| *** radez` is now known as radez_g0n3 | 00:21 | |
| *** matsuhashi has joined #openstack-trove | 00:23 | |
| *** shakayumi has joined #openstack-trove | 00:31 | |
| *** shakayumi has quit IRC | 00:31 | |
| *** ramashri has quit IRC | 00:40 | |
| *** demorris has joined #openstack-trove | 00:41 | |
| *** annashe__ has joined #openstack-trove | 00:51 | |
| *** Barker has joined #openstack-trove | 00:53 | |
| *** sbfox has joined #openstack-trove | 01:01 | |
| *** ViswaV has quit IRC | 01:10 | |
| *** ViswaV has joined #openstack-trove | 01:28 | |
| *** nosnos has joined #openstack-trove | 01:38 | |
| *** yidclare has quit IRC | 01:39 | |
| *** amcrn has quit IRC | 01:48 | |
| *** sbfox has quit IRC | 01:53 | |
| *** ViswaV_ has joined #openstack-trove | 02:02 | |
| *** ViswaV has quit IRC | 02:05 | |
| *** sbfox has joined #openstack-trove | 02:09 | |
| *** ViswaV_ has quit IRC | 02:15 | |
| *** demorris has quit IRC | 02:15 | |
| *** ramishra has joined #openstack-trove | 02:21 | |
| *** ramishra has quit IRC | 02:48 | |
| *** ramishra has joined #openstack-trove | 02:58 | |
| NavsterG | what is the best repository for installing trove on centos | 03:04 |
|---|---|---|
| NavsterG | ?? | 03:04 |
| *** rwsu has quit IRC | 03:10 | |
| *** Barker has quit IRC | 03:15 | |
| *** nosnos has quit IRC | 03:29 | |
| *** achampion has joined #openstack-trove | 03:29 | |
| *** matsuhashi has quit IRC | 03:31 | |
| *** Longgeek has joined #openstack-trove | 03:33 | |
| *** robertmyers has quit IRC | 03:35 | |
| *** haomaiwang has joined #openstack-trove | 03:40 | |
| *** haomaiwang has quit IRC | 03:48 | |
| *** haomaiwang has joined #openstack-trove | 03:48 | |
| *** achampion has quit IRC | 03:56 | |
| *** haomai___ has joined #openstack-trove | 04:05 | |
| *** haomaiwang has quit IRC | 04:05 | |
| *** ramishra has quit IRC | 04:19 | |
| *** haomaiwang has joined #openstack-trove | 04:21 | |
| *** haomai___ has quit IRC | 04:23 | |
| *** matsuhashi has joined #openstack-trove | 04:25 | |
| *** haomaiwa_ has joined #openstack-trove | 04:25 | |
| *** haomaiwang has quit IRC | 04:25 | |
| *** nosnos has joined #openstack-trove | 04:26 | |
| *** ramishra has joined #openstack-trove | 04:33 | |
| *** haomai___ has joined #openstack-trove | 04:34 | |
| *** haomaiwa_ has quit IRC | 04:36 | |
| *** achampion has joined #openstack-trove | 04:53 | |
| *** Longgeek has quit IRC | 04:56 | |
| *** Longgeek has joined #openstack-trove | 05:01 | |
| NavsterG | question for all.. I used http://docs.openstack.org/icehouse/install-guide/install/yum/content/trove-install.html to install trove on my centos 6.5 and nothing is working. | 05:02 |
| NavsterG | is there better install instructions out there somewhere | 05:03 |
| *** shakamunyi has joined #openstack-trove | 05:04 | |
| *** yogeshmehra has joined #openstack-trove | 05:17 | |
| *** coolsvap|afk is now known as coolsvap | 05:30 | |
| *** sgotliv has joined #openstack-trove | 05:30 | |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Add a new column and indexes to agent_heartbeats https://review.openstack.org/81682 | 05:36 |
| *** NavsterG has quit IRC | 05:38 | |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Partially implements guest agent upgrade strategy https://review.openstack.org/85225 | 05:41 |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Partially implements guest agent upgrade strategy https://review.openstack.org/85225 | 05:42 |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Add a new column and indexes to agent_heartbeats https://review.openstack.org/81682 | 05:42 |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Add a new column and indexes to agent_heartbeats https://review.openstack.org/81682 | 05:45 |
| *** haomai___ has quit IRC | 05:49 | |
| *** haomaiwa_ has joined #openstack-trove | 05:49 | |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Partially implements guest agent upgrade strategy https://review.openstack.org/85225 | 05:52 |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Add a new column and indexes to agent_heartbeats https://review.openstack.org/81682 | 05:52 |
| *** coolsvap is now known as coolsvap|afk | 05:52 | |
| *** coolsvap|afk is now known as coolsvap | 05:58 | |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Partially implements guest agent upgrade strategy https://review.openstack.org/85225 | 06:03 |
| esp | ^ testing gerrit :) | 06:03 |
| *** esp has left #openstack-trove | 06:04 | |
| *** sgotliv has quit IRC | 06:05 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/trove: Imported Translations from Transifex https://review.openstack.org/97008 | 06:06 |
| *** yogeshmehra has quit IRC | 06:24 | |
| *** yogeshmehra has joined #openstack-trove | 06:25 | |
| *** yogeshme_ has joined #openstack-trove | 06:26 | |
| *** amytron has quit IRC | 06:29 | |
| *** yogeshmehra has quit IRC | 06:30 | |
| *** coolsvap is now known as coolsvap|afk | 06:31 | |
| *** sbfox has quit IRC | 06:40 | |
| *** coolsvap|afk is now known as coolsvap | 06:47 | |
| *** yogeshme_ has quit IRC | 06:52 | |
| *** yogeshmehra has joined #openstack-trove | 06:53 | |
| *** yogeshmehra has quit IRC | 06:57 | |
| *** yogeshmehra has joined #openstack-trove | 06:59 | |
| *** freyes has joined #openstack-trove | 07:04 | |
| *** eghobo has joined #openstack-trove | 07:08 | |
| *** denis_makogon has joined #openstack-trove | 07:09 | |
| openstackgerrit | Yongli He proposed a change to openstack/trove: Apply six for metaclass https://review.openstack.org/73136 | 07:09 |
| *** matsuhashi has quit IRC | 07:17 | |
| *** matsuhashi has joined #openstack-trove | 07:18 | |
| *** coolsvap is now known as coolsvap|afk | 07:21 | |
| *** eghobo has quit IRC | 07:24 | |
| *** coolsvap|afk is now known as coolsvap | 07:29 | |
| *** mattgriffin has quit IRC | 07:36 | |
| *** boden has joined #openstack-trove | 07:59 | |
| *** shakamunyi has quit IRC | 08:10 | |
| *** SushillKM has joined #openstack-trove | 08:14 | |
| *** annashe__ has quit IRC | 08:20 | |
| *** ramishra has quit IRC | 08:26 | |
| *** Longgeek has quit IRC | 08:27 | |
| *** Longgeek has joined #openstack-trove | 08:28 | |
| *** Longgeek has quit IRC | 08:29 | |
| *** Longgeek has joined #openstack-trove | 08:30 | |
| *** flaper87|afk is now known as flaper87 | 08:35 | |
| *** matsuhashi has quit IRC | 08:42 | |
| *** openstackgerrit has quit IRC | 08:48 | |
| *** global__ has quit IRC | 08:48 | |
| *** number80 has quit IRC | 08:48 | |
| *** global__ has joined #openstack-trove | 08:48 | |
| *** openstackgerrit has joined #openstack-trove | 08:49 | |
| *** number80 has joined #openstack-trove | 08:50 | |
| *** matsuhashi has joined #openstack-trove | 08:50 | |
| *** freyes has quit IRC | 09:02 | |
| *** freyes has joined #openstack-trove | 09:05 | |
| *** shakamunyi has joined #openstack-trove | 09:06 | |
| *** shakamunyi has quit IRC | 09:10 | |
| *** ramishra has joined #openstack-trove | 09:23 | |
| *** ramishra has quit IRC | 09:33 | |
| *** shakamunyi has joined #openstack-trove | 09:36 | |
| *** ramishra has joined #openstack-trove | 09:40 | |
| *** shakamunyi has quit IRC | 09:41 | |
| *** freyes has quit IRC | 09:44 | |
| *** yogeshmehra has quit IRC | 09:48 | |
| *** yogeshmehra has joined #openstack-trove | 09:48 | |
| *** yogeshmehra has quit IRC | 09:53 | |
| *** SushillKM has quit IRC | 10:03 | |
| *** haomaiwa_ has quit IRC | 10:05 | |
| *** haomaiwang has joined #openstack-trove | 10:06 | |
| *** denis_makogon has quit IRC | 10:06 | |
| *** nosnos has quit IRC | 10:08 | |
| *** haomai___ has joined #openstack-trove | 10:09 | |
| *** haomaiwang has quit IRC | 10:12 | |
| *** ramishra has quit IRC | 10:17 | |
