| *** matsuhashi has joined #openstack-trove | 00:12 | |
| *** matsuhashi has quit IRC | 00:31 | |
| *** rhodgin has joined #openstack-trove | 00:36 | |
| *** matsuhashi has joined #openstack-trove | 00:52 | |
| *** nosnos has joined #openstack-trove | 01:30 | |
| *** shakamunyi has joined #openstack-trove | 01:38 | |
| *** robertmyers has quit IRC | 01:51 | |
| *** robertmyers has joined #openstack-trove | 01:52 | |
| *** haomaiw__ has quit IRC | 02:17 | |
| *** haomaiwa_ has joined #openstack-trove | 02:18 | |
| *** haomai___ has joined #openstack-trove | 02:31 | |
| *** haomaiwa_ has quit IRC | 02:35 | |
| *** erkules_ has joined #openstack-trove | 02:38 | |
| *** erkules has quit IRC | 02:40 | |
| *** matsuhashi has quit IRC | 02:48 | |
| *** matsuhas_ has joined #openstack-trove | 02:52 | |
| *** dkehn has joined #openstack-trove | 03:04 | |
| *** dkehn_ has quit IRC | 03:06 | |
| *** ViswaV has joined #openstack-trove | 03:09 | |
| *** mrsnivvel has joined #openstack-trove | 03:16 | |
| *** matsuhas_ has quit IRC | 03:22 | |
| *** dkehn is now known as dkehn_ | 03:23 | |
| *** ViswaV has quit IRC | 03:45 | |
| *** robertmyers has quit IRC | 03:47 | |
| *** ViswaV has joined #openstack-trove | 03:49 | |
| *** rhodgin has quit IRC | 03:56 | |
| *** achampion has quit IRC | 04:10 | |
| *** ViswaV has quit IRC | 04:23 | |
| openstackgerrit | SlickNik proposed a change to openstack/python-troveclient: Get rid of XML related trove client bindings https://review.openstack.org/69999 | 04:34 |
|---|---|---|
| *** achampion has joined #openstack-trove | 04:36 | |
| *** matsuhashi has joined #openstack-trove | 04:43 | |
| *** achampion has quit IRC | 04:47 | |
| *** erkules_ is now known as erkules | 04:50 | |
| *** achampion has joined #openstack-trove | 05:43 | |
| *** achampion has quit IRC | 05:47 | |
| *** nosnos has quit IRC | 05:54 | |
| *** nosnos has joined #openstack-trove | 05:54 | |
| *** matsuhashi has quit IRC | 05:56 | |
| *** matsuhashi has joined #openstack-trove | 05:57 | |
| *** matsuhashi has quit IRC | 05:57 | |
| *** matsuhashi has joined #openstack-trove | 06:00 | |
| *** matsuhashi has quit IRC | 06:09 | |
| *** matsuhashi has joined #openstack-trove | 06:10 | |
| *** matsuhas_ has joined #openstack-trove | 06:12 | |
| *** matsuhashi has quit IRC | 06:14 | |
| *** nosnos has quit IRC | 06:24 | |
| *** nosnos has joined #openstack-trove | 06:25 | |
| *** matsuhas_ has quit IRC | 06:25 | |
| *** matsuhashi has joined #openstack-trove | 06:25 | |
| *** matsuhashi has quit IRC | 06:39 | |
| *** matsuhashi has joined #openstack-trove | 06:40 | |
| *** saju_m has joined #openstack-trove | 06:47 | |
| *** rongze has joined #openstack-trove | 07:21 | |
| *** rongze has quit IRC | 07:23 | |
| *** flaper87|afk is now known as flaper87 | 07:30 | |
| *** wrongfan has joined #openstack-trove | 07:50 | |
| *** nosnos has quit IRC | 08:13 | |
| *** nosnos has joined #openstack-trove | 08:14 | |
| *** matsuhashi has quit IRC | 08:17 | |
| *** matsuhashi has joined #openstack-trove | 08:17 | |
| *** openstack has quit IRC | 08:21 | |
| *** openstack has joined #openstack-trove | 08:29 | |
| *** openstackstatus has joined #openstack-trove | 08:30 | |
| *** nosnos has quit IRC | 08:34 | |
| *** matsuhas_ has quit IRC | 08:56 | |
| *** matsuhashi has joined #openstack-trove | 09:06 | |
| *** IvanZ has joined #openstack-trove | 09:10 | |
| *** nosnos has joined #openstack-trove | 09:14 | |
| *** dukhlov has joined #openstack-trove | 09:28 | |
| *** wrongfan has quit IRC | 09:49 | |
| *** wrongfan has joined #openstack-trove | 09:50 | |
| *** nosnos_ has joined #openstack-trove | 09:55 | |
| *** nosnos has quit IRC | 09:58 | |
| *** tattabbum has joined #openstack-trove | 10:04 | |
| *** nosnos_ has quit IRC | 10:10 | |
| *** nosnos has joined #openstack-trove | 10:12 | |
| *** saju_m has quit IRC | 10:17 | |
| *** iartarisi has joined #openstack-trove | 10:24 | |
| *** saju_m has joined #openstack-trove | 10:32 | |
| *** matsuhashi has quit IRC | 10:38 | |
| *** nosnos has quit IRC | 10:46 | |
| *** nosnos has joined #openstack-trove | 10:47 | |
| *** matsuhashi has joined #openstack-trove | 10:54 | |
| *** matsuhashi has quit IRC | 11:05 | |
| *** openstackgerrit has quit IRC | 11:10 | |
| *** openstackgerrit has joined #openstack-trove | 11:10 | |
| *** matsuhashi has joined #openstack-trove | 11:13 | |
| *** saju_m has quit IRC | 11:15 | |
| *** tattabbum has left #openstack-trove | 11:21 | |
| *** asterix has joined #openstack-trove | 11:22 | |
| *** matsuhashi has quit IRC | 11:31 | |
| *** saju_m has joined #openstack-trove | 11:36 | |
| asterix | Hello all. I would build a trove image (like ubuntu_mysql.qcow2 image) without using redstack. I would upload this image in an exting OpenStack environment. Is this scenario possible' | 12:00 |
| asterix | ? | 12:00 |
| *** pdmars has joined #openstack-trove | 12:12 | |
| *** pdmars has quit IRC | 12:14 | |
| *** pdmars has joined #openstack-trove | 12:20 | |
| *** matsuhashi has joined #openstack-trove | 12:27 | |
| *** matsuhashi has quit IRC | 12:37 | |
| *** nosnos_ has joined #openstack-trove | 12:37 | |
| *** nosnos has quit IRC | 12:37 | |
| *** matsuhashi has joined #openstack-trove | 12:37 | |
| *** matsuhashi has quit IRC | 12:42 | |
| *** asterix has quit IRC | 12:50 | |
| openstackgerrit | A change was merged to openstack/trove: Hide Datastores In List With No Active Versions https://review.openstack.org/79513 | 12:56 |
| *** matsuhashi has joined #openstack-trove | 13:03 | |
| *** matsuhashi has quit IRC | 13:05 | |
| *** matsuhashi has joined #openstack-trove | 13:05 | |
| *** radez_g0n3 is now known as radez | 13:07 | |
| *** matsuhashi has quit IRC | 13:10 | |
| *** wrongfan has quit IRC | 13:12 | |
