| grapex | See everyone tomorrow | 00:00 |
|---|---|---|
| *** grapex has quit IRC | 00:01 | |
| *** jcru has quit IRC | 00:08 | |
| *** yogesh has joined #openstack-trove | 00:11 | |
| *** cweid has quit IRC | 00:18 | |
| *** demorris has joined #openstack-trove | 00:19 | |
| *** matsuhashi has joined #openstack-trove | 00:20 | |
| *** jmontemayor has quit IRC | 00:22 | |
| *** yogesh has quit IRC | 00:34 | |
| openstackgerrit | A change was merged to openstack/trove: Fixes resizes for volumes attached to active Nova servers https://review.openstack.org/65365 | 00:45 |
| *** demorris has quit IRC | 00:47 | |
| *** esp has left #openstack-trove | 00:52 | |
| *** grapex has joined #openstack-trove | 00:53 | |
| *** nosnos has joined #openstack-trove | 00:56 | |
| *** grapex has quit IRC | 01:09 | |
| *** igor has joined #openstack-trove | 01:15 | |
| *** igor____ has quit IRC | 01:16 | |
| *** esmute has quit IRC | 01:22 | |
| *** tanisdl has quit IRC | 01:23 | |
| *** saurabhs has quit IRC | 01:33 | |
| *** freyes has joined #openstack-trove | 01:33 | |
| *** amcrn has quit IRC | 01:43 | |
| *** saurabhs has joined #openstack-trove | 01:43 | |
| *** michael-yu has quit IRC | 01:49 | |
| *** grapex has joined #openstack-trove | 01:57 | |
| *** grapex has quit IRC | 02:02 | |
| *** amytron has joined #openstack-trove | 02:03 | |
| *** saurabhs has quit IRC | 02:12 | |
| *** michael-yu has joined #openstack-trove | 02:12 | |
| *** amrith has quit IRC | 02:14 | |
| *** amrith has joined #openstack-trove | 02:15 | |
| *** michael-yu has quit IRC | 02:30 | |
| *** erkules has quit IRC | 02:34 | |
| *** esp has joined #openstack-trove | 02:35 | |
| *** harlowja has quit IRC | 02:44 | |
| *** erkules has joined #openstack-trove | 02:48 | |
| *** SushilKM has quit IRC | 02:52 | |
| *** grapex has joined #openstack-trove | 02:53 | |
| *** harlowja has joined #openstack-trove | 02:55 | |
| *** freyes has quit IRC | 02:57 | |
| *** doug_shelley66 has quit IRC | 03:00 | |
| *** yogesh has joined #openstack-trove | 03:20 | |
| *** yogesh has quit IRC | 03:25 | |
| openstackgerrit | Craig Vyvial proposed a change to openstack/trove: adding configuration group support https://review.openstack.org/53168 | 03:28 |
| *** amrith has quit IRC | 04:03 | |
| openstackgerrit | Mat Lowery proposed a change to openstack/trove: Get service endpoints from catalog https://review.openstack.org/68015 | 04:14 |
| *** jcooley_ has quit IRC | 04:43 | |
| openstackgerrit | Craig Vyvial proposed a change to openstack/trove: adding configuration group support https://review.openstack.org/53168 | 04:47 |
| *** michael-yu has joined #openstack-trove | 04:49 | |
| *** doug_shelley66 has joined #openstack-trove | 04:57 | |
| *** doug_shelley66 has quit IRC | 05:01 | |
| *** nosnos_ has joined #openstack-trove | 05:02 | |
| *** harlowja has quit IRC | 05:05 | |
| *** nosnos has quit IRC | 05:06 | |
| *** harlowja has joined #openstack-trove | 05:14 | |
| *** ViswaV has quit IRC | 05:26 | |
| *** edmund1 has quit IRC | 05:29 | |
| *** jcooley_ has joined #openstack-trove | 05:49 | |
| *** denis_makogon has joined #openstack-trove | 06:00 | |
| *** amytron has quit IRC | 06:03 | |
| *** matsuhashi has quit IRC | 06:07 | |
| *** matsuhashi has joined #openstack-trove | 06:07 | |
| *** matsuhas_ has joined #openstack-trove | 06:09 | |
| *** matsuhashi has quit IRC | 06:09 | |
| *** jcooley_ has quit IRC | 06:11 | |
| *** igor has quit IRC | 06:18 | |
| *** igor has joined #openstack-trove | 06:18 | |
| *** grapex has quit IRC | 06:24 | |
| *** jcooley_ has joined #openstack-trove | 06:24 | |
| *** denis_makogon_ has joined #openstack-trove | 06:24 | |
| *** denis_makogon has quit IRC | 06:26 | |
| *** krow has quit IRC | 06:39 | |
| *** nosnos has joined #openstack-trove | 06:40 | |
| *** yogesh has joined #openstack-trove | 06:42 | |
| *** nosnos_ has quit IRC | 06:42 | |
| *** krow has joined #openstack-trove | 06:50 | |
| *** krow has quit IRC | 06:54 | |
| *** michael-yu_ has joined #openstack-trove | 06:56 | |
| *** krow has joined #openstack-trove | 06:56 | |
| *** michael-yu has quit IRC | 06:57 | |
| *** michael-yu_ is now known as michael-yu | 06:57 | |
| *** grapex has joined #openstack-trove | 06:59 | |
| *** krow has quit IRC | 07:00 | |
| *** krow has joined #openstack-trove | 07:01 | |
| *** yogesh has quit IRC | 07:03 | |
| *** krow1 has joined #openstack-trove | 07:03 | |
| *** grapex has quit IRC | 07:04 | |
| *** esmute has joined #openstack-trove | 07:04 | |
| *** krow has quit IRC | 07:06 | |
| *** PradeepChandani_ has quit IRC | 07:17 | |
| *** PradeepChandani_ has joined #openstack-trove | 07:17 | |
| *** harlowja is now known as harlowja_away | 07:23 | |
| *** flaper87|afk is now known as flaper87 | 07:39 | |
| *** denis_makogon_ has quit IRC | 07:55 | |
| *** grapex has joined #openstack-trove | 08:00 | |
| *** grapex has quit IRC | 08:04 | |
| *** grapex has joined #openstack-trove | 09:00 | |
| *** grapex has quit IRC | 09:05 | |
| *** igor_ has joined #openstack-trove | 09:30 | |
| *** igor has quit IRC | 09:32 | |
| *** jcooley_ has quit IRC | 09:32 | |
| *** michael-yu has quit IRC | 09:42 | |
| *** michael-yu has joined #openstack-trove | 09:47 | |
| *** michael-yu has quit IRC | 09:48 | |
| *** grapex has joined #openstack-trove | 10:01 | |
| *** krow1 has quit IRC | 10:03 | |
| *** grapex has quit IRC | 10:05 | |
| *** krow has joined #openstack-trove | 10:14 | |
| *** krow has quit IRC | 10:19 | |
| *** krow has joined #openstack-trove | 10:20 | |
| *** krow1 has joined #openstack-trove | 10:24 | |
| *** krow has quit IRC | 10:24 | |
| *** denis_makogon has joined #openstack-trove | 10:46 | |
| *** PradeepChandani_ has quit IRC | 11:01 | |
| *** grapex has joined #openstack-trove | 11:01 | |
| *** grapex has quit IRC | 11:06 | |
| *** krow1 has quit IRC | 11:10 | |
| *** AndroUser has joined #openstack-trove | 11:21 | |
| *** AndroUser has quit IRC | 11:55 | |
| *** AndroUser has joined #openstack-trove | 11:56 | |
| *** AndroUser has quit IRC | 11:56 | |
| *** SnowDust has joined #openstack-trove | 11:57 | |
| *** AndroUser has joined #openstack-trove | 12:01 | |
| *** grapex has joined #openstack-trove | 12:02 | |
| *** doug_shelley66 has joined #openstack-trove | 12:04 | |
| *** grapex has quit IRC | 12:06 | |
| *** amrith has joined #openstack-trove | 12:08 | |
| *** AndroUser has quit IRC | 12:20 | |
| *** AndroUser has joined #openstack-trove | 12:21 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Security groups workflow update https://review.openstack.org/50944 | 12:25 |
| *** robertmy_ has joined #openstack-trove | 12:25 | |
| *** robertmyers has quit IRC | 12:25 | |