| *** sgotliv has joined #openstack-trove | 10:24 | |
| *** matsuhashi has quit IRC | 10:24 | |
| *** haomai___ has quit IRC | 10:24 | |
| *** haomaiwang has joined #openstack-trove | 10:25 | |
| *** Longgeek has quit IRC | 10:34 | |
| *** Longgeek has joined #openstack-trove | 10:35 | |
| *** shakamunyi has joined #openstack-trove | 10:37 | |
| *** shakamunyi has quit IRC | 10:41 | |
| *** matsuhashi has joined #openstack-trove | 10:47 | |
| *** freyes has joined #openstack-trove | 10:53 | |
| *** demorris has joined #openstack-trove | 10:56 | |
| *** ramishra has joined #openstack-trove | 11:05 | |
| *** ramishra has quit IRC | 11:21 | |
| *** ramishra has joined #openstack-trove | 11:21 | |
| *** coolsvap is now known as coolsvap|afk | 11:23 | |
| *** matsuhashi has quit IRC | 11:25 | |
| *** flaper87 is now known as flaper87|afk | 11:25 | |
| *** matsuhashi has joined #openstack-trove | 11:26 | |
| *** ramishra has quit IRC | 11:26 | |
| *** demorris has quit IRC | 11:27 | |
| *** r0j4z0 has joined #openstack-trove | 11:29 | |
| *** matsuhashi has quit IRC | 11:29 | |
| *** matsuhashi has joined #openstack-trove | 11:29 | |
| *** shakamunyi has joined #openstack-trove | 11:38 | |
| *** shakamunyi has quit IRC | 11:43 | |
| *** PeterSchwaller has quit IRC | 11:46 | |
| *** demorris has joined #openstack-trove | 11:46 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix updated timestamp for SecurityGroup model https://review.openstack.org/97510 | 11:50 |
| *** PeterSchwaller has joined #openstack-trove | 11:51 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix calls for proboscis methods https://review.openstack.org/97478 | 11:53 |
| *** PeterSchwaller has quit IRC | 11:56 | |
| openstackgerrit | Denis M. proposed a change to openstack/python-troveclient: Fix shell view for secgroup-* commands https://review.openstack.org/85314 | 11:57 |
| *** coolsvap|afk is now known as coolsvap | 11:58 | |
| *** PeterSchwaller has joined #openstack-trove | 12:03 | |
| *** coolsvap is now known as coolsvap|afk | 12:04 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix poll_until retriever when heat provisioning is enabled https://review.openstack.org/97194 | 12:19 |
| *** ramishra has joined #openstack-trove | 12:21 | |
| *** radez_g0n3 is now known as radez | 12:23 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix tracking of SG's provisioned by Heat https://review.openstack.org/96795 | 12:23 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Use tcp/udp ports from config for heat templates https://review.openstack.org/96780 | 12:23 |
| *** achampion has quit IRC | 12:24 | |
| *** demorris has quit IRC | 12:25 | |
| *** kevinconway has joined #openstack-trove | 12:31 | |
| *** shakamunyi has joined #openstack-trove | 12:39 | |
| *** haomaiwang has quit IRC | 12:40 | |
| *** haomaiwang has joined #openstack-trove | 12:40 | |
| *** shakamunyi has quit IRC | 12:44 | |
| *** haomaiw__ has joined #openstack-trove | 12:46 | |
| *** haomaiwang has quit IRC | 12:49 | |
| *** shakamunyi has joined #openstack-trove | 13:03 | |
| *** tvoran has joined #openstack-trove | 13:04 | |
| *** tvoran has quit IRC | 13:04 | |
| *** tvoran has joined #openstack-trove | 13:05 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Add backup size validation at restoring https://review.openstack.org/86300 | 13:06 |
| *** ramishra has quit IRC | 13:07 | |
| *** ramishra has joined #openstack-trove | 13:10 | |
| *** pdmars has joined #openstack-trove | 13:16 | |
| *** demorris has joined #openstack-trove | 13:16 | |
| *** Longgeek has quit IRC | 13:18 | |
| *** Longgeek has joined #openstack-trove | 13:18 | |
| *** Longgeek has quit IRC | 13:19 | |
| *** Longgeek has joined #openstack-trove | 13:19 | |
| *** Longgeek has quit IRC | 13:19 | |
| *** Longgeek_ has joined #openstack-trove | 13:19 | |
| *** Longgeek_ has quit IRC | 13:21 | |
| *** Longgeek has joined #openstack-trove | 13:24 | |
| *** achampion has joined #openstack-trove | 13:25 | |
| *** radez is now known as radez_g0n3 | 13:26 | |
| *** freyes has quit IRC | 13:36 | |
| *** jcru has joined #openstack-trove | 13:38 | |
| *** mattgriffin has joined #openstack-trove | 13:43 | |
| *** ramishra has quit IRC | 13:47 | |
| *** ramishra has joined #openstack-trove | 13:47 | |
| *** ramishra has quit IRC | 13:52 | |
| *** NehaV has joined #openstack-trove | 13:55 | |
| *** robertmyers has joined #openstack-trove | 13:55 | |
| *** NehaV has quit IRC | 13:55 | |
| *** NehaV has joined #openstack-trove | 13:56 | |
| *** NehaV1 has joined #openstack-trove | 13:57 | |
| *** NehaV has quit IRC | 14:00 | |
| *** pdmars has quit IRC | 14:02 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix updated timestamp for SecurityGroup model https://review.openstack.org/97510 | 14:02 |
| *** pdmars has joined #openstack-trove | 14:02 | |
| *** rubick_ has joined #openstack-trove | 14:05 | |
| *** Barker has joined #openstack-trove | 14:08 | |
| *** grapex has joined #openstack-trove | 14:17 | |
| *** esp has joined #openstack-trove | 14:17 | |
| openstackgerrit | Dan Nguyen proposed a change to openstack/trove: Partially implements guest agent upgrade strategy https://review.openstack.org/85225 | 14:18 |
| *** grapex has quit IRC | 14:19 | |
| *** grapex has joined #openstack-trove | 14:19 | |
| *** amytron has joined #openstack-trove | 14:20 | |
| *** Longgeek has quit IRC | 14:27 | |
| *** coolsvap|afk is now known as coolsvap | 14:30 | |
| *** rwsu has joined #openstack-trove | 14:40 | |
| *** thedodd has joined #openstack-trove | 14:44 | |
| *** matsuhashi has quit IRC | 14:49 | |
| *** tvoran_ has joined #openstack-trove | 14:49 | |
| *** tvoran has quit IRC | 14:52 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Add backup size validation at restoring https://review.openstack.org/86300 | 14:53 |
| *** haomaiw__ has quit IRC | 14:58 | |
| *** ramishra has joined #openstack-trove | 14:58 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Add logs validation rules for mysql datastore https://review.openstack.org/97814 | 14:58 |
| *** haomaiwang has joined #openstack-trove | 14:58 | |
| *** ramishra has quit IRC | 15:03 | |
| *** sbfox has joined #openstack-trove | 15:03 | |
| *** annashe__ has joined #openstack-trove | 15:07 | |
| *** radez_g0n3 is now known as radez | 15:13 | |
| *** spiffxp- has quit IRC | 15:13 | |
| *** SlickNik has quit IRC | 15:14 | |
| *** SlickN1k is now known as SlickNik | 15:14 | |
| *** coolsvap is now known as coolsvap|afk | 15:14 | |
| *** aginghipster has quit IRC | 15:15 | |
| *** sbfox has quit IRC | 15:26 | |
| *** coolsvap|afk is now known as coolsvap | 15:28 | |
| *** jmontemayor has joined #openstack-trove | 15:29 | |
| *** yidclare has joined #openstack-trove | 15:30 | |
| *** ViswaV has joined #openstack-trove | 15:37 | |
| *** tvoran has joined #openstack-trove | 15:38 | |
| *** annashe__ has quit IRC | 15:39 | |
| *** tvoran__ has joined #openstack-trove | 15:41 | |
| *** tvoran has quit IRC | 15:42 | |
| *** tvoran_ has quit IRC | 15:42 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Fix updated timestamp for SecurityGroup model https://review.openstack.org/97510 | 15:49 |
| *** dmakogon_ is now known as denis_makogon | 15:50 | |
| denis_makogon | amrith, ping, take a look at https://review.openstack.org/97472 (it's related to your comments on https://review.openstack.org/77551) | 15:50 |
| *** sbfox has joined #openstack-trove | 15:52 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/trove: Updated from global requirements https://review.openstack.org/96275 | 15:53 |
| *** ramishra has joined #openstack-trove | 15:59 | |
