk in this way | 17:27 |
| fsaad | cloudnull: around? I believe you mentioned you'd actually prefer craton team to not work on the ansible plugin, looks like the 'devices' endpoint should give the data needed | 17:27 |
| fsaad | to build said inventory module from what I can tell from the above ^ | 17:27 |
| cloudnull | i am . | 17:28 |
| * cloudnull reading | 17:28 | |
| cloudnull | fsaad: correct. | 17:29 |
| * cloudnull just sent an email back to the group toughing on that | 17:30 | |
| fsaad | cloudnull: good so looks like from a server standpoint it's done based on the answers I got, we can discuss who tackles the client side/ansible inventory | 17:30 |
| cloudnull | ++ | 17:30 |
| fsaad | unless we know you were thinking of doing it already (or have it done as it's your modus operandi :P) | 17:30 |
| cloudnull | i've not. | 17:31 |
| fsaad | it's cool, I can bring it up in the tiger team and discuss | 17:32 |
| * cloudnull has been working my many other tiger team/rpco/upstream tasks. | 17:32 | |
| fsaad | yeah understood, this will be part of that too | 17:32 |
| cloudnull | cool | 17:32 |
| fsaad | thomasem: thanks for the uc to bug# mapping | 17:33 |
| *** openstackgerrit has joined #craton | 17:36 | |
| openstackgerrit | Ian Cordasco proposed openstack/python-cratonclient master: Add Betamax for testing https://review.openstack.org/442165 | 17:36 |
| sigmavirus | ^ *could* be split up, but I'll leave that to y'all to decide | 17:36 |
| sigmavirus | A lot of the line count is in cassettes which I don't expect you to actually read/review | 17:37 |
| thomasem | fsaad: sure thing! | 17:37 |
| sigmavirus | Outside of that, it's just over 220 lines of code or so | 17:37 |
| * sigmavirus forgot to add a couple files | 17:38 | |
| * sigmavirus apologizes again for such fast-follow updates | 17:38 | |
| openstackgerrit | Ian Cordasco proposed openstack/python-cratonclient master: Add Betamax for testing https://review.openstack.org/442165 | 17:39 |
| -openstackstatus- NOTICE: restarting gerrit to address performance problems | 17:43 | |
| thomasem | YEah, gerrit's driving me nuts | 17:43 |
| sigmavirus | thomasem: heh | 17:44 |
| sigmavirus | "Working as intended' | 17:44 |
| thomasem | I meant the performance problems. | 17:44 |
| thomasem | FYI, not your fast follows... I promise I'm far noisier. | 17:44 |
| fsaad | betamax and cassettes, what is this?! | 17:45 |
| thomasem | The 80s? | 17:45 |
| fsaad | lol | 17:45 |
| thomasem | ... 70s? :D | 17:45 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: Variable search for resources now uses resolved variables. https://review.openstack.org/440929 | 18:03 |
| thomasem | Hmm, weird. functional test environment seems to be getting SQLAlchemy < 1.1.0... I need to run to an appointment, bbiab | 18:37 |
| sigmavirus | thomasem: I think I can guess why | 19:00 |
| sigmavirus | I will check | 19:00 |
| sigmavirus | yep | 19:00 |
| sigmavirus | submitting fix for you | 19:00 |
| sigmavirus | (in a separate patch) | 19:00 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Remove tox usage of upper-constraints https://review.openstack.org/445054 | 19:02 |
| sigmavirus | thomasem: you'll want to rebase on top of ^ or have jimbaker / sulo / git-harry fast approve that | 19:02 |
| sulo | just for my own understanding why arent we bumping sqlalchemy version in global ? | 19:26 |
| sulo | ah nevermind .. found this: https://review.openstack.org/#/c/423192/ | 19:34 |
| sigmavirus | sulo: yeah, that's last week's discussion =P | 19:38 |
| sigmavirus | we unsubscribed from openstack/requirements's mailing list too | 19:38 |