| *** nosnos_ has quit IRC | 13:12 | |
| *** mrsnivvel has quit IRC | 13:14 | |
| *** aterix has joined #openstack-trove | 13:18 | |
| *** rhodgin has joined #openstack-trove | 13:21 | |
| *** robertmyers has joined #openstack-trove | 13:34 | |
| *** robertmyers has quit IRC | 13:35 | |
| *** ViswaV has joined #openstack-trove | 13:37 | |
| *** robertmy_ has joined #openstack-trove | 13:39 | |
| openstackgerrit | Ionut Artarisi proposed a change to openstack/trove: fix default rabbitmq configuration values in sample cfgs https://review.openstack.org/74638 | 13:40 |
| *** robertmy_ has quit IRC | 13:40 | |
| *** robertmy_ has joined #openstack-trove | 13:40 | |
| *** robertmy_ has quit IRC | 13:41 | |
| *** robertmyers has joined #openstack-trove | 13:41 | |
| *** robertmyers has quit IRC | 13:42 | |
| *** robertmyers has joined #openstack-trove | 13:43 | |
| *** robertmy_ has joined #openstack-trove | 13:44 | |
| *** robertmyers has quit IRC | 13:44 | |
| *** achampion has joined #openstack-trove | 13:47 | |
| *** jcru has joined #openstack-trove | 13:48 | |
| *** Barker has joined #openstack-trove | 13:57 | |
| *** rramirez has joined #openstack-trove | 13:57 | |
| *** amytron has quit IRC | 14:00 | |
| *** rwsu has joined #openstack-trove | 14:07 | |
| *** kevinconway has joined #openstack-trove | 14:13 | |
| *** spredzy has joined #openstack-trove | 14:14 | |
| *** grapex has joined #openstack-trove | 14:16 | |
| *** grapex has quit IRC | 14:18 | |
| *** grapex has joined #openstack-trove | 14:19 | |
| *** flaper87 has quit IRC | 14:26 | |
| *** amytron has joined #openstack-trove | 14:26 | |
| *** flaper87|afk has joined #openstack-trove | 14:31 | |
| *** flaper87|afk is now known as flaper87 | 14:32 | |
| *** flaper87 has quit IRC | 14:32 | |
| *** flaper87 has joined #openstack-trove | 14:32 | |
| *** datsun180b has joined #openstack-trove | 14:34 | |
| grapex | uh-oh, yet another wealthy benefactor is trying to get in touch with me: https://gist.github.com/TimSimpsonR/9600609 | 14:46 |
| *** Barker has quit IRC | 14:59 | |
| *** thedodd has joined #openstack-trove | 15:01 | |
| *** saju_m has quit IRC | 15:03 | |
| *** iartarisi has quit IRC | 15:08 | |
| *** saju_m has joined #openstack-trove | 15:20 | |
| aterix | Hello all. I would build a trove image (like ubuntu_mysql.qcow2 image) without using redstack. I would upload this image in an exting OpenStack environment. Is this scenario possible? | 15:22 |
| *** shakamunyi has quit IRC | 15:31 | |
| *** ViswaV has quit IRC | 15:33 | |
| *** flaper87 is now known as flaper87|afk | 15:34 | |
| cp16net | hello all | 15:45 |
| *** Barker has joined #openstack-trove | 15:53 | |
| *** jmontemayor has joined #openstack-trove | 15:57 | |
| *** jmontemayor_ has joined #openstack-trove | 16:05 | |
| *** jmontemayor_ has quit IRC | 16:06 | |
| *** SnowDust has joined #openstack-trove | 16:06 | |
| *** ViswaV has joined #openstack-trove | 16:06 | |
| *** jmontemayor_ has joined #openstack-trove | 16:06 | |
| *** ViswaV_ has joined #openstack-trove | 16:07 | |
| *** ViswaV has quit IRC | 16:07 | |
| *** jmontemayor has quit IRC | 16:08 | |
| *** ViswaV_ has quit IRC | 16:08 | |
| *** ViswaV has joined #openstack-trove | 16:08 | |
| *** saju_m has quit IRC | 16:15 | |
| vgnbkr | Happy Saint Patty's Day! | 16:16 |
| hub_cap | aterix: it should be possible, sure, just upload it to glance, and name it aptly, like ubuntu_{datastore_type} (ubuhtu_mysql) in glance | 16:19 |
| hub_cap | oh and howdy all, im back from the mountains | 16:19 |
| hub_cap | whre teh wind was plentiful and he internet was none | 16:19 |
| grapex | hub_cap: Sounds awful | 16:21 |
| vgnbkr | hub_cap: I'm working on what you describe above right now. So how would the trove-guestagent (etc) get installed when you boot the trove instance? | 16:22 |
| hub_cap | vgnbkr: well thats a scenario we solve in the dev env by ssh keys and rsync | 16:24 |
| hub_cap | if u look into how the image is made, we use the same username as the host system, and we put a ssh key into the vm so the developer / rsync can get in | 16:25 |
| vgnbkr | Right, but how is it supposed to work in a real OpenStack environment. | 16:25 |
| hub_cap | but this is only for dev purposes :) | 16:25 |
| hub_cap | vgnbkr: most people ive talked to package the guest in the vm | 16:25 |
| vgnbkr | Right, but there's nothing in Trove to tell you what you need to do for that, or any tools to help you, correct? | 16:26 |
| hub_cap | correct. we have little/no deployment docs | 16:27 |
| hub_cap | there may be something in the manual install docs | 16:27 |
| vgnbkr | OK, that's basically what I surmised, but just wanted to make sure I hadn't missed something. | 16:28 |
| hub_cap | but vgnbkr we dont really build packages in openstack | 16:28 |
| hub_cap | its up to deployers to do that, or external community peoples | 16:28 |
| hub_cap | weve also wanted to split out the guest for a while, makign this less of a problem (guest gets its own pkg as well) | 16:28 |
| hub_cap | there is also tarballs.o.o | 16:28 |
| hub_cap | http://tarballs.openstack.org/trove/ | 16:29 |
| hub_cap | and that has trove-master, so u could pull it from there and just pip install it, or use virtualenv | 16:29 |
| hub_cap | ther are a few ways to skin that cat hehe | 16:29 |
| hub_cap | vgnbkr: do u know if anyone on yer end is starting any deployment docs? this kind of thing would be good to put in the docs id think | 16:31 |
| vgnbkr | My understanding is that we would like to, but someone would have to tell the doc writer how to deploy it, which is why I'm trying to figure it out. | 16:32 |
| openstackgerrit | Mat Lowery proposed a change to openstack/trove: Remove IDENTIFIED BY clause for MySQL REVOKE https://review.openstack.org/80694 | 16:33 |
