| *** eghobo has quit IRC | 00:03 | |
| *** sbfox has quit IRC | 00:04 | |
| *** sbfox has joined #openstack-trove | 00:05 | |
| *** Barker has joined #openstack-trove | 00:13 | |
| *** khyati_ has joined #openstack-trove | 00:15 | |
| *** jcru has joined #openstack-trove | 00:18 | |
| *** michael-yu has quit IRC | 00:21 | |
| *** jcru has quit IRC | 00:23 | |
| *** michael-yu has joined #openstack-trove | 00:23 | |
| *** achampio1 has joined #openstack-trove | 00:24 | |
| *** michael-yu has quit IRC | 00:26 | |
| *** achampion has quit IRC | 00:26 | |
| *** matsuhashi has joined #openstack-trove | 00:27 | |
| *** sbfox has quit IRC | 00:29 | |
| *** ViswaV has quit IRC | 00:34 | |
| *** grapex has quit IRC | 00:36 | |
| *** michael-yu has joined #openstack-trove | 00:38 | |
| *** amcrn has quit IRC | 00:49 | |
| *** amcrn has joined #openstack-trove | 00:55 | |
| *** zigo has quit IRC | 00:59 | |
| *** zigo has joined #openstack-trove | 01:01 | |
| *** robertmyers has joined #openstack-trove | 01:03 | |
| *** michael-yu has quit IRC | 01:04 | |
| *** khyati_ has quit IRC | 01:17 | |
| *** rmyers has joined #openstack-trove | 01:24 | |
| *** nosnos has joined #openstack-trove | 01:26 | |
| *** rmyers has quit IRC | 01:30 | |
| *** Barker has quit IRC | 01:33 | |
| *** sbfox has joined #openstack-trove | 01:49 | |
| *** achampion has joined #openstack-trove | 02:01 | |
| *** achampio1 has quit IRC | 02:02 | |
| *** grapex has joined #openstack-trove | 02:06 | |
| *** grapex has quit IRC | 02:14 | |
| *** sbfox has quit IRC | 02:15 | |
| *** haomaiw__ has quit IRC | 02:31 | |
| *** haomaiwang has joined #openstack-trove | 02:32 | |
| *** amcrn has quit IRC | 02:42 | |
| *** jasonb365 has joined #openstack-trove | 02:50 | |
| *** harlowja is now known as harlowja_away | 02:59 | |
| *** haomai___ has joined #openstack-trove | 02:59 | |
| *** haomaiwang has quit IRC | 03:03 | |
| *** grapex has joined #openstack-trove | 03:07 | |
| *** harlowja_away is now known as harlowja | 03:09 | |
| *** khyati_ has joined #openstack-trove | 03:10 | |
| *** grapex has quit IRC | 03:11 | |
| *** matsuhashi has quit IRC | 03:22 | |
| *** robertmyers has quit IRC | 03:34 | |
| *** robertmyers has joined #openstack-trove | 03:35 | |
| *** nosnos has quit IRC | 03:37 | |
| *** mattgriffin has joined #openstack-trove | 03:41 | |
| *** robertmy_ has joined #openstack-trove | 03:42 | |
| *** esp has left #openstack-trove | 03:43 | |
| *** eghobo has joined #openstack-trove | 03:44 | |
| *** robertmy_ has quit IRC | 03:45 | |
| *** robertmyers has quit IRC | 03:45 | |
| openstackgerrit | Nikhil Manchanda proposed a change to openstack/python-troveclient: Remove ports and proto from secgrp create call https://review.openstack.org/84335 | 04:09 |
|---|---|---|
| openstackgerrit | Nikhil Manchanda proposed a change to openstack/python-troveclient: Remove ports and proto from secgrp create call https://review.openstack.org/84335 | 04:12 |
| *** matsuhashi has joined #openstack-trove | 04:14 | |
| openstackgerrit | Nikhil Manchanda proposed a change to openstack/python-troveclient: Remove ports and proto from secgrp create call https://review.openstack.org/84335 | 04:14 |
| *** jasonb365 has quit IRC | 04:16 | |
| *** nosnos has joined #openstack-trove | 04:23 | |
| *** mattgrif_ has joined #openstack-trove | 04:25 | |
| *** mattgriffin has quit IRC | 04:25 | |
| *** sbfox has joined #openstack-trove | 05:01 | |
| *** matsuhas_ has joined #openstack-trove | 05:02 | |
| *** matsuhas_ has quit IRC | 05:02 | |
| *** matsuhashi has quit IRC | 05:05 | |
| *** matsuhashi has joined #openstack-trove | 05:05 | |
| *** harlowja is now known as harlowja_away | 05:11 | |
| *** michael-yu has joined #openstack-trove | 05:13 | |
| *** michael-yu has quit IRC | 05:17 | |
| *** michael-yu has joined #openstack-trove | 05:18 | |
| *** michael-yu has quit IRC | 05:18 | |
| *** khyati_ has quit IRC | 05:22 | |
| *** michael-yu has joined #openstack-trove | 05:22 | |
| *** michael-yu has quit IRC | 05:24 | |
| *** SushilKM has joined #openstack-trove | 05:24 | |
| *** sbfox has quit IRC | 05:27 | |
| *** sbfox has joined #openstack-trove | 05:31 | |
| *** yidclare has joined #openstack-trove | 05:40 | |
| *** mattgrif_ has quit IRC | 05:41 | |
| *** mattgriffin has joined #openstack-trove | 05:52 | |
| *** mattgriffin has quit IRC | 05:53 | |
| *** michael-yu has joined #openstack-trove | 05:57 | |
| *** sbfox has quit IRC | 06:02 | |
| *** sbfox has joined #openstack-trove | 06:03 | |
| openstackgerrit | A change was merged to openstack/trove: Remove mockito, and replace with mock https://review.openstack.org/82888 | 06:12 |
| *** saju_m has joined #openstack-trove | 06:21 | |
| openstackgerrit | Jenkins proposed a change to openstack/trove: Imported Translations from Transifex https://review.openstack.org/82721 | 06:23 |
| *** yogesh has joined #openstack-trove | 06:29 | |
| *** demorris has joined #openstack-trove | 06:43 | |
| *** nosnos has quit IRC | 06:45 | |
| *** nosnos has joined #openstack-trove | 06:46 | |
| *** demorris has left #openstack-trove | 06:50 | |
| *** demorris has joined #openstack-trove | 06:50 | |
| *** flaper87|afk is now known as flaper87 | 06:50 | |
| *** michael-yu has quit IRC | 06:57 | |
| *** matsuhashi has quit IRC | 06:58 | |
| *** matsuhashi has joined #openstack-trove | 06:58 | |
| *** matsuhashi has quit IRC | 07:03 | |
| *** matsuhashi has joined #openstack-trove | 07:06 | |
| *** sbfox has quit IRC | 07:10 | |
| *** demorris has quit IRC | 07:28 | |
| *** michael-yu has joined #openstack-trove | 07:31 | |
| *** eghobo has quit IRC | 07:45 | |
| *** flaper87 has left #openstack-trove | 07:49 | |
| *** SushilKM has quit IRC | 07:56 | |
| *** michael-yu has quit IRC | 08:00 | |
| *** matsuhashi has quit IRC | 08:03 | |
| *** matsuhas_ has joined #openstack-trove | 08:05 | |
| *** tanisdl has joined #openstack-trove | 08:16 | |
| *** derez has joined #openstack-trove | 08:32 | |
| *** derez is now known as Guest39748 | 08:32 | |
| *** matsuhas_ has quit IRC | 08:33 | |
| *** sgotliv has joined #openstack-trove | 08:33 | |
| sgotliv | Guys, do you know how to add trove to already installed Openstack? | 08:35 |
| *** matsuhashi has joined #openstack-trove | 08:38 | |
| *** saju_m has quit IRC | 08:43 | |
| *** SushilKM has joined #openstack-trove | 08:45 | |
| *** saju_m has joined #openstack-trove | 08:56 | |
| *** yogesh has quit IRC | 08:58 | |
| *** yogesh has joined #openstack-trove | 08:58 | |
| *** yogesh has quit IRC | 09:03 | |
| *** dmakogon_ is now known as denis_makogon | 09:25 | |
| denis_makogon | sgotliv, you need to create project inside the keystone (add appropriate endpoints, etc), then launch all trove services with correct configs | 09:26 |
| sgotliv | denis_makogon, Hi Denis, I saw your presentation about Trove yesterday and it was awesome | 09:27 |
| denis_makogon | sgotliv, thanks | 09:28 |
| sgotliv | denis_makogon, but my opinion that most basic staff like howto install Trove is missing, I searched over the net. So now I kind of thrilled by your preso but can't install trove :-) | 09:29 |
| denis_makogon | sgotliv, take a look at https://review.openstack.org/#/c/78608/ | 09:29 |
| denis_makogon | sgotliv, it works for havanna release | 09:30 |
| denis_makogon | sgotliv, also, as the contribution for Trove, you could verify this doc | 09:30 |
| denis_makogon | sgotliv, about my presentation, it wasn't about howto install Trove on OS evn, it was about overview =) | 09:30 |
| sgotliv | denis_makogon, It starts with Ubuntu, I have RHEL or Fedora. I tried to find a clear statement about support of other distributions | 09:31 |
| denis_makogon | sgotliv, trove works on Fedora | 09:31 |
| denis_makogon | sgotliv, we haven't any docs for RHEL/Fedora, yet, because all our tests are running over Ubuntu 12.04-cloudinit | 09:32 |
| denis_makogon | sgotliv, we (you and me) could try to write doc for the RHEL/Fedora | 09:32 |
| sgotliv | denis_makogon, 1. Your preso is great!!! and it exactly what it is. 2. I'll try to look into Fedora and then we'll write a doc | 09:33 |
| sgotliv | denis_makogon, thanks!!! | 09:33 |