| openstackgerrit | Greg Lucas proposed a change to openstack/python-troveclient: Add 'slave_of' option for enabling replication https://review.openstack.org/97840 | 16:00 |
| *** ramishra has quit IRC | 16:04 | |
| *** tvoran__ has quit IRC | 16:05 | |
| *** tvoran has joined #openstack-trove | 16:06 | |
| *** haomaiwang has quit IRC | 16:10 | |
| *** Gordon has joined #openstack-trove | 16:11 | |
| *** Gordon is now known as Guest10752 | 16:12 | |
| *** Guest10752 has quit IRC | 16:12 | |
| *** ramishra has joined #openstack-trove | 16:12 | |
| *** demorris has quit IRC | 16:13 | |
| openstackgerrit | Greg Lucas proposed a change to openstack/trove: Add slave_of_id to instance model https://review.openstack.org/97846 | 16:19 |
| *** demorris has joined #openstack-trove | 16:20 | |
| *** ramishra has quit IRC | 16:23 | |
| *** yidclare has quit IRC | 16:24 | |
| *** ramishra has joined #openstack-trove | 16:24 | |
| *** ramishra has quit IRC | 16:27 | |
| *** ramishra has joined #openstack-trove | 16:27 | |
| *** ramishra has quit IRC | 16:27 | |
| *** yidclare has joined #openstack-trove | 16:28 | |
| *** rubick_ has quit IRC | 16:30 | |
| *** ramishra has joined #openstack-trove | 16:34 | |
| anteaya | hi trove was reddwarf at some point way back when, yes? | 16:38 |
| anteaya | there is a gerrit automated account by name of reddwarf and I was wondering if anyone knew who started or maintains the account | 16:39 |
| anteaya | reddwarf reddwarf reddwarf-jenkins@googlegroups.com | 16:39 |
| *** demorris has quit IRC | 16:42 | |
| *** demorris has joined #openstack-trove | 16:42 | |
| *** annashe__ has joined #openstack-trove | 16:44 | |
| *** jcru has quit IRC | 16:46 | |
| *** amytron has quit IRC | 16:47 | |
| *** jcru has joined #openstack-trove | 16:47 | |
| *** ramishra has quit IRC | 16:50 | |
| *** saurabhs has joined #openstack-trove | 16:53 | |
| *** yogeshmehra has joined #openstack-trove | 16:57 | |
| denis_makogon | anteaya, yes, we have reddwarf-ci account | 16:59 |
| anteaya | denis_makogon: great | 16:59 |
| denis_makogon | anteaya, any complains with that ? | 17:00 |
| anteaya | we need to rename it, so the name fits the new standardization of format | 17:00 |
| anteaya | denis_makogon: no complaints | 17:00 |
| anteaya | we just need to select a new name | 17:00 |
| anteaya | so do you want to stay with reddwarf? | 17:00 |
| *** zuqiang has joined #openstack-trove | 17:01 | |
| anteaya | and actually -ci is what we need to add to the name, since according to gerrit that is lacking | 17:02 |
| anteaya | denis_makogon: | 17:02 |
| *** sgotliv has quit IRC | 17:02 | |
| denis_makogon | anteaya, i think you need to talk with SlickNik and vipul, they have direct access to that jenkins | 17:04 |
| anteaya | great thanks | 17:04 |
| anteaya | SlickNik: let me know when you are available | 17:04 |
| anteaya | or vipul | 17:04 |
| *** shakamunyi has quit IRC | 17:13 | |
| *** amytron has joined #openstack-trove | 17:13 | |
| SlickNik | anteaya: Around now | 17:17 |
| SlickNik | anteaya: We use the reddwarf account for integration tests running on a different jenkins box (using the gerrit trigger plugin) | 17:19 |
| anteaya | hey | 17:19 |
| anteaya | cool | 17:19 |
| anteaya | we need to standardize names of automated gerrit accounts | 17:20 |
| anteaya | does the reddwarf account vote? | 17:20 |
| anteaya | or never ever will vote? | 17:20 |
| SlickNik | anteaya: Yes, I'm good with renaming the account if we need to. Perhaps something like trove-ci? | 17:20 |
| SlickNik | Yes, the reddwarf account votes. | 17:20 |
| anteaya | trove-ci works for me | 17:20 |
| anteaya | do you imagine there ever will be any other accounts testing trove? | 17:20 |
| anteaya | or just trove-ci? | 17:20 |
| SlickNik | We could possibly have accounts for third-party testing of datastores in the future. | 17:21 |
| anteaya | sure | 17:22 |
| anteaya | okay so | 17:22 |
| anteaya | trove-ci may or may not cause some naming issues | 17:22 |
| anteaya | if other folks want to call theirs trove-ci as well | 17:23 |
| anteaya | trove-internal-ci? | 17:23 |
| anteaya | or we could just go with trove-ci and deal with fallout later | 17:23 |
| anteaya | I'm curious | 17:23 |
| anteaya | why do you have your own internal testing? | 17:24 |
| SlickNik | How about trove-integration-ci? | 17:24 |
| anteaya | what does it test? | 17:24 |
| anteaya | we can call it that? | 17:24 |
| anteaya | what does it test that the gate doesn't? | 17:24 |
| SlickNik | yes, it runs the trove integration tests. | 17:24 |
| SlickNik | I'm working on adding these tests to the gate, so it's temporary. | 17:24 |
| SlickNik | It's likely not going to be needed after Juno. | 17:25 |
| anteaya | kk | 17:26 |
| SlickNik | FYI: https://wiki.openstack.org/wiki/Governance/TechnicalCommittee/Trove_Gap_Coverage#Concern_.233:_Deprecating_Reddwarf_CI | 17:26 |
| *** sbfox has quit IRC | 17:27 | |
| denis_makogon | SlickNik, i updated patch, could you please take a look at it?https://review.openstack.org/85314 | 17:27 |
| anteaya | SlickNik: awesome that was my next question | 17:29 |
| SlickNik | anteaya: What's the timeline for the renaming? (Asking so I know to be around to update it on the jenkins end :)) | 17:29 |
| anteaya | I don't know | 17:30 |
| anteaya | it will be announced | 17:30 |
| anteaya | we are still working on coming to agreement on format | 17:30 |
| *** sbfox has joined #openstack-trove | 17:30 | |
| anteaya | https://etherpad.openstack.org/p/automated-gerrit-account-naming-format | 17:31 |
| *** rueben has joined #openstack-trove | 17:31 | |
| *** sbfox has quit IRC | 17:31 | |
| SlickNik | anteaya: Okay, sounds good. Let's go with trove-integration-ci as the new name. | 17:32 |
| * anteaya nods | 17:33 | |
| anteaya | thanks | 17:33 |
| SlickNik | anteaya: Thank You! :) | 17:33 |
| *** SushillKM has joined #openstack-trove | 17:35 | |
| *** SushilKM has joined #openstack-trove | 17:37 | |
| *** sbfox has joined #openstack-trove | 17:40 | |
| *** yogeshmehra has quit IRC | 17:40 | |
| *** SushillKM has quit IRC | 17:41 | |
| *** yogeshmehra has joined #openstack-trove | 17:41 | |
| *** yogeshme_ has joined #openstack-trove | 17:43 | |
| *** yogeshmehra has quit IRC | 17:43 | |
| cp16net | so i guess there isnt much to talk about | 17:43 |
| *** eghobo has joined #openstack-trove | 17:44 | |
| esp | yay :) | 17:46 |
| * esp just kidding... | 17:46 | |
| *** cweid has quit IRC | 17:50 | |
| SlickNik | cp16net: Yes, I think it's gonna be a quick meeting. | 17:50 |
| SlickNik | cp16net: A couple of FYIs about the release. | 17:51 |
| SlickNik | cp16net: (Juno-1) | 17:51 |
| cp16net | cool | 17:51 |
| *** cweid has joined #openstack-trove | 17:52 | |
| cp16net | i heard some of the thangs mentioned | 17:52 |
| *** sbfox has quit IRC | 17:52 | |
| *** esp2 has joined #openstack-trove | 17:54 | |
| *** esp2 has quit IRC | 17:55 | |
| *** demorris has quit IRC | 17:55 | |
| amrith | denis_makogon, yt? | 17:56 |
| *** robertmyers has quit IRC | 17:59 | |
| SlickNik | reminder: Weekly trove meeting in #openstack-meeting-alt | 17:59 |
| *** atomic77 has joined #openstack-trove | 18:00 | |
| *** amcrn has joined #openstack-trove | 18:06 | |
| *** sbfox has joined #openstack-trove | 18:07 | |
| *** michael-yu has joined #openstack-trove | 18:08 | |
| *** shakayumi has joined #openstack-trove | 18:08 | |
| *** shakayumi has quit IRC | 18:08 | |
| *** shakayumi has joined #openstack-trove | 18:09 | |
| *** shakayumi has quit IRC | 18:09 | |
| *** sgotliv has joined #openstack-trove | 18:25 | |
| *** atomic77 has left #openstack-trove | 18:28 | |
| *** sbfox has quit IRC | 18:29 | |