| hub_cap | vgnbkr: :D | 16:36 |
| grapex | SlickNik: What command are you using to run the Tempest tests? | 16:37 |
| grapex | "./run_tempest.sh" works, but I can't seem to pass a module path to it and have it work as I expect. | 16:37 |
| *** eghobo has joined #openstack-trove | 16:39 | |
| grapex | SlickNik: Ditto with testr run --subunit tempest.api.database.flavors.test_flavors | subunit-2to1 | tools/colorizer.py | 16:39 |
| *** eghobo has quit IRC | 16:39 | |
| *** ramashri has joined #openstack-trove | 16:46 | |
| SnowDust | Happy HOLI to all .. from india ! | 16:47 |
| *** amcrn has joined #openstack-trove | 16:49 | |
| aterix | hub_cap: how should I configure the image out of redstack? Are there particular parameters that should I be careful? | 16:50 |
| hub_cap | aterix: for all of this to work, we assume some things in the dev env, id suggest you look @ the ubuntu-mysql elements we use | 16:53 |
| aterix | hub_cap: do you mean this? https://github.com/openstack/trove-integration/tree/master/scripts/files/elements/ubuntu-mysql | 16:56 |
| hub_cap | aye aterix | 16:57 |
| *** jmontemayor_ has quit IRC | 16:57 | |
| juice | good morning (afternoon or evening) all | 16:58 |
| aterix | hub_cap: but, that link allows to install the only DBMS! What should I build the image in order to run the trove guest-agent? | 16:59 |
| amcrn | mornin' juice | 17:01 |
| juice | good morning amcrn | 17:04 |
| aterix | * That link allows to install the only DBMS! How should I build the image in order to run the trove guest-agent? | 17:05 |
| aterix | good afternoon juice | 17:06 |
| *** eghobo has joined #openstack-trove | 17:07 | |
| SnowDust | aterix: disk image builder elements = ubuntu-mysql | 17:07 |
| SnowDust | aterix: juice is the disk image builder best help :) | 17:07 |
| juice | aterix: it's been a while but I'd be happy to help | 17:08 |
| *** harlowja_away is now known as harlowja | 17:09 | |
| hub_cap | aterix: there should be a *firstboot* file (let me find) | 17:10 |
| hub_cap | aterix: this is incorrectly named, but here is the magic for teh guest | 17:11 |
| hub_cap | https://github.com/openstack/trove-integration/blob/master/scripts/files/bootstrap_init-mysql.sh | 17:11 |
| hub_cap | so effectively, u need to get the trove source into the vm, and potentially pip install or venv the pip's in requirments.txt | 17:11 |
| hub_cap | then make sure the init script is in the right place for the guest (trove-guest.init) | 17:12 |
| aterix | juice: Well, I would build a trove image (like ubuntu_mysql.qcow2 image) without using redstack. I would upload this image in an exting OpenStack environment. I don't know what I should specify (variables, parameters, ssh key, user) disk-image-builder in order to build the image | 17:13 |
| juice | aterix: and we use these elements to build the image (${DISTRO} ${EXTRA_ELEMENTS} vm heat-cfntools ${DISTRO}-guest ${DISTRO}-${SERVICE_TYPE}) (i.e. ubuntu vm heat-cnftools ubuntu-guest ubuntu-mysql). Disk Image Builder aggregates all of those elements to shape the final image | 17:14 |
| juice | aterix: each of those elements (found in trove-integration/files/elements) adds configurations, packages to the user. They answer the question you are asking. | 17:17 |
| juice | For example (https://github.com/openstack/trove-integration/blob/master/scripts/files/elements/ubuntu-guest/extra-data.d/62-ssh-key) will show you how the ssh keys are added | 17:17 |
| juice | aterix: function build_vm() in https://github.com/openstack/trove-integration/blob/master/scripts/functions_qemu has most if not all parameters that are used in disk-image-builder script | 17:19 |
| *** aterix has quit IRC | 17:19 | |
| SlickNik | grapex: testr run --parallel tempest.api.database | 17:19 |
| grapex | SlickNik: Thanks. I could have sworn I tried that. I'm in between VMs now but I'll attempt that when I get a shot. | 17:21 |
| SlickNik | grapex: no worries. | 17:22 |
| *** ramashri has quit IRC | 17:22 | |
| *** eguz has joined #openstack-trove | 17:25 | |
| *** eguz has quit IRC | 17:25 | |
| *** eguz has joined #openstack-trove | 17:25 | |
| *** eghobo has quit IRC | 17:28 | |
| *** ViswaV has quit IRC | 17:32 | |
| *** ViswaV has joined #openstack-trove | 17:33 | |
| *** ViswaV has quit IRC | 17:34 | |
| *** michael-yu has joined #openstack-trove | 17:34 | |
| *** Barker has quit IRC | 17:34 | |
| *** michael-yu has quit IRC | 17:35 | |
| *** yidclare has joined #openstack-trove | 17:36 | |
| *** ViswaV has joined #openstack-trove | 17:38 | |
| *** tanisdl has joined #openstack-trove | 17:38 | |
| *** ViswaV_ has joined #openstack-trove | 17:39 | |
| *** ViswaV has quit IRC | 17:39 | |
| *** mattgriffin has joined #openstack-trove | 17:39 | |
| *** michael-yu has joined #openstack-trove | 17:40 | |
| *** yogesh has joined #openstack-trove | 17:42 | |
| *** khyati has joined #openstack-trove | 17:47 | |
| *** esp has left #openstack-trove | 17:55 | |
| *** esp has joined #openstack-trove | 17:55 | |
| *** jmontemayor has joined #openstack-trove | 17:56 | |
| *** isviridov has quit IRC | 17:58 | |
| SlickNik | bp review meeting in 2 mins | 17:58 |
| SlickNik | Just a couple of bps to talk about | 17:59 |
| SlickNik | Agenda: | 17:59 |
| SlickNik | https://blueprints.launchpad.net/trove/+spec/upgrade-guestagent | 17:59 |
| SlickNik | https://blueprints.launchpad.net/trove/+spec/refactoring-datastore-options-in-cfg | 17:59 |
| SlickNik | esp: you around to follow up on https://blueprints.launchpad.net/trove/+spec/upgrade-guestagent | 18:00 |
| *** SnowDust has quit IRC | 18:00 | |
| esp | hey | 18:01 |
| esp | so does anyone have any questions regarding the bp? | 18:01 |
| SlickNik | We mostly had agreement. Just a couple of open regarding the RPC API and tracking history, IIRC. | 18:01 |