| *** iartarisi has joined #openstack-trove | 09:33 | |
| denis_makogon | sgotliv, you can ping me, if you need any help | 09:33 |
| sgotliv | denis_makogon, great, I | 09:33 |
| denis_makogon | sgotliv, my daytime is UTC + 2 | 09:33 |
| sgotliv | denis_makogon, my either | 09:34 |
| denis_makogon | SlickNik, saw your patch, it absorbs more than secgroups fix, and mine patch was first https://review.openstack.org/#/c/84115/. | 09:36 |
| denis_makogon | SlickNik, and your patch wouldn't work if datastore has more than one rule | 09:38 |
| SlickNik | denis_makogon: Sure np, I don't care whose patch lands as long as the issue is fixed. | 09:40 |
| denis_makogon | SlickNik, i'll try to beat this beast today | 09:41 |
| denis_makogon | i almost fixed server-side and client | 09:41 |
| denis_makogon | SlickNik, would be great if you take a look at both patches | 09:41 |
| SlickNik | denis_makogon: btw, your patch will not pass the Jenkins tests, you might want to take a look at https://review.openstack.org/#/c/84335/3/troveclient/tests/test_secgroups.py | 09:42 |
| SlickNik | And incorporate some of those changes. | 09:42 |
| SlickNik | denis_makogon: Why would my patch not work for multiple datastores? | 09:42 |
| denis_makogon | SlickNik, not multiple datastores | 09:43 |
| denis_makogon | SlickNik, multiple rules | 09:43 |
| denis_makogon | SlickNik, take a look how view formed | 09:43 |
| denis_makogon | SlickNik, mysql has one rule, cassandra has 4, couchbase has alot | 09:44 |
| denis_makogon | SlickNik, view formed by the only one rule | 09:44 |
| denis_makogon | SlickNik, and client expects that income response contains only one object | 09:44 |
| denis_makogon | SlickNik, suppose we create new rules with new cidr, as the result (at trove side) we have four rules that should be used for forming the view | 09:45 |
| denis_makogon | SlickNik, my suggestion is to change the way the view is prepared | 09:46 |
| denis_makogon | SlickNik, it'll be the list of groups, instead of single group, | 09:46 |
| denis_makogon | SlickNik, and it would work for every datastore, since we could have the list of the one rule | 09:46 |
| denis_makogon | one or many, no matter | 09:47 |
| SlickNik | I think what you're saying is that the issue is with the representation of the rule during a list rules command? | 09:50 |
| denis_makogon | SlickNik, no, i'm talking about the representation of the create call | 09:50 |
| denis_makogon | SlickNik, this is how it will be https://review.openstack.org/#/c/83768/3/trove/extensions/security_group/views.py at trove | 09:51 |
| denis_makogon | SlickNik, and this is how it'll be in client https://review.openstack.org/#/c/84115/1/troveclient/v1/shell.py | 09:51 |
| SlickNik | gotcha | 09:53 |
| denis_makogon | SlickNik, cool, and this will work in any cases | 09:53 |
| SlickNik | Sounds good. | 09:54 |
| SlickNik | btw, denis_makogon can you make sure to update the launchpad bug/bp when you start working on something? The reason I picked this up was because it wasn't updated. | 09:54 |
| SlickNik | (for the client side) | 09:55 |
| denis_makogon | SlickNik, i though infra should've do that, don't know why this didn't happen | 09:55 |
| SlickNik | Strange. Infra didn't pick it up (bug?) | 09:57 |
| denis_makogon | SlickNik, yes | 09:57 |
| denis_makogon | SlickNik, i branched the fix, i mentioned the Closes-Bug | 09:57 |
| openstackgerrit | Denis M. proposed a change to openstack/python-troveclient: Fix create call for security group rules https://review.openstack.org/84115 | 10:00 |
| SlickNik | denis_makogon: gonna head out. Will you be able to get done with the fixes for the secgroups issue like we talked about? | 10:08 |
| denis_makogon | SlickNik, i'll try to fix it, in 2-3h | 10:09 |
| SlickNik | Thanks! I'll test it out in the morning when I'm back up. | 10:09 |
| SlickNik | Hoping to get it done soon so we can cut RC1, and move on to Juno. | 10:10 |
| SlickNik | See you later. | 10:11 |
| denis_makogon | SlickNik, cu 2 | 10:12 |
| *** matsuhashi has quit IRC | 10:14 | |
| openstackgerrit | Pierre Rambaud proposed a change to openstack/trove-integration: Improve README.md https://review.openstack.org/84398 | 10:25 |
| *** sgotliv has quit IRC | 10:30 | |
| denis_makogon | SlickNik, done, fix is ready | 10:33 |
| openstackgerrit | Denis M. proposed a change to openstack/python-troveclient: Fix create call for security group rules https://review.openstack.org/84115 | 10:37 |
| *** nosnos has quit IRC | 10:40 | |
| *** sgotliv has joined #openstack-trove | 10:43 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix create call for security group rules https://review.openstack.org/83768 | 10:46 |
| *** sgotliv has quit IRC | 10:51 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Support full backup/restore for Cassandra https://review.openstack.org/77551 | 10:58 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Add point in time recovery https://review.openstack.org/77222 | 10:59 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Refactoring strategies implementation https://review.openstack.org/77899 | 11:04 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Refactor mysql and redis apps https://review.openstack.org/80020 | 11:11 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Track security group provisioned by heat https://review.openstack.org/71040 | 11:15 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Allow db instance conditional logging https://review.openstack.org/63789 | 11:20 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Allow log files audit https://review.openstack.org/64302 | 11:20 |
| *** SushilKM has quit IRC | 11:24 | |
| *** SushilKM has joined #openstack-trove | 11:25 | |
| *** tanisdl has quit IRC | 11:34 | |
| openstackgerrit | Denis M. proposed a change to openstack/python-troveclient: Fix create call for security group rules https://review.openstack.org/84115 | 11:40 |
| *** saju_m has quit IRC | 11:44 | |
| *** SushilKM has quit IRC | 11:49 | |
| *** radez_g0n3 is now known as radez | 11:51 | |
| *** saju_m has joined #openstack-trove | 11:56 | |
| openstackgerrit | Jenkins proposed a change to openstack/python-troveclient: Updated from global requirements https://review.openstack.org/79646 | 11:57 |
| openstackgerrit | Jenkins proposed a change to openstack/trove: Updated from global requirements https://review.openstack.org/79648 | 11:57 |
| *** pdmars has joined #openstack-trove | 12:01 | |
| *** saju_m has quit IRC | 12:02 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Refactoring strategies implementation https://review.openstack.org/77899 | 12:10 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix create call for security group rules https://review.openstack.org/83768 | 12:12 |
| denis_makogon | SlickNik, this crazy bug appears again, http://logs.openstack.org/38/76538/6/check/gate-trove-python27/4f7a876/console.html#_2014-04-01_10_59_27_787 | 12:13 |
| *** achampio1 has joined #openstack-trove | 12:19 | |
| *** achampion has quit IRC | 12:20 | |
| *** saju_m has joined #openstack-trove | 12:21 | |
| *** saju_m has quit IRC | 12:23 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Allow log files audit https://review.openstack.org/64302 | 12:27 |
| *** saju_m has joined #openstack-trove | 12:39 | |
| *** Barker has joined #openstack-trove | 12:58 | |
| *** sgotliv_ has joined #openstack-trove | 13:07 | |
| *** achampio1 has quit IRC | 13:07 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix create call for security group rules https://review.openstack.org/83768 | 13:12 |
| *** saju_m has quit IRC | 13:15 | |
| *** mattgriffin has joined #openstack-trove | 13:21 | |
| *** kevinconway has joined #openstack-trove | 13:27 | |
| *** saju_m has joined #openstack-trove | 13:28 | |
| *** jasonb365 has joined #openstack-trove | 13:32 | |
| *** jasonb365 has quit IRC | 13:34 | |
| *** demorris has joined #openstack-trove | 13:53 | |
| *** achampion has joined #openstack-trove | 13:53 | |
| *** jasonb365 has joined #openstack-trove | 14:04 | |
| *** radez is now known as radez_g0n3 | 14:07 | |
