*** VW has quit IRC | 00:00 | |
cloudnull | sorry was away. fsaad, sigmavirus is right I used cassandra in my POC and it was because I wanted to be able to offload as much as I could elsewhere and cassandra let me do exactly that. | 00:54 |
---|---|---|
cloudnull | I was recently thinking about how i could get craton to use solr instead of trying to build all the search things into the app | 00:54 |
cloudnull | though i've not progressed beyond just a thought at this moment | 00:55 |
cloudnull | however i suspect indexing in solr w/ craton should not be too much of a massive lift | 00:55 |
cloudnull | for reference http://mysolr.readthedocs.io/en/latest/user/userguide.html | 01:03 |
*** Syed__ has quit IRC | 01:05 | |
*** VW has joined #craton | 03:00 | |
*** VW has quit IRC | 03:01 | |
*** VW has joined #craton | 03:02 | |
*** pwnall1337 is now known as zz_pwnall1337 | 03:03 | |
*** Mitschke has quit IRC | 03:43 | |
*** Mitschke has joined #craton | 04:28 | |
*** anonymike has quit IRC | 06:26 | |
*** anonymike has joined #craton | 06:28 | |
*** tojuvone has quit IRC | 06:55 | |
*** tojuvone has joined #craton | 07:26 | |
*** sigmavirus has quit IRC | 09:23 | |
*** sigmavirus has joined #craton | 09:30 | |
*** sigmavirus is now known as Guest81350 | 09:30 | |
*** openstackgerrit has quit IRC | 10:33 | |
*** VW has quit IRC | 13:21 | |
sulo | alright here is our final req: https://gist.github.com/sulochan/347e70dabb12e66670590ae003a15dcd | 13:45 |
*** VW has joined #craton | 13:51 | |
thomasem | o/ | 13:52 |
thomasem | sulo: yay! | 13:53 |
antonym | nice, got nested vars working? | 14:06 |
tojuvone | nice | 14:06 |
Guest81350 | open laptop... forget to log onto IRC... get stuck in quagmire | 14:08 |
*** Guest81350 is now known as sigmavirus | 14:08 | |
*** sigmavirus has quit IRC | 14:09 | |
*** sigmavirus has joined #craton | 14:09 | |
*** ChanServ sets mode: +o sigmavirus | 14:09 | |
sigmavirus | sulo: well done! | 14:09 |
sulo | heh, thanks but i didnt do much :) the work was already done here: https://github.com/cloudnull/cruton/blob/master/playbooks/populate-data.yml | 14:31 |
sulo | i only modified it a little bit | 14:31 |
sulo | antonym: no nested search is not working yet | 14:31 |
antonym | ah, looks like you had some nested vars injected in there | 14:36 |
sulo | ah injection works fine | 14:36 |
sulo | its only searching those nested vars ... but once we fulfill this last req | 14:36 |
sulo | i am going to put a spec to do it on python side | 14:36 |
sulo | so it might be alittle slow for large vars set | 14:37 |
sulo | but will have it atleast | 14:37 |
sulo | well, actually depending on where that bit of work falls in priority list i guess | 14:38 |
sigmavirus | sulo: IOW, we're planning to knowingly create a way for someone to issue several requests against craton that would grind it to a halt? | 14:39 |
sulo | huh ? | 14:39 |
sulo | sigmavirus: oh for nested var search you mean ? | 14:40 |
sigmavirus | yes | 14:40 |
sulo | sigmavirus: heh, well i dont think it would be that extreme .. but i ack your point | 14:41 |
sulo | sigmavirus: we need to find some way to do it | 14:41 |
sigmavirus | I agree we need to find a way to do it | 14:42 |
sigmavirus | I'm just wary of introducing potential DoS vectors to the project to support specific behaviour | 14:42 |
sulo | yes, very valid point .. especially if done across tenants etc | 14:42 |
sulo | i guess we can discuss this once we set out the priority items | 14:43 |
sulo | mabye switch to > mysql5.7 | 14:43 |
sulo | and use raw query for json search with index | 14:43 |
sulo | sqlalchemy utils with json has no support for json search afaik | 14:44 |
sigmavirus | sulo: I don't have the right answer, and I don't expect us to find that right now, just voicing a concern over what you proposed :) | 14:45 |