| SlickNik | That came up last time. | 18:01 |
| *** SnowDust has joined #openstack-trove | 18:02 | |
| *** isviridov has joined #openstack-trove | 18:02 | |
| esp | I wanted to give hub_cap a little time to look at it (the event_history table) | 18:02 |
| hub_cap | :o | 18:03 |
| esp | not sure what the open question on the RPC API was..can you remind me? | 18:03 |
| esp | sorry hub_cap, surprise! there is an event history table | 18:03 |
| grapex | hub_cap: So the discussion on Wednesday was that event_history is something we could use on a lot of different things in Trove, and it resembled the action log or event or whatever it was called that you looked into in the past | 18:03 |
| hub_cap | esp: link me to the bp plz | 18:03 |
| esp | k | 18:03 |
| esp | #link https://wiki.openstack.org/wiki/Trove-Guest-Agent-Upgrades | 18:04 |
| SlickNik | esp: I think it was around whether implementing the RPC API versioning was being proposed as part of this BP, or something that would be done outside the scope of this BP. | 18:04 |
| esp | SlickNik: #link https://wiki.openstack.org/wiki/Trove-Guest-Agent-Upgrades#Future_Follow_up_Blueprints | 18:05 |
| amcrn | from the meeting last week (regarding history/audit): http://eavesdrop.openstack.org/meetings/trove/2014/trove.2014-03-12-17.59.log.txt 18:06:21 | 18:05 |
| grapex | I'm for pushing that to next time, did anyone disagree? | 18:05 |
| SlickNik | grapex: nope. I think esp's on the same page regarding that. | 18:06 |
| esp | thx amcrn and grapex :) | 18:06 |
| grapex | Now- the "event audit" stuff- | 18:06 |
| esp | hehe | 18:06 |
| grapex | rather than say "let's punt" how about we make that into a fast-follow blueprint for this one? | 18:07 |
| amcrn | from the log it looks like we all agreed to a single row for now, and someone in parallel investigates what it means to solve audit/versioning generically (either looking @ kevinconway's suggestion, or finding another alternative) | 18:07 |
| vipul | +1 i personally don't think an upgrade needs to be recorded as an event | 18:07 |
| grapex | amcrn vipul: +1 | 18:07 |
| vipul | its enough to record the version somewhere | 18:07 |
| vipul | which could be the instance table | 18:08 |
| esp | I looked into kevinconway suggestion I don't think that will work for what I was trying to do. | 18:08 |
| grapex | esp: I think kevinconway's idea was good but maybe not a total match for this particular problem. It seems worth considering at some point though. | 18:09 |
| hub_cap | yea im not sure what we are trying to solve, ill have to catch up | 18:09 |
| esp | grapex: yep | 18:09 |
| SlickNik | I'm good with single update for now, and look into audit as a separate bp (follow up on hub_cap's actions?) | 18:09 |
| SlickNik | So given those answers, are we good with approving this bp? | 18:10 |
| grapex | Aye | 18:10 |
| esp | so let me ask this | 18:10 |
| esp | if we look at the back up feature | 18:10 |
| hub_cap | assuming all that stuff u just mentioned is pending and wont be worked on | 18:10 |
| esp | is the backups table useful at all? | 18:10 |
| esp | could we drop the backups table all together? | 18:11 |
| grapex | esp: It is, as backups is a query-able resource. | 18:11 |
| vipul | nope.. it's a resource | 18:11 |
| grapex | Same thing with the instance version- we'll be able to query for the current version | 18:11 |
| grapex | but historical records of the different states a resource was in has no precedent in Trove and could be used for multiple things. | 18:12 |
| esp | What I'm asking is, has anyone found the backups table useful in troubleshooting backups | 18:12 |
| vipul | i don't quite understand the point.. it's not about troubleshooting.. a Backup exists as a first-class thing in Trove | 18:12 |
| grapex | correct me if I'm wrong, but the backups table stores the current state of the backup, right? The one returned by the API get call for a backup? Or is there a second table I'm forgetting? | 18:12 |
| amcrn | esp: extending your question, couldn't security_groups and security_group_rules be dropped as well? | 18:13 |
| vipul | amcrn: Yes | 18:13 |
| esp | amcrn: +1 | 18:13 |
| amcrn | i actually agree with the premise, but it's a bit tangential to this blueprint | 18:13 |
| vipul | that is one area that we should drop, since it's actually duplicated | 18:13 |
| vipul | but backups do have a status like grapex says | 18:13 |
| SlickNik | Okay, I think we're going off on a bit of a tangent here. | 18:14 |
| amcrn | ah, fair point, while it's in-flight you need some sort of indicator | 18:14 |
| esp | I'm making the argument that an upgrade has enough moving parts to warrant multiple statuses | 18:14 |
| grapex | esp: I think it's a fair argument. My counterpoint is it will complicate the code, so if we make something like that we need to see how it could apply in multiple areas | 18:15 |
| grapex | I also think upgrades should be as simple as possible and should allow for different versions | 18:15 |
| grapex | I mean, allow for different strategies | 18:16 |
| vipul | yea in my mind.. it either succeeds and you have the new version recorded.. or fails and you don't get version update | 18:16 |
| vipul | like i don't care if it failed downloading a package | 18:17 |
| vipul | the upgrade just failed | 18:17 |
| grapex | vipul: Agreed. I don't want us to have a ton of Trove code designed around creating a new package manager / deployment tool | 18:17 |
| hub_cap | ++ | 18:17 |
| esp | So when something fails you will have to go dig through logs. | 18:17 |
| grapex | esp: You need to talk to our operations guys here. That is pretty much how it works for every complex thing in Trove today and they do not like it. :p | 18:18 |
