| *** Syed__ has quit IRC | 00:35 | |
| jimbaker | antonym, let me double check that. hopefully we got the logic right here! | 00:40 |
|---|---|---|
| antonym | yeah, i tried a call without the project_id and it went into the first project created, i'm guessing that call should fail if you don't specify username and project_id | 02:33 |
| jimbaker | antonym, so this bug was filed earlier, but i have bumped its priority: https://bugs.launchpad.net/craton/+bug/1666695 | 04:41 |
| openstack | Launchpad bug 1666695 in craton "project ID not honored in requests" [Critical,Confirmed] | 04:41 |
| jimbaker | i also took a bit of time updating all my wrapper functions for craton, in light of the new dbsync bootstrap stuff: https://gist.github.com/jimbaker/9089e381af38d9db6bc4af59baed6d47 | 04:42 |
| jimbaker | now with convenience support for starting up docker, setting env vars, etc | 04:43 |
| jimbaker | i guess i'm learning something about bash programming... | 04:44 |
| *** openstackgerrit has quit IRC | 10:18 | |
| *** openstackgerrit has joined #craton | 10:44 | |
| openstackgerrit | lidong proposed openstack/craton master: Correct some typo errors https://review.openstack.org/445872 | 10:44 |
| openstackgerrit | Merged openstack/craton master: Correct some typo errors https://review.openstack.org/445872 | 11:16 |
| openstackgerrit | Ian Cordasco proposed openstack/python-cratonclient master: Add integration tests for clouds client https://review.openstack.org/445909 | 12:27 |
| sigmavirus | (morning btw) | 12:46 |
| sigmavirus | (that should have been proposed yesterday but I forgot to tell it that I did want to submit everything) | 12:48 |
| sigmavirus | thomasem: do you want to "pair" on the launchpad stuff today? Figure it's probably a good idea to do some knowledge transfer on this topic | 13:27 |
| thomasem | sigmavirus: sure! | 13:37 |
| fsaad | morning guys | 13:39 |
| thomasem | hey hey | 13:40 |
| sigmavirus | morning fsaad | 13:40 |
| fsaad | hi sigmavirus | 13:40 |
| antonym | jimbaker: cool, thanks | 13:40 |
| thomasem | FYI, I think jimbaker is out today. Also, they removed the cap on SQLAlchemy: https://review.openstack.org/#/c/423192/ | 13:41 |
| fsaad | thanks for reminder thomasem | 13:41 |
| thomasem | You bet | 13:42 |
| sulo | thomasem: ah thats nice, that means we can go back to syncing with global | 13:43 |
| sulo | and o/ all | 13:43 |
| thomasem | Yep | 13:44 |
| thomasem | brb | 13:50 |
| sulo | cloudnull: so shared devices re: http://cdn.pasteraw.com/90abpeqkkw8sa0zxaa1vj3jm03ld5rl | 13:51 |
| sulo | have a few questions when you are around | 13:51 |
| cloudnull | o/ | 13:51 |
| cloudnull | what's up ? | 13:51 |
| sulo | so these shared hosts .. how do they inherit properties | 13:52 |
| sulo | when you do some deploy it is takign cell1's properties | 13:52 |
| sulo | or cell2 | 13:52 |
| cloudnull | we've never used cells. | 13:52 |
| sulo | how is it shared right now ... on region level ? | 13:53 |
| cloudnull | I just added Cell.* to keep within the craton spec | 13:53 |
| sulo | gotcha | 13:53 |
| cloudnull | it's regional, yes | 13:53 |
| cloudnull | we'd do something like a keystone between two regions | 13:54 |
| cloudnull | or a db, | 13:54 |
| * cloudnull over simplifying | 13:54 | |
| sulo | so mostly top level | 13:54 |
| sulo | so when you run some playbook on this shared keystone | 13:54 |
| sulo | i am guessing we've never had to worry about if it takes some property of one or the other region its shared between | 13:55 |
| cloudnull | no, not to my knowledge | 13:56 |
| sulo | how about shared containers | 13:56 |
| cloudnull | keystone is deployed within a container | 13:57 |
| sulo | gotcha, there is nothing regin specific in these deploys ? | 13:57 |
| sulo | not even ip blocks etc ? | 13:57 |
| cloudnull | there wasn't | 13:58 |
| cloudnull | we can define anything that was already consumed using the "used_ips" interface | 13:59 |
| sulo | ok, thx | 13:59 |