| *** jcooley_ has joined #openstack-trove | 12:26 | |
| *** SnowDust has quit IRC | 12:28 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Security groups workflow update https://review.openstack.org/50944 | 12:29 |
| openstackgerrit | Denis M. proposed a change to openstack/trove-integration: Add support for minimal MongoDB testing https://review.openstack.org/53378 | 12:30 |
| *** jcooley_ has quit IRC | 12:31 | |
| *** AndroUser has quit IRC | 12:40 | |
| *** AndroUser has joined #openstack-trove | 12:40 | |
| *** jcru has joined #openstack-trove | 12:44 | |
| *** igor has joined #openstack-trove | 12:45 | |
| *** igor_ has quit IRC | 12:46 | |
| *** SnowDust has joined #openstack-trove | 12:52 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Allow db instance conditional logging https://review.openstack.org/63789 | 13:01 |
| *** grapex has joined #openstack-trove | 13:03 | |
| *** pdmars has joined #openstack-trove | 13:04 | |
| *** freyes has joined #openstack-trove | 13:06 | |
| *** grapex has quit IRC | 13:07 | |
| denis_makogon | http://vk.com/makphotos?w=wall-36488060_237 | 13:08 |
| *** AndroUser has quit IRC | 13:09 | |
| *** jimbobhickville has joined #openstack-trove | 13:10 | |
| *** amrith has quit IRC | 13:15 | |
| *** freyes has quit IRC | 13:16 | |
| openstackgerrit | shivam shukla proposed a change to openstack/trove: Tests for heat based instance workflow https://review.openstack.org/66499 | 13:29 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Security groups workflow update https://review.openstack.org/50944 | 13:46 |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Track security group provisioned by heat https://review.openstack.org/71040 | 13:47 |
| *** SnowDust has quit IRC | 13:50 | |
| *** doug_shelley66 has quit IRC | 13:50 | |
| *** yidclare has quit IRC | 13:54 | |
| denis_makogon | sorry for link above, wrong chat | 13:55 |
| *** jcooley_ has joined #openstack-trove | 13:57 | |
| *** jcooley_ has quit IRC | 14:03 | |
| *** grapex has joined #openstack-trove | 14:03 | |
| *** robertmy_ has quit IRC | 14:03 | |
| *** grapex has quit IRC | 14:08 | |
| *** SnowDust has joined #openstack-trove | 14:08 | |
| *** edmund has joined #openstack-trove | 14:12 | |
| *** demorris has joined #openstack-trove | 14:15 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Allow db instance conditional logging https://review.openstack.org/63789 | 14:16 |
| *** michael-yu has joined #openstack-trove | 14:29 | |
| *** Barker has joined #openstack-trove | 14:29 | |
| *** robertmyers has joined #openstack-trove | 14:31 | |
| *** edmund has quit IRC | 14:32 | |
| *** edmund has joined #openstack-trove | 14:35 | |
| *** doug_shelley66 has joined #openstack-trove | 14:52 | |
| *** amrith has joined #openstack-trove | 14:58 | |
| *** amrith has quit IRC | 15:02 | |
| *** Barker has quit IRC | 15:03 | |
| *** amrith has joined #openstack-trove | 15:03 | |
| *** grapex has joined #openstack-trove | 15:04 | |
| *** SnowDust has quit IRC | 15:08 | |
| *** ViswaV has joined #openstack-trove | 15:15 | |
| *** matsuhas_ has quit IRC | 15:17 | |
| *** nosnos has quit IRC | 15:21 | |
| *** kevinconway has joined #openstack-trove | 15:22 | |
| *** ViswaV has quit IRC | 15:24 | |
| *** datsun180b has joined #openstack-trove | 15:33 | |
| *** Barker has joined #openstack-trove | 15:38 | |
| *** tanisdl has joined #openstack-trove | 15:40 | |
| *** jcooley_ has joined #openstack-trove | 15:46 | |
| *** jcooley_ has quit IRC | 15:53 | |
| *** michael-yu has quit IRC | 15:54 | |
| *** shakayumi has joined #openstack-trove | 15:54 | |
| *** michael-yu has joined #openstack-trove | 15:59 | |
| *** doug_shelley66 has quit IRC | 16:04 | |
| *** plodronio_ has joined #openstack-trove | 16:05 | |
| *** ViswaV has joined #openstack-trove | 16:05 | |
| *** doug_shelley66 has joined #openstack-trove | 16:06 | |
| *** plodronio has quit IRC | 16:06 | |
| *** plodronio_ is now known as plodronio | 16:06 | |
| *** ViswaV has quit IRC | 16:06 | |
| *** ViswaV has joined #openstack-trove | 16:06 | |
| *** ViswaV has quit IRC | 16:09 | |
| *** ViswaV_ has joined #openstack-trove | 16:09 | |
| *** ViswaV_ has quit IRC | 16:10 | |
| *** ViswaV has joined #openstack-trove | 16:10 | |
| *** jmontemayor has joined #openstack-trove | 16:13 | |
| *** jmontemayor has quit IRC | 16:14 | |
| *** shakayumi has quit IRC | 16:14 | |
| *** jmontemayor has joined #openstack-trove | 16:15 | |
| *** jcooley_ has joined #openstack-trove | 16:30 | |
| *** jasonb365 has joined #openstack-trove | 16:30 | |
| *** amrith has quit IRC | 16:42 | |
| *** michael-yu has quit IRC | 16:48 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Security groups workflow update https://review.openstack.org/50944 | 16:50 |
| *** jcooley_ has quit IRC | 16:56 | |
| denis_makogon | mat-lowery, ping | 16:56 |
| mat-lowery | hello denis_makogon | 16:57 |
| denis_makogon | mat-lowery, hi | 16:57 |
| *** jcooley_ has joined #openstack-trove | 16:57 | |
| denis_makogon | last patch landed, and i suppose, that this one is the last | 16:57 |
| denis_makogon | mat-lowery, btw, just letting you know TroveError object doesn't contain status code attribute | 16:58 |
| denis_makogon | mat-lowery, there's a policy for exceptions, dictated by oslo-incubator(OpenStackException) | 16:59 |
| *** jmontemayor has quit IRC | 17:00 | |
| *** jmontemayor has joined #openstack-trove | 17:00 | |
| *** michael-yu has joined #openstack-trove | 17:01 | |
| *** freyes has joined #openstack-trove | 17:03 | |
| mat-lowery | denis_makogon: Just saw your comments. I was proposing a status code in the new MalformedSecurityGroupRuleError as a compromise between two extremes: one MalformedSecurityGroupRuleError class everywhere with different error messages vs. Error class per fine-grained problem (in the case of above patch set: 3 Error classes). This compromise keeps the class count down while avoiding the parsing of error messages in tests. If | 17:04 |
| mat-lowery | OpenStackException forbids status codes, then that's another story. | 17:04 |
| denis_makogon | mat-lowery, as i said not so long ago, populating error classes - bad idea | 17:06 |
| denis_makogon | we need to keep everything simple and fast | 17:07 |
| denis_makogon | mat-lowery, its tests only, you run it once, no big deal | 17:07 |
| *** kiall has quit IRC | 17:08 | |
| mat-lowery | denis_makogon: My point is only that ideally, changing an error message shouldn't break tests. And if by "populating error classes" you mean an explosion in the number of exception classes, then that is precisely what a status code is meant to address. My two cents. Thanks for the feedback! | 17:11 |