| iccha1 | SlickNik: so here are my updates , Associate instance flavors with datastores needs to be bumped to juno 2, https://blueprints.launchpad.net/trove/+spec/list-datastore-type-and-versions needs to be included in juno 1 and and hope to get this done this week https://blueprints.launchpad.net/trove/+spec/update-instance-name | 18:33 |
| SlickNik | iccha1: just updated https://blueprints.launchpad.net/trove/+spec/list-datastore-type-and-versions | 18:33 |
| iccha1 | thanks SlickNik ! | 18:34 |
| SlickNik | iccha1: and bumped https://blueprints.launchpad.net/trove/+spec/associate-flavors-datastores to Juno-2. | 18:35 |
| SlickNik | Thank you! | 18:35 |
| *** r0j4z0 has quit IRC | 18:41 | |
| *** CAP1089 has joined #openstack-trove | 18:41 | |
| *** CAP1089 is now known as r0j4z0 | 18:41 | |
| *** denis_makogon_ has joined #openstack-trove | 18:44 | |
| *** denis_makogon has quit IRC | 18:45 | |
| *** denis_makogon_ is now known as denis_makogon | 18:45 | |
| *** dmakogon_ has joined #openstack-trove | 18:45 | |
| denis_makogon | vipul, ping | 18:45 |
| vipul | denis_makogon: what's up | 18:45 |
| denis_makogon | vipul, need to talk to you | 18:46 |
| denis_makogon | vipul, about https://review.openstack.org/96795 | 18:47 |
| vipul | uh oh | 18:47 |
| vipul | oh ok | 18:47 |
| denis_makogon | vipul, do you have like 5 mins ? | 18:47 |
| vipul | yep | 18:47 |
| denis_makogon | cool | 18:47 |
| denis_makogon | vipul, i've read your comments | 18:47 |
| denis_makogon | vipul, the only one answer for you that i have - nobody actually running heat-base integration tests | 18:47 |
| denis_makogon | vipul, inside nova/cinder-based int-tests everything is ok | 18:48 |
| vipul | denis_makogon: Ok, so it seems like we already have a Test that checks that a security group exists after an instance has been created | 18:49 |
| denis_makogon | vipul, yes | 18:49 |
| vipul | and when you turn on Heat, that specific test fails? | 18:49 |
| denis_makogon | vipul, yes | 18:49 |
| denis_makogon | setUp fails | 18:49 |
| vipul | Alright.. then we're good.. | 18:49 |
| vipul | the tests caught something :) | 18:49 |
| denis_makogon | test caught missing security group for instance | 18:50 |
| denis_makogon | at least in Trove backend | 18:50 |
| vipul | ok cool, i'm good with it.. | 18:50 |
| denis_makogon | thanks | 18:50 |
| denis_makogon | vipul, last question, i guess | 18:52 |
| *** yidclare has quit IRC | 18:52 | |
| denis_makogon | vipul, could you ask HP guys what is happening with gate? | 18:52 |
| denis_makogon | 17:30:39 Connection reset by peer | 18:52 |
| denis_makogon | at least 4 times in a row | 18:52 |
| denis_makogon | full log https://rdjenkins.dyndns.org/job/Trove-Gate/3947/console | 18:53 |
| *** zuqiang has quit IRC | 18:53 | |
| vipul | denis_makogon: I'll look into what's going | 18:54 |
| vipul | on | 18:54 |
| denis_makogon | vipul, thanks again | 18:54 |
| denis_makogon | grapex, ping | 18:57 |
| *** demorris has joined #openstack-trove | 18:57 | |
| vipul | denis_makogon: might have been a one-time issue.. it failed to delete the Devstack VM.. I manually cleaned it up.. you should re-trigger the job | 18:58 |
| denis_makogon | vipul, i already did that | 18:58 |
| denis_makogon | vipul, let me find another log with the same issue | 18:58 |
| vipul | denis_makogon: ok let me know if it happens again.. | 18:59 |
| denis_makogon | vipul, ok | 18:59 |
| denis_makogon | vipul, https://rdjenkins.dyndns.org/job/Trove-Gate/3946/console | 19:00 |
| denis_makogon | 15:26:44 URLError: <urlopen error [Errno 104] Connection reset by peer | 19:01 |
| SlickNik | denis_makogon: That seems like a pypi issue while fetching the package. | 19:02 |
| denis_makogon | SlickNik, i though jenkins pulling packages from OS mirror ? | 19:03 |
| denis_makogon | SlickNik, isn't it ? | 19:03 |
| *** SushilKM has quit IRC | 19:04 | |
| SlickNik | denis_makogon: rdjenkins is still pulling from pypi; it doesn't use nodepool and puppet to do the same infra config. | 19:04 |
| denis_makogon | SlickNik, makes sense, thanks | 19:04 |
| SlickNik | One of the many reasons we're trying to move away from rdjenkins :) | 19:05 |
| denis_makogon | SlickNik, vipul, saw this bug-report, https://bugs.launchpad.net/trove/+bug/1324995, what do you think about it ? | 19:05 |
| vipul | i would personally like to see guest_info go away.. since that relies on file_injection | 19:08 |
| denis_makogon | vipul, me neither | 19:09 |
| *** kevinconway has quit IRC | 19:10 | |
| denis_makogon | vipul, but actually the question is not about in which conf file we have to put packages description | 19:10 |
| denis_makogon | vipul, question is to pass them through config in general | 19:11 |
| denis_makogon | i guess this bug-report is going to be revisited when we would have full heat integration | 19:11 |
| denis_makogon | since heat can do software deployment through cfn-tools | 19:11 |
| denis_makogon | then we would be able to put packages directly to heat templates | 19:12 |
| openstackgerrit | Christian Berendt proposed a change to openstack/python-troveclient: Overwrite HelpFormatter constructur to extend argument column https://review.openstack.org/97886 | 19:17 |
| *** sgotliv has quit IRC | 19:20 | |
| *** sgotliv has joined #openstack-trove | 19:21 | |
| SlickNik | denis_makogon: Also, we should get grapex's input when discussing this. | 19:22 |
| SlickNik | He might have actually run into some issues in production with this, that we might be overlooking. | 19:22 |
| denis_makogon | SlickNik, i pinged him, seems that he's out | 19:22 |
| *** sbfox has joined #openstack-trove | 19:22 | |
| grapex | denis_makogon SlickNik: I'm back | 19:23 |
| denis_makogon | grapex, cool | 19:23 |
| grapex | sorry... let me catch up- unless you guys want to give me a TL;dr; | 19:23 |
| denis_makogon | grapex, we were talking about your bug-report, seems it's very interesting | 19:24 |
| grapex | Oh yeah- to put the package name somewhere the guest could see it on start up? | 19:24 |
| *** sbfox1 has joined #openstack-trove | 19:25 | |
| denis_makogon | grapex, can we actually split prepare call into two ? | 19:25 |
| grapex | I logged this- I would've done it myself but I'm knea-deep in the Rax guest, Sneaky Pete, at the moment, and I want to make sure it works well for Reffy too. But I do think the fundamental problem- that there's this race condition that only doesn't affect us because the prepare call is normally there- should be fixed as its bad design. | 19:26 |
| *** sbfox has quit IRC | 19:26 | |
| grapex | vipul: As for your concern, currently the guest *does* depend on file injection as it needs to know its own ID on start up no matter what. | 19:27 |
| grapex | When we move away from file injection I'm sure adding one more property to that file won't add technical debt. | 19:27 |
| denis_makogon | grapex, i can track down that bug, but need your help with that | 19:27 |
| grapex | The root issue is the guest has to know its ID and other things at start up. Currently that's done with file injection, but we may find other methods in the future. The issue will still remain though. | 19:28 |
| grapex | even if we change how we solve it | 19:28 |
| denis_makogon | grapex, so, you're telling that guest should deploy packages before RPC service starts ? | 19:28 |
| grapex | denis_makogon: No | 19:28 |
| grapex | I'm just saying the package is important enough for the guest to know about, so it can query if the datastore has been installed. | 19:28 |
| grapex | and if not behave appropriately | 19:29 |
| denis_makogon | grapex, ok, get that, but is it ok to have two ways to mention packages (from datastore registry and CONFIG) ? | 19:30 |