| cloudnull | the biggest thing I'd say craton should be concerned with as it pertains to shared hosts would be shared devices | 14:00 |
| cloudnull | like an f5 spanning two regions | 14:00 |
| cloudnull | or a switch | 14:00 |
| cloudnull | we can circle back on shared hosts and such | 14:00 |
| sulo | yeah, i mean the problem comes where you have a shared device with parent(s) | 14:01 |
| sulo | top level stuff is probably a bit easier to handle | 14:01 |
| sulo | mostly because of how variablrs are handled in craton | 14:01 |
| sulo | in terms of child device or cell etc inheriting parent variables | 14:01 |
| *** VW has joined #craton | 14:09 | |
| sulo | git-harry: ^ this might also help us decide on why we might want to keep network-devices around .. but separate it from devices maybe so its not constrained to a region | 14:21 |
| fsaad | did we really take 90mins on yesterday's call? do we want to schedule that much time or 1 hour does it for the every tuesday window ? | 14:40 |
| fsaad | I'll do 90 and we can always cut short if necessary | 14:43 |
| sulo | fsaad: i vote for just 1 .. i dont think more time is any useful here | 14:44 |
| sigmavirus | fsaad: we didn't *need* 90minutes yesterday, we just used it because it was available to us | 14:45 |
| sigmavirus | I honestly think a retro can run in 30minutes | 14:45 |
| sulo | sigmavirus: +1 | 14:45 |
| fsaad | allright let's do 60 mins | 14:45 |
| sigmavirus | why not 30? | 14:46 |
| fsaad | I think once we have a good cadence and get down to nitty gritty on those calls we can definitely do 30 or even less | 14:51 |
| fsaad | the first few I expect to have more open conversation like yesterday, but if y'all think we can get started with 30 we can give that a try | 14:51 |
| fsaad | I don't have anything after so can run over if necessary | 14:51 |
| jimbaker | thomasem, i'm actually here today | 14:56 |
| jimbaker | but i will be out tomorrow & friday on PTO | 14:56 |
| thomasem | jimbaker: gotcha | 14:56 |
| jimbaker | sulo, thomasem, so we need to fix this bug that antonym is running against: https://bugs.launchpad.net/craton/+bug/1666695 | 14:56 |
| openstack | Launchpad bug 1666695 in craton "project ID not honored in requests" [Critical,Confirmed] | 14:56 |
| fsaad | jimbaker: didn't end up taking the half today ? | 14:57 |
| jimbaker | in a nutshell, this means that we cannot create new users in new projects | 14:57 |
| jimbaker | fsaad, i will do that some other time :) | 14:57 |
| fsaad | boo, allright | 14:57 |
| sulo | jimbaker: ah yikes .... but is a simple bug to fix | 14:58 |
| jimbaker | i'm going to be in aspen tomorrow/friday to ski with a college friend, so that's something i cannot push back. but enough outstanding stuff | 14:58 |
| jimbaker | sulo, yeah, that's what i expected | 14:59 |
| antonym | jimbaker: should just be able to go back to the sql creates for now now that i know what was missing, but yeah, i'll switch over once that bug is squashed | 14:59 |
| antonym | jimbaker: that sounds fun | 14:59 |
| jimbaker | antonym, it's my colorado life :) | 14:59 |
| sulo | jimbaker: iam unassigning you from this bug https://bugs.launchpad.net/craton/+bug/1662574 | 14:59 |
| openstack | Launchpad bug 1662574 in craton "Support device sharing between clouds, regions, and cells" [High,New] - Assigned to Jim Baker (jimbaker) | 14:59 |
| antonym | nice | 14:59 |
| sulo | unless you have already stated it | 15:00 |
| jimbaker | sulo, yeah, that's definitely a good thing to do - let me check if i'm squatting on anything else | 15:00 |
| jimbaker | indeed i am. will unassign | 15:01 |
| sulo | i will take that, but first let me take a crack at this user create issue | 15:01 |
| jimbaker | sulo, so device sharing is interesting. did you have a chance to take look at my comments in https://etherpad.openstack.org/p/craton-sprint-planning line 40 and on? | 15:05 |
| sulo | looking | 15:06 |