| *** mattgriffin has quit IRC | 14:08 | |
| *** saju_m has quit IRC | 14:14 | |
| *** amytron has joined #openstack-trove | 14:14 | |
| *** spredzy has left #openstack-trove | 14:16 | |
| *** michael-yu has joined #openstack-trove | 14:18 | |
| *** tanisdl has joined #openstack-trove | 14:26 | |
| *** jcru has joined #openstack-trove | 14:26 | |
| *** michael-yu has quit IRC | 14:26 | |
| *** tanisdl has quit IRC | 14:29 | |
| *** tanisdl has joined #openstack-trove | 14:31 | |
| *** BobBall has joined #openstack-trove | 14:32 | |
| BobBall | Hey - I've probably got a configuration issue... http://paste.openstack.org/show/74753/ | 14:33 |
| BobBall | happened all the time since devstack added trove :) | 14:34 |
| BobBall | Can anyone suggest what I should look at? | 14:34 |
| BobBall | omg... I think that the dodgy DNS resolver is preferring IPv6 over IPv4 so it's trying to get to localhost with IPv6 | 14:40 |
| *** tanisdl has quit IRC | 14:45 | |
| *** tanisdl has joined #openstack-trove | 14:45 | |
| *** thedodd has joined #openstack-trove | 14:56 | |
| *** sbfox has joined #openstack-trove | 15:02 | |
| denis_makogon | BobBall, wrong nova url mentioned | 15:10 |
| denis_makogon | BobBall, does nova runs at localhost ? | 15:10 |
| BobBall | yes | 15:12 |
| BobBall | this is devstack - so everything is localhost | 15:12 |
| BobBall | I think it might be because trove is setting IPv4... | 15:12 |
| BobBall | trove/guestagent/common/operating_system.py: s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM) | 15:12 |
| BobBall | but I've not tracked that through | 15:12 |
| BobBall | so that might be completely unrelated :D | 15:13 |
| *** shivamshukla has joined #openstack-trove | 15:13 | |
| denis_makogon | devstack's nova runs at 10.0.0.1, i guess | 15:13 |
| BobBall | why 10.0.0.1? Everything is set up to localhost / 127.0.0.1 for me | 15:14 |
| denis_makogon | BobBall, that's only my guess, and you're right, it runs at 127.0.0.1 | 15:14 |
| BobBall | hmmm - worse - localhost seems to be the default option somewhere.. it's not specified as localhost in localrd | 15:14 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix create call for security group rules https://review.openstack.org/83768 | 15:15 |
| BobBall | stack.sh: RABBIT_HOST=${RABBIT_HOST:-localhost} | 15:16 |
| denis_makogon | try to check if nova is alive ? | 15:17 |
| *** RockKuo has joined #openstack-trove | 15:18 | |
| BobBall | yeah, it's all fine | 15:19 |
| denis_makogon | BobBall, hm, strange | 15:21 |
| denis_makogon | BobBall, take a look at nova logs ? | 15:22 |
| *** amytron has quit IRC | 15:23 | |
| BobBall | nova's quite happy with using "localhost"; it's only the few tempest tests that make use of trove that are failing | 15:26 |
| BobBall | but I'm going to propose we change devstack to set RABBIT_HOST=${RABBIT_HOST:-$SERVICE_HOST} | 15:27 |
| BobBall | in which case it'll avoid my problem and trove ipv6-ness can be saved for another day | 15:27 |
| *** tanisdl has quit IRC | 15:27 | |
| *** tanisdl has joined #openstack-trove | 15:28 | |
| *** jasonb365 has quit IRC | 15:29 | |
| *** grapex has joined #openstack-trove | 15:31 | |
| *** grapex has quit IRC | 15:32 | |
| *** grapex has joined #openstack-trove | 15:33 | |
| *** denis_makogon has quit IRC | 15:34 | |
| *** shivamshukla has quit IRC | 15:34 | |
| *** jasonb365 has joined #openstack-trove | 15:35 | |
| *** demorris has quit IRC | 15:37 | |
| *** Guest39748 has quit IRC | 15:38 | |
| *** ViswaV has joined #openstack-trove | 15:41 | |
| *** jmontemayor has joined #openstack-trove | 15:49 | |
| *** amytron has joined #openstack-trove | 15:53 | |
| *** amytron_ has joined #openstack-trove | 15:56 | |
| *** jmontemayor has quit IRC | 15:56 | |
| *** jmontemayor has joined #openstack-trove | 15:56 | |
| *** amytron has quit IRC | 15:58 | |
| *** amytron_ is now known as amytron | 15:58 | |
| *** iartarisi has quit IRC | 15:59 | |
| *** jasonb365 has quit IRC | 15:59 | |
| *** jasonb365 has joined #openstack-trove | 16:01 | |
| *** Barker has quit IRC | 16:02 | |
| *** Barker has joined #openstack-trove | 16:03 | |
| openstackgerrit | sai krishna proposed a change to openstack/trove: Added new variable to validate CA_file.pem of keystone and cinder. https://review.openstack.org/84484 | 16:03 |
| *** eghobo has joined #openstack-trove | 16:04 | |
| *** sbfox has quit IRC | 16:09 | |
| hub_cap | BobBall: :) | 16:10 |
| *** SushilKM has joined #openstack-trove | 16:13 | |
| *** sbfox has joined #openstack-trove | 16:13 | |
| *** RockKuo has quit IRC | 16:17 | |
| *** SushilKM has quit IRC | 16:18 | |
| *** SushilKM has joined #openstack-trove | 16:19 | |
| *** sgotliv_ has quit IRC | 16:24 | |
| *** sgotliv has joined #openstack-trove | 16:25 | |
| *** SushilKM has quit IRC | 16:28 | |
| *** sbfox1 has joined #openstack-trove | 16:30 | |
| *** sbfox has quit IRC | 16:33 | |
| *** Barker has quit IRC | 16:36 | |
| openstackgerrit | Viswa Vutharkar proposed a change to openstack/trove: Adds backup/restore support for mongodb https://review.openstack.org/78339 | 16:37 |
| *** sgotliv has quit IRC | 16:39 | |
| *** SushilKM has joined #openstack-trove | 16:44 | |
| *** grapex has quit IRC | 16:53 | |
| *** khyati has joined #openstack-trove | 16:58 | |
| *** esp has joined #openstack-trove | 16:59 | |
| *** michael-yu has joined #openstack-trove | 17:04 | |
| openstackgerrit | Daniel Salinas proposed a change to openstack/trove: Add instance metadata functionality to trove https://review.openstack.org/82123 | 17:07 |
| *** sbfox1 has quit IRC | 17:12 | |
| *** grapex has joined #openstack-trove | 17:14 | |
| *** harlowja_away is now known as harlowja | 17:14 | |
| *** grapex_ has joined #openstack-trove | 17:14 | |
| *** grapex has quit IRC | 17:18 | |
| *** mattgriffin has joined #openstack-trove | 17:20 | |
| *** michael-yu has quit IRC | 17:23 | |
| *** radez_g0n3 is now known as radez | 17:26 | |
| *** Barker has joined #openstack-trove | 17:29 | |
| *** thedodd has quit IRC | 17:29 | |
| *** SushilKM has quit IRC | 17:36 | |
| *** michael-yu has joined #openstack-trove | 17:41 | |
| *** tanisdl has quit IRC | 17:44 | |
| grapex_ | Btw- is Gerrit not showing the last four patchsets for a reason? https://review.openstack.org/#/c/80061/ | 17:47 |
| hub_cap | its gerrits april fools grapex_ | 17:50 |
| hub_cap | i see 5 patchsets | 17:50 |
| grapex_ | hub_cap: LOL! | 17:50 |
| grapex_ | Looks like I just got joked! | 17:51 |
| hub_cap | but 4 of them are draft, so mayb eu had to be on the list? | 17:51 |
| hub_cap | try now, i added u to the list of reviewers | 17:51 |
| grapex_ | hub_cap: bingo | 17:51 |
| *** amcrn has joined #openstack-trove | 17:51 | |
| * hub_cap is a gerrit master | 17:51 | |
| *** yogesh has joined #openstack-trove | 17:51 | |
| grapex_ | So... can anyone explain why this isn't a circle import issue still? | 17:51 |
| hub_cap | magic | 17:52 |
| hub_cap | the more i look at this, the less i like it heh | 17:52 |
| grapex_ | hub_cap: Does the import_group defer the import until it's needed? | 17:52 |
| hub_cap | nope | 17:52 |
| hub_cap | it imposrts them all at load time | 17:52 |
| hub_cap | the problem is such | 17:52 |
| hub_cap | the taskmgr needs to use these | 17:52 |
| hub_cap | which i thik is bad code | 17:52 |
| hub_cap | not his fault | 17:53 |
| grapex_ | So, by "magic" are you saying "it actually imports them right away and in fact does cause the import error you're mentioning?" | 17:53 |
| hub_cap | but just bad design | 17:53 |
| grapex_ | hub_cap: Agreed- which vars is it using? | 17:53 |
| hub_cap | we shouldnt need "mysql" specifics in the taskmgr | 17:53 |
| hub_cap | ports for secgrps | 17:53 |
| grapex_ | Oh yeah | 17:53 |
| hub_cap | and mounts stuff | 17:53 |
| grapex_ | duuuuuhhh | 17:53 |
| hub_cap | but i looked at the mounts stuff | 17:53 |
| grapex_ | hub_cap: Because they share configs | 17:53 |
| hub_cap | it actually puts it in a hash | 17:53 |
| hub_cap | and gives it to the guest | 17:53 |
| grapex_ | Oh man | 17:53 |
| hub_cap | so it doesnt need it | 17:53 |
| hub_cap | but the secgrps goes direct to nova | 17:53 |