| sigmavirus | (automated updates of our requirements, including security bumps) | 19:38 |
| sulo | yeah, would have been nice to have like a tag to bypass one module rather than the whole thing .. but maybe that defeats the whole constraint idea | 19:40 |
| fsaad | jimbaker: ok the box has a craton container running now, no great progress since I've been back to back meetings with tiger team | 19:41 |
| fsaad | haven't put any data into it though | 19:41 |
| sigmavirus | sulo: openstack/requirements, it's all in or all out, you decide =P | 19:43 |
| jimbaker | sigmavirus, ahh, thanks for that | 19:43 |
| jimbaker | makes sense there was separate usage | 19:44 |
| sigmavirus | jimbaker: ? | 19:44 |
| jimbaker | sigmavirus, tox and its use of upper constraints | 19:44 |
| jimbaker | sulo, do you mind if we output ProjectId: ..., etc in the docker log - i think it's OK if it's just these keys, although it might be better if it were Craton-ProjectId, etc, but having to workaround spaces (or not) seems awkward for any automation | 19:47 |
| jimbaker | also there seems to be some question of nonprintable characters in the docker log output, but still looking into that | 19:47 |
| jimbaker | obvious usage of grep and awk seem to fail, that's all i know at this point | 19:47 |
| sulo | jimbaker: eh ? it should have printed it in docker log | 19:49 |
| sulo | did docker log not show project id etc ? | 19:49 |
| jimbaker | sulo, it is in the docker log | 19:49 |
| jimbaker | but i want it to be automated | 19:49 |
| sulo | jimbaker: automated ? | 19:50 |
| jimbaker | sure, anytime we output, we should be able to automate in a chain | 19:50 |
| jimbaker | just the next obvious step | 19:50 |
| sulo | yeah still not following | 19:51 |
| jimbaker | sulo, i'm just trying to revisit this sort of script: https://gist.github.com/jimbaker/fda7b48c925e2c568781bf272a6fc99b | 19:51 |
| jimbaker | so i expect i can extract OS_PROJECT_ID from say "Project Id: d4606e22-f09a-462e-af2d-cec95e477eef | 19:52 |
| jimbaker | " | 19:52 |
| sulo | jimbaker: ah gotcha, so did that not work ? | 19:52 |
| jimbaker | using something like docker logs craton-api | grep 'Project Id' | awk '{print $3}' | 19:52 |
| jimbaker | first: i rather use {print $2} | 19:53 |
| jimbaker | because sometimes it's $2, sometimes it's $3 | 19:53 |
| jimbaker | so to be consistent, just $2 by using ProjectId would be preferrable | 19:53 |
| jimbaker | (for example) | 19:54 |
| jimbaker | the next is, it doesn't quite work, and it seems to be nonprintable characters. not certain why, going to use python to explore | 19:54 |
| sulo | grep 'ID' | cut -d ':' -f2 | xarg | 19:55 |
| sulo | well "Id" | 19:55 |
| jimbaker | sulo, i don't think that invocation is quite right, but i haven't used cut for a while | 19:57 |
| sigmavirus | later y'all | 19:58 |
| sulo | sigmavirus: later | 20:00 |
| sulo | jimbaker: hmm let me check | 20:00 |
| jimbaker | sulo, here's what i'm working against: https://gist.github.com/jimbaker/fb0c845c8018408cb122eb41253819fc | 20:03 |
| sulo | jimbaker: https://gist.github.com/sulochan/097b30a77ea500c782527007118ad5c4 | 20:04 |
| jimbaker | sulo, presumably we don't want to grep on 'Id', but that's besides the point for the nonprintable question | 20:04 |
| jimbaker | anyway, try my bash function | 20:05 |
| jimbaker | get-project-id | 20:05 |
| sulo | you should be able to do "Project Id" | 20:05 |
| sulo | ok, let me try | 20:05 |
| jimbaker | and attempt to use it | 20:05 |
| jimbaker | just by itself, it seems to work | 20:05 |
| sulo | which part didnt work ? | 20:06 |
| jimbaker | looks like it's an embedded \r | 20:07 |
| jimbaker | >>> import os | 20:07 |