| sulo | jimbaker: so .. there are a few parts to this .. see my chat with cloudnull above, i think the first step is to allow top level devices like switch, router to be in multiple resources ... it seems that the use case that is | 15:08 |
| sulo | being looked for .. second is device sharing .. and while your approach looks promising to me .. i feel like a bit more info on what the users are looking to do with this is needed | 15:08 |
| jimbaker | right, and this makes perfect sense given that layout | 15:08 |
| *** Syed__ has joined #craton | 15:09 | |
| jimbaker | agreed about more user info | 15:09 |
| jimbaker | i sketched out the three alternatives mostly to be able to ask that question | 15:09 |
| sulo | jimbaker: rgr, good points to start | 15:09 |
| jimbaker | vs building something that would not make sense for real usage | 15:09 |
| sulo | jimbaker: so i am planning to spec this .. ill start with your points | 15:09 |
| sulo | but will find the use case etc first before doing anything | 15:10 |
| jimbaker | +1 | 15:10 |
| sulo | from what i understand most of the use cases have no parent to inherit from | 15:10 |
| sulo | example f5 or keystone for two regions etc | 15:10 |
| jimbaker | right, and it's interesting it's *keystone* - a service, not a physical device per se | 15:11 |
| sulo | well, inside a container | 15:12 |
| sulo | the container could be shared from what i understand | 15:12 |
| sulo | but has nothing in common to the region or cell .. no properties inherited | 15:12 |
| jimbaker | i see. so model as being in some "shared service region" | 15:13 |
| jimbaker | i assume that's what we do. although if it's keystone, i would assume some standard HA setup could also happen for such a shared service | 15:15 |
| *** VW has quit IRC | 15:15 | |
| jimbaker | just trying to figure out what exactly we are trying to model here, that's all | 15:16 |
| jimbaker | sulo, so here's a question: is the sharing of resource the same as the management of the resource? | 15:17 |
| jimbaker | the fact that someone stands up a HA keystone, with galera, etc - this seems different than i'm using keystone | 15:18 |
| sulo | if mgt = confiuratino also then looks like not .. but need more info | 15:19 |
| sulo | jimbaker: but what i see so far is shared is more of service level thing | 15:20 |
| sulo | except for network devices .. which makes sense | 15:20 |
| jimbaker | exactly | 15:20 |
| thomasem | fsaad: Does that invite you sent have the timezone set? | 15:33 |
| thomasem | I usually see a little note on it when it does, but that could be because it was set to my timezone, too. | 15:34 |
| fsaad | exchange should take care of that tmk, so yeah should have that for you, can someone else confirm plz? | 15:39 |
| fsaad | I used the same start time as our meeting yesterday, with 60 min length | 15:40 |
| fsaad | (vs 90 it took us yesterday) | 15:40 |
| thomasem | sulo: ^^ | 15:52 |
| sulo | checking | 15:52 |
| thomasem | Thank you! | 15:52 |
| sulo | i got 60 mins | 15:52 |
| jimbaker | fsaad, also checked. but any reason why we are moving the meeting time earlier than before? | 15:53 |
| sulo | yeah it starts 3:30 GMT | 15:53 |
| sulo | vs 4:00 GMT | 15:53 |
| jimbaker | it was pretty difficult to settle on this time, so i think we should keep the old start time | 15:53 |
| jimbaker | also the tues and thurs start times have been aligned | 15:54 |
| fsaad | wait is the time this invite start not the exact same as we did yesterday ad-hoc ? | 15:56 |
| fsaad | I can move, just need to confirm that, I was aiming to match what we did yesterday (minus the last 30 mins) | 15:56 |
| jimbaker | fsaad, but we ad hoc decided to make it 90 minutes by extending 30 min earlier | 15:56 |
| fsaad | oh ok cool, I just matched yesterday cause I had no other invite | 15:57 |
| jimbaker | fsaad, ahh, i need to add you to the invite i have | 15:57 |
| fsaad | and seemed all wanted that yesterday. Cool, so want me to keep the length and push 30 mins later ? | 15:57 |
| jimbaker | i actually have a recurring calendar item | 15:57 |