| hub_cap | so its not so easy | 17:53 |
| hub_cap | but its a flaw in our design really | 17:53 |
| hub_cap | if we dont use those "datastore specifics" in tasmgr we dont need to impor thtem | 17:54 |
| hub_cap | so u get this thing where u have to "import" files u never use | 17:54 |
| hub_cap | so u can just use the config value, hence the reason for the "preload" | 17:54 |
| hub_cap | so rather than adding from datastoreX import config for EVERY datatstore in taskmgr | 17:54 |
| hub_cap | its magic'd in cfg | 17:54 |
| grapex_ | So wait- the config values in those modules are actually being loaded only to be passed to the guest over RPC? | 17:55 |
| hub_cap | the mount point yes | 17:55 |
| hub_cap | but the secgrps ports no | 17:55 |
| hub_cap | so we can fix mntpoint | 17:55 |
| grapex_ | That's another problem | 17:55 |
| hub_cap | thts easy... | 17:55 |
| grapex_ | the mount point should not be passed to the guest | 17:55 |
| grapex_ | unless we always pass it to the guest | 17:55 |
| hub_cap | but the fact that [mysql]tcp_port=3306 is there is a problem | 17:55 |
| hub_cap | cuz nova needs em | 17:55 |
| grapex_ | there have been recent commits where RPC arguments for the mount point where removed | 17:55 |
| hub_cap | yea but its in a hash still | 17:56 |
| hub_cap | in the block device mapping | 17:56 |
| hub_cap | or something | 17:56 |
| hub_cap | i cant rmeember, i looked at the code, but it escapes me at present | 17:56 |
| grapex_ | We have to seperate the guest | 17:56 |
| hub_cap | but the code basically grabs mount point, never uses it, puts it in a hash, and gives to guest | 17:56 |
| grapex_ | it just breeds confusion like this | 17:56 |
| hub_cap | rather than, does nada, and then lets the guest do the work | 17:56 |
| hub_cap | yea that will actually help | 17:56 |
| grapex_ | Well the guest is also responsible for remembering the mount point | 17:56 |
| hub_cap | but still if we are from guestagent impoirt blah, we are failing | 17:56 |
| grapex_ | only the guest has no persistent storage to call its own to remember things like that if it ever has to be restarted | 17:57 |
| hub_cap | at least we can restrict it w/ library unlinking | 17:57 |
| hub_cap | yea but its in the gust config | 17:57 |
| hub_cap | *guest | 17:57 |
| hub_cap | the datastore config | 17:57 |
| grapex_ | so the guest also grabs the mount point from the config file | 17:57 |
| hub_cap | right | 17:57 |
| grapex_ | yeah- point being we should remove it from the RPC call | 17:57 |
| amcrn | grapex_: long-term, the mount-point shouldn't be hardcoded anyway (see https://bugs.launchpad.net/nova/+bug/1246189) | 17:57 |
| hub_cap | amcrn: u better not stand me up on thr for lunch | 17:57 |
| hub_cap | im telling hte whole channel | 17:57 |
| hub_cap | so they all know if u do | 17:57 |
| amcrn | lol | 17:58 |
| hub_cap | EVERYONE IS MY WITNESS | 17:58 |
| grapex_ | I guess taskmanager does need to know the mount point so it can pass it to the file that gets created on the server | 17:58 |
| *** sbfox has joined #openstack-trove | 17:58 | |
| grapex_ | amcrn: That will be nice | 17:59 |
| *** sbfox has quit IRC | 18:01 | |
| grapex_ | brb | 18:03 |
| grapex_ | bbq | 18:21 |
| *** sbfox has joined #openstack-trove | 18:22 | |
| esp | grapex_: making me hungry | 18:27 |
| *** yogesh has quit IRC | 18:31 | |
| *** sbfox has quit IRC | 18:32 | |
| *** demorris has joined #openstack-trove | 18:33 | |
| *** yogesh has joined #openstack-trove | 18:35 | |
| *** sbfox has joined #openstack-trove | 18:42 | |
| *** thedodd has joined #openstack-trove | 18:48 | |
| *** grapex_ has quit IRC | 18:55 | |
| *** ViswaV has quit IRC | 19:02 | |
| *** ViswaV has joined #openstack-trove | 19:03 | |
| *** sgotliv has joined #openstack-trove | 19:03 | |
| *** ViswaV_ has joined #openstack-trove | 19:09 | |
| *** ViswaV has quit IRC | 19:09 | |
| *** ViswaV has joined #openstack-trove | 19:10 | |
| *** ViswaV_ has quit IRC | 19:10 | |
| *** eguz has joined #openstack-trove | 19:11 | |
| *** shivamshukla has joined #openstack-trove | 19:12 | |
| *** denis_makogon has joined #openstack-trove | 19:14 | |
| *** eghobo has quit IRC | 19:15 | |
| denis_makogon | amcrn, ping | 19:28 |
| *** mattgriffin has quit IRC | 19:29 | |
| denis_makogon | amcrn, please tell me, if proposed commit message is more appropriate for that patch | 19:32 |
| amcrn | denis_makogon: since i haven't looked at the bug or patch, is the idea that since datastores have their own cfg.py now, that if a user calls to create a security-group-rule, that only the port ranges and protocols defined in the cfg.py will be honored? | 19:33 |
| amcrn | meaning, users are not allowed to specify port ranges or protocols, they can only say "i want what you've defined, or I don't" | 19:34 |
| denis_makogon | amcrn, yes, only config ports are will be used on create call | 19:34 |
| amcrn | so the new flow would be (1) security_groups is turned off by default (2) user creates a new security-group, it starts off with zero rules (3) user asks to create rules, and the rules pre-defined in the cfg.py are created | 19:35 |
| amcrn | the alternate flow would be (1) security_groups is turned on by default (2) user creating a security-group is pointless | 19:35 |
| *** shivamshukla has quit IRC | 19:36 | |
| amcrn | correct? | 19:36 |
| denis_makogon | amcrn, yes, it is, but user cannot create security group through the API | 19:37 |
| *** ramashri has joined #openstack-trove | 19:37 | |
| denis_makogon | amcrn, user can create only set of the new rules, that has the same CIDR | 19:38 |
| amcrn | ah, that's right, thought for a second you could create a secgroup via api as well | 19:38 |
| amcrn | so it's really a purge-replace now, with the CIDR being the only change'able field | 19:39 |
| denis_makogon | amcrn, no, at least now, user cannot create group | 19:39 |
| denis_makogon | amcrn, yes, it is | 19:39 |
| amcrn | gotcha, i'm caught up now :) | 19:39 |
| denis_makogon | amcrn, cool =)) | 19:39 |
| amcrn | long-term the "create rule" operation in the cli should be removed then, and renamed to something more intuitive like "edit", because you can't really create any rules, you can only modify the existing set | 19:40 |
| denis_makogon | amcrn, basically it means, that user can define who can access instances, but not through wich ports | 19:40 |
| amcrn | but changing the bindings right now is a bit too much, so i'm on-board with what you've done i suppose | 19:41 |
| denis_makogon | amcrn, no, you're actually creating new rules, you cannot modify existing | 19:41 |
| denis_makogon | amcrn, nova-network doesn't allow to modify rules | 19:41 |
| amcrn | well, under the covers they're new rules, but from a user vantage point: they're the same rules, different cidr | 19:41 |
| amcrn | if you're a silly user, doesn't understand openstack | 19:41 |
| amcrn | net-net, the only thing that's changed is the cidr, you aren't able to *create* a rule | 19:42 |
| amcrn | so, my only point was the wording is a bit misleading as to what you can actually accomplish | 19:42 |
| denis_makogon | amcrn, yes, user cannot fully control the content of the rule, yes, only the part of it | 19:42 |
| amcrn | in v2 i suppose we can fix this up | 19:42 |
| amcrn | alright, cool. let me review both the api and the cli patches together throughout the day, and i'll add to the review any comments (if deemed necessary) | 19:43 |
| *** ViswaV has quit IRC | 19:43 | |
| amcrn | thanks for the quick explanation | 19:43 |
| denis_makogon | amcrn, hard to say about v2, at least now | 19:43 |
| denis_makogon | amcrn, does rules has quotas per group ? | 19:44 |
| denis_makogon | amcrn, i mean, are there any limitations on how many rules can i add to the group | 19:45 |
| denis_makogon | amcrn, if that so, i guess, we need to add create security group API (create method) | 19:45 |
| *** ViswaV has joined #openstack-trove | 19:48 | |
| *** mattgriffin has joined #openstack-trove | 19:50 | |
| *** michael-yu has quit IRC | 19:54 | |
| *** michael-yu has joined #openstack-trove | 19:55 | |
| *** ramashri has quit IRC | 19:59 | |