| grapex | two ways? | 19:31 |
| grapex | There wouldn't be two ways | 19:31 |
| grapex | Trove writes the guest config already. So the datastore registry would still be the source of truth. | 19:32 |
| denis_makogon | ah | 19:32 |
| denis_makogon | finally i get that | 19:32 |
| denis_makogon | grapex, last question, how should we deal this guest API ? | 19:32 |
| SlickNik | grapex: Yeah, we probably should fix that to switch to metadata (instead of file injection) at some point, but I agree that it probably makes sense for this (package) to be included on the guest from the outset. | 19:33 |
| SlickNik | and by that, I mean the file injection piece | 19:33 |
| grapex | SlickNik: Agreed | 19:33 |
| grapex | denis_makogon: I don't understand- which part of the guest API? | 19:33 |
| denis_makogon | grapex, prepare call | 19:33 |
| denis_makogon | if we would have packages in config file then we should ignore packages from parameters and use from CONFIG | 19:34 |
| grapex | denis_makogon: Agreed | 19:34 |
| denis_makogon | grapex, that's what i was asking for =) | 19:34 |
| denis_makogon | grapex, can i take that bug ? | 19:35 |
| grapex | denis_makogon: Sure. | 19:35 |
| denis_makogon | grapex, thanks | 19:36 |
| denis_makogon | grapex, last thing, could you re-approve https://review.openstack.org/#/c/96780/ it again ? | 19:36 |
| grapex | denis_makogon: Just to make sure we're on the same page, this bug has a very small objective- add the package name to that config, and check to see if the package is installed on start up- that's all there is. Right? | 19:37 |
| grapex | denis_makogon: Approved | 19:37 |
| denis_makogon | grapex, yes, put packages(could be more than 1 for datastore version) into CONFIG, check if packages installed on guest start up | 19:39 |
| grapex | Cool | 19:39 |
| grapex | denis_makogon: Awesome! Thanks | 19:39 |
| denis_makogon | grapex, last thing | 19:39 |
| denis_makogon | grapex, "on start up" - when RPC service starts or when prepare comes ? | 19:39 |
| *** rueben has quit IRC | 19:40 | |
| grapex | denis_makogon: When it starts up | 19:41 |
| denis_makogon | grapex, "it" means guest =) | 19:41 |
| grapex | Yes | 19:41 |
| grapex | denis_makogon: Hold on... digging through reference guest code | 19:42 |
| *** mattgriffin has quit IRC | 19:43 | |
| denis_makogon | the only thing that is not so clear for me is behavior of the guest when packages are installed or not | 19:43 |
| *** rueben has joined #openstack-trove | 19:44 | |
| grapex | denis_makogon: Your right. It isn't clear because currently the thing that updates status makes no sense | 19:45 |
| grapex | So, trove.guestagent.datastore.service BaseDbStatus needs to be updated so its initializer takes an argument "is_installed" | 19:45 |
| denis_makogon | let me take a look | 19:46 |
| *** robertmyers has joined #openstack-trove | 19:46 | |
| grapex | There is already a property "is_installed" which is True only if its not certain statuses... | 19:46 |
| grapex | how is this even working? | 19:46 |
| grapex | SlickNik / vipul: What happens if the guest agent is restarted? Doesn't it return "True" for is_installed? | 19:47 |
| grapex | I mean "False" for is_installed? | 19:47 |
| denis_makogon | grapex, each time RPC service gets restarted, new status object appers | 19:50 |
| denis_makogon | that's why it's actually never None | 19:50 |
| grapex | Yes- that's what makes no sense to me. With the introduction to Conductor, the guest agent can't retrieve its status, so it can't know the prepare call was already received. | 19:51 |
| denis_makogon | grapex, guest doesn't retrieveing its status | 19:51 |
| grapex | denis_makogon: I know | 19:51 |
| grapex | I am commenting that I don't know how it works | 19:52 |
| grapex | because the way the code is written is confusing | 19:52 |
| grapex | denis_makogon: Maybe I should look at this... | 19:52 |
| denis_makogon | grapex, of course, if you want | 19:54 |
| denis_makogon | grapex, the vely last thing for today, could you please re-review https://review.openstack.org/#/c/86300/ this one ? All your suggestions i took into account | 19:56 |
| *** fifieldt_ has joined #openstack-trove | 20:00 | |
| *** yogeshme_ has quit IRC | 20:01 | |
| *** yogeshmehra has joined #openstack-trove | 20:02 | |
| *** mattgriffin has joined #openstack-trove | 20:03 | |
| *** yogeshme_ has joined #openstack-trove | 20:04 | |
| *** fifieldt-afk has quit IRC | 20:04 | |
| *** yogeshmehra has quit IRC | 20:04 | |
| *** mattgrif_ has joined #openstack-trove | 20:05 | |
| *** mattgriffin has quit IRC | 20:06 | |
| *** mattgrif_ has quit IRC | 20:10 | |
| *** mattgriffin has joined #openstack-trove | 20:10 | |
| grapex | denis_makogon: Done, a few more nit picks though | 20:10 |
| denis_makogon | grapex, thanks | 20:11 |
| *** yidclare has joined #openstack-trove | 20:11 | |
| *** denis_makogon has quit IRC | 20:12 | |
| *** PeterSchwaller_ has joined #openstack-trove | 20:13 | |
| *** PeterSchwaller has quit IRC | 20:15 | |
| *** mattgriffin has quit IRC | 20:18 | |
| *** sbfox1 has quit IRC | 20:19 | |
| *** yidclare has quit IRC | 20:23 | |
| *** jcromer has joined #openstack-trove | 20:27 | |
| jcromer | hello, is anyone around? I have a problem that i need some help on. Would like to ask a quick question. | 20:28 |
| jcromer | I am seeing Cannot find compute instance None for instance in my taskmanager log | 20:30 |
| juice | jcromer: how did you build your stack? devstack? | 20:30 |
| jcromer | juice: no built from ubuntu packages | 20:31 |
| juice | depending on the context of that error message it may be non issue | 20:31 |
| *** michael-yu has quit IRC | 20:31 | |
| jcromer | juice: it is on its own node | 20:31 |
| juice | can you paste the entire and exact log message | 20:31 |
| juice | jcromer: if I understand what you are saying, you spun up a vm via nova and on that vm did and apt-get install trove-taskmanager | 20:32 |
| jcromer | juice: 2014-06-04 16:30:08.902 12717 ERROR trove.common.exception [req-b32f4af6-80d3-4f43-9d57-cb31eda9a11e 8d456978afa849feb46ea64f494107ab 2cce88b4c1944f7cb18701c6fc5e36f7 - - -] Cannot find compute instance None for instance 8db902c0-676a-48ac-8e78-fe854ecd6605. | 20:32 |
| jcromer | no | 20:32 |
| *** boden has quit IRC | 20:32 | |
| jcromer | sorry | 20:32 |
| jcromer | juice: we spun up a machine and installed trove package on it | 20:33 |
| juice | jcromer: what configuration/installation guide did you follow? | 20:33 |
| *** michael-yu has joined #openstack-trove | 20:34 | |
| jcromer | juice: http://docs.openstack.org/icehouse/install-guide/install/yum/content/trove-install.html | 20:34 |
| juice | jcromer: did you configure/customize /etc/trove/trove*.conf | 20:34 |
| jcromer | juice: yes | 20:34 |
| jcromer | juice: sorry used http://docs.openstack.org/icehouse/install-guide/install/apt/content/trove-install.html | 20:35 |
| jcromer | juice: it is on ubuntu 14.04 | 20:35 |
| juice | ok - if you do a nova list, with the same creds that you used to create the trove instance do you get any instances listed | 20:35 |
| esp | SlickNik: ping | 20:35 |
| juice | jcromer: afaik we only test and verify on 12.04 | 20:35 |
| juice | jcromer: also can you do a "trove flavor-list" | 20:36 |
| SlickNik | esp: 'sup? | 20:36 |
| jcromer | juice: well i don't have a nova-client installed on the trove node, but if i used the same user on nova controller i do get instances just not database | 20:36 |
| esp | SlickNik: I just pulled down a new trove-integration and ran the tests. | 20:36 |
| esp | looks like they are stuck on a couple of tests | 20:37 |
| esp | https://gist.github.com/dan1/c4194d0a8de9d027041d | 20:37 |