| jimbaker | everyone else is on this. except you :) | 15:57 |
| fsaad | hah, ok let's use that and I can just drop this one | 15:57 |
| *** VW has joined #craton | 15:58 | |
| fsaad | jimbaker: plz add tim and dusty if they're not there | 15:58 |
| jimbaker | yeah, there are a couple of people to add | 15:58 |
| fsaad | thanks! | 15:58 |
| jimbaker | will also add michael porras | 15:59 |
| fsaad | sweet, thanks | 15:59 |
| sulo | did i send my ooo as free ? | 15:59 |
| jimbaker | i will also update our irc topic and the meeting etherpad to correspond with these decisions. and i will update how we track these things so anyone can do as needed in the future - just the collection of resources/process | 16:04 |
| fsaad | sulo: yes sir looks like it | 16:04 |
| sulo | fsaad: thanks | 16:04 |
| fsaad | yep | 16:05 |
| * fsaad wonders about the signifficance of sulo's PTO date | 16:06 | |
| sulo | nothing from my side :) | 16:07 |
| *** VW has quit IRC | 16:07 | |
| *** VW has joined #craton | 16:08 | |
| fsaad | hehe | 16:09 |
| sulo | :) yeah i've been mostly looking holidays every year becasue i dont take them during the year and they pile up in december | 16:10 |
| sulo | so making a point to take all my days off this year | 16:10 |
| jimbaker | plan to be doing the same | 16:11 |
| jimbaker | more even distribution of relaxation/recharging, will do me good | 16:11 |
| *** VW has quit IRC | 16:12 | |
| jimbaker | thomasem, btw, here are the new versions of the wrappers i put together - https://gist.github.com/jimbaker/9089e381af38d9db6bc4af59baed6d47 | 16:14 |
| thomasem | jimbaker: thanks | 16:14 |
| jimbaker | it's not quite perfect because of the hard-coded dependency on the location of the conf file. maybe replace with an env var, better bet, replace with some of sort of temp file usage | 16:15 |
| jimbaker | but that's only for the direct case, which i find to be handy when i'm hacking on sql stuff. or in general because i want a server i can relaunch immediately and not pay the docker overhead | 16:15 |
| thomasem | buhh, another bug in our resolved vars search patch | 16:17 |
| jimbaker | sulo, you might find them of use as well. note this command: eval $(craton-docker-env) - introspects docker logs for the env vars, and sets them. handy | 16:17 |
| thomasem | I guess we solved this one in the JSON path patch, but not here... might need to do the approximate_json there instead. | 16:18 |
| thomasem | It's not handling values other than strings. | 16:18 |
| jimbaker | thomasem, ahh, yes | 16:18 |
| jimbaker | so are you planning on pushing that functionality down (or is it up?) into the parent? | 16:19 |
| thomasem | And since I don't think I can easily take the existing query for unresolved, we'll need to break apart the resolved code path a bit to make it usable for both scenarios. | 16:19 |
| thomasem | up | 16:19 |
| thomasem | into the parent | 16:20 |
| jimbaker | thomasem, but this is just about patch management? or actual functionality? | 16:20 |
| jimbaker | they are related patches. but one handles resolution and conjunction; the other, better matching | 16:21 |
| jimbaker | so if we just say the json patch is really the "better matching" patch, i think that it's fine | 16:21 |
| jimbaker | at least if i understand the problem correctly | 16:22 |
| *** valw has joined #craton | 16:23 | |
| thomasem | It's trivial where I put the fix... I'd prefer start with a thing that works well before adding on more complexity. | 16:23 |
| thomasem | It's a bug in master, btw. | 16:23 |
| jimbaker | thomasem, yes, it's in master - we can only match strings | 16:24 |
| thomasem | Okay | 16:24 |
| thomasem | Alright, that's cool. I'll just fix that in JSON path patch, then | 16:25 |
| thomasem | and test with strings for now. | 16:25 |
| thomasem | for this parent patch | 16:25 |
| thomasem | It just keeps getting in my way when I'm testing, is all. | 16:25 |
| thomasem | But, I'll just use strings and stop complaining. :) | 16:25 |