| *** sbfox1 has joined #openstack-trove | 19:59 | |
| *** sbfox has quit IRC | 20:00 | |
| *** mattgriffin has quit IRC | 20:02 | |
| *** sbfox has joined #openstack-trove | 20:03 | |
| *** sbfox1 has quit IRC | 20:03 | |
| *** radez is now known as radez_g0n3 | 20:04 | |
| *** yogesh has quit IRC | 20:05 | |
| *** yogesh has joined #openstack-trove | 20:13 | |
| *** robertmyers has joined #openstack-trove | 20:15 | |
| *** sbfox has quit IRC | 20:18 | |
| *** robertmyers has quit IRC | 20:25 | |
| *** eguz has quit IRC | 20:27 | |
| *** eghobo has joined #openstack-trove | 20:28 | |
| *** haomai___ has quit IRC | 20:32 | |
| *** jmontemayor has quit IRC | 20:34 | |
| *** haomaiwang has joined #openstack-trove | 20:36 | |
| *** sbfox has joined #openstack-trove | 20:45 | |
| *** grapex has joined #openstack-trove | 20:46 | |
| *** michael-yu has quit IRC | 20:47 | |
| *** grapex has quit IRC | 20:47 | |
| *** grapex has joined #openstack-trove | 20:48 | |
| *** sbfox has quit IRC | 20:49 | |
| *** sbfox has joined #openstack-trove | 20:49 | |
| *** Barker has quit IRC | 20:58 | |
| *** michael-yu has joined #openstack-trove | 21:00 | |
| *** laurelm has quit IRC | 21:02 | |
| *** laurelm has joined #openstack-trove | 21:03 | |
| BobBall | denis_makogon: ping? | 21:11 |
| denis_makogon | BobBall, 'sup ? | 21:11 |
| *** Barker has joined #openstack-trove | 21:11 | |
| BobBall | Trying to understand trove... and why it's not working :) | 21:12 |
| BobBall | Seems there are a lot of hardcoded 10.0.0.1's in guestagent's conf file | 21:12 |
| BobBall | are they used in devstack? | 21:12 |
| BobBall | I can't see a reference to the file in the installed setup, but devstack does mess with it | 21:12 |
| *** sbfox has quit IRC | 21:13 | |
| BobBall | I ask you because you were helpful about the 10.0.0.1 earlier today :P | 21:13 |
| kevinconway | i don't know if this is the case, there is a 10.0.0.* bridge set up | 21:13 |
| kevinconway | those IPs may also be rewritten during setup | 21:13 |
| denis_makogon | kevinconway, we're talking about manual trove setup | 21:13 |
| BobBall | Not always - that's my problem. My setup of devstack on xenserver definitely does not use 10.0.0.1 but the lib/trove code doens't seem to replace them | 21:13 |
| BobBall | no - not manual - devstack with something other than libvirt ;) | 21:13 |
| *** pdmars has quit IRC | 21:14 | |
| denis_makogon | BobBall, hm, didn't tried devstack on xen | 21:15 |
| BobBall | Don't blame you :) | 21:15 |
| BobBall | but I'm trying to understand what's going on | 21:15 |
| BobBall | because the introduction of trove into devstack seems to have broken the xenserver CI entirely | 21:15 |
| denis_makogon | BobBall, here's the gap for contribution =)) | 21:17 |
| kevinconway | i would say that the trove sample configs are fairly specific | 21:17 |
| *** jcru has quit IRC | 21:17 | |
| kevinconway | and get rewritten to work with a default devstack | 21:17 |
| kevinconway | so they may require a healthy amount of customization for anything other than that | 21:18 |
| *** jcru has joined #openstack-trove | 21:18 | |
| BobBall | That's a real bugger... | 21:18 |
| BobBall | hmmm | 21:18 |
| denis_makogon | kevinconway, true | 21:18 |
| kevinconway | BobBall: deployment documentation is one of the ongoing tasks since our mid-cycle meetup a in february | 21:19 |
| hub_cap | BobBall: thats impressive :) | 21:19 |
| hub_cap | vgnbkr: is working on it iirc | 21:19 |
| hub_cap | right amrith ? | 21:19 |
| BobBall | how far out is customisation based on devstack's config? | 21:20 |
| denis_makogon | hub_cap, i saw two patches with docs, and it means that we have more than one dev. who works on it | 21:21 |
| hub_cap | denis_makogon: not those docs | 21:21 |
| hub_cap | jaeger is working on docs | 21:21 |
| hub_cap | and so is (i foget her name) | 21:21 |
| hub_cap | but im speaking specifially of deployment docs | 21:21 |
| hub_cap | and tesora was working on the demorris | 21:21 |
| hub_cap | did densi just go ffline? | 21:22 |
| denis_makogon | get it | 21:22 |
| hub_cap | *offline | 21:22 |
| BobBall | Example failure that I believe is related to trove is http://70dcd673047f0ebed6a7-dda4e480e005aaa13ec303551d2d8155.r49.cf5.rackcdn.com/14/57614/12/results.html | 21:22 |
| SlickNik | BobBall: The guestagent config is only used for the trove Guest to talk to the server. And we don't currently have any guest related tests. | 21:22 |
| openstackgerrit | Justin Hopper proposed a change to openstack/trove-integration: Remove the iptables workaround https://review.openstack.org/84555 | 21:22 |
| BobBall | A number of tests that are failing because the server cannot be contacted - e.g. *** Not Whitelisted *** 2014-04-01 19:10:04.924 16555 TRACE trove.common.wsgi ConnectionError: HTTPConnectionPool(host='localhost', port=8774): Max retries exceeded with url: /v2/05c49d69349149f89d31aad72bdddd97/flavors/1 (Caused by <class 'socket.gaierror'>: [Errno -9] Address family for hostname not supported) | 21:22 |
| hub_cap | so BobBall there is some config stuff, and it talks to nova | 21:23 |
| hub_cap | maybe those need to be updated? | 21:23 |
| hub_cap | are they all on the same machine/network/etc. in these tests? | 21:23 |
| BobBall | all in one, yes | 21:23 |
| hub_cap | cuz its just using novaclient to chat to it | 21:23 |
| hub_cap | is there a .gz that has the configs in them BobBall ? | 21:24 |
| BobBall | sure - but how? e.g. where is the config that's getting "localhost"? | 21:24 |
| BobBall | which configs? from /etc? no.. that doesn't work currently - but I have a host that I'm looking at ATM | 21:24 |
| BobBall | SERVICE_HOST is set to 192.168.33.1 in localrc | 21:25 |
| juice | core guys - you might want to check this one out. openstackgerrit: Justin Hopper proposed a change to openstack/trove-integration: Remove the iptables workaround https://review.openstack.org/84555 | 21:25 |
| BobBall | so _everything_ should be using that, not "localhost" | 21:25 |
| hub_cap | BobBall: so assuming we accurately set this, yer right | 21:25 |
| hub_cap | the "default" is set to localhost, so if we dont set it in the devstack job then it might not be working proper | 21:26 |
| hub_cap | https://github.com/openstack/trove/blob/master/etc/trove/trove.conf.sample#L50 | 21:26 |
| hub_cap | thats what we would have to set in devstack, let me see if it does | 21:26 |
| hub_cap | yea its not BobBall | 21:26 |
| hub_cap | https://github.com/openstack-dev/devstack/blob/master/lib/trove | 21:26 |
| hub_cap | so given that, its likely using the in memory version, which is set to http://localhost.... | 21:27 |
| BobBall | ah | 21:27 |
| hub_cap | so we prolly just need to iniset it in lib/trove | 21:27 |
| SlickNik | hub_cap: Yup, I think that's the exact issue. | 21:27 |
| hub_cap | :) | 21:27 |
| * hub_cap drops the mic | 21:27 | |
| hub_cap | jkjk | 21:27 |
| hub_cap | actually | 21:27 |
| BobBall | perfect | 21:27 |
| * hub_cap runs from the issue | 21:27 | |
| * hub_cap puts thumb on forhead | 21:27 | |
| hub_cap | NOT IT | 21:27 |
| denis_makogon | guys, i've got one question to all | 21:28 |
| hub_cap | feel free to ask it | 21:28 |
| kevinconway | denis_makogon: 100 duck sized horses | 21:28 |
| hub_cap | id think horse sized ducks would be better | 21:29 |
| hub_cap | id love to ride a giant duck | 21:29 |
| denis_makogon | i'm working on updating the databases dashboard | 21:29 |
| hub_cap | http://usatthebiglead.files.wordpress.com/2012/05/389606_408713175839651_2112741493_n.jpeg | 21:29 |
| BobBall | Guess itneeds to set cinder_url and swift_url too... | 21:29 |
| hub_cap | BobBall: quite likely | 21:29 |
| SlickNik | BobBall: yeah, probably makes sense. | 21:29 |
| hub_cap | those shold be all of em tho | 21:29 |
| denis_makogon | and i came into the issue of the trove's views | 21:29 |
| denis_makogon | https://github.com/openstack/trove/blob/master/trove/instance/views.py#L39-L40 | 21:29 |
| kevinconway | hub_cap: great choice of images | 21:30 |
| hub_cap | well.. how does it talk to glance too? | 21:30 |
| hub_cap | kevinconway: ;) | 21:30 |