| jimbaker | >>> os.environ["OS_PROJECT_ID"] | 20:07 |
| jimbaker | 'd4606e22-f09a-462e-af2d-cec95e477eef\r' | 20:07 |
| jimbaker | hmmm, my bash is pretty weak... | 20:08 |
| sulo | ah interestesting | 20:08 |
| tojuvone | Hi, finally at hotel with working internet | 20:08 |
| sulo | why it has this \r ... it should habe been just unixstype \n | 20:09 |
| * sulo goes to google | 20:09 | |
| sulo | jimbaker: https://gist.github.com/sulochan/097b30a77ea500c782527007118ad5c4 | 20:11 |
| sulo | so it looks like docker is doing somethign weird with carriage return | 20:12 |
| jimbaker | sulo, yeah, something crazy | 20:12 |
| sulo | get-project-id() { docker logs craton-api | grep 'Project Id' | cut -d ':' -f2 | xargs| tr -d '\r'; } | 20:14 |
| sulo | that should work | 20:14 |
| sulo | jimbaker: ^ that works for me .. Its docker doing windoes stype carriage new line return | 20:14 |
| sulo | https://github.com/docker/docker/issues/6674 | 20:15 |
| jimbaker | sulo, got it | 20:16 |
| sulo | jimbaker: ah its tty doing this | 20:17 |
| sulo | https://github.com/docker/docker/issues/8513#issuecomment-216191236 | 20:17 |
| jimbaker | it's very interesting how messed up getting a \r in output is | 20:17 |
| sulo | yeah | 20:17 |
| sulo | anyway whe should be able to use it with | 20:18 |
| sulo | get-project-id() { docker logs craton-api | grep 'Project Id' | cut -d ':' -f2 | xargs| tr -d '\r'; } | 20:18 |
| jimbaker | totally messes up scripting, but looks fine for interactive. truly WTF | 20:18 |
| jimbaker | sulo, ok, i'm going to just suggest changing the output to ProjectId | 20:19 |
| jimbaker | to simplify scripting | 20:19 |
| sulo | how will that help | 20:19 |
| jimbaker | because awk is still nice | 20:19 |
| sulo | you'll still have exactlay the same | 20:19 |
| jimbaker | yes, i understand docker has its uses | 20:19 |
| jimbaker | issues | 20:19 |
| jimbaker | (uses?) | 20:19 |
| jimbaker | but at least we don't need to make it more complex than necessary | 20:20 |
| jimbaker | white space in keys? ;) | 20:20 |
| jimbaker | sure, but no need to make things more complex than necessary | 20:20 |
| jimbaker | i guess i really wanted to hammer that point :) | 20:21 |
| sulo | how does your awk line look like now ? | 20:21 |
| sulo | i dont mind changing it but just wonding if we are chaing on personal perference of awk vs cut or whatever | 20:22 |
| sulo | my bash is pretty weak .. so ill take anyone word on what is good | 20:22 |
| jimbaker | sure, we do use awk elsewhere in our docs | 20:24 |
| jimbaker | which is how i got the invocation | 20:24 |
| sulo | also do we need API Key to be APIKey | 20:25 |
| tojuvone | I used to do a lot of bash before, but didn't use awk | 20:26 |
| tojuvone | Now I keep it very nice way of ppinting out when use any openstack cli | 20:27 |
| jimbaker | but in a nutshell, it would be something like the following: get-username() { | 20:27 |
| jimbaker | docker logs craton-api | grep 'Username' | awk '{print $2}' | tr -d '\r' | 20:27 |
| jimbaker | } | 20:27 |
| jimbaker | sulo, yeah, we should rename it to something like | 20:28 |
| jimbaker | ProjectId, APIKey, Username | 20:28 |
| jimbaker | or possibly better | 20:28 |
| jimbaker | Craton-ProjectId | 20:28 |
| jimbaker | or maybe CratonProjectId | 20:28 |
| jimbaker | i don't care | 20:28 |
| jimbaker | just no spaces, and minimal ambiguity for automation | 20:29 |
| jimbaker | tojuvone, yep, hopefully it's something we can avoid for working with craton however | 20:29 |
| tojuvone | sure, bash is nightmare overall to make anything more complex | 20:30 |