sulo | sigmavirus: ack, i am going to propose a spec and discuss etc | 14:45 |
tojuvone | I would enhance docker-install.rst. Do I need a bug for it? | 14:47 |
git-harry | sigmavirus: how is that different from adding the detail and all flags to the client? | 14:57 |
git-harry | Regarding sulo's suggestion of using the client to search vars. | 14:57 |
cloudnull | sulo: whats tha t? | 15:01 |
sulo | cloudnull: ? | 15:01 |
cloudnull | [09:00] <sulo> [14:36:37] heh, thanks but i didnt do much :) the work was already done here: https://github.com/cloudnull/cruton/blob/master/playbooks/populate-data.yml | 15:02 |
cloudnull | ? | 15:02 |
sulo | oh, i just used that script to populate data to craton | 15:02 |
sulo | so was saying that i didnt do much other than point thing at craton | 15:03 |
sulo | in that script | 15:03 |
sulo | playbook i shoud say | 15:03 |
sigmavirus | git-harry: I expected sulo to be doing all the python searching server-side, not client side | 15:05 |
sigmavirus | Client-side would be orders of magnitude worse but would not be a DoS vector | 15:06 |
sigmavirus | git-harry: so if your sole aim is to make everything worse for the user, then by all means, let's never have bulk import or proper variable search functionality | 15:06 |
git-harry | sigmavirus: they are separate things and do not necessarily need to be treated the same way. Regarding variable searching, I wasn't suggesting it was ideal, I was simply trying to understand how it is worse than providing 'craton host-list --details --all' | 15:08 |
sulo | hmm .. not sure about bulk import .. i am only talking about variable search | 15:08 |
sigmavirus | git-harry: --details --all is a bit of python glue code (if I remember implementation correctly) as opposed to what would likely need to be some recursive searching through dictionaries | 15:11 |
cloudnull | sulo: ah . cool | 15:12 |
sigmavirus | --details --all is not client-side glue code either | 15:12 |
sigmavirus | So having recursive search of a large group of variables (on each host) would be kind of awful server-side and even worse client-side | 15:12 |
sulo | are we supposed to have the team sync now ? | 15:44 |
thomasem | I thought it was in 45 min? | 15:44 |
sulo | ah cool. it says 3:30 pm | 15:45 |
thomasem | Oh | 15:45 |
sulo | but i dont see it on my cal | 15:45 |
thomasem | My invite says 10:30AM CST | 15:45 |
sigmavirus | yeah 45m | 15:45 |
thomasem | Mmm, also DST is coming up. How fun. | 15:49 |
sigmavirus | thomasem: no | 15:53 |
sigmavirus | I will not change my clocks | 15:53 |
sigmavirus | never agai | 15:53 |
thomasem | :P | 15:56 |
*** Syed__ has joined #craton | 15:58 | |
*** openstackgerrit has joined #craton | 16:00 | |
jimbaker | the good thing about DST is that it occurs at different dates. so for our team: mar 12 (US) and mar 26 (UK). enjoy! :) | 16:00 |
thomasem | Lol, yep! | 16:01 |
thomasem | Gotta spread the fun around. Make people wake up at odd times and feel jet lagged for a few days. | 16:02 |
jimbaker | also i noticed the calendar invite is wrong for thurs meetings starting next week | 16:02 |
jimbaker | we agreed to tie this to US DST, so it coincides with tues time | 16:02 |
jimbaker | i will adjust accordingly | 16:02 |
jimbaker | meanwhile our #openstack-meeting-4 is on UTC... | 16:03 |
thomasem | jimbaker: And we're honoring UTC, correct? | 16:07 |
jimbaker | ye | 16:07 |
jimbaker | yes | 16:07 |
thomasem | Okay, good. | 16:07 |
jimbaker | that meeting took some work to get scheduled in the first place, because it's an official meeting | 16:07 |
jimbaker | so basically set in stone | 16:07 |
thomasem | Lol okay | 16:08 |
thomasem | Gotcha | 16:08 |
sigmavirus | jimbaker: it's not that hard to schedule, but they don't take kindly to moving meetings around | 16:09 |