| vipul | esp: sure.. and that's what the expectation shoudl be.. you try to issue the upgrade again | 18:18 |
| vipul | you don't care why it failed | 18:18 |
| esp | grapex: lol | 18:18 |
| grapex | esp: That's why if we take on the code to create an audit thing in the db I'd like it to be general enough to handle some other stuff as well. | 18:18 |
| esp | grapex: I already created the new table. | 18:19 |
| hub_cap | dear god its too late *rimshot* | 18:19 |
| esp | #link https://wiki.openstack.org/wiki/Trove-Guest-Agent-Upgrades#Schema_Changes | 18:19 |
| esp | hub_cap: lol | 18:20 |
| esp | I think maybe folks haven't had enough time to absorb the changes to the bp maybe? | 18:20 |
| esp | should I kill the event_history table? sounds like not right? | 18:20 |
| vipul | i think we're saying yes kill it | 18:20 |
| esp | I hear grapex telling me to make it generic? | 18:21 |
| SlickNik | esp: I think the earlier discussion suggested that event history should be addressed as a separate bp. | 18:21 |
| vipul | and come up with a general auditing solution later, that handles more than upgrades | 18:21 |
| hub_cap | yes lets not solve that in this bp | 18:21 |
| amcrn | +1 | 18:21 |
| esp | k, I'm fine with this | 18:22 |
| esp | will make the event history a separate table | 18:23 |
| esp | table --> bp | 18:23 |
| vipul | oh good :) | 18:23 |
| hub_cap | lol i was like huuuuh | 18:23 |
| SlickNik | Okay, so back to approving this. Given that "event history" will be a different BP, do people feel good with the upgrade bp? | 18:23 |
| SlickNik | +1 | 18:23 |
| vipul | did you guys discuss agent hearbeats table? | 18:24 |
| grapex | +1 | 18:24 |
| hub_cap | i will review and give esp any thoughts so +0 | 18:24 |
| grapex | So... maybe this is a question of how much we should be nailing down here? | 18:24 |
| esp | k | 18:24 |
| grapex | vipul: What was your thought on the agent heartbeats table? I just worry maybe now we might get too prescriptive | 18:24 |
| grapex | it seems easier to sort some questions out during the pull request | 18:24 |
| vipul | well we have service_status for mysql status.. can the agent_status be anything but 'running'? | 18:24 |
| hub_cap | grapex: still good to discuss so we at least remember, somewhat, when reading the review | 18:25 |
| grapex | vipul: I suppose it could be "unknown" | 18:25 |
| vipul | seems like if the agent is down.. well then nothing is going to send that hearbeat | 18:25 |
| vipul | but what will change it to unknown | 18:25 |
| grapex | vipul: Good point | 18:25 |
| SlickNik | grapex: I think intent, scope and high level design of bp. We don't have to get all the design details fleshed out in one sitting. | 18:25 |
| grapex | vipul: I thought the purpose of that new status would be if the agent tries to do an upgrade, then gets itself into a jacked up state, it can at least report it | 18:26 |
| hub_cap | wouldnt an old time be considered unknown? | 18:26 |
| hub_cap | like active, last heartbeat 200s ago | 18:26 |
| grapex | hub_cap: True... I guess we should trust conductor | 18:26 |
| esp | wrt to setting the guest_agent_status I think it will be separate message. I did not want to send it with every heart beat. | 18:26 |
| vipul | is this something we want to report to the end user? or is this purely for admin? | 18:26 |
| grapex | vipul: I figured it was for an admin / trove code | 18:27 |
| amcrn | admin | 18:27 |
| esp | vipul: the idea originally was to only allow admin to see it | 18:27 |
| vipul | esp: ok, are we concerned about the chattiness if we add another heartbeat | 18:27 |
| hub_cap | admin it is :) | 18:27 |
| grapex | vipul: I'm starting to agree | 18:28 |
| hub_cap | im not sure we need another heartbeat, but just a larger message, right? | 18:28 |
| esp | vipul: I was only planning to send the guest_agent_status on startup/restart of the guest agent | 18:28 |
| vipul | well then now we don't have a good way to say it's 'unknown' | 18:28 |
| hub_cap | the heartbeat should be checking all status and sendign that up in a single msg | 18:28 |
| SlickNik | hub_cap: yes, I think same message, just another field (or two). | 18:29 |
| esp | hub_cap: we could do this, I was just trying to keep the logic separate | 18:29 |
| grapex | So does this mean the version gets put into every heart beat? | 18:29 |
| SlickNik | not a completely different message. | 18:29 |
| hub_cap | well, if the msg is current, the guest is online | 18:29 |
| hub_cap | so that can be assumed | 18:29 |
| hub_cap | if needed | 18:29 |
| hub_cap | and if the message is crrent and datastore status is "dead yo" | 18:30 |
| vipul | right.. so if it was bolted on to existing heartbeat, then it might be an useful status | 18:30 |
| esp | yeah, it's less work. | 18:30 |
| grapex | So I think we should also consider passing the guest agent version back on it's own message | 18:30 |
| vipul | grapex: like a one time thing? | 18:31 |
| grapex | vipul: Yeah- | 18:31 |
| *** asterix_ has joined #openstack-trove | 18:31 | |
| vipul | the concern i'd have about that is.. once the guests are upgraded.. you might be upgrading the control plane | 18:31 |
| esp | grapex: I like that | 18:31 |
| vipul | and if you happen to miss that one-time message while you're upgrading | 18:31 |
| grapex | vipul: That's a good point | 18:31 |
| grapex | vipul: Ok, you've sold me on it then | 18:31 |
| esp | not reason to hammer the instances table for every heartbeat | 18:31 |
| vipul | grapex: that was easy! | 18:31 |
| grapex | I wonder in this case if long-term we might want to send the backup status on the heartbeat call as well | 18:32 |
| hub_cap | grapex: id prefer one msg if possible for all of these statii | 18:32 |
| esp | so are we saying 1 message for everything? | 18:32 |