| SlickNik | BobBall: I take it you're proposing a patchset to devstack for this then? | 21:30 |
| * hub_cap puts thumb on forehead (again) | 21:30 | |
| hub_cap | NOT IT | 21:30 |
| denis_makogon | to add new fields into the main table, i need to be sure that i have such attributes inside each response(get/create, etc.) | 21:31 |
| hub_cap | k denis_makogon | 21:31 |
| *** robertmyers has joined #openstack-trove | 21:31 | |
| *** jcru__ has joined #openstack-trove | 21:31 | |
| *** jcru has quit IRC | 21:31 | |
| BobBall | SlickNik: https://review.openstack.org/84557 | 21:31 |
| BobBall | SlickNik: now just letting it run to see if it fixes the issues | 21:31 |
| hub_cap | BobBall: BobBaddass is more like it | 21:32 |
| denis_makogon | can i re-do view output of the get and create call ? | 21:32 |
| hub_cap | denis_makogon: u might want to ask robertmyers his opinion | 21:32 |
| denis_makogon | i need to have separate attributes for datastore and datastore_version | 21:32 |
| hub_cap | i think he did most of the orig work, im not familair | 21:32 |
| hub_cap | so my answer would not be a good one | 21:32 |
| hub_cap | but i agree, separate ds/dsversin | 21:32 |
| denis_makogon | hub_cap, i also worked with Robert on that | 21:33 |
| *** sbfox has joined #openstack-trove | 21:33 | |
| hub_cap | cool | 21:33 |
| hub_cap | then your answer will be better than mine denis_makogon ;) | 21:33 |
| denis_makogon | hub_cap, i'm asking you 'cuz i need to do changes in trove and troveclient | 21:33 |
| hub_cap | ohhhh | 21:33 |
| hub_cap | i guess im confused then | 21:33 |
| SlickNik | BobBall: Cool, thanks! | 21:33 |
| hub_cap | they are already 2 fields | 21:33 |
| hub_cap | why do we need "changes" to trove+client ? | 21:34 |
| robertmyers | hub_cap: denis_makogon: what are we talking about... just got here | 21:34 |
| BobBall | SlickNik: as sdague said in #openstack-qa it needs a tweak | 21:34 |
| hub_cap | crap BobBall we might need to add those to the taskmgr too | 21:34 |
| hub_cap | hey robertmyers u down in the wasteland? | 21:34 |
| robertmyers | yup | 21:34 |
| hub_cap | ill be there tomorrow, round lunch i think | 21:35 |
| robertmyers | someone is giving a tutorial on Trove here | 21:35 |
| denis_makogon | hub_cap, https://github.com/openstack/trove/blob/master/trove/instance/views.py#L32-L46 - as i can see, view doesn't have separate field for datastore_version, it's the part of the datastore field | 21:35 |
| robertmyers | not sure who | 21:35 |
| hub_cap | BobBall: https://github.com/openstack/trove/blob/master/etc/trove/trove-taskmanager.conf.sample#L33 | 21:35 |
| hub_cap | well its 2 sub-fields right? | 21:35 |
| denis_makogon | robertmyers, i'm working on dashboard update | 21:35 |
| robertmyers | Florin Haas? | 21:35 |
| denis_makogon | hub_cap, yes | 21:35 |
| hub_cap | robertmyers: yup | 21:36 |
| hub_cap | from hastexo | 21:36 |
| denis_makogon | hub_cap, but horizon cannot investigate sub-fields | 21:36 |
| BobBall | don't understand that hub_cap ... | 21:36 |
| hub_cap | BobBall: we might be doing the ccommunication to nova/glance etc via the taskmgr, not the api server | 21:36 |
| robertmyers | denis_makogon: what are you adding? | 21:36 |
| hub_cap | so i wonder, if we need to, instead, iniset the taskmgr conf | 21:36 |
| hub_cap | robertmyers: denis_makogon says he needs to break up the trove api datastore into datastore and datastore_version | 21:37 |
| hub_cap | instead of datastore {type, version} | 21:37 |
| hub_cap | denis_makogon: we cant do that | 21:37 |
| denis_makogon | robertmyers, datastore and datastore version columns at main table + selecting datastore version at provisioning | 21:37 |
| hub_cap | we are integrated now | 21:37 |
| hub_cap | we cant just _change_ the api | 21:37 |
| hub_cap | u know this :) | 21:37 |
| BobBall | :( that sounds harder than I can do hub_cap ... I know nothing about trove! only devstack etc :) | 21:37 |
| hub_cap | BobBall: iniset the 2nd file instead of trove.conf | 21:37 |
| hub_cap | sec | 21:37 |
| hub_cap | oh wait BobBall they use the same in devstack? | 21:38 |
| denis_makogon | hub_cap, how does response body affects API ? | 21:38 |
| hub_cap | cuz its the RESPONSE | 21:38 |
| hub_cap | which is 1/2 of the api contract | 21:38 |
| BobBall | oh - that simple? cool | 21:38 |
| hub_cap | the other have is... request | 21:38 |
| hub_cap | thats _is_ the contract denis_makogon | 21:38 |
| hub_cap | just changing something change sthe contract | 21:38 |
| hub_cap | BobBall: let me double check | 21:38 |
| hub_cap | ;) | 21:38 |
| denis_makogon | well, it's bad | 21:38 |
| hub_cap | denis_makogon: its impossible to change | 21:39 |
| hub_cap | until we have a V2 a;i | 21:39 |
| hub_cap | api | 21:39 |
| hub_cap | weve said this for a long time, when we do things, we have to live with them :) | 21:39 |
| hub_cap | BobBall: bad news! there is a trove-taskmgr.conf | 21:39 |
| hub_cap | line 161 | 21:39 |
| hub_cap | so u might have to add those all there too | 21:39 |
| robertmyers | denis_makogon: why can't you work with the response, can't you have a special filter on the table view? | 21:39 |
| hub_cap | https://github.com/openstack-dev/devstack/blob/master/lib/trove#L161 | 21:40 |
| hub_cap | BobBall: ^ ^ | 21:40 |
| denis_makogon | hub_cap, then how do we let such mistake inside the response ? | 21:40 |
| hub_cap | denis_makogon: who cares | 21:40 |
| hub_cap | whats done is done | 21:40 |
| hub_cap | no use trying to figure out _how_ we did it | 21:40 |
| SlickNik | hub_cap / BobBall: Yeah, we need to iniset those values in both confs (trove.conf and trove-taskmanager.conf). | 21:40 |
| hub_cap | when andrew designed the spec denis_makogon , we all allowed it | 21:40 |
| denis_makogon | hub_cap, we should __avoid__ such things with new features | 21:40 |
| hub_cap | denis_makogon: duh | 21:40 |
| hub_cap | :) | 21:40 |
| hub_cap | but to say "we should never make a mistake" is silly | 21:41 |
| hub_cap | id love ot say that | 21:41 |
| hub_cap | but that wont happen... | 21:41 |
| hub_cap | so we deal w/ it | 21:41 |
| hub_cap | like robertmyers said, do a custom filter | 21:41 |
| hub_cap | if thats possible | 21:41 |
| hub_cap | if not, why not fix the _limitation_ in horizon instead? | 21:41 |
| robertmyers | its not ideal, but as long as the data you need is there it shouldnt be a blocker | 21:42 |
| hub_cap | ++ robertmyers | 21:42 |
| hub_cap | its somethign we fix in v2 | 21:42 |
| hub_cap | simple as that :) | 21:42 |
| *** sgotliv has quit IRC | 21:42 | |
| hub_cap | and we deal w it for now | 21:42 |
| denis_makogon | hub_cap, do we have filed plan for the API v2 ? | 21:43 |
| hub_cap | nope | 21:43 |
| *** sbfox has quit IRC | 21:43 | |
| hub_cap | i dont thikn a v2 api is worht filing until we need it | 21:43 |
| hub_cap | and this is not a _neeD_ for a v2 api | 21:43 |
| hub_cap | to fix 1 thing in horizon | 21:43 |
| hub_cap | so thats a long time down the road | 21:44 |
| denis_makogon | hub_cap, we could start filling it, and use it when we need | 21:44 |
| hub_cap | itll just rot | 21:44 |
| *** sbfox has joined #openstack-trove | 21:44 | |
| hub_cap | just like the other "filed but not needed" bps | 21:44 |
| hub_cap | feel free to | 21:44 |
| hub_cap | ill kick it to wishlist | 21:44 |
| hub_cap | we can start storing info on it that suxk about v1 | 21:44 |
| hub_cap | like this | 21:44 |
| hub_cap | and like the string vs int for flavors limitation | 21:45 |
| hub_cap | that is the only one i know of as of now | 21:45 |
| *** mgriffin has joined #openstack-trove | 21:46 | |
| BobBall | hub_cap: https://review.openstack.org/#/c/84557/ ? | 21:46 |
| hub_cap | BobBall: that seems safe :) | 21:46 |
| hub_cap | awesome work | 21:46 |
| hub_cap | i guess we will chat in 6h | 21:46 |
| hub_cap | :P | 21:47 |
| BobBall | ok - hopefully jenkins will have done it's magic by then ;) | 21:47 |
| hub_cap | hehe hopefully | 21:47 |
| hub_cap | thx again BobBall u da man | 21:47 |
| vipul | horizon can't handle a json response that has sub attributes??! | 21:48 |
| vipul | that seems like a pretty big restrictions to me denis_makogon | 21:48 |