| jcromer | juice: trove flavor-list works | 20:37 |
| juice | so it looks like at some point you tried to create a trove instance, trove started the process of provisioning the instance but failed to connect to nova in some fashion and now you have a half-baked instance in trove | 20:37 |
| jcromer | juice: so does trove list | 20:37 |
| juice | trove list shows your instance? | 20:37 |
| jcromer | juice: i do have one in error status and two in build | 20:37 |
| juice | if you do a "trove show <instance id>" on any of those instances do you they return and do they show a server id? | 20:39 |
| jcromer | juice: they return and show an id, but status of error on the one that is error | 20:40 |
| jcromer | juice: let me try one of the ones stuck perpetually in build | 20:40 |
| juice | jcromer: what flavor are you using to boot the instances? | 20:40 |
| jcromer | juice: 2 | 20:40 |
| jcromer | juice: the ones that are stuck in build state I am unable to delete | 20:41 |
| SlickNik | esp: strange. Can you check your trove-api.log to see what the actual ERROR was when the test_configuration_is_applied_to_instance test was executed? | 20:41 |
| esp | SlickNik: yep | 20:42 |
| * esp looking ... | 20:42 | |
| *** yidclare has joined #openstack-trove | 20:42 | |
| SlickNik | esp: I've kicked off a pure rdjenkins run to verify, but it looks like there were successful runs of the int-tests from earlier this morning (https://rdjenkins.dyndns.org/job/Trove-Gate/3949/) | 20:43 |
| juice | jcromer: that is unfortunately the case unless you set a flag in task manager to convert them to ERROR in these conditions | 20:43 |
| *** yidclare has quit IRC | 20:43 | |
| SlickNik | Not 100% sure but perhaps the issue may be timeout / environment related. | 20:43 |
| juice | jcromer: the way I get rid of those is to manually mark them as ERROR in the trove db and then through the trove command line, delete them | 20:43 |
| juice | there is a status flag on the instances table that can be updated | 20:44 |
| jcromer | juice: i will check the db | 20:44 |
| juice | this is a PITA...but its what I do | 20:44 |
| juice | perhaps someone else has a better solution | 20:44 |
| esp | SlickNik: 2014-06-04 13:30:53.761 ^[[01;31mERROR trove.extensions.mgmt.instances.models [^[[00;36m-^[[01;31m] ^[[01;35m^[[01;31mResource 22b46dcd-c4cb-4e4c-8f6d-58fe1015cff2 can not be retrieved.^[[00m | 20:45 |
| *** mattgriffin has joined #openstack-trove | 20:45 | |
| *** sbfox has joined #openstack-trove | 20:45 | |
| juice | jcromer: the best place to check is in the logs (nova and trove) to see what is the reason for the error | 20:46 |
| esp | SlickNik: k, let me try to find the secret time out :) | 20:46 |
| SlickNik | esp: It's a decorator on the test class. But I'd first look in the log to see if that's indeed the issue. | 20:47 |
| *** sbfox has quit IRC | 20:50 | |
| *** sbfox has joined #openstack-trove | 20:50 | |
| jcromer | juice: what are you setting to error in the instances table | 20:53 |
| juice | I believe it is status or status_id | 20:53 |
| juice | you will see a status(_id) and description | 20:53 |
| juice | description is not important | 20:53 |
| juice | I believe the value 8 is valid for a type of ERROR | 20:53 |
| jcromer | i see server_status column | 20:53 |
| juice | jcromer: is it numeric? | 20:54 |
| jcromer | varchar | 20:54 |
| juice | let me take a look | 20:54 |
| openstackgerrit | A change was merged to openstack/trove: pep8: Ignore locale files https://review.openstack.org/97087 | 20:55 |
| esp | SlickNik: k, thx! | 20:56 |
| juice | sorry the correct table is server_statuses jcromer | 20:57 |
| juice | service_statuses (sheesh) | 20:57 |
| esp | SlickNik: 2014-06-04 13:27:50.289 ERROR trove.guestagent.api [-] Timeout while waiting on RPC response - topic: "guestagent.a485beef-b7ec-480d-8815-48ccc5621572", RPC method: "get_filesystem_stats" info: "<unknown>" | 20:57 |
| esp | SlickNik: I do see a lot of these. | 20:57 |
| esp | I’ll see if messing with the timeout helps | 20:58 |
| *** sbfox has quit IRC | 20:58 | |
| *** pdmars has quit IRC | 20:58 | |
| *** michael-yu has quit IRC | 21:05 | |
| *** michael-yu has joined #openstack-trove | 21:07 | |
| *** amcrn_ has joined #openstack-trove | 21:07 | |
| *** sbfox has joined #openstack-trove | 21:08 | |
| *** amcrn has quit IRC | 21:09 | |
| openstackgerrit | Peter Stachowski proposed a change to openstack/python-troveclient: Added optional marker/limit to list command https://review.openstack.org/97936 | 21:10 |
| jcromer | juice: 8 doesn't seem to work | 21:11 |
| jcromer | juice: still in build status | 21:11 |
| *** yidclare has joined #openstack-trove | 21:12 | |
| *** sgotliv has quit IRC | 21:13 | |
| *** yogeshme_ has quit IRC | 21:15 | |
| *** yogeshmehra has joined #openstack-trove | 21:16 | |
| *** yogeshmehra has quit IRC | 21:19 | |
| *** yogeshmehra has joined #openstack-trove | 21:20 | |
| *** sbfox has quit IRC | 21:20 | |
| *** rueben has quit IRC | 21:26 | |
| *** sgotliv has joined #openstack-trove | 21:29 | |
| *** jcromer has quit IRC | 21:30 | |
| *** rueben has joined #openstack-trove | 21:32 | |
| *** demorris has quit IRC | 21:33 | |
| *** sgotliv has quit IRC | 21:37 | |
| *** demorris has joined #openstack-trove | 21:38 | |
| *** radez is now known as radez_g0n3 | 21:45 | |
| *** yogeshmehra has quit IRC | 21:46 | |
| *** yogeshmehra has joined #openstack-trove | 21:47 | |
| *** rueben_ has joined #openstack-trove | 21:49 | |
| *** yogeshmehra has quit IRC | 21:51 | |
| *** rueben has quit IRC | 21:53 | |
| *** rueben_ has quit IRC | 21:53 | |
| *** Barker has quit IRC | 21:58 | |
| openstackgerrit | Peter Stachowski proposed a change to openstack/trove: Added an int-test for user-update-attributes https://review.openstack.org/97952 | 21:59 |
| *** sbfox has joined #openstack-trove | 22:01 | |
| *** sbfox has quit IRC | 22:11 | |
| *** sbfox has joined #openstack-trove | 22:11 | |
| *** demorris has quit IRC | 22:11 | |
| *** sbfox has quit IRC | 22:11 | |
| *** achampion has quit IRC | 22:11 | |
| *** yogeshmehra has joined #openstack-trove | 22:12 | |
| *** amcrn_ has quit IRC | 22:24 | |
| *** jmontemayor has quit IRC | 22:29 | |
| *** thedodd has quit IRC | 22:30 | |
| *** sbfox has joined #openstack-trove | 22:32 | |
| grapex | esmute SlickNik vipul: I've got another concern nagging me on this pull request to add datastore metadata to backups | 22:37 |
| grapex | So we're associating the datastore version name, and datastore name - two strings- to a backup, due to cross region funtimes | 22:38 |
| grapex | I guess my problem is we are treating that tuple as the definitive ID for a datastore | 22:38 |
| grapex | amcrn: ^^ | 22:38 |
| grapex | So honestly, what purpose is the ID serving at this point? | 22:39 |
| vipul | grapex: Yep, that's what it ends up being.. I don't see another way to do this.. unless we add region:id or something.. and enforce that Trove goes to that region to lookup the actual datastore type | 22:39 |
| grapex | If you work with datastores in a region, you use the nasty ID, but if working across regions, we use strings? | 22:40 |
| grapex | Yeah | 22:40 |
| hub_cap | id prefer the opposite | 22:40 |
| grapex | vipul: So the root problem (which won't get solved anytime soon) is the IDs should maybe, somehow, be globally unique amongst regions | 22:40 |
| *** amcrn has joined #openstack-trove | 22:40 | |
| hub_cap | uuid across region, string name shorthand within a region | 22:40 |
| vipul | grapex: You mean ID's shoudl be the same across regions? | 22:41 |
| esmute | how does other openstacks project solve this cross-region ID issues? | 22:41 |