| jimbaker | makes sense - the parent patch only is solving a more abstract problem, in terms of set stuff and all | 16:26 |
| thomasem | Yeah. The conjunctive + resolution | 16:26 |
| jimbaker | so it's pretty irrelevant to it what the primitive matches are | 16:26 |
| jimbaker | cool, glad the road is simplified! | 16:26 |
| thomasem | heh, if it can so be called. :P | 16:26 |
| jimbaker | well, maybe the road has potholes | 16:27 |
| thomasem | Here and there. Dancing around them. | 16:27 |
| jimbaker | and robbers. i don't know | 16:27 |
| thomasem | LOL | 16:27 |
| jimbaker | goblins and trolls preying on the unaware | 16:27 |
| thomasem | Bringing my torch and decoy pouch of gold. | 16:28 |
| jimbaker | i believe gandalf has a good saying or two | 16:28 |
| thomasem | fsaad: I don't anticipate having the nested search code ready as soon as originally thought. The branch can be pulled and used, but it's going to be missing the ability to search for unresolved variables until I can get this broken up in the parent patch and refactor the json path patch accordingly. | 16:31 |
| thomasem | I thought it was going to be a simple addition to the original query to get unresolved working, but I was wrong. | 16:31 |
| thomasem | jimbaker: can I get your invite for the Craton meeting(s)? I don't have them. | 16:32 |
| jimbaker | thomasem, you too will be added to the list | 16:32 |
| thomasem | Thanks! | 16:32 |
| jimbaker | just a comprehensive combing of loose ends here! | 16:33 |
| thomasem | lol yep | 16:34 |
| jimbaker | really a metaphor for this project i guess. i'm just glad that once some of this infra is solid, we can just keep on re-using | 16:34 |
| jimbaker | or at least that's the theory. but it feels well founded. variables in particular, we plan to use everywhere, for all sorts of capabilities | 16:36 |
| *** valw has quit IRC | 16:41 | |
| *** valw has joined #craton | 16:46 | |
| fsaad | thanks for hedas up thomasem | 16:49 |
| sigmavirus | thomasem: send me a cal invite to do the lp stuff? | 16:56 |
| sigmavirus | I don't need to do all of it with you, but just want to go over it with you | 16:56 |
| fsaad | jimbaker: the missing functionality (unresolved variables) shouldn't show-stop the demo since it's only pending the import and filtering of imported data right? we don't need to re-do the unresolved vars demo piece | 16:56 |
| jimbaker | fsaad, i don't believe it should stop the demo | 16:57 |
| fsaad | ok we'll discuss monday when you're back then, focus on closing the week for now ;) | 16:57 |
| jimbaker | fsaad, thanks! | 16:57 |
| fsaad | yup | 16:57 |
| thomasem | sigmavirus: sent! | 17:05 |
| thomasem | sigmavirus: What about 2:30? | 17:07 |
| sigmavirus | wfm | 17:07 |
| thomasem | I have an appointment at 1 and may have traffic getting back | 17:08 |
| thomasem | Excellent! | 17:08 |
| sigmavirus | heh | 17:08 |
| fsaad | the etherpad from yesterday has LP #'s for the ones I know about, but missing some that I believe need to be created (since they even need a spec) | 17:09 |
| tojuvone | Evening | 17:09 |
| fsaad | I'm not able to create a new milestone btw, so I think jimbaker can do that at some point, plz lemme know if I can help with any lp crud I can get out of y'alls way. | 17:10 |
| fsaad | hi tojuvone | 17:10 |
| jimbaker | fsaad, i will add you to the list | 17:10 |
| jimbaker | and get this milestone added as well | 17:10 |
| fsaad | jimbaker: good deal thanks! | 17:10 |
| thomasem | Alright, think I fixed it. | 17:13 |
| thomasem | Basically, don't get_descendants if resolved=False. So, return empty list | 17:13 |
| thomasem | Then it'll only include the matching variables for the resource in question | 17:13 |
| tojuvone | First ops day behind | 17:15 |
| tojuvone | Config mgmt session today where mentioned Craton, but tomorrow the 2 main sessions | 17:16 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: Variable search for resources now uses resolved variables. https://review.openstack.org/440929 | 17:17 |