| esp | I'm confused | 18:32 |
| grapex | Ok, sorry for the tangent, I'm good with passing it back with the heartbeat | 18:32 |
| hub_cap | is that a plural of status? /me thinks so | 18:32 |
| grapex | hub_cap: A group of status is technically referred to as a "murder" | 18:32 |
| vipul | i think we extend the heartbeat with the version + agent_status.. that seems to be the safest | 18:32 |
| hub_cap | grapex: nice | 18:33 |
| grapex | vipul: +1 | 18:33 |
| esp | k, so 1 message in heartbeat (version + status) | 18:33 |
| hub_cap | vipul: ++ | 18:33 |
| vipul | esp: +1 | 18:33 |
| esp | version --> instances table | 18:33 |
| esp | status --> agent_heartbeats table | 18:33 |
| esp | anything else? | 18:33 |
| amcrn | what's your favorite color? | 18:33 |
| grapex | esp: I thought we decided there wouldn't be a status for the agent in the db- we'd just infer it from the time of the heartbeat? | 18:33 |
| amcrn | that's the pressing question here. | 18:33 |
| hub_cap | bluond amcrn | 18:34 |
| esp | amcrn: black | 18:34 |
| *** isviridov has quit IRC | 18:34 | |
| SlickNik | okay, let's keep moving. | 18:34 |
| hub_cap | grapex: i think thats all we need for now | 18:34 |
| SlickNik | I think we're good with upgrades | 18:34 |
| SlickNik | next in the pipeline: | 18:34 |
| SlickNik | https://blueprints.launchpad.net/trove/+spec/refactoring-datastore-options-in-cfg | 18:35 |
| SlickNik | SnowDust: here? | 18:35 |
| SnowDust | yes | 18:35 |
| SnowDust | shall i detail the work ? | 18:36 |
| SlickNik | Yes, go ahead. You've got the floor. | 18:36 |
| SnowDust | ok .. so | 18:36 |
| SnowDust | if we see the recent changes to trove/common/cfg.py | 18:36 |
| vipul | do we have anyone here that's wokring on capabilities? isn't there some intersection here? | 18:37 |
| SnowDust | i have been suggesting how difficult it is to keep up with hardcoded changes for database options there | 18:37 |
| SnowDust | my solution keeps datastore options in their own containers | 18:38 |
| SnowDust | so that trove is nutral to external or built in database support | 18:38 |
| grapex | SnowDust: Can you define "containers"? | 18:38 |
| SnowDust | grapex: i meant for trove the datastore should be a component | 18:38 |
| hub_cap | i think hes saying use entry points to load the datastore impls | 18:39 |
| SnowDust | and the way the managers are loaded , a datastore may be implemented as separate python packages | 18:39 |
| SnowDust | thus they are separate containers. | 18:39 |
| amcrn | correct, as an example cassandra lives by itself: https://review.openstack.org/#/c/80061/5/trove/guestagent/datastore/cassandra/options.py (vs. being in the same file as all other datastore cfgs) | 18:39 |
| SnowDust | amcrn : thanks u got it right | 18:39 |
| SlickNik | hub_cap: He's saying use entry points to load the datastore _configs_ I think. | 18:40 |
| vipul | so is this purely refactoring, and moving things around? Or are we changing behavior | 18:40 |
| SlickNik | #link https://review.openstack.org/#/c/80061/5 | 18:40 |
| grapex | So far I like this idea. Now if we could just move all the guest code into it's own directory before Juno my life would be complete. :) | 18:40 |
| SnowDust | vipul: its enhancing the current implementations 99% refactoring 1% more smartness :) | 18:40 |
| amcrn | vipul: see http://eavesdrop.openstack.org/irclogs/%23openstack-trove/%23openstack-trove.2014-03-10.log, specifically 2014-03-10T20:40:01 to 2014-03-10T20:44:10 | 18:41 |
| vipul | amcrn: ok thx | 18:41 |
| grapex | 1% more smartness? I hope you spread that across at least three commits, my brain can handle only 0.33% smartness on any given pull request. | 18:41 |
| SlickNik | vipul: purely refactoring around how the datastore configs are loaded. | 18:41 |
| SlickNik | lol @ grapex | 18:42 |
| hub_cap | ok so heres the thing | 18:42 |
| SnowDust | :D | 18:42 |
| SnowDust | lol | 18:42 |
| hub_cap | at the end of the day, its still looking for [cassandra]... optgroups right? | 18:42 |
| hub_cap | [redis]... etc | 18:43 |
| SnowDust | hub_cap: its loaded differently and an external datastore can implement in similar pattern without a change in cfg.py | 18:43 |
| grapex | hub_cap: is your point that it still needs to be baked into an image? | 18:43 |
| SnowDust | similar to cassandra/redis as in the #80061/5 | 18:43 |
| hub_cap | i agree w/ the split in the config file into things like | 18:43 |
| *** asterix_ has quit IRC | 18:43 | |
| hub_cap | [cassandra]... some options ... [redis]... some options ... | 18:44 |
| hub_cap | as opposed to keeping things in default | 18:44 |
| vipul | So the issue amcrn raised last time still woudl be an issue.. where you wanted to reuse an existing manager, but have different cfg opts | 18:44 |
| SnowDust | vipul: which issue ? | 18:45 |
| *** IvanZ has quit IRC | 18:46 | |
| SnowDust | datastore_options = <datastore_manager> : class.from.which.to.load.the.options ( gets in conf files) | 18:46 |
| vipul | SInce you'll be loading datastore specific options.py, you can't have the scenario where you can reuse a manager, but supply different config | 18:46 |
| SnowDust | and then is loaded from cfg.py .. | 18:46 |
| vipul | oh i see | 18:46 |
| SlickNik | hub_cap: It's still keeping that structure. It's splitting out the config into different files (one per datastore), and loading those files dynamically based on a default config dict value. | 18:47 |
| vipul | yea i don't really see anything contentious about this BP | 18:48 |
| SlickNik | Yeah, I'm good with it too. | 18:48 |
| SnowDust | vipul: additionally to that to keep the current behaviour i have used a dict in cfg.py for already available datastores | 18:48 |
| hub_cap | will this require operators to change their configs? | 18:49 |