| openstackgerrit | Sulochan Acharya proposed openstack/craton master: Adds project/user bootstrap command to dbsync https://review.openstack.org/443170 | 20:30 |
| jimbaker | sulo, i still could see it being the following fwiw: eval $(docker exec -it craton-api /craton/bin/craton-dbsync --format=env --config-file=etc/craton-api-conf.sample bootstrap) | 20:31 |
| jimbaker | --format=env could say - output OS_USERNAME, OS_PROJECT_ID, OS_PASSWORD settings | 20:31 |
| jimbaker | less grep, more power | 20:32 |
| jimbaker | :) | 20:32 |
| jimbaker | tojuvone, that's my hope in removing awk from the picture | 20:32 |
| jimbaker | but we can do this in a separate pass | 20:33 |
| jimbaker | sulo, thanks for that update | 20:33 |
| jimbaker | i think we are all set. let me check | 20:33 |
| tojuvone | I am updating the etherpad right now | 20:34 |
| jimbaker | sulo, cool, it looks reasonable to automate | 20:37 |
| jimbaker | and +2 and workflow | 20:41 |
| jimbaker | sulo, thanks! | 20:41 |
| jimbaker | the note i appended in my review should go into some docs i would think | 20:42 |
| jimbaker | tojuvone, sounds good about updating etherpad. we have been busy in prep of a demo tomorrow, but we are going to help here! | 20:42 |
| jimbaker | just a little too much right now, but we are getting through things | 20:43 |
| jimbaker | and with that, dog walking time | 20:43 |
| thomasem | Boy... that took way longer than I meant. | 20:43 |
| thomasem | sigmavirus: OHHHHHHHHHHHHH | 20:44 |
| thomasem | got it. Thanks! | 20:44 |
| thomasem | Holy scrollback, Batman. | 20:47 |
| openstackgerrit | Merged openstack/craton master: Adds project/user bootstrap command to dbsync https://review.openstack.org/443170 | 20:48 |
| thomasem | There, that gets tests passing again. | 20:48 |
| thomasem | FYI, now that that's merged, here's the deploy script I use to deal with the new bootstrap stuff: https://gist.github.com/jimbaker/e74a7b98bc60519033fd455a22163ad2#gistcomment-2023979 | 20:49 |
| tojuvone | jimbaker, Thanks, I totally understand. Just that it was "empty" whiel others "ready" | 20:50 |
| thomasem | Basically source it and use the environment variables from the grep'ed docker logs | 20:50 |
| thomasem | Ugh... and this is why I dislike using SQLite in tests... AttributeError: 'SQLiteTypeCompiler' object has no attribute 'visit_JSON' | 20:51 |
| tojuvone | Btw, Ops stuff is on thursday. So there is time. | 20:52 |
| tojuvone | I could see the demo and prepare to show similar, taking it uses the latest stuff from git | 20:55 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: Move to MySQL 5.7 and SQLAlchemy>=1.1.0 https://review.openstack.org/443186 | 21:05 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: Variable search for resources now uses resolved variables. https://review.openstack.org/440929 | 21:05 |
| jimbaker | thomasem, yep, you are certainly conservative in what you grep for! | 21:23 |
| jimbaker | anyway, i think the use of docker exec is probably a nicer way of doing this. i'm going to propose --format=env so it can be done without any such grepping - i'm a big fan of eval for CLI stuff | 21:24 |
| jimbaker | ssh does this aspect well for example | 21:24 |
| jimbaker | tojuvone, good to know about thurs, we will be ready! | 21:25 |
| openstackgerrit | Merged openstack/craton master: Remove tox usage of upper-constraints https://review.openstack.org/445054 | 21:25 |
| jimbaker | nice, glad to see that change come in. thanks sigmavirus for that! | 21:26 |
| *** jovon has quit IRC | 21:32 | |
| tojuvone | Catch you tomorrow, buona notte! | 21:35 |