| vipul | the problem is that another region doesn't have that ID.. | 22:41 |
| hub_cap | aww hell vipul why u gotta be right | 22:41 |
| grapex | also, if these IDs are guids, shouldn't they be unique? | 22:42 |
| grapex | What if we had alias ids? | 22:42 |
| vipul | the only way to solve it is amcrn's approach where.. we have a template like <region>:<id> -- and Trove is aware of the other regions and does lookups | 22:42 |
| esmute | unless the database is shared across regions | 22:42 |
| amcrn | look at what conversation i just dropped into ;D | 22:43 |
| vipul | esmute: which we don't want to enforce | 22:43 |
| grapex | So here's my issue- until now, we could, if we wanted, change the names of datastores or versions | 22:43 |
| *** cweid has quit IRC | 22:43 | |
| SlickNik | hi amcrn, what have we here. | 22:43 |
| grapex | I mean yes, users would have to get used to different names- but everything else would work | 22:43 |
| grapex | this new thing is saying they have to be eternally constant | 22:43 |
| grapex | so if you use a name for something you later badly wish you used for something else, too bad | 22:44 |
| hub_cap | amcrn: GET A BOUNCER!!! | 22:44 |
| vipul | grapex: not necessarily.. it works as long as you change it in all regions | 22:44 |
| amcrn | hub_cap: i'm caught up, i read the eavesdrop | 22:44 |
| hub_cap | ;) | 22:45 |
| grapex | vipul: Change it in all regions, and then update every existing backup record in the database... in all regions | 22:45 |
| hub_cap | amcrn: bt srsly, et a bouncer | 22:45 |
| hub_cap | and ill buy a U and a G alex | 22:45 |
| grapex | I feel like we're violating a law here by using these names for the ID. We're making renaming them close to impossible | 22:45 |
| vipul | grapex: nah.. the db records are fkey'd by IDs.. just gotta make sure the names are consistent across regions | 22:45 |
| grapex | I think aliases really would solve this problem easier as a short term solution | 22:45 |
| grapex | maybe we need a global UUID | 22:46 |
| grapex | for datastore versions | 22:46 |
| grapex | if you want cross-region functionality, make sure they align | 22:46 |
| hub_cap | i think thats sane | 22:46 |
| hub_cap | lets call it "mysql" | 22:46 |
| hub_cap | and "redis" | 22:46 |
| grapex | we'll call it a UUUID | 22:46 |
| grapex | or maybe U3ID for short | 22:46 |
| hub_cap | U^3ID | 22:46 |
| vipul | how about call it a tuple of datastore name and version :) | 22:46 |
| SlickNik | better yet, have a tuple | 22:47 |
| hub_cap | vipul: too wordy | 22:47 |
| vipul | honestly isn't that the sme thing | 22:47 |
| grapex | vipul: Then we hit the problem I've been describing | 22:47 |
| vipul | sure.. but same arument.. if you care about cross region.. can't you just enforce the name is consistent? | 22:47 |
| vipul | instead of introducing yet another thing | 22:47 |
| esmute | we can have "datastore_name:version_name" as the ID... a tuple just like it was suggested | 22:48 |
| grapex | I guess my problem is we *should* be introducing another thing instead of relying on a kind of idiom that becomes a rule people have to know to follow | 22:48 |
| grapex | and they don't realize they didn't follow it until its too late | 22:48 |
| esmute | keep in mind that this are just meta data for backups... no where else in trove will they care or refer to datastore by their name | 22:49 |
| vipul | yea i get your point.. it's a little sneaky.. | 22:49 |
| esmute | once the backups are copied, their records will contains id for datastore | 22:49 |
| vipul | esmute: won't this metadata be used in some way when you create an instance from a backup in another region? | 22:49 |
| *** NehaV1 has quit IRC | 22:49 | |
| esmute | vipul: Nop... when the backup is 'copied', a new backup record is created, which contains datastore id (instead of names) | 22:50 |
| esmute | and then everything would be the same after | 22:50 |
| vipul | ok.. so the backup-copy call requires the metadata? | 22:50 |
| esmute | that datstore name is only needed for the metadata which is stored in swift | 22:50 |
| esmute | vipul: correct | 22:50 |
| esmute | that datastore name, (or tuple or GUUID ) wont queried or saved in the trove DB | 22:51 |
| esmute | wont be queried* | 22:51 |
| grapex | Maybe the compromise here that keeps us moving forward is to... also save the ID | 22:52 |
| vipul | but the tuple has to be understood by the Trove.. so it can associate the correct datastore id? | 22:52 |
| grapex | so we save datastore name, version name (the string tuple) and the datastore version ID. | 22:52 |
| SlickNik | So another option is to require the backup-copy call to require the user to input which local datastore the remote backup corresponds too. | 22:52 |
| esmute | vipul: only the backupagent needs to understand that... | 22:53 |
| vipul | SlickNik: then the issue is they could totally screw the backup up by choosing the wrong datastore | 22:53 |
| SlickNik | And if the user screws that up, and there is a mismatch, we'll fail horribly, and well *shrug* PEBCAK. | 22:53 |
| esmute | the backup agent download backup from swift and parse metadata file in swift (which contains the datastore name), and then it create a new backup record with the datastore ID | 22:53 |
| vipul | grapex: what does saving the ID give us.. i guess i'm trying ot figure out what the BackupAgent would do with that ID | 22:54 |
| SlickNik | vipul: I said it was an option; I didn't necessarily say it was a *good* one. :) | 22:55 |
| grapex | esmute: Why would the agent do that and not Trove? | 22:55 |
| vipul | yea, actualy the copy shouldn't go to any Agnet.. it's a server side thing | 22:55 |
| grapex | vipul: I am worried we'll not like having this one feature of Trove not use the datastore_version.id | 22:56 |
| grapex | so it gives us an escape hatch for when things improve | 22:56 |
| esmute | oops.. you are right... the copy is done by the TM... ignore what i said | 22:56 |
| grapex | esmute vipul: So quick questions about you use case. Sorry I'm rusty on how backups work. The idea is someone can make a backup in region A, then restore to region B, right? | 22:57 |
| vipul | grapex: yea that would be fine.. except that you may poentially have a ID as part of metadata that doesn't exist in that region after the copy is done | 22:57 |
| esmute | the agent creates the metadata that gets put in swift... | 22:57 |
| vipul | grapex: that's the goal | 22:57 |
| esmute | grapex: yes and no.. there is a intermediate work | 22:57 |
| *** demorris has joined #openstack-trove | 22:57 | |
| grapex | Do they do this by passing in a swift url that came from region A? | 22:57 |
| esmute | the backup in A is copied to region B.... when the user can just restore normarly as if the backup was created in region B | 22:58 |
| *** yogeshmehra has quit IRC | 22:58 | |
| vipul | I think there are two approaches you can take.. 1) directly pass a ID/ swift URL from region A to region B .. or 2) copy the backup to region B | 22:58 |
| vipul | esmute: was tehre a reasion 2) was chosen? | 22:58 |
| esmute | the backup-copy will stream backup in region A to the swift in region B and create the backup record in region B | 22:58 |
| *** yogeshmehra has joined #openstack-trove | 22:58 | |
| grapex | Also, how are checksums handled? | 22:59 |
| *** shakayumi has joined #openstack-trove | 22:59 | |
| esmute | vipul: i didnt want to make changes in the restore.... | 22:59 |
| *** jcru has quit IRC | 22:59 | |
| SlickNik | grapex: The idea is that someone makes a backup X in region A. Then creates a "copied" backup Y in region B (copied from the swift files and metadata of backup X), | 22:59 |