| denis_makogon | mat-lowery, ideally, we need to have separate class for taskmanager exception, api exception, guest exceptions | 17:12 |
| denis_makogon | Error message are what we need | 17:13 |
| mat-lowery | denis_makogon: gotcha. thanks | 17:15 |
| *** krow has joined #openstack-trove | 17:15 | |
| denis_makogon | mat-lowery, thanks for keeping eye on patches | 17:16 |
| *** michael-yu has quit IRC | 17:17 | |
| *** Barker has quit IRC | 17:26 | |
| *** doug_shelley66 has quit IRC | 17:27 | |
| *** krow has quit IRC | 17:28 | |
| *** doug_shelley66 has joined #openstack-trove | 17:37 | |
| *** krow has joined #openstack-trove | 17:38 | |
| *** jcooley_ has quit IRC | 17:41 | |
| *** jcooley_ has joined #openstack-trove | 17:42 | |
| *** krow1 has joined #openstack-trove | 17:43 | |
| *** krow has quit IRC | 17:44 | |
| *** jmontemayor has quit IRC | 17:57 | |
| *** harlowja_away is now known as harlowja | 17:59 | |
| *** jasonb365 has quit IRC | 17:59 | |
| *** jmontemayor has joined #openstack-trove | 18:02 | |
| *** jasonb365 has joined #openstack-trove | 18:03 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove: Security groups workflow update https://review.openstack.org/50944 | 18:05 |
| *** amcrn has joined #openstack-trove | 18:06 | |
| *** Barker has joined #openstack-trove | 18:08 | |
| juice | eh hem (https://review.openstack.org/#/c/66063/) | 18:17 |
| juice | hub_cap, grapex ^ ^ | 18:17 |
| hub_cap | hey juice i need to spend a day focused on fixing my email client | 18:18 |
| hub_cap | so im layin low on reviews till this evernin | 18:18 |
| hub_cap | denis_makogon: hey dude | 18:18 |
| juice | okie doke | 18:18 |
| juice | this patch may need to be rebased based on pdmars recently merged patch | 18:18 |
| hub_cap | i coudlnt get the trove-integration code to publish the glance image last night, i rebased your trove-integration c* review too | 18:19 |
| hub_cap | can u look into it? | 18:19 |
| hub_cap | ill try again this afternoon denis_makogon | 18:19 |
| *** amrith has joined #openstack-trove | 18:20 | |
| *** jcooley_ has quit IRC | 18:21 | |
| openstackgerrit | Denis M. proposed a change to openstack/trove-integration: Initial support for single instance Cassandra Database https://review.openstack.org/52666 | 18:24 |
| denis_makogon | hub_cap, ok, i changed commit description | 18:24 |
| openstackgerrit | Denis M. proposed a change to openstack/trove-integration: Initial support for single instance Cassandra Database https://review.openstack.org/52666 | 18:25 |
| *** yogesh has joined #openstack-trove | 18:32 | |
| *** igor has quit IRC | 18:40 | |
| *** yidclare has joined #openstack-trove | 18:41 | |
| *** igor has joined #openstack-trove | 18:44 | |
| *** Ranjitha has joined #openstack-trove | 18:53 | |
| *** tanisdl has quit IRC | 18:55 | |
| *** jcooley_ has joined #openstack-trove | 18:58 | |
| *** kiall has joined #openstack-trove | 19:24 | |
| openstackgerrit | Craig Vyvial proposed a change to openstack/trove: adding configuration group support https://review.openstack.org/53168 | 19:29 |
| *** tanisdl has joined #openstack-trove | 19:31 | |
| *** jcooley_ has quit IRC | 19:35 | |
| *** jcooley_ has joined #openstack-trove | 19:36 | |
| *** jcooley_ has quit IRC | 19:40 | |
| *** jcooley_ has joined #openstack-trove | 19:41 | |
| *** SushilKM has joined #openstack-trove | 19:42 | |
| SushilKM | hey | 19:43 |
| SushilKM | was luking into root-show api and ..... I think that though it says that is_root_enabled but it means actually checking any historical root enablements .... | 19:44 |
| *** jcooley_ has quit IRC | 19:45 | |
| SushilKM | also guestagent_api.is_root_enabled is not called anywhere to decide root enable status | 19:47 |
| SushilKM | so wisely, would we like to add a date to confirm that historical root enabled, hw do we see that .... | 19:49 |
| *** jcooley_ has joined #openstack-trove | 19:50 | |
| hub_cap | SushilKM: im not sure what you are proposing, but it cannot break the current api :) | 19:54 |
| * hub_cap goes back to setting up gnus | 19:54 | |
| SushilKM | let me expalin a bit | 19:54 |
| *** jcooley_ has quit IRC | 19:56 | |
| SushilKM | root-show api uses https://github.com/openstack/trove/blob/master/trove/extensions/mysql/models.py#L189 to find root has been enabled or not | 19:56 |
| SushilKM | this checks basically in the root-history if root was ever enabled | 19:57 |
| *** jcooley_ has joined #openstack-trove | 19:57 | |
| SushilKM | because if someone enables it and then deletes the root user from system, root-show wud still show as true, I agree that behind the scens before deleting the root user the user must have done something to make his way | 19:58 |
| SushilKM | so it seems we want to show using root-show api is if root was ever enabled on this instance | 20:00 |
| SushilKM | in history | 20:00 |
| *** jasonb365 has quit IRC | 20:02 | |
| denis_makogon | hub_cap, i hope, cassandra and mongo would reach codebase soon | 20:04 |
| *** jmontemayor has quit IRC | 20:04 | |
| *** jasonb365 has joined #openstack-trove | 20:05 | |
| *** Barker has quit IRC | 20:05 | |
| denis_makogon | guys, i'm out, gonna be reachable soon, in 1-2h | 20:05 |
| *** demorris has quit IRC | 20:07 | |
| *** jasonb365 has quit IRC | 20:09 | |
| *** jmontemayor has joined #openstack-trove | 20:12 | |
| *** jcooley_ has quit IRC | 20:12 | |
| *** michael-yu has joined #openstack-trove | 20:12 | |
| *** jcooley_ has joined #openstack-trove | 20:12 | |
| SushilKM | how do u see that hub_cap | 20:14 |
| *** Barker has joined #openstack-trove | 20:15 | |
| hub_cap | denis_makogon: once i get things working well it will be merged :) | 20:15 |
| datsun180b | is_root_enabled is did_i_find_at_least_one_instance_of_the_user_enabling_root | 20:16 |
| *** grapex_ has joined #openstack-trove | 20:17 | |
| SushilKM | so, how do u see if we show a date to when was it enabled last | 20:17 |
| hub_cap | im not sure it matters | 20:18 |
| datsun180b | i forget but i thought the mgmt api gave it in higher detail | 20:18 |
| hub_cap | it does datsun180b | 20:18 |
| hub_cap | it gives the date as well | 20:18 |
| datsun180b | so there you go | 20:18 |
| SushilKM | oh really | 20:18 |
| hub_cap | i dont think weve had any users who have asked when they had it enabled tho | 20:18 |
| datsun180b | the purpose was to function like a tamper-proof seal for support purposes | 20:19 |
| SushilKM | yeah, but i was doing it the way that i enabled root then dropped root user back in database, and whn i used root-show it was still showing me root is enabled | 20:19 |
| SushilKM | and i was complaininng for that | 20:19 |
| *** demorris has joined #openstack-trove | 20:19 | |