| thomasem | jimbaker: saw your recheck on the MySQL 5.7 patch... all unit DB tests are going to fail until we either forklift them into functional tests, or sort out what we want to do with SQLite. | 21:45 |
| thomasem | The rest of the tests seem to pass fine. It's just because the DB unit tests use SQLite. | 21:46 |
| thomasem | I don't think it'd take much work to forklift them into functional tests so they can use a real DB. | 21:48 |
| thomasem | The moment you diverge from the most basic functionality SQLite all of a sudden becomes just a PITA to work around. | 21:48 |
| jimbaker | thomasem, got it | 21:53 |
| jimbaker | yeah, if it's a question of sqlite, obviously we don't care about that platform specifically | 21:53 |
| thomasem | Right. And I don't want our implementation to have to care about it for the sake of tests, either. | 21:55 |
| jimbaker | given that the main thing we get in tox -e functional is mysql, the forklift is the only thing that makes sense | 21:56 |
| thomasem | Agreed. That's 158 tests, so. | 21:56 |
| jimbaker | until we do, we are going to have to disable | 21:56 |
| jimbaker | and selectively re-enable | 21:56 |
| thomasem | Eh, I think I can move them over in this patch. Is that a problem? | 21:56 |
| jimbaker | thomasem, well, more work | 21:57 |
| jimbaker | but if you are good for that, i am | 21:57 |
| jimbaker | the only other thing i could think of was having some sort of env var setting say - here's my mysql setup, use it | 21:58 |
| thomasem | I only worry about the surface that's tested by functional tests. If those make up for it, I don't mind disabling. | 21:58 |
| jimbaker | the unit db tests do cover some things not seen in the functional tests | 21:58 |
| jimbaker | pretty minimal, and mostly stuff that has not been exposed to the api. one thing iirc - blame | 21:59 |
| thomasem | How about I take half a day to a day max, and if I run into too many snags and it will take longer, we just disable for a short time? | 21:59 |
| jimbaker | over time, we will see more of this exposed | 21:59 |
| jimbaker | because i fully expect that blame will be exposed through the vars rest api | 21:59 |
| thomasem | Gotcha | 21:59 |
| thomasem | Yeah, I don't doubt we need to move these into functional tests, just a matter of when. | 22:00 |
| jimbaker | it will be an option to emit more interesting data. and you basically can almost handle in the var side of things you dif | 22:00 |
| jimbaker | did | 22:00 |
| jimbaker | we will see similar things happen when we get namespaces, secrets, rbac, etc. hopefully most of it just uses the existing api, with minor to no extensions | 22:01 |
| thomasem | Gotcha. So, are you cool w/ me time boxing to half a day or so? | 22:01 |
| jimbaker | thomasem, yeah. +1 | 22:01 |
| thomasem | And if it's going to take longer, we disable and punt? | 22:01 |
| thomasem | Cool. | 22:01 |
| jimbaker | a very good plan | 22:01 |
| thomasem | Excellent | 22:01 |
| thomasem | That'll be my morning coffee task. | 22:01 |
| jimbaker | :) | 22:01 |
| thomasem | Hopefully it'll just be a matter of exposing the craton-functionl-test-api's mysql port. | 22:02 |
| thomasem | And pointing the tests at it | 22:02 |
| thomasem | Or use the same container management to set up a clean mysql container. | 22:03 |
| thomasem | Just depends on how decoupled things are. I'd prefer the latter option for things that don't need Craton API. | 22:04 |
| thomasem | Anyway, I just heard the dinner bell and don't want to annoy the cook. Have the loveliest of evenings! | 22:04 |
| jimbaker | i prefer the latter option as well. have a great evening! | 22:07 |
| fsaad | laters thomasem have a good evening | 22:13 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!