| * sigmavirus was writing tests and found that the client doesn't support /v1/network-devices | 17:18 | |
| sigmavirus | sad | 17:18 |
| thomasem | saddyface | 17:19 |
| sigmavirus | I guess I'll fix it at some point | 17:19 |
| jimbaker | fsaad, here's the milestone https://launchpad.net/craton/+milestone/v0.2.0 | 17:20 |
| jimbaker | presumably the last one we will create in launchpad. to be seen, i will check in with the migration to storyboard beginning of next week | 17:20 |
| sigmavirus | jimbaker: was going to go over that work with thomasem =/ | 17:21 |
| sigmavirus | can thomasem how to do it in python-cratonclient | 17:21 |
| sigmavirus | *show | 17:21 |
| thomasem | lol :) sounds good | 17:21 |
| jimbaker | tojuvone, hopefully milan is going well. i was planning to work on etherpads, but other stuff. now there's no more time, so will help out. also feel free to chat about any ideas that are coming up | 17:24 |
| tojuvone | jimbaker, thanks, there is not much activity in etherpad | 17:25 |
| tojuvone | I put there like topic of "what tools do you use" | 17:26 |
| tojuvone | as normal session your would talk about different users experiences | 17:26 |
| tojuvone | but in the other hand I think that was already today with config mgmt session | 17:26 |
| tojuvone | and they will just have that ansbile, puppet, chef,... but nothing like this | 17:28 |
| tojuvone | This is something written there currently: Retirement of hypervisors at end of life | 17:28 |
| tojuvone | but that is inventory with some nice filtering from Craton hosts | 17:29 |
| tojuvone | there you have the hosts | 17:29 |
| tojuvone | and then workflow to do it | 17:30 |
| tojuvone | comments? | 17:30 |
| tojuvone | I have also looked to show craton API for similar | 17:31 |
| tojuvone | I mean devices | 17:32 |
| tojuvone | relations to parent | 17:32 |
| tojuvone | getting by cell, region | 17:32 |
| tojuvone | own variable | 17:32 |
| jimbaker | ok, sounds good | 17:33 |
| tojuvone | If there will not come things people want to discuss in common about inventory and fleet management | 17:34 |
| tojuvone | I should go trough more the Craton | 17:34 |
| tojuvone | and let them then come up with questions | 17:34 |
| jimbaker | one key thing: being able to show things in a stateful fashion, such as the maintenance we have discussed | 17:34 |
| jimbaker | also the specific workflow steps that are in the workflow engine spec might be useful here | 17:35 |
| tojuvone | yes, that is nice one to show as example | 17:35 |
| jimbaker | another thing that came up recently is specific support for ip address assignment that antonym asked us in a potential feature request | 17:36 |
| tojuvone | shortly said about workflow in todays conifg mgmt | 17:36 |
| jimbaker | it's an interesting variant on the state machine we talked about encoding in a namespace for maintenance | 17:37 |
| tojuvone | You mean that assign in craton and it will automatically go in place | 17:37 |
| jimbaker | effectively, it's an authoritative (so ACID compliant) smart counter (knows the semantics of ip address addition) that says the next ip is this; and better yet, can be assigned to another variable | 17:38 |
| jimbaker | so as i understand antonym's ask - i could assign a range of IP addresses at some logical level, then make claims against it | 17:39 |
| tojuvone | ok | 17:39 |
| jimbaker | yes, there are all sorts of tooling to do this. but maybe it's nice if we can just get a simple bit of functionality like this in the cmdb itself | 17:40 |
| fsaad | jimbaker: thanks! | 17:40 |
| antonym | yea, example of use case is say i want new servers booting up to auto inventory and assign themselves a static address for next boot, i might have two cidrs, one for out of band management, and management ips | 17:40 |
| antonym | i'd like it to auto assign ips as they are registering records based on something unique like a mac addy | 17:40 |
