Friday, 2013-12-06

*** amytron has quit IRC00:00
openstackgerritDan Nguyen proposed a change to openstack/trove: Validate databases for user grants  https://review.openstack.org/5952900:00
*** flaper87 is now known as flaper87|afk00:12
openstackgerritDan Nguyen proposed a change to openstack/trove: Validate databases for user grants  https://review.openstack.org/5952900:21
*** harlowja has quit IRC00:40
*** harlowja has joined #openstack-trove00:42
openstackgerritDan Nguyen proposed a change to openstack/trove: Validate databases for user grants  https://review.openstack.org/5952900:46
*** esp has left #openstack-trove00:49
*** vipul has quit IRC00:51
*** vipul has joined #openstack-trove00:52
*** jcooley_ has joined #openstack-trove00:56
*** yogeshmehra has joined #openstack-trove01:03
*** yidclare has quit IRC01:08
openstackgerritSlickNik proposed a change to openstack/trove: Remove comments from empty __init__.py files  https://review.openstack.org/6039301:11
openstackgerritSlickNik proposed a change to openstack/trove: Remove comments from empty __init__.py files  https://review.openstack.org/6039301:13
*** yogeshmehra has quit IRC01:22
*** mmcdaris has joined #openstack-trove01:37
*** nosnos has joined #openstack-trove01:42
*** tanisdl has quit IRC02:01
*** jcooley_ has quit IRC02:03
*** jcooley_ has joined #openstack-trove02:04
*** mmcdaris has quit IRC02:07
*** jcooley_ has quit IRC02:08
*** rongze has joined #openstack-trove02:12
*** robertmyers has joined #openstack-trove02:16
*** jcooley_ has joined #openstack-trove02:19
*** mmcdaris has joined #openstack-trove02:26
*** mmcdaris has quit IRC02:40
*** erkules_ has joined #openstack-trove02:44
*** erkules has quit IRC02:47
*** erkules_ has quit IRC02:48
*** erkules_ has joined #openstack-trove02:50
*** vipul is now known as vipul-away03:05
*** mmcdaris has joined #openstack-trove03:07
*** amcrn has quit IRC03:08
*** vipul-away is now known as vipul03:11
*** robertmyers has quit IRC03:24
*** jcooley_ has quit IRC03:29
*** jcooley_ has joined #openstack-trove03:30
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: .gitignore some extra generated files and test user creds for cli  https://review.openstack.org/5952703:31
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: fake mode script needs to be updated after changed were merged in trove  https://review.openstack.org/5844103:31
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: changes to support configuration groups  https://review.openstack.org/5844503:31
*** jcooley_ has quit IRC03:34
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: fake mode script needs to be updated after changed were merged in trove  https://review.openstack.org/5844103:39
*** krast has joined #openstack-trove03:43
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: .gitignore some extra generated files and test user creds for cli  https://review.openstack.org/5952703:43
*** shalini has quit IRC03:47
openstackgerritCraig Vyvial proposed a change to openstack/trove: adding configuration group support  https://review.openstack.org/5316803:53
*** shalini has joined #openstack-trove04:02
*** erkules_ is now known as erkules04:23
*** mmcdaris has quit IRC04:25
*** mmcdaris has joined #openstack-trove04:28
*** rongze has quit IRC04:50
*** rongze has joined #openstack-trove04:54
*** jcooley_ has joined #openstack-trove04:55
*** esp has joined #openstack-trove04:56
*** esp has left #openstack-trove04:59
*** shalini has quit IRC05:06
*** jcooley_ has quit IRC05:22
*** jcooley_ has joined #openstack-trove05:22
*** jcooley_ has quit IRC05:27
*** vipul has quit IRC05:52
*** vipul has joined #openstack-trove05:53
*** yogeshmehra has joined #openstack-trove05:59
*** yogeshmehra has quit IRC06:00
*** yogeshmehra has joined #openstack-trove06:06
*** denis_makogon has joined #openstack-trove06:27
*** yogeshmehra has joined #openstack-trove06:30
*** nosnos_ has joined #openstack-trove06:30
*** mmcdaris has quit IRC06:31
*** nosnos has quit IRC06:33
*** yogeshmehra has quit IRC06:40
*** jumping_Bond has joined #openstack-trove06:41
*** jumping_Bond has quit IRC06:41
*** Flying_Bond has joined #openstack-trove06:42
*** Flying_Bond has quit IRC06:43
*** Flying_Bond has joined #openstack-trove06:43
*** mmcdaris has joined #openstack-trove06:44
*** yogesh has joined #openstack-trove06:44
*** mmcdaris has quit IRC06:55
*** denis_makogon has quit IRC07:14
*** yogesh has quit IRC07:22
*** yogesh has joined #openstack-trove07:40
*** yogesh_ has joined #openstack-trove07:43
*** yogesh has quit IRC07:45
*** yogesh has joined #openstack-trove07:46
*** yogesh_ has quit IRC07:46
*** mmcdaris has joined #openstack-trove07:51
*** yogeshmehra has joined #openstack-trove07:51
*** yogeshmehra has quit IRC07:56
*** flaper87|afk is now known as flaper8707:56
*** Flying_Bond has quit IRC08:20
*** mmcdaris_ has joined #openstack-trove08:27
*** krast_ has joined #openstack-trove08:30
*** hub_cap_ has joined #openstack-trove08:31
*** esmute_ has joined #openstack-trove08:33
*** esmute has quit IRC08:34
*** lifeless has quit IRC08:34
*** isviridov has quit IRC08:34
*** mmcdaris has quit IRC08:34
*** krast has quit IRC08:34
*** hub_cap has quit IRC08:34
*** isviridov has joined #openstack-trove08:34
*** mmcdaris_ is now known as mmcdaris08:35
*** rongze_ has joined #openstack-trove08:35
*** esmute_ is now known as esmute08:35
*** lifeless has joined #openstack-trove08:36
*** rongze has quit IRC08:37
*** haomaiwang has joined #openstack-trove08:44
*** nosnos_ has quit IRC09:29
*** nosnos has joined #openstack-trove09:29
*** yogesh has quit IRC09:30
*** Flying_Bond has joined #openstack-trove09:32
*** yogeshmehra has joined #openstack-trove09:32
*** yogeshmehra has quit IRC09:59
*** bot_bot is now known as denis_makogon10:09
*** krast_ has quit IRC10:34
*** nosnos_ has joined #openstack-trove10:44
*** nosnos has quit IRC10:48
*** Flying_Bond has quit IRC10:49
*** rongze_ has quit IRC11:17
*** rongze has joined #openstack-trove11:28
*** pdmars has joined #openstack-trove12:48
*** pdmars has quit IRC12:49
*** pdmars has joined #openstack-trove12:50
*** abramley has joined #openstack-trove13:02
*** nosnos_ has quit IRC13:11
*** demorris has joined #openstack-trove13:29
*** mmcdaris has quit IRC13:36
*** mmcdaris has joined #openstack-trove13:59
*** radez_g0n3 is now known as radez14:14
*** rnirmal has joined #openstack-trove14:32
*** russellb is now known as rustlebee14:38
*** cweid has joined #openstack-trove14:53
*** robertmyers has joined #openstack-trove14:53
*** ashestakov has joined #openstack-trove14