sigmavirus | and it'd be easier if we picked friday as a meeting day instead of Monday | 16:09 |
sigmavirus | that said | 16:09 |
sigmavirus | I like it sticking to UTC (like every other openstack team) | 16:09 |
jimbaker | yeah... we are not moving the monday meeting | 16:10 |
sigmavirus | It at least gives us teh appearance of trying to participate in good faith in the big tent | 16:10 |
jimbaker | and we will do a better job of that participation over time | 16:10 |
jimbaker | looks like i will be late to our meeting in 8 min. be there asap | 16:21 |
* sigmavirus ran into problems with our testing plan | 16:28 | |
sigmavirus | https://review.openstack.org/#/c/442516/ | 16:28 |
sigmavirus | https://review.openstack.org/#/c/442536/ | 16:28 |
sigmavirus | hopefully those will be fixed/released today | 16:28 |
fsaad | sulo / jimbaker are you able to join the slightly earlier scheduled meeting? :) | 16:32 |
git-harry | fsaad: sulo is afk | 16:34 |
fsaad | thanks git-harry | 16:34 |
*** zz_pwnall1337 is now known as pwnall1337 | 16:42 | |
pwnall1337 | how do we feel about importing ansible vars with key/value in this format so it can be parsed by ansible if included in a playbook? https://gist.github.com/pwnall1337/0c1389b07258e24bad155437242285d5 | 16:49 |
pwnall1337 | then on top of those host vars, the craton client can parse cpu information | 16:49 |
pwnall1337 | fsaad: ^ | 16:50 |
pwnall1337 | "ansible_processor " | 16:50 |
pwnall1337 | "ansible_virtualization_type" | 16:51 |
sigmavirus | pwnall1337: having those up one level makes sense | 16:59 |
pwnall1337 | cool ill update the code here https://github.com/osic/ops-generic/tree/master/tools/osa_import_for_craton | 17:00 |
fsaad | room #craton-recordable | 17:01 |
openstackgerrit | Ian Cordasco proposed openstack/python-cratonclient master: Add Betamax for testing https://review.openstack.org/442165 | 17:07 |
openstackgerrit | Syed Ahsan Shamim Zaidi proposed openstack/python-cratonclient master: Add Betamax for testing https://review.openstack.org/442165 | 17:52 |
*** david-lyle_ has joined #craton | 17:52 | |
*** david-lyle has quit IRC | 17:54 | |
*** david-lyle_ is now known as david-lyle | 17:55 | |
*** david-lyle has quit IRC | 18:04 | |
*** david-lyle has joined #craton | 18:05 | |
*** david-lyle has quit IRC | 18:23 | |
*** david-lyle has joined #craton | 18:25 | |
*** david-lyle has quit IRC | 18:29 | |
*** david-lyle has joined #craton | 18:36 | |
*** david-lyle has quit IRC | 19:04 | |
*** david-lyle has joined #craton | 19:06 | |
cloudnull | when do we think https://github.com/openstack/craton/blob/master/tools/docker_run.sh#L38-L46 will be rolled into the inital db-sync? | 19:14 |
cloudnull | sulo: jimbaker: ? | 19:15 |
cloudnull | or maybe something we could add as an additional cli command to create those users? | 19:17 |
pwnall1337 | created a bug report, for host show failing after importing vars https://bugs.launchpad.net/craton/+bug/1670802 | 19:25 |
openstack | Launchpad bug 1670802 in craton "host show fails after importing variables to host" [Undecided,New] | 19:25 |
cloudnull | fsaad: RE: db mungery we talked about earlier. | 19:35 |
antonym | pwnall1337: yeah, once you run those few commands from the docker image, it corrects that | 19:46 |
cloudnull | pwnall1337: ++ | 19:46 |
pwnall1337 | ? | 19:48 |
pwnall1337 | sorry i just uninstalled craton, pulled code and re installed and restarted the api | 19:49 |
pwnall1337 | what am i missing :) | 19:49 |
pwnall1337 | i upgraded db too* | 19:49 |
pwnall1337 | sorry | 19:49 |
antonym | pwnall1337: see update to bug | 19:50 |
fsaad | pwnall1337: https://github.com/openstack/craton/blob/master/tools/docker_run.sh | 19:51 |
fsaad | oh, thanks antonym :) brb foods | 19:51 |
antonym | np, figured it might be good to put it there in case someone else runs across that | 19:51 |