| esmute | it was much easier to just stream copy the backup from one swift to another and create a DB record | 22:59 |
| vipul | SlickNik, esmute: does this issue go away if we just did option 1? | 23:00 |
| vipul | we talked about a URI scheme before when reviwing this BP | 23:00 |
| SlickNik | grapex: Once the new "copied" backup Y exists in region B, you can do a restore using backup Y (as you can today). | 23:00 |
| vipul | where we'd use that region:ID / region:url to a Instance-create call | 23:00 |
| grapex | SlickNik: What code is making the copy from region A to region B? | 23:00 |
| esmute | grapex: i havent put it up.. let me get a gist | 23:01 |
| SlickNik | grapex: Some code that esmute was working on, I think. :) | 23:01 |
| grapex | At some point it seems you have to copy the checksum from region A to region B- you can't assume the Swift metadata's checksum is accurate as the user can manipulate it | 23:01 |
| grapex | unless I'm wrong, at that point the Trove databases in regions A and B have to communicate with each other | 23:02 |
| *** saurabhs has left #openstack-trove | 23:02 | |
| esmute | grapex: you are right.. if users manipulate the backup or metadata, the backup copy will be error.. or the restore will error out | 23:02 |
| esmute | https://gist.github.com/kokhang/59f725df728854e2f3fd | 23:04 |
| grapex | if the db's are talking, Trove code could sort out the mapping of datastore_version.IDs between A and B. At that one small point in time, they could even do this using the datastore name and datastore version version strings until a more perfect solution is found (which admittedly will be never) | 23:04 |
| esmute | that is the code that does the copy.. it is done in the TM | 23:04 |
| esmute | grapex: db are not the one talking... swift are the one talking.. so we assume that swift is globally shared across regions | 23:05 |
| SlickNik | esmute: I think what grapex is saying is that you can't get away with just pulling the data from swift since that data could be compromised. | 23:05 |
| grapex | But Swift can't be trusted 100% because the user has access to and can change any of the metadata. | 23:05 |
| esmute | what this code does is just pull backup from region A in swift, parse the metadata in region A, streams to region B with new metadata, and creates a backup record in the DB | 23:06 |
| SlickNik | esmute: And without verifying the checksum, there's really no way you can tell. | 23:06 |
| *** tvoran has quit IRC | 23:06 | |
| amcrn | but in short, you'll guaranteed need at a minimum *some* replication of state; take clusters for instance. the cluster-id itself should be understood, irrespective of the region endpoint. | 23:07 |
| esmute | the checksum is read from the source metadata | 23:08 |
| grapex | Can't we solve this by introducing a mgmt API to Trove that could fetch this data? | 23:08 |
| grapex | Region B gets a request for a restore, using a backup ID from region A | 23:08 |
| grapex | so it calls region A on its mgmt super-admin hotline to get the skinny on what the backup from A looks like: what the datastore originally involved was, the checksum, and the swift URL | 23:09 |
| esmute | grapex: but the idea was that the users would use this backup-copy... having it in mgmt would prevent that | 23:09 |
| grapex | then the Trove on region B copies things as esmute describes but applies the original checksum and other equivalent metadata to the metadata of its fresh copy in Swift's region B | 23:09 |
| grapex | esmute: it sounds like this isn't a backup copy then but an import of arbitrary data in Swift | 23:10 |
| *** yogeshmehra has quit IRC | 23:10 | |
| grapex | which is a good idea but it introduces difference | 23:10 |
| amcrn | esmute: "so we assume that swift is globally shared across regions" => we can't make that assumption | 23:10 |
| esmute | amcrn: if we dont, then how would one access the objects in another swift.. im assuming that the token in keystones are shared across region | 23:12 |
| amcrn | this is going to have to behave a lot like federated identity in keystone; either you'll need a galera-like cluster to replicate certain 99% read-only metadata tables (like datastore, datastore-version, etc.), or something within trove will have to initiate multi-writes to every known registered endpoint | 23:12 |
| amcrn | esmute: identity can be the same across regions, but that doesn't mean that the swift deployments are | 23:12 |
| amcrn | with the multi-writes going to a mgmt-api | 23:13 |
| *** demorris has quit IRC | 23:14 | |
| esmute | if that is the case, it wouldnt matter whether we use datastore id or name... this is a new problem we need to discuss | 23:14 |
| grapex | esmute: So... getting back to the pull request, I guess I'm ok with sticking the version, name, whatever into the metadata. It looks like the datastore_version.id is being associated with the backup in the database which is enough to make me happy with this pull request. | 23:14 |
| grapex | esmute: Although- I wonder if we should be returning the datastore version ID in the view with the backup. | 23:15 |
| *** shakayumi has quit IRC | 23:15 | |
| grapex | I know UI writers will want the name and version though so maybe I should stop talking. | 23:15 |
| amcrn | grapex: that's inferrable from the uuid though :) | 23:16 |
| grapex | amcrn: agreed | 23:16 |
| grapex | amcrn: I know right away though our UI guys will complain. But I can't help but 100% agree with you. | 23:16 |
| amcrn | grapex: well, on list-view, having to make an additional call does bring down responsiveness a bit, so i kind of see their angle | 23:16 |
| esmute | amcrn: i think we need to have a separate talk about the whole cross-regions comunication between troves... | 23:17 |
| esmute | grapex: i thought names would be more readble :P | 23:17 |
| grapex | esmute: We do. I just wanted to bring it up to understand your motivations. | 23:17 |
| amcrn | esmute: 100% agree my friend, it's definitely a tricky problem :) | 23:17 |
| grapex | esmute: Thanks for explaining things. | 23:17 |
| grapex | esmute: So I am ok with the views returning that extra data. I think anyone writing a UI will be thrilled. | 23:18 |
| grapex | I do think we should add the datastore version ID though | 23:18 |
| esmute | grapex, amcrn: im pretty much done with the cross-region feature... i havent put it up since i am waiting for https://review.openstack.org/#/c/86242/ to get there | 23:19 |
| esmute | once we see the review, we can discuss this further. | 23:19 |
| amcrn | grapex: for what it's worth, configuration-group returns the datastore_version_id, and not the names | 23:19 |
| grapex | amcrn: yeah. I'm going to -1 this until the datastore_version_id is returned for the view. | 23:20 |
| grapex | I think it should be in addition to the name / version tuple though, since anyone writing a UI will want that. | 23:21 |
| SlickNik | amcrn: Do we have a bug for that? If not let me open one. | 23:23 |
| *** sbfox has quit IRC | 23:26 | |
| *** amytron has quit IRC | 23:29 | |
| *** amytron has joined #openstack-trove | 23:29 | |
| SlickNik | amcrn: We do now. FYI https://bugs.launchpad.net/trove/+bug/1326584 | 23:31 |
| amcrn | SlickNik: cool, thanks (was grabbin' a redbull, you beat me to the punch) | 23:31 |
| SlickNik | amcrn: No worries. My turn to go grab some caffeine. bbiab :) | 23:35 |
| grapex | talk to you all later! | 23:43 |
| *** grapex has quit IRC | 23:43 | |
| *** sbfox has joined #openstack-trove | 23:48 | |
| *** amytron has quit IRC | 23:49 | |
| *** michael-yu has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!