| datsun180b | right, if you open a jar of pickles once you can't hermetically seal the jar again unless you own a pickle factory | 20:19 |
| SushilKM | then i found it was to show my historical root enable .... | 20:19 |
| *** grapex has quit IRC | 20:20 | |
| datsun180b | if you enable root and then erase all trace of ever having enabled root it makes it difficult to support an instance | 20:20 |
| SushilKM | so, wat i wanted to was to add a date for last enable so that i do not complain it .... | 20:20 |
| datsun180b | iirc "Never" was the default | 20:21 |
| datsun180b | and "Nobody" to go with it | 20:21 |
| datsun180b | sorry, would you explain what you meant with that last bit? | 20:21 |
| SushilKM | i meant when we use show-root it shows is_root_enable = true | 20:22 |
| datsun180b | oh so you just want to muck with the return type for that method? | 20:23 |
| datsun180b | weeeeell i wonder if it doesn't count as violating the contract if you return _more_ than what is stated | 20:24 |
| datsun180b | so if you returned {enabled: True, when=<some date>} i can't see any client exploding for that | 20:24 |
| SushilKM | cool, so shall i make that | 20:25 |
| datsun180b | i'm working on a similar change to user access/grant/revoke | 20:25 |
| *** grapex_ has quit IRC | 20:25 | |
| datsun180b | and you probably want core to lean in on that to make sure before you start | 20:25 |
| SushilKM | so cores, how do u see that | 20:25 |
| *** grapex has joined #openstack-trove | 20:26 | |
| SushilKM | http://paste.openstack.org/show/62872/ is what i wanna do | 20:29 |
| SushilKM | please suggest, if that looks good to go ahead | 20:29 |
| *** krast has quit IRC | 20:30 | |
| *** freyes has quit IRC | 20:31 | |
| datsun180b | how would that look in json? do we have any other examples of cli output in three columns? | 20:32 |
| SushilKM | did not checked that | 20:33 |
| SushilKM | but research cud be mad | 20:33 |
| hub_cap | yea cli output is generally not what i want to see when talking about api changes | 20:33 |
| hub_cap | fwiw, i think we all know yer talking about adding a field to the root call | 20:34 |
| SushilKM | huh | 20:34 |
| datsun180b | it's really the api's schema we have the contract with | 20:34 |
| hub_cap | to show the enabled date | 20:34 |
| datsun180b | the cli can do whatever we like | 20:34 |
| hub_cap | exactly | 20:34 |
| SushilKM | or if column does not works with then we can think to add in form of a row, my first idea is to have a date | 20:35 |
| datsun180b | so when do we start talking about the next version of the api so i can take a chainsaw to the users methods | 20:35 |
| hub_cap | datsun180b: :D | 20:35 |
| datsun180b | SushilKM: how about {'enabled': True, 'when': <ten minutes ago>} | 20:35 |
| hub_cap | SushilKM: http://docs.rackspace.com/cdb/api/v1.0/cdb-devguide/content/GET_isRootEnabled__version___accountId__instances__instanceId__root_Database_Instances.html | 20:36 |
| hub_cap | look at the examples there | 20:36 |
| hub_cap | we do not discuss thigns in terms of how the cli outputs it | 20:36 |
| hub_cap | we show like how datsun180b is displaying it | 20:36 |
| hub_cap | i guess i dont see any harm in adding the date | 20:36 |
| hub_cap | and calling it, say, 'enabled' | 20:37 |
| hub_cap | but we shoudl get buyin from other people as well :) | 20:37 |
| datsun180b | well if the date's there then clearly it's been enabled | 20:37 |
| datsun180b | similar to 'created' fields elsewhere | 20:37 |
| hub_cap | sure but lets say for audit purposes | 20:37 |
| hub_cap | like if a user wants to know when his admin enabled it...? | 20:37 |
| hub_cap | i dunno | 20:37 |
| datsun180b | well in that case we better add a "by" field | 20:38 |
| datsun180b | even if it was done on behalf of another user | 20:38 |
| datsun180b | and on that note why not slide all the way down the slippery slope and keep the user-agent, and i'm being serious btw | 20:38 |
| hub_cap | datsun180b: we already have the date yo, we can display that :) | 20:39 |
| datsun180b | well if you want to go full audit | 20:39 |
| hub_cap | and we dont allow trusts / impersonation | 20:39 |
| hub_cap | so we cant do 'by' | 20:39 |
| datsun180b | _yet_ | 20:39 |
| hub_cap | i think keystone v3 can | 20:39 |
| SushilKM | we have the dates in the tables | 20:39 |
| datsun180b | that's true | 20:39 |
| hub_cap | but i dont kjnwo fo sure | 20:39 |
| hub_cap | i tend to not kjnwo things, in general | 20:39 |
| datsun180b | one must first unkjnwo everying before one can truly begin to kjnwo oneself | 20:40 |
| datsun180b | typos abound | 20:40 |
| datsun180b | well unless i'm mistaken SushilKM the rule i've seen so far is you can't remove anything from a response or add new requirements to a request | 20:41 |
| SushilKM | hey hub_cap r u playing with ur kjnwoledge :P | 20:41 |
| datsun180b | compatibility is the concern | 20:43 |
| ViswaV | SushilKM: the data is there in the table, the mgmt api returns it, are you only asking whether to enhance to cli output to also show an extra column with the date information (to know when it was enabled on? ) ? | 20:44 |
| ViswaV | for examples of CLI showing more than two columns look at "nova list" output. | 20:44 |
| ViswaV | It shows 6 columns | 20:44 |
| openstackgerrit | Paul Lodronio proposed a change to openstack/trove: Adding additional datastore tests https://review.openstack.org/70296 | 20:45 |
| datsun180b | cool, there you go | 20:46 |
| SushilKM | i wud say yes @VishwaV, it could be column or a row but wanted to see the date as a user .... | 20:47 |
| ViswaV | SushilKM: The second part of your concern (which is why you want to show this 'enabled on' information) is because you wiped out the root user by logging into DB in the background….and the trove root-show does not reflect that. Not sure if that is a valid thing to 'sync' up to trove DB. We do report back service status…. but root user status… probably not. The same thing probably applies for DB … you can "trove database-create" and | 20:49 |
| ViswaV | then in the background go behind the scenes and delete the db from mysql. I don't think "trove database-list" will probably still show you that db which no longer exists in the mysql now. Should it reflect the accurate state for operations done outside the context of Trove? I donno... | 20:49 |
| ViswaV | SushilKM: It should be a column since it is 'peer' field to the others coming from the same table. | 20:50 |
| datsun180b | dbs and users and grants are all pulled from the instance, we don't store them in a second place | 20:50 |
| datsun180b | if you delete a user or db you delete it | 20:50 |