| tojuvone | so the maintenance was currently expecting something like: craton host-set-vars compute.host.example.com /host_details | 17:40 |
| fsaad | I did get the administrator permission granted email | 17:40 |
| tojuvone | craton host-set-vars compute.host.example.com /host_details/maintenance_state="PLANNED_MAINTENANCE" | 17:40 |
| tojuvone | + some other vars | 17:41 |
| jimbaker | antonym, right, in principle this makes sense, and i think we can add through some sort of plugin that is based on namespace | 17:42 |
| antonym | yeah, it just takes out the needing a config file to config craton with your inventory, it'd be awesome if it could auto populate | 17:42 |
| jimbaker | antonym, just want to sketch out the exact flow so we can implement. would be highly motivating for the namespace support that tojuvone also needs | 17:43 |
| jimbaker | but i think we can do this in such way that leverages everything that is already been built out | 17:44 |
| jimbaker | namespace support basically becomes like the fuse filesystem | 17:44 |
| tojuvone | What should I advertise the next steps: workflow, inventory, namespace, notification, rbac | 17:45 |
| jimbaker | can layer on custom semantics on top of the existing key-value support in variables | 17:45 |
| jimbaker | tojuvone, immediate next steps are described here: https://etherpad.openstack.org/p/craton-sprint-planning | 17:46 |
| jimbaker | but this is too low level i think | 17:46 |
| jimbaker | let me add on another doc i have to that to help motivate | 17:46 |
| tojuvone | ok | 17:47 |
| * tojuvone finished sushi plate | 17:49 | |
| tojuvone | Wrong country, but fast food :) | 17:49 |
| jimbaker | tojuvone, i'm pretty sure sushi is to be found in most places in the world today. i have eaten it in brazil, south africa, taiwan, ... | 17:50 |
| tojuvone | jimbaker, yeah, but should take a pizza or pasta ;) | 17:51 |
| jimbaker | that would be more typically italian, i do agree | 17:52 |
| jimbaker | it's not really milan per se, but i'm sure one can find friulian cuisine there (friuli is next to austria/slovenia in italy. love the food from that area!) | 17:55 |
| tojuvone | jimbaker, yes, maybe tomorrow can relax and have a longer supper with some local food | 17:57 |
| tojuvone | Just that I am going to be hit by air strike on friday | 17:57 |
| tojuvone | I think it should fly, but there is nobody to empty bags | 17:58 |
| tojuvone | so it should do for me | 17:58 |
| tojuvone | if I rush I'd might change to late thursday and forget the good night sleep | 17:59 |
| jimbaker | which might be a good idea, considering that travel interruptions generally are so unpleasant | 17:59 |
| tojuvone | yep, on friday planning to take taxi home and head straight to play floorball | 18:00 |
| tojuvone | back to presenting Craton | 18:01 |
| tojuvone | I think one can get very good points whay and what it is from the Barcelona video | 18:02 |
| tojuvone | s/whay/why/ | 18:02 |
| tojuvone | To say some words, if people do not know. | 18:03 |
| tojuvone | jimbaker, sulo You did a very nice show there | 18:04 |
| jimbaker | tojuvone, thanks! | 18:13 |
| *** valw has quit IRC | 19:11 | |
| openstackgerrit | Ian Cordasco proposed openstack/python-cratonclient master: Add integration tests for regions https://review.openstack.org/446134 | 19:23 |
| thomasem | sigmavirus: just got back, need a couple mins! | 19:29 |
| sigmavirus | Okay :) | 19:29 |
| sigmavirus | I'll hop into the video room | 19:29 |
| thomasem | sulo: Would you be able to add me to craton-drivers in LP? :D | 19:33 |
| *** valw has joined #craton | 19:44 | |
| *** chrisspencer has quit IRC | 19:45 | |
| fsaad | thomasem: think you mentioned mysql 5.7 is needed for the json search correct? asking cause pwnall138 will be loading inventory and attempting filtering when that's ready, so wanted to give heads up | 20:50 |
| thomasem | fsaad: that's correct | 20:51 |
| fsaad | o/ thanks, is a separate install necessary or from ubuntu repo packages ? | 20:51 |