| vipul | i bet you most of the openstack api responses are not a flat dictionary | 21:48 |
| hub_cap | vipul: thts why i said to fix the limitation in horizon, instead of rev'ing the api for trove LOL | 21:48 |
| vipul | hub_cap: yea we can't be changing api just to make horizon integration easier | 21:49 |
| robertmyers | vipul: it can | 21:49 |
| robertmyers | it is just not straight forward | 21:49 |
| robertmyers | but there are plenty of examples | 21:49 |
| vipul | robertmyers: ah ok.. problem solved :) | 21:49 |
| robertmyers | https://github.com/openstack/horizon/blob/master/openstack_dashboard/dashboards/project/databases/tables.py#L138 | 21:50 |
| hub_cap | vipul: robertmyers "flavor": {"id":1....} nuff said | 21:50 |
| hub_cap | in novas get server details | 21:50 |
| robertmyers | basically a helper method to extract the data | 21:50 |
| hub_cap | so um.... im pretty sure its possible, right robertmyers ;) | 21:50 |
| robertmyers | yes | 21:51 |
| *** kevinconway_ has joined #openstack-trove | 21:51 | |
| denis_makogon | ok | 21:51 |
| denis_makogon | now, cores, please review this one, which is critical https://review.openstack.org/#/c/83768/ https://review.openstack.org/#/c/84115/ for RC1 | 21:52 |
| hub_cap | SlickNik: is testing it now denis_makogon | 21:52 |
| *** mayfield has joined #openstack-trove | 21:53 | |
| *** mayfield has joined #openstack-trove | 21:53 | |
| *** cweid_ has joined #openstack-trove | 21:54 | |
| denis_makogon | hub_cap, i know that, but didn't received any response from SlickNik, that is why i'm asking you to take a look at the code | 21:54 |
| hub_cap | denis_makogon: we know abou tit | 21:54 |
| *** openstackgerrit has quit IRC | 21:54 | |
| *** kevinconway has quit IRC | 21:54 | |
| *** mayfield_ has quit IRC | 21:54 | |
| *** cweid has quit IRC | 21:54 | |
| *** kevinconway_ has quit IRC | 21:54 | |
| *** kevinconway_ has joined #openstack-trove | 21:54 | |
| *** cweid_ has quit IRC | 21:54 | |
| *** cweid_ has joined #openstack-trove | 21:54 | |
| *** kevinconway_ is now known as kevinconway | 21:54 | |
| hub_cap | *about it | 21:54 |
| hub_cap | its the highest prio for core now to get rc1 out | 21:54 |
| SlickNik | denis_makogon: I didn't know you were waiting on a response from me for something? Did you ask me a question? | 21:55 |
| denis_makogon | SlickNik, i mean "the response" after testing | 21:55 |
| denis_makogon | SlickNik, is it ok or not | 21:55 |
| SlickNik | denis_makogon: Still testing, like hub_cap said. | 21:56 |
| hub_cap | denis_makogon: it seems ther are some updates for you to do this.. SlickNik are you wanting him to update it? | 21:56 |
| hub_cap | the things mat-lowery pointed out in particular | 21:56 |
| *** demorris has quit IRC | 21:56 | |
| denis_makogon | hub_cap, i'm waiting for the full feedback and only then i'm going to update review | 21:56 |
| hub_cap | denis_makogon: im not sure we can afford to do that... u might be asleep when we do it | 21:57 |
| hub_cap | we have to get this fix in by tomorrow morning your time | 21:57 |
| hub_cap | plz update it | 21:57 |
| hub_cap | SlickNik: thoughts? | 21:57 |
| *** Barker has quit IRC | 21:58 | |
| SlickNik | I'd like to see comments in reviews addressed in a timely manner. | 21:58 |
| hub_cap | i think thats a general rule of thumb | 21:59 |
| SlickNik | It's really hard to tell when you've received "full feedback". | 21:59 |
| *** demorris has joined #openstack-trove | 22:00 | |
| denis_makogon | SlickNik, for me "full feedback" means that all cores reviewed that patch | 22:02 |
| SlickNik | But denis_makogon: All cores don't have to review a patch. | 22:02 |
| hub_cap | all cores _never_ review a patch denis_makogon | 22:03 |
| hub_cap | u only need 2 cores to review to land | 22:03 |
| denis_makogon | hub_cap, SlickNik isn't this some kind of policy ? | 22:03 |
| hub_cap | denis_makogon: its never been so | 22:03 |
| denis_makogon | hub_cap, infra expects to have two approves | 22:03 |
| denis_makogon | hub_cap, but the patches deserves to be reviewed by the core team | 22:04 |
| hub_cap | denis_makogon: show me where you see that in writing | 22:05 |
| hub_cap | and ill agree w/ you | 22:05 |
| hub_cap | talk to nova | 22:05 |
| hub_cap | aks them if every core reviews ever patch | 22:05 |
| hub_cap | and ill agre w/ you | 22:05 |
| hub_cap | thats silly | 22:05 |
| hub_cap | its _neveR_ been like that | 22:06 |
| hub_cap | ever | 22:06 |
| hub_cap | no team | 22:06 |
| hub_cap | i really dont know where this is coming from man... ive never said that, nor have any cores agreed to this notion of "full feedback" | 22:06 |
| denis_makogon | fine, patch will be ready by the 8-11AM PST | 22:07 |
| denis_makogon | cu tomorrow | 22:08 |
| *** denis_makogon has quit IRC | 22:08 | |
| hub_cap | man that came out of left field... | 22:08 |
| juice | speaking of Florian Haas... | 22:09 |
| juice | :) | 22:09 |
| juice | I tried out his squid proxy setup | 22:09 |
| juice | but moved it to my laptop/host so I get the benefit between boots | 22:09 |
| hub_cap | sweet | 22:09 |
| juice | I would like some comparison but my own test is about 15 minute time savings | 22:10 |
| robertmyers | So how is he connected to Trove? | 22:10 |
| robertmyers | just never heard his name before | 22:10 |
| juice | vagrant up through kick-start mysql | 22:10 |
| hub_cap | robertmyers: he works for (owns?) hastexo | 22:10 |
| juice | he works for an openstack education company | 22:10 |
| juice | they do training | 22:10 |
| robertmyers | ahh | 22:10 |
| hub_cap | they do cloud tra... dammit juice ;) | 22:10 |
| juice | lol :) | 22:11 |
| hub_cap | u beeeet me | 22:11 |
| juice | I am wearing my pro keds today | 22:11 |
| vipul | yea it's kinda interesting that they've been building the tooling without really being involved in this room | 22:11 |
| juice | they make me type faster | 22:11 |
| *** achampion has quit IRC | 22:11 | |
| *** dmitryme has joined #openstack-trove | 22:11 | |
| juice | I don't think he is really doing any tooling - he just wanted something he can give the audience so that they can follow along for his demo | 22:12 |
| dougshelley66 | we have been working with them and announced a partnership with hastexo the other day | 22:12 |
| hub_cap | dougshelley66: killer!! | 22:12 |
| vipul | oh cool | 22:12 |
| juice | dougshelley66 but I thought WE were your partners :`( | 22:12 |
| dougshelley66 | of course you are | 22:12 |
| hub_cap | sans that ass whos stepping down | 22:12 |
| hub_cap | :P | 22:12 |
| juice | yeah except THAT guy | 22:13 |
| dougshelley66 | hey, we don't discriminate | 22:13 |
| juice | sheesh | 22:13 |
| hub_cap | LOL | 22:13 |
| juice | yeah we are all losers... | 22:13 |
| juice | I mean winners | 22:13 |
| dougshelley66 | have to go play taxi ...back in a bit | 22:14 |
| juice | so the reason for bringing up the squid thingy is that I think it would make a wonderful enhancement to our gate tests on rdjenkins | 22:14 |
| juice | ...and I am sad that it took me so long to hop on the vagrant-bus | 22:14 |
| robertmyers | juice: how so? | 22:15 |
| hub_cap | robertmyers: cuz its squishy | 22:15 |
| robertmyers | I get that caching is faster, but what do we need to cache? | 22:15 |
| *** Barker has joined #openstack-trove | 22:16 | |
| juice | all the packages that we download for each every build | 22:16 |
| juice | robertmyers | 22:16 |
| robertmyers | ahh | 22:16 |
| juice | ohhhh yeahhhh | 22:16 |
| juice | :) | 22:16 |
| robertmyers | well, yes, or we could use a preloaded instance | 22:16 |
| robertmyers | setup devstack and image it | 22:17 |
| robertmyers | update that once a day or so | 22:17 |
| juice | robertmyers: I am good with anything that will help speed up the "full feedback" cycle :) | 22:17 |
| robertmyers | hahahaha | 22:18 |
| hub_cap | ZOMGGGG | 22:19 |
| juice | robertmyers: this appeared to be simple enough to implement so I thought I would suggest that here | 22:19 |
| esmute | Hi all, i am looking ways to get backups from one region to another. Here is a BP that i just put it. Let me know what you guys think. https://blueprints.launchpad.net/trove/+spec/cross-region-backup-migration | 22:19 |