| SushilKM | ur point about databases create-delete-list is fine ... but with root what happens is i can do anything in background and do a cleanup, but the provider needs to ensure and suggest that root was enabled | 20:51 |
| SushilKM | which currently show root api does | 20:51 |
| datsun180b | the root history is kept as a table on the host specifically to prevent someone from enabling root, mucking about, and then 'disabling' root | 20:51 |
| SushilKM | wanted it to show the date also about when was it enabled | 20:52 |
| ViswaV | datsun180b: makes sense about the databases, users. | 20:52 |
| *** demorris has quit IRC | 20:52 | |
| *** Barker has quit IRC | 20:54 | |
| *** mattgriffin has quit IRC | 20:54 | |
| ViswaV | On contrary example to SushilKM's situation…… if someone accidentally (or purposefully) enabled root user after directly logged into DB instance and not using trove… the trove root-show I am guessing would not reflect that fact and give out false info (that it is not enabled) , thusly not giving a key insight to DB admins on a critical hole that was left open (accidentally or willfully) on their DB instance. | 20:55 |
| *** jimbobhickville has quit IRC | 20:55 | |
| datsun180b | sorry, got to take off. you guys sound like you've got a handle on things | 20:56 |
| *** datsun180b has quit IRC | 20:56 | |
| ViswaV | Should the guest agent status monitoring , actively monitor all root users and report them back on a periodic basis? | 20:56 |
| amcrn | there's already been discussions on having a true root enabled history (multiple rows vs. updating a single), and there's also been discussion on unifying created/updated/deleted_at across the tables to provide a consistent experience | 20:56 |
| amcrn | not sure if they made it into bps/issues though | 20:56 |
| *** Ranjitha has quit IRC | 20:57 | |
| SushilKM | one cannot enable directly the root user | 20:57 |
| SushilKM | i suppose that | 20:57 |
| *** ViswaV has quit IRC | 20:57 | |
| *** ViswaV has joined #openstack-trove | 20:58 | |
| ViswaV | amcrn: what about active monitoring of root users (by the guest agents) to monitor/detect root (or other) user creations not done via trove… is there any value in such a feature? | 20:58 |
| *** demorris has joined #openstack-trove | 20:58 | |
| amcrn | ViswaV: i don't see any, but i haven't heard anyone ask for it | 20:59 |
| ViswaV | ok | 20:59 |
| *** amytron has joined #openstack-trove | 20:59 | |
| *** denis_makogon_ has joined #openstack-trove | 21:00 | |
| *** ViswaV has quit IRC | 21:00 | |
| *** denis_makogon has quit IRC | 21:01 | |
| *** denis_makogon_ is now known as denis_makogon | 21:01 | |
| denis_makogon | i'm in | 21:01 |
| *** ViswaV has joined #openstack-trove | 21:01 | |
| *** dmakogon_ has joined #openstack-trove | 21:01 | |
| *** ViswaV has quit IRC | 21:02 | |
| *** amrith has quit IRC | 21:02 | |
| *** ViswaV has joined #openstack-trove | 21:03 | |
| *** amrith has joined #openstack-trove | 21:03 | |
| denis_makogon | vipul, SlickNik , ping, i'd like to talk about https://review.openstack.org/#/c/70773/ | 21:04 |
| *** Barker has joined #openstack-trove | 21:05 | |
| denis_makogon | imsplitbit, ping | 21:06 |
| vipul | denis_makogon | 21:06 |
| vipul | denis_makogon: what's up | 21:06 |
| denis_makogon | goodday | 21:06 |
| denis_makogon | nice 2 see you here | 21:06 |
| vipul | likewise denis_makogon :) | 21:06 |
| denis_makogon | vipul, i hope you read last comments | 21:06 |
| SlickNik | denis_makogon: Sure, let me check the review. | 21:07 |
| denis_makogon | yes, sure | 21:07 |
| SlickNik | So denis_makogon: I'm trying to understand what's driving the requirement for these validators. | 21:08 |
| *** mattgriffin has joined #openstack-trove | 21:09 | |
| denis_makogon | SlickNik, basicaly validation based upon mysql models | 21:09 |
| denis_makogon | like User | 21:10 |
| denis_makogon | Schema | 21:10 |
| SlickNik | You're adding validators for redis; but then they're not really needed or implemented for redis. (All they are doing is returning None). | 21:10 |
| denis_makogon | Yes | 21:11 |
| *** amytron has quit IRC | 21:11 | |
| SlickNik | So then I'm not entirely sure what the point of the change is. | 21:11 |
| vipul | I still feel like this sort of datastore specific validation should not be a concern of the Trove API -- leaving it further down at the guest allows a deployer to override what is considered valid | 21:11 |
| denis_makogon | redis doesnt support users/schemes | 21:11 |
| *** mattgriffin has quit IRC | 21:11 | |
| denis_makogon | vipul, as i remember we decided to keep guest simple | 21:11 |
| SlickNik | denis_makogon: If redis doesn't support it, why add an artificial construct (validators) for it? | 21:12 |
| denis_makogon | SlickNik, before capabilities, users are allowed to pass any data to API service | 21:12 |
| vipul | sure but the sort of validation that we do at the API layer should be json schema based.. nothing more | 21:12 |
| *** ViswaV has quit IRC | 21:12 | |
| vipul | if you want to make a change to the json schema, then i might be ok with that | 21:13 |
| denis_makogon | and without pluggablity create call would fail | 21:13 |
| *** ViswaV has joined #openstack-trove | 21:13 | |
| vipul | but we need to provide an easy way to extend what is considered valid | 21:13 |
| denis_makogon | if there's incorrect format for users/schemes | 21:13 |
| SlickNik | Sure, and the datastore can decide to honor the extension args or not depending on whether it supports it or not. | 21:13 |
| denis_makogon | that is why i made such change, basicaly redis validation will pass in any way, even if data is malformed' | 21:14 |
| denis_makogon | SlickNik, unfortunately, capabilities is far-far away from now | 21:14 |
| denis_makogon | not even in I release | 21:15 |
| SlickNik | denis_makogon: How would you consider this malformed data. It would be an extra, unsupported argument, but that hardly constitutes malformed data. | 21:15 |
| *** jasonb365 has joined #openstack-trove | 21:15 | |
| SlickNik | Vipul, I tend to agree with you in that I think the datastore should decide whether it needs to do anything with this information or not. It's probably overkill to have a separate extension decide which arguments are supported for which datastores. | 21:18 |
| denis_makogon | SlickNik, each datastore will have a set of validators for ingress data | 21:18 |
| SlickNik | denis_makogon: Why? | 21:19 |
| denis_makogon | SlickNik, because mysql already have them | 21:19 |
| denis_makogon | SlickNik, mysql validators will be the problem for mongo, postgressql users/schemes | 21:21 |
| denis_makogon | SlickNik, in general this topic came from unflexible extenstions | 21:21 |