| SnowDust | hub_cap: no for default datastores no changes from their end | 18:49 |
| hub_cap | cool | 18:49 |
| amcrn | one divergence from a pattern already in existence is https://github.com/openstack/trove/blob/fba8cabea326527bacdeca56760a97e14cdcc18f/trove/guestagent/dbaas.py#L34-L51 | 18:50 |
| esmute | instead of defining the 'datastore_options_definition' in the cfg, can that be somehow obtained based on the datastore? | 18:50 |
| amcrn | where the defaults are force loaded, but the operator can add extensions | 18:50 |
| amcrn | whereas this patch allows you to override everything | 18:50 |
| SnowDust | hub_cap: this adds the flexibility to the external datastore implementation ( where operators CAN kick in ) | 18:50 |
| SnowDust | amcrn : right | 18:50 |
| esmute | like the config would be obtained from trove.guestagent.datastore.[datastore].options | 18:50 |
| SnowDust | esmute: we can have them in the conf file itself | 18:50 |
| SnowDust | esmute: not like that i suggest .. | 18:51 |
| esmute | that way the cfg file is totally decoupled from the datastore config | 18:51 |
| hub_cap | if u put that conf value in the file | 18:51 |
| hub_cap | itll override it, period | 18:52 |
| hub_cap | the trove.blah... wont matter | 18:52 |
| vipul | amcrn: ++ that would also address what esmute is saying.. have sane defaults already configured | 18:52 |
| SnowDust | vipul: cfg.py is loading the sane defaults if u see | 18:53 |
| amcrn | SnowDust: but your implementation is a purge-replace, see datastore_options @ https://review.openstack.org/#/c/80061/5/trove/common/cfg.py | 18:54 |
| SnowDust | but to decouple completely, we may set datastore_options variable to default=dict() | 18:54 |
| SnowDust | and datastore_options may be defined in sample.conf for sane defaults | 18:54 |
| amcrn | the point is, even if you decouple there, https://github.com/openstack/trove/blob/fba8cabea326527bacdeca56760a97e14cdcc18f/trove/guestagent/dbaas.py#L34-L51 still exists | 18:54 |
| grapex | amcrn: So you're saying "choose a path"? | 18:55 |
| amcrn | i'm saying that the in-flight patch as-is is dangerous, because it allows you to sever the registration of a datastore without cleaning up https://github.com/openstack/trove/blob/fba8cabea326527bacdeca56760a97e14cdcc18f/trove/guestagent/dbaas.py#L34-L51; it's an incomplete solution. | 18:56 |
| SnowDust | amcrn: its about the options not the datastore manager class | 18:56 |
| grapex | What if this self-contained, datastore specific code could also append that "defaults" dict? | 18:56 |
| amcrn | grapex: right, all i'm asking for is consistency; that would be one option. | 18:56 |
| grapex | amcrn: What if "defaults" was made into an empty dict, then this extra code, when loaded, would add to the dict? | 18:56 |
| amcrn | that's an option | 18:57 |
| SlickNik | amcrn: agreed. | 18:57 |
| grapex | amcrn: I get it. I second making that a requirement and enforcing it during the pull request. | 18:57 |
| *** esp has left #openstack-trove | 18:57 | |
| grapex | So... are we good with this? | 18:58 |
| SlickNik | grapex: sounds good. | 18:58 |
| SnowDust | i am ready to take notes | 18:58 |
| amcrn | have to bounce, bbl. | 18:59 |
| grapex | Me too- I've got a meeting in a minute or so. | 18:59 |
| SlickNik | That's all we had for today anyhow. | 18:59 |
| hub_cap | https://github.com/openstack/cinder/blob/master/cinder/volume/configuration.py | 18:59 |
| SlickNik | Thanks guys. Catch you later. | 18:59 |
| grapex | Ok, cool. ttyl | 19:00 |
| SnowDust | guys: is the BP approved ? | 19:00 |
| hub_cap | i think its still pending, we should disduss later today | 19:00 |
| hub_cap | when the peoples get back who had to leave | 19:01 |
| SlickNik | SnowDust: still needs some discussion. | 19:01 |
| SnowDust | SlickNik: sure | 19:01 |
| *** jcru has quit IRC | 19:01 | |
| *** robertmy_ has quit IRC | 19:02 | |
| *** robertmyers has joined #openstack-trove | 19:02 | |
| *** kevinconway_ has joined #openstack-trove | 19:03 | |
| *** kevinconway has quit IRC | 19:05 | |
| *** kevinconway_ is now known as kevinconway | 19:05 | |
| *** ViswaV_ has quit IRC | 19:05 | |
| *** ViswaV has joined #openstack-trove | 19:07 | |
| *** ViswaV_ has joined #openstack-trove | 19:08 | |
| *** ViswaV has quit IRC | 19:08 | |
| *** ViswaV has joined #openstack-trove | 19:09 | |
| *** ViswaV_ has quit IRC | 19:09 | |
| *** demorris has joined #openstack-trove | 19:10 | |
| spredzy | Is it possible to specify a keypair when creating an instance with trove-cli ? | 19:35 |
| *** michael-yu has quit IRC | 19:37 | |
| *** jcru has joined #openstack-trove | 19:38 | |
| *** esp has joined #openstack-trove | 19:49 | |
| *** SnowDust has quit IRC | 19:53 | |
| *** michael-yu has joined #openstack-trove | 19:57 | |
| *** datsun180b has quit IRC | 20:04 | |
| *** ViswaV has quit IRC | 20:23 | |
| *** IvanZ has joined #openstack-trove | 20:26 | |
| *** ViswaV has joined #openstack-trove | 20:26 | |
| *** ViswaV_ has joined #openstack-trove | 20:29 | |
| *** ViswaV has quit IRC | 20:29 | |
| *** radez is now known as radez_g0n3 | 20:40 | |
| openstackgerrit | SlickNik proposed a change to openstack/trove: Update sqlalchemy migrate scripts for postgres https://review.openstack.org/81073 | 20:44 |
| *** harlowja is now known as harlowja_away | 20:44 | |
| openstackgerrit | SlickNik proposed a change to openstack/trove: Update sqlalchemy migrate scripts for postgres https://review.openstack.org/81073 | 20:45 |
| SlickNik | spredzy: No. You don't get ssh access to the trove instance (by design). You're only allowed to access the datastore (eg. mysql) on the appropriate port(s) (eg. 3306). | 20:48 |
| *** jasonb365 has joined #openstack-trove | 20:48 | |