| juice | esmute: I am talking now! you have to wait your turn :) | 22:19 |
| esmute | juice: i hope i followed the right way to create BP | 22:19 |
| vipul | sounds like node pool :) | 22:20 |
| juice | hub_cap: is the blueprint connoisseur | 22:20 |
| esmute | haha .. sorry juice... didnt mean to hijack your convo | 22:20 |
| robertmyers | juice: it wouldn't hurt to try it out | 22:20 |
| vipul | well at some point if we run this stuff in infra like we had planned.. this wouldn't be as painful | 22:20 |
| vipul | i guess the question is how much longer does rdjenkins need to be around | 22:20 |
| robertmyers | well seeings how it is RD | 22:21 |
| robertmyers | it should go away | 22:21 |
| vipul | lol yea or at least renamed | 22:22 |
| hub_cap | vipul: infra said once we have gating jobs we can move it under their umbrella | 22:22 |
| hub_cap | so maybe if ew ask nicely we can "fix" it | 22:22 |
| hub_cap | juice: do i detect a hint of peet in that blueprint | 22:22 |
| hub_cap | so cherries, bing of course | 22:22 |
| hub_cap | *some | 22:22 |
| hub_cap | maybe sandlewood | 22:22 |
| SlickNik | I'm inclined to create a non-tempest gate job that does what rdjenkins is doing today, and move it under CI infra. | 22:23 |
| hub_cap | SlickNik: ask first, i got railed last time i asked | 22:23 |
| hub_cap | "not before u do the other stuff" | 22:23 |
| hub_cap | was the answer ;) | 22:23 |
| SlickNik | hub_cap: Yeah, will definitely ask first. | 22:23 |
| juice | hub_cap: oak maybe and some speyside peet yes | 22:23 |
| hub_cap | wtf is monty not online?!?!? | 22:23 |
| hub_cap | juice: maybe a hint of purp.... oh wait | 22:24 |
| SlickNik | juice: islay, please... | 22:24 |
| SlickNik | That's where all the peat is at. | 22:24 |
| hub_cap | u konw SlickNik , the super peety ones i cant do | 22:24 |
| hub_cap | smells like the simpsons tire fire in my nose before i drink it | 22:24 |
| SlickNik | hub_cap: You gotta work your way up. | 22:25 |
| *** jcru__ has quit IRC | 22:25 | |
| juice | hub_cap: naughty naughty | 22:25 |
| hub_cap | haha | 22:25 |
| hub_cap | to both of u | 22:25 |
| *** kevinconway has quit IRC | 22:28 | |
| *** amytron has quit IRC | 22:32 | |
| *** demorris has quit IRC | 22:33 | |
| robertmyers | well my beer is almost out, I have a fridge in my room here that needs to cool something | 22:36 |
| robertmyers | talk to you all later! | 22:36 |
| *** Barker has quit IRC | 22:36 | |
| *** khyati has quit IRC | 22:36 | |
| *** jasonb365 has quit IRC | 22:37 | |
| *** BobBall_ has joined #openstack-trove | 22:38 | |
| *** BobBall has quit IRC | 22:38 | |
| *** BobBall_ has quit IRC | 22:41 | |
| *** khyati has joined #openstack-trove | 22:42 | |
| *** thedodd has quit IRC | 22:43 | |
| *** BobBall has joined #openstack-trove | 22:43 | |
| *** BobBall is now known as Guest86798 | 22:44 | |
| *** jdbarry_ has joined #openstack-trove | 22:50 | |
| *** grapex has quit IRC | 22:51 | |
| hub_cap | cu robertmyers | 22:52 |
| *** khyati has quit IRC | 22:56 | |
| *** yogesh has quit IRC | 22:57 | |
| dmitryme | Hey Trove devs! I am working on unified guest agent proposal. Here is the link: https://wiki.openstack.org/wiki/UnifiedGuestAgent | 23:10 |
| dmitryme | I know you guys already have a guest agent implemented and working | 23:10 |
| dmitryme | so I'd like to hear your opinion on my current design | 23:11 |
| dmitryme | so that maybe someday the design could be adopted by Trove as well | 23:12 |
| dmitryme | how about discussing that at the next Trove team meeting? | 23:12 |
| dmitryme | As far as I understand, the next meeting is going to happen tomorrow at 18 UTC, right? | 23:14 |
| hub_cap | llets give us time to review first dmitryme | 23:16 |
| dmitryme | hub_cap: sure thing | 23:16 |
| dmitryme | I also wanted to give you guys a little background: initially I started developing an agent for Sahara in a design pretty close to Trove's agent. The main difference is that I based agent on oslo.messaging, not on oslo.rpc | 23:19 |
| dmitryme | but after some time in a discussion | 23:20 |
| dmitryme | http://lists.openstack.org/pipermail/openstack-dev/2014-March/029194.html | 23:20 |
| dmitryme | we realized that such a design has a flaw which is hard to fix | 23:20 |
| dmitryme | oslo.messaging use RabbitMQ or Qpid as a backend | 23:21 |
| hub_cap | yea we dislike this :) | 23:21 |
| dmitryme | ao the agent running on the VM works directly with the queue | 23:21 |
| hub_cap | yes weve had this conversation in trove as well | 23:22 |
| hub_cap | what is yoru proposal for a different communication protocol? | 23:22 |
| hub_cap | im curious | 23:22 |
| dmitryme | actually in the discussion it was proposed to use Marconi as a backend | 23:22 |
| dmitryme | and after some thinking I realized that it is indeed a fine candidate | 23:23 |
| dmitryme | it supports multi-tenancy out of the box | 23:23 |
| hub_cap | makeit pluggable | 23:23 |
| hub_cap | we want to use xmpp :) | 23:23 |
| dmitryme | hub_cap: lol :-) | 23:23 |
| hub_cap | xmpp is well suited for this use case | 23:24 |
| dmitryme | yes, make it pluggable is exactly my thinking | 23:24 |
| hub_cap | awesome | 23:25 |
| dmitryme | I've made a stupid marconi driver for oslo.messaging | 23:25 |
| hub_cap | lol | 23:26 |
| dmitryme | the only problem is that it is not interchangeable with the other drivers | 23:26 |
| dmitryme | which is not surprising, since you need to pass at tenant and credentials as a part of the Target | 23:27 |
| dmitryme | while 'canonical' drivers take creds directly from config file | 23:27 |
| hub_cap | sure | 23:27 |
| hub_cap | im not sure we care too much about multi tenant tho personally | 23:27 |
| hub_cap | or at least i dont | 23:27 |
| dmitryme | hub_cap: from what I've heard Trove don't give users direct access on the VMs, right? | 23:28 |
| hub_cap | whats the need/use case | 23:28 |
| hub_cap | right | 23:28 |
| dmitryme | that is not the same for Savanna, so that is why we care | 23:28 |
| hub_cap | see thats my concern for a "generic guest" | 23:29 |
| hub_cap | is that every project has its own requirements and all of the guests will be drastically diffrerent | 23:29 |
| hub_cap | ill definitely read the proposal | 23:29 |
| hub_cap | i just dont want it to be "the guest takes arbitrary commands from the infrastructure and does it" | 23:29 |
| hub_cap | and is really dumb | 23:29 |
| dmitryme | I am not that pessimistic :-) | 23:30 |
| hub_cap | our guest has to understand state of the vm | 23:30 |
| hub_cap | and fix things when they go wrong | 23:30 |
| juice | dmitryme - I read the proposal but I am not getting what/how this is different than what we are doing? | 23:30 |
| hub_cap | what ive seen in developing large scaple systems is that if i make the "infra" decide everything, then i have a huge bottleneck | 23:30 |
| hub_cap | mind you dmitryme , i have not read your propsal... | 23:30 |
| hub_cap | *proposal | 23:30 |
| juice | dmitryme: I'm not being rude - just not sure what the benefits would be or how it would be better than what we have? | 23:31 |
| juice | dmitryme:...other than reuse of some building blocks. | 23:32 |
| dmitryme | juice: sure, let me explain. The idea is not to have a fixed binary working for all the projects, but to use the same approach across the projects | 23:32 |
| juice | ...and yes different transports | 23:32 |
| juice | dmitryme: can you give me something concrete where it would be the same approach? | 23:33 |
| dmitryme | juice: use the same transport across several projects | 23:33 |
| dmitryme | I think deployers would value that | 23:33 |
| dmitryme | I mean they would value not keeping a zoo of different backends with different settings | 23:34 |
| hub_cap | if thats the only reason i think we can agree on an approace and take it | 23:34 |
| dmitryme | ok guys, thanks for talking | 23:40 |
| juice | dmitryme: I am with hub_cap on that one, I think the consistency from a user perspective would be great value | 23:43 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