| denis_makogon | users/schemes | 21:21 |
| denis_makogon | how is that possible to work with mongo users through mysql service | 21:22 |
| denis_makogon | same problem with mysql validation/population on prepare call | 21:23 |
| denis_makogon | i mean create call | 21:23 |
| SlickNik | We can cross that when we need to do mongo validation. I don't see a mongo validator in that changeset. | 21:23 |
| vipul | how about refactoring this code to not validate at the API layer | 21:23 |
| SlickNik | Right now, redis doesn't _need_ validation, so this doesn't buy us anything. | 21:23 |
| denis_makogon | vipul, how ? | 21:24 |
| vipul | it's actually pretty bad right now.. the extensions.mysql.common.populate_validated_databases() uses the guest agent models to validate user name | 21:24 |
| vipul | which means that once guest is split out.. this code will need to be duplicated | 21:24 |
| vipul | so let's not validate the username at the API layer.. and let's defer the validation to the Guest.. | 21:24 |
| SlickNik | Yeah, I agree vipul. I think we need to think about how to support this per datastore and not as extensions (common across datastores). | 21:25 |
| SlickNik | And the easiest way to support this per datastore is to actually have the datastore (guestagent) pieces do the validation. | 21:25 |
| denis_makogon | SlickNik, so, you mean to pass data from cli through api, taskmanager to guest without any validations ? | 21:26 |
| vipul | SlickNik, denis_makogon: the only side effect is on the create all-in-one call where we might fail to create due to invalid names | 21:26 |
| vipul | should that be an error or just partially provisioned | 21:27 |
| hub_cap | this is why we should just remove users | 21:27 |
| hub_cap | :) | 21:27 |
| denis_makogon | seems like it should be partially provisioned | 21:27 |
| *** jcooley_ has quit IRC | 21:28 | |
| denis_makogon | hub_cap, and schemes =) | 21:28 |
| hub_cap | :) | 21:28 |
| hub_cap | but that cant happen for a while | 21:28 |
| hub_cap | we just wont put em in the v2 api i think heh | 21:28 |
| denis_makogon | yeah | 21:28 |
| denis_makogon | al least 2 release cycles | 21:28 |
| denis_makogon | hub_cap, i agree | 21:29 |
| *** jcooley_ has joined #openstack-trove | 21:29 | |
| denis_makogon | it would be easier to drop them from guest | 21:29 |
| denis_makogon | i know we talked about this topic in general, but what do you think about this one https://review.openstack.org/#/c/70742/ ? | 21:30 |
| denis_makogon | without taking into account future capabilities | 21:31 |
| denis_makogon | this two reviews are related to heat stuff in trove - https://review.openstack.org/#/c/67873/ |||| https://review.openstack.org/#/c/71040/ | 21:32 |
| hub_cap | well denis_makogon i wonder, why should we continue to develop on the extensions if we all know we are removing it? | 21:32 |
| hub_cap | aka, shoudl we ever have a c* /users or a mongo /users | 21:32 |
| hub_cap | especially if we are going to remove future functionality | 21:32 |
| denis_makogon | hub_cap, c* would never have users =) | 21:33 |
| denis_makogon | hub_cap, but it's in future | 21:33 |
| kevinconway | denis_makogon: cassandra has user/pass auth | 21:33 |
| denis_makogon | but now it would cause error | 21:33 |
| denis_makogon | kevinconway, have you tried to register a user in cassandra ? | 21:33 |
| kevinconway | you mean edit YAML? | 21:33 |
| hub_cap | denis_makogon: thats why im saying to u, if nothing else has users, why make it pluggable? | 21:34 |
| *** Ranjitha has joined #openstack-trove | 21:34 | |
| hub_cap | thats just more code to make something we will never use | 21:34 |
| denis_makogon | kevinconway, not yaml, access.properties, security.properties | 21:34 |
| kevinconway | its a flat file with USER=PASS | 21:34 |
| kevinconway | just curious why you don't use it is all | 21:35 |
| denis_makogon | kevinconway, after each addition server _must_ be restarted | 21:35 |
| kevinconway | so it's not that cassandra doesn't *have* users. you just don't want to use them? | 21:35 |
| denis_makogon | kevinconway, in production cassandra ACL organized with LDAP, EC2, Trusts | 21:35 |
| *** amrith has quit IRC | 21:36 | |
| denis_makogon | kevinconway, have you ever seen production cassandra cluster deployment ? downtime is not acceptable | 21:36 |
| denis_makogon | and access.properties should be updated on each node | 21:36 |
| denis_makogon | suppose you have 1k nodes | 21:37 |
| kevinconway | denis_makogon: easy, restart them all at once | 21:37 |
| denis_makogon | Gossip protocol works only with data | 21:37 |
| kevinconway | like removing a bandage | 21:37 |
| denis_makogon | kevinconway, lol, this is not production way at all | 21:37 |
| denis_makogon | downtime is huge, data migration, version management will take a lot time | 21:38 |
| *** amrith has joined #openstack-trove | 21:38 | |
| kevinconway | so it's that the user/pass auth is not used for cassandra except for testing | 21:38 |
| denis_makogon | yes | 21:38 |
| denis_makogon | simple auth never used in production | 21:39 |
| kevinconway | what is the default auth system then? | 21:39 |
| kevinconway | if i created a trove cassandra, how would i get in? | 21:39 |
| denis_makogon | it accepts all connections | 21:39 |
| kevinconway | wow, so *no* auth | 21:39 |
| denis_makogon | use LDAP | 21:40 |
| denis_makogon | thats all | 21:40 |
| kevinconway | so is it fair to say that cassandra can't have users at all? | 21:40 |
| kevinconway | what if i really wanted it and made the user calls pull from and save to LDAP? | 21:40 |
| denis_makogon | anyway, production grade cluster deployment doesn't use cassandra ACL, because it bad | 21:43 |
| denis_makogon | production also accepts firewall based access | 21:44 |
| denis_makogon | hub_cap, so, you say leave it as it is before it dropped ? | 21:45 |
| hub_cap | denis_makogon: the extensions stuff? | 21:45 |
| denis_makogon | hub_cap, yeag | 21:45 |
| denis_makogon | yeah | 21:45 |
| hub_cap | id rather it _fixed_ to be like other extensions | 21:45 |
| hub_cap | aka | 21:45 |
| hub_cap | fix the common code | 21:45 |
| hub_cap | that runs / exposes / turns on/off extensions | 21:46 |
| hub_cap | nova/cinder/etc.. have a good plugin model | 21:46 |
| hub_cap | but i dont want to really tackle that before other teams have tackled wsme | 21:46 |
| kevinconway | are db calls an extension like users? | 21:46 |
| denis_makogon | yes | 21:47 |
| hub_cap | ya | 21:47 |
| kevinconway | and our plan is to keep users and db extensions? | 21:47 |
| kevinconway | not make them part of the core api? | 21:47 |
| denis_makogon | kevinconway, yes, until API v2 | 21:47 |