pwnall1337 | ill test it out, but if it does fix my issue, any plans to include that into the bootstrap upgrade command? | 19:52 |
pwnall1337 | wait | 19:52 |
pwnall1337 | i already have projects and users | 19:53 |
pwnall1337 | i created them as part of my reset-db script | 19:53 |
pwnall1337 | and i create region and cloud | 19:53 |
pwnall1337 | ill re read ,maybe im missing something | 19:53 |
pwnall1337 | antonym: that fixed my issue | 20:03 |
pwnall1337 | should be part of craton docs imo | 20:03 |
antonym | or abstracted away if it's required | 20:03 |
pwnall1337 | ^ | 20:03 |
antonym | seems pretty critical to not ending up with a broken deployment | 20:04 |
pwnall1337 | https://github.com/osic/ops-generic/tree/master/tools/osa_import_for_craton has been updated for the craton demo requirements | 20:06 |
cloudnull | antonym: pwnall1337: +1 i too ran into that, and have been ranting about it. sadly its been radio silence. | 20:08 |
cloudnull | I've posted that in here a few times now. | 20:09 |
sigmavirus | jimbaker: see above conversation | 20:57 |
jimbaker | cloudnull, i opened a bug for this: https://bugs.launchpad.net/craton/+bug/1670561 | 21:17 |
jimbaker | thomasem, ^^^ note the integration with craton-dbsync | 21:17 |
openstack | Launchpad bug 1670561 in craton "Create bootstrap user and project only via SQLAlchemy models" [Undecided,New] | 21:17 |
cloudnull | ^ when will then be now ? | 21:17 |
thomasem | Thanks, jimbaker | 21:26 |
jimbaker | well, i suppose ASAP is the ask? right now it has not been prioritized. more importantly it has not been discussed yet in terms of whether this specific approach is the right one | 21:26 |
jimbaker | but assuming it's the right way to do this, it could be done very quickly. implementation is certainly very small - it's comparable to what is done in the cited test | 21:28 |
jimbaker | sigmavirus, sulo, thomasem, so what do you think of this bug ( https://bugs.launchpad.net/craton/+bug/1670561), so we can get bootstrap construction moved out of sql? | 21:34 |
openstack | Launchpad bug 1670561 in craton "Create bootstrap user and project only via SQLAlchemy models" [Undecided,New] | 21:34 |
thomasem | jimbaker: This is referring to our discussion over the past week or so regarding moving bootstrapping into the migrations? | 21:38 |
thomasem | Although this suggests parameterizing it, which makes sense. | 21:39 |
jimbaker | thomasem, it's not strictly speaking a migration step | 21:39 |
jimbaker | but dbsync is for db setup | 21:40 |
thomasem | Right, it could be out-of-band, given the parameterization. | 21:40 |
jimbaker | exactly' | 21:40 |
thomasem | And just adhere to HEAD as the version it assumes. | 21:40 |
jimbaker | you need to create a new bootstrap user/project (for some reason? hard to know why) | 21:40 |
jimbaker | thomasem, correct, because it uses model code | 21:40 |
jimbaker | which makes sense, because it's not a schema change, it's just an object construction in the db | 21:41 |
thomasem | I think that makes good sense and has the added flexibility that deployers will probably want. | 21:41 |
thomasem | And then they won't have to go through the whole "delete the original root after you make your own" thing. | 21:41 |
jimbaker | yeah, if you want to do so, you can | 21:41 |
jimbaker | but more importantly, because we will get it a crypto strong secret, it doesn't matter | 21:42 |
jimbaker | up to the point of being able to read/write the db. wait, if you can do that, it doesn't matter :) | 21:42 |
jimbaker | at all | 21:42 |
thomasem | Okay, I'm burnt for the day. Have a lovely evening, everyone! | 22:59 |
jimbaker | thomasem, enjoy, and do come back refreshed! thanks! | 23:00 |
*** VW has quit IRC | 23:37 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!