| thomasem | fsaad: this is how I did it: https://review.openstack.org/#/c/443186/9/Dockerfile | 20:54 |
| *** valw has quit IRC | 20:54 | |
| thomasem | separate package, I believe. | 20:54 |
| fsaad | thomasem: thanks for the pointer! | 20:54 |
| thomasem | Sure thing! | 20:54 |
| *** valw has joined #craton | 20:54 | |
| *** valw has quit IRC | 20:59 | |
| jimbaker | yeah, hopefully mariadb 10 will catch up on the json support, but at least we can use mysql 5.7 for now | 21:07 |
| thomasem | That's a good point | 21:07 |
| thomasem | Yeah, I didn't realize that implication entirely. | 21:07 |
| thomasem | Hmmmmm | 21:08 |
| thomasem | PostgreSQL? | 21:08 |
| thomasem | ... doubt it. | 21:08 |
| jimbaker | jsonpath is pretty standardized, but there are a few things we may need to consider | 21:09 |
| jimbaker | i don't think it's a big stretch to support pgsql and most recent mysql | 21:09 |
| thomasem | does pgsql fit the bill? | 21:09 |
| thomasem | for CDC? | 21:09 |
| jimbaker | thomasem, i'm thinking upstream | 21:09 |
| jimbaker | for rackspace, it's mysql as far as i know | 21:10 |
| thomasem | Right :\ | 21:10 |
| thomasem | yeah | 21:10 |
| jimbaker | pgsql is great. except for management and clustering. other than minor things like that | 21:10 |
| thomasem | mariadb has the support, it's just not released as stable. | 21:11 |
| thomasem | afaik. | 21:11 |
| jimbaker | yeah, they are working on it. everyone knows json support is important | 21:11 |
| jimbaker | so we are very much mainstream, as we have discussed before | 21:11 |
| thomasem | https://mariadb.com/kb/en/mariadb/json-functions/ | 21:11 |
| thomasem | It's got the two we're using | 21:12 |
| jimbaker | yep. so it will be released soon enough | 21:12 |
| * thomasem crosses fingers | 21:12 | |
| jimbaker | https://mariadb.com/kb/en/mariadb/what-is-mariadb-102/ | 21:14 |
| jimbaker | i see we are now in RC for mariadb 10.2 | 21:14 |
| thomasem | Excellent | 21:14 |
| jimbaker | let me check the other mysql clone i have been tracking - https://github.com/pingcap/tidb/blob/master/docs/ROADMAP.md | 21:16 |
| jimbaker | json support is still in the works | 21:16 |
| jimbaker | anyway, if they get that out, and it really works, that will be very nice | 21:16 |
| jimbaker | i would like a F1 clone | 21:16 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: Move to MySQL 5.7 and SQLAlchemy>=1.1.0 https://review.openstack.org/443186 | 21:16 |
| jimbaker | https://research.google.com/pubs/pub41344.html | 21:16 |
| jimbaker | http://highscalability.com/blog/2013/10/8/f1-and-spanner-holistically-compared.html | 21:17 |
| thomasem | Haha, nice | 21:21 |
| tojuvone | Project On-Boarding Rooms going in dev list | 21:22 |
| tojuvone | 90min slots in Boston | 21:24 |
| tojuvone | Good night, afternoon and thanks jimbaker for the etherpad | 21:28 |
| thomasem | Take it easy tojuvone! Thanks for doing this for us. | 21:33 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: JSON Path-like querying for variables https://review.openstack.org/443941 | 21:38 |
| thomasem | Regarding SQLAlchemy 1.1.0 cap being removed in global requirements, it's still not set to the same definition as in our requirements.txt in my patch. I still think it makes sense to leave that definition there as-is to describe the actual support Craton has, but what about making that requirements gate check experimental until the definition in global-requirements.txt forces >= 1.1.0 so we can use | 21:43 |
| thomasem | the same definition? | 21:43 |
| thomasem | That they don't match is causing that job to fail, I think | 21:44 |
| *** valw has joined #craton | 21:52 | |
| *** valw_ has joined #craton | 21:59 | |
| *** valw has quit IRC | 22:02 | |
| thomasem | Time for me to run. Have a lovely evening/day, everyone! | 22:02 |
| jimbaker | tojuvone, plan to spend more time on that etherpad. good luck tomorrow! | 22:03 |
| *** valw_ has quit IRC | 22:21 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!