| denis_makogon | kevinconway, no, they are extensions | 21:47 |
| denis_makogon | kevinconway, DBaaS should not deal with users/schemes | 21:48 |
| denis_makogon | kevinconway, take a look at RDS | 21:48 |
| denis_makogon | and take a look at DynamoDB | 21:48 |
| kevinconway | hub_cap: i'm trying really hard not to talk about this | 21:48 |
| hub_cap | kevinconway: i know :) | 21:48 |
| hub_cap | lets go w/ dont toucn anything till v | 21:48 |
| hub_cap | v2 | 21:48 |
| hub_cap | we can discuss it all in the future :) | 21:48 |
| kevinconway | yeah i've got no complain with extensions in v1 | 21:49 |
| denis_makogon | but the future is far away =) | 21:49 |
| denis_makogon | it means that i need to abandon this patches, sad =( | 21:49 |
| denis_makogon | hub_cap, any luck with mongo/cassandra testing ? | 21:51 |
| amcrn | denis_makogon: i'm seeing prepare() finish, but it's still flipping to error for Cassandra | 21:51 |
| denis_makogon | and how about enabling gating for test group | 21:51 |
| denis_makogon | amcrn, link ? | 21:51 |
| denis_makogon | amcrn, could you please post a log file | 21:52 |
| denis_makogon | amcrn, which flavor did you used ? | 21:53 |
| amcrn | https://gist.github.com/anonymous/f49a27fc75adfb682d29 | 21:55 |
| amcrn | my first hunch is the timeout is the problem | 21:55 |
| denis_makogon | amcrn, no | 21:55 |
| denis_makogon | please post /var/log/cassandra/system | 21:56 |
| denis_makogon | /var/log/cassandra/system.log | 21:56 |
| *** demorris has quit IRC | 21:57 | |
| *** doug_shelley66 has quit IRC | 21:57 | |
| denis_makogon | amcrn, so ? | 21:59 |
| denis_makogon | due to slow env, cassandra service getting up slowly | 21:59 |
| denis_makogon | how many RAM did you reserve for redstack VM ? | 22:00 |
| amcrn | see my truncated reply to the original gist, cassandra came up just fine | 22:00 |
| denis_makogon | instance should be ACTIVE | 22:00 |
| denis_makogon | there's no stacktraces, guest didn't marked instance with ERROR | 22:01 |
| denis_makogon | exec timeout doesn't causes exceptions, it's only timeout | 22:02 |
| denis_makogon | i'd suggest to reserv for redstack VM at leat 8GB RAM or above | 22:03 |
| denis_makogon | with 2 CPU cores | 22:03 |
| hub_cap | http://developer.rackspace.com/blog/welcome-to-performance-cloud-servers-have-some-benchmarks.html | 22:04 |
| hub_cap | i use Performance1-8 denis_makogon | 22:04 |
| hub_cap | thas good eh? 8gb, 8 vcpu | 22:04 |
| vgnbkr | Sorry, just catching up, but thought I'd throw in my $0.02 about the user thing. | 22:05 |
| *** shakayumi has joined #openstack-trove | 22:05 | |
| hub_cap | vgnbkr: shoot :) | 22:05 |
| vgnbkr | Seems like a lot of the friction is because the issue isn't really about users. Users can be managed by third party tools. | 22:05 |
| * hub_cap runs, arms flailing | 22:05 | |
| kevinconway | ... | 22:05 |
| hub_cap | for sure vgnbkr | 22:05 |
| vgnbkr | The real issue is about access control. For mysql, access may be controlled by users, but for other databases, other access methods are required. | 22:06 |
| denis_makogon | vgnbkr, agreed | 22:06 |
| hub_cap | so the bigger issue | 22:06 |
| hub_cap | is | 22:06 |
| hub_cap | how do we allow a managed service by just giving out _root_ users | 22:06 |
| hub_cap | aka its hard to provide a SLA | 22:06 |
| hub_cap | now im on you guy's side, but that will be the arguement to overcome | 22:07 |
| vgnbkr | Trove would need a mechanism to prevent a Cassandra db from being "open to the public" upon creation. | 22:07 |
| hub_cap | thats the higest level arguement, aka, how do we provide a sla and a managed datastore giving root access | 22:07 |
| hub_cap | sure but if i give u root mysql, u can screw things up and make, say, automated backup stop working | 22:07 |
| *** michael-yu has quit IRC | 22:07 | |
| hub_cap | so u end up having a hard to support sla | 22:07 |
| hub_cap | again, im w/ you but im arguing the _other_ side ;) | 22:07 |
| vgnbkr | Could it be that for mysql, trove just manages root and leaves it to the dba to manage users. For some dbs, trove might manage by controlling security group access? | 22:09 |
| *** pdmars has quit IRC | 22:09 | |
| denis_makogon | vgnbkr, security groups already managed by trove | 22:10 |
| hub_cap | vgnbkr: sure, thats what ive proposed before | 22:11 |
| hub_cap | and some agree w me, and some dont :) | 22:11 |
| vgnbkr | denis_makogon: trove "manages" security groups, or allows one to be attached to the nova instance? | 22:11 |
| vgnbkr | (and I'm not saying one is right, just asking) | 22:11 |
| denis_makogon | vgnbkr, it creates and attaches to instances | 22:11 |
| *** shakayumi has quit IRC | 22:11 | |
| denis_makogon | vgnbkr, we call it association | 22:12 |
| kevinconway | (-_-) | 22:12 |
| vgnbkr | denis_makogon: so how does trove know (at present) who to restrict access to? | 22:12 |
| denis_makogon | vgnbkr, cidr and ports | 22:13 |
| vgnbkr | But, anyways, it sounds like I'm rehashing something that was covered before. Didn't mean to do that. Sorry all. | 22:13 |
| *** jcru has quit IRC | 22:15 | |
| denis_makogon | hub_cap, let me know with the results of testing, please | 22:15 |
| hub_cap | yup ill test again tonight denis_makogon | 22:18 |
| vgnbkr | Was just chatting with the guys around the office. When is the last time that anyone ever saw a database with more than one user (i.e., the one that the web server connects to)? | 22:27 |
| vgnbkr | For DBaaS as databases are used in modern deployments (ORM typically), are users even relevant any more? | 22:29 |
| hub_cap | vgnbkr: a TON | 22:31 |
| hub_cap | read only users for reportin | 22:31 |
| hub_cap | or access to only certian tables for non priv'd users | 22:31 |
| vgnbkr | I couldn't remember seeing users used for many years. Asked around the office - all our customers just had one account. | 22:32 |
| *** Ranjitha has quit IRC | 22:32 | |
| vgnbkr | For access to certain tables, is it in the trove plan to add access controls by table? | 22:33 |
| denis_makogon | vgnbkr, no, this is not DBaaS functionality | 22:33 |
| *** doug_shelley66 has joined #openstack-trove | 22:34 | |
| denis_makogon | vgnbkr, maybe you need to search for OpenStack MangetoDB | 22:34 |
| denis_makogon | it's NoSQL key-value storage | 22:34 |
| denis_makogon | with HTTP interface | 22:35 |
| denis_makogon | its almost the same as Amazon DynamoDB | 22:35 |
| vgnbkr | The point that I was getting at is that if the objective is to offer a certain SLA, there must be a path. Does support for multiple users need to be on that path? | 22:36 |
| kevinconway | denis_makogon: there you go the the "A" word again | 22:36 |