| juice | is anyone using the "notifier_queue*" configurations in etc/trove/trove-taskmanager.conf? | 20:49 |
| spredzy | SlickNik, Thanks. The by-design was the information I was looking for | 20:49 |
| *** jasonb365 has quit IRC | 20:49 | |
| spredzy | aslo I was wondering how one can create datastore when not using redstack ? | 20:50 |
| spredzy | trove cli and trove-manage cli does not seems to be willing to do it. Does it has to be done manually ? | 20:50 |
| SlickNik | spredzy: Yes, you'd have to use diskimage-builder to build a datastore image. | 20:51 |
| spredzy | By datastore image, an image with trove-guestagent + the proper db packages is meant ? | 20:51 |
| spredzy | I've been reading the doc but I couldn't get a clear picture of that point | 20:52 |
| SlickNik | spredzy: This image would have to have the datastore and trove-guestagent installed, have connectivity to the trove rabbitmq queues, and start the guestagent on bootup. | 20:52 |
| spredzy | and the entry in datastores and datastore_versions tables needs to be done manually, right (as for now)? | 20:52 |
| SlickNik | spredzy: Yes, that is correct (for now). | 20:53 |
| SlickNik | juice: I think nobody's using them. There's a review to clean them up in-flight. | 20:53 |
| SlickNik | juice: https://review.openstack.org/#/c/74638/10/etc/trove/trove-taskmanager.conf.sample | 20:54 |
| spredzy | SlickNik, thanks for all those info it does make sense now | 20:54 |
| *** grapex has quit IRC | 20:54 | |
| juice | thanks slicknik | 20:54 |
| SlickNik | spredzy: You're welcome. Ping us in the channel if you need more help with any of this. | 20:54 |
| spredzy | Will do for sure. Thanks again for the help | 20:55 |
| *** pdmars has quit IRC | 21:02 | |
| openstackgerrit | Viswa Vutharkar proposed a change to openstack/trove-integration: Adds support to kickstart diff datastore versions https://review.openstack.org/72804 | 21:09 |
| openstackgerrit | SlickNik proposed a change to openstack/trove: Turning off XML fake-mode tests https://review.openstack.org/70000 | 21:12 |
| *** rwsu has quit IRC | 21:16 | |
| openstackgerrit | Auston McReynolds proposed a change to openstack/python-troveclient: Remove Duplicate ClientExceptions and Merge https://review.openstack.org/80328 | 21:22 |
| *** rwsu has joined #openstack-trove | 21:31 | |
| *** asterix has joined #openstack-trove | 21:34 | |
| *** achampion has quit IRC | 21:44 | |
| *** tanisdl_ has joined #openstack-trove | 21:47 | |
| *** tanisdl has quit IRC | 21:50 | |
| *** tanisdl_ is now known as tanisdl | 21:50 | |
| *** yogesh has quit IRC | 21:58 | |
| *** IvanZ has quit IRC | 22:04 | |
| *** thedodd has quit IRC | 22:08 | |
| *** robertmyers has quit IRC | 22:09 | |
| *** rhodgin has quit IRC | 22:12 | |
| *** harlowja_away is now known as harlowja | 22:17 | |
| *** datsun180b has joined #openstack-trove | 22:17 | |
| openstackgerrit | Giuseppe Galeota proposed a change to openstack/trove: Trove Manual Installation guide https://review.openstack.org/78608 | 22:24 |
| *** michael-yu has quit IRC | 22:29 | |
| *** tanisdl_ has joined #openstack-trove | 22:30 | |
| *** tanisdl has quit IRC | 22:32 | |
| *** tanisdl_ is now known as tanisdl | 22:32 | |
| *** openstackgerrit has quit IRC | 22:39 | |
| *** openstackgerrit has joined #openstack-trove | 22:39 | |
| *** cweid has quit IRC | 22:41 | |
| *** michael-yu has joined #openstack-trove | 22:42 | |
| *** mattgriffin has quit IRC | 22:51 | |
| *** rhodgin has joined #openstack-trove | 22:52 | |
| *** tanisdl_ has joined #openstack-trove | 22:59 | |
| *** tanisdl has quit IRC | 23:02 | |
| cp16net | hmm anyone used the trove cli to list more than the paginated limit? | 23:03 |
| cp16net | i dont see a way of passing in the marker | 23:04 |
| cp16net | or a new limit | 23:04 |
| *** tanisdl_ has quit IRC | 23:04 | |
| *** demorris has quit IRC | 23:05 | |
| *** michael-yu has quit IRC | 23:08 | |
| *** michael-yu has joined #openstack-trove | 23:09 | |
| *** jcru has quit IRC | 23:16 | |
| *** eguz_ has joined #openstack-trove | 23:16 | |
| SlickNik | cp16net: is this for backups? | 23:17 |
| *** eguz has quit IRC | 23:17 | |
| SlickNik | cp16net: I haven't tried it myself. robertmyers would be the best person to ask. | 23:17 |
| *** yidclare has quit IRC | 23:18 | |
| cp16net | no this is for instances | 23:18 |
| cp16net | looks like the cli doesnt support this | 23:18 |
| cp16net | with optional args | 23:19 |
| cp16net | but the client lib does | 23:19 |
| cp16net | need a bug for this. | 23:19 |
| SlickNik | cp16net: gotcha. Can you open one for this, please? | 23:19 |
| *** kevinconway has quit IRC | 23:22 | |
| openstackgerrit | Giuseppe Galeota proposed a change to openstack/trove: Trove Manual Installation guide https://review.openstack.org/81112 | 23:23 |
| *** rramirez has quit IRC | 23:26 | |
| asterix | please don't consider the review I posted https://review.openstack.org/81112 | 23:27 |
| *** achampion has joined #openstack-trove | 23:30 | |
| juice | asterix: is that an attempt at reverse-psychology | 23:32 |
| juice | asterix: did you know there was a person on this channel name aterix earlier? | 23:32 |
| asterix | juice: no, I didn't | 23:33 |
| *** ViswaV_ has quit IRC | 23:33 | |
| *** amytron has quit IRC | 23:33 | |
| openstackgerrit | Giuseppe Galeota proposed a change to openstack/trove: Trove Manual Installation guide https://review.openstack.org/81117 | 23:38 |
| asterix | OK, this review is good | 23:39 |
| asterix | So, now I change my nickname | 23:40 |
| *** asterix has quit IRC | 23:41 | |
| *** tattabbum has joined #openstack-trove | 23:41 | |
| *** tattabbum has quit IRC | 23:46 | |
| *** jmontemayor has quit IRC | 23:48 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!