| *** Ranjitha has joined #openstack-trove | 22:36 | |
| denis_makogon | kevinconway, yes, i could use kumu logic example | 22:36 |
| vgnbkr | Aren't most people using a cloud DBaaS just going to point their Apache/PHP at it? Are multiple users necessary for that use case? | 22:38 |
| denis_makogon | kevinconway, AWS is the greates stack of cloud services, we need to keep an eye on it, all the time | 22:38 |
| denis_makogon | vgnbkr, it depens on how application interracts with DB | 22:38 |
| vgnbkr | Absolutely. | 22:39 |
| denis_makogon | vgnbkr, but mostly, no | 22:39 |
| *** michael-yu has joined #openstack-trove | 22:39 | |
| vgnbkr | But if we're not going to provide access to access controls, what would people do with separate users? | 22:39 |
| denis_makogon | vgnbkr, twitter is a good example | 22:40 |
| *** Barker has quit IRC | 22:40 | |
| vgnbkr | denis_makogon: can you provide a short description of why twitter is a good example (I don't know what you're getting at). | 22:41 |
| kevinconway | twitter uses a database | 22:41 |
| denis_makogon | i mean single user access | 22:41 |
| kevinconway | QED | 22:41 |
| vgnbkr | Ah, sorry, I thought you meant twitter used multiple users. | 22:42 |
| *** jasonb365 has quit IRC | 22:42 | |
| denis_makogon | backend is driver driven | 22:42 |
| kevinconway | vgnbkr: multiple users is common in reporting/analytics databases | 22:42 |
| *** jasonb365 has joined #openstack-trove | 22:43 | |
| vgnbkr | kevinconway: is that because the reporting apps are given read only access? | 22:44 |
| vipul | even for mysql I think that multiple users with potentially different grants is totally valid | 22:44 |
| kevinconway | well a lot of analytics that leverage RDBMs don't start with an app | 22:44 |
| kevinconway | you test/discover in sql | 22:45 |
| kevinconway | then move your findings over into a reporting app | 22:45 |
| vipul | you may have a read-only slave you want to grant access to to a different app potentially | 22:45 |
| kevinconway | depending on the scope of the db you may have multiple reporting groups that need access to different subsets of data | 22:45 |
| vgnbkr | kevinconway: sure, but at that point wouldn't the dba have to open root access and go in and muck with grants anyways? | 22:46 |
| vipul | are we talking about reasons to keep user management in Trove or to remove it | 22:46 |
| kevinconway | vipul: neither. just talking about use cases | 22:46 |
| kevinconway | vgnbkr: you may also have multiple writers to the db that need write to only their portion | 22:46 |
| kevinconway | like using a trove instance to run an openstack deploy | 22:47 |
| kevinconway | nova writes to nova schema while keystone writes to keystone schema | 22:47 |
| vgnbkr | kevinconway: agreed, but is that something that trove would ever be able to configure? | 22:47 |
| kevinconway | same db instance | 22:47 |
| kevinconway | it does now | 22:47 |
| vipul | also i would hesistate to go with 'let the DBA do it' type of cases, because the value add of a managed service like DBaaS is that you don't necessarily need DBAs to figure this stuff out | 22:47 |
| openstackgerrit | A change was merged to openstack/python-troveclient: adds support for configurations management https://review.openstack.org/53169 | 22:48 |
| kevinconway | oh neat, config mgmt is in | 22:48 |
| vgnbkr | I was just thinking about hub_cap's SLA comment, which is solid. What if trove gave you just a write connect string and a read connect string by default with a high SLA. | 22:54 |
| vgnbkr | You could click "Allow Root Access" and do whatever you needed, but the SLA dropped accordingly. | 22:55 |
| hub_cap | vgnbkr: i think scenarios like that are worth exploring | 22:56 |
| hub_cap | when we start gettin into v2 | 22:56 |
| vgnbkr | Would that unify the appearance of disparate dbs from a management perspective? | 22:56 |
| *** mattgriffin has joined #openstack-trove | 22:57 | |
| vgnbkr | OK, np. | 22:58 |
| *** amrith has quit IRC | 22:59 | |
| cp16net | hub_cap: btw i see that trove config group is approved it will require the trove-integration as well | 23:01 |
| cp16net | client is merged already :) | 23:01 |
| *** amrith has joined #openstack-trove | 23:02 | |
| *** amcrn has quit IRC | 23:03 | |
| *** amcrn has joined #openstack-trove | 23:06 | |
| hub_cap | cp16net: link | 23:16 |
| hub_cap | vgnbkr: fwiw, i think thats a decent idea | 23:16 |
| cp16net | hub_cap: tim just approved | 23:18 |
| cp16net | i got +11 | 23:18 |
| cp16net | i think thats a record | 23:18 |
| SlickNik | cp16net: nice | 23:19 |
| SlickNik | Are all pieces merged? | 23:19 |
| SlickNik | Just making sure rdjenkins doesn't keel over. | 23:19 |
| cp16net | SlickNik: they are about to be merged | 23:21 |
| SlickNik | sweet. | 23:22 |
| openstackgerrit | A change was merged to openstack/trove-integration: changes to support configuration groups https://review.openstack.org/58445 | 23:22 |
| cp16net | OMG | 23:22 |
| cp16net | there it is | 23:22 |
| hub_cap | :P | 23:22 |
| hub_cap | woop | 23:22 |
| cp16net | that was fast | 23:22 |
| cp16net | i thought the merge took a while | 23:22 |
| *** robertmyers has quit IRC | 23:25 | |
| *** jasonb365 has quit IRC | 23:25 | |
| SlickNik | cp16net: trove-integration merges are super fast. | 23:32 |
| SlickNik | You'll probably have to wait a little longer for the other patch. :) | 23:33 |
| hub_cap | yea there is only a no-op gate for t-i | 23:33 |
| hub_cap | lets just call it t-i | 23:33 |
| hub_cap | or ti | 23:33 |
| openstackgerrit | A change was merged to openstack/trove: adding configuration group support https://review.openstack.org/53168 | 23:33 |
| hub_cap | t-i i think works | 23:33 |
| *** kevinconway has quit IRC | 23:34 | |
| hub_cap | thoughts? | 23:34 |
| SlickNik | I'm down with that. | 23:34 |
| *** jmontemayor has quit IRC | 23:35 | |
| *** plodronio has quit IRC | 23:35 | |
| *** jmontemayor has joined #openstack-trove | 23:35 | |
| *** yogesh has quit IRC | 23:36 | |
| *** jmontemayor has quit IRC | 23:39 | |
| hub_cap | word | 23:41 |
| hub_cap | lets do it | 23:41 |
| hub_cap | lets make it a thing | 23:41 |
| *** denis_makogon has quit IRC | 23:41 | |
| hub_cap | t-i is how we refer to trove-integration from now on | 23:41 |
| hub_cap | yeesh, another fresh sync of offlineimap... /me crys | 23:42 |
| hub_cap | and leaves for ~6 hrs | 23:42 |
| hub_cap | heh | 23:42 |
| SlickNik | heh, I know that feeling | 23:42 |
| *** tanisdl has quit IRC | 23:47 | |
| *** amrith has quit IRC | 23:48 | |
| *** Barker has joined #openstack-trove | 23:49 | |
| *** Barker has quit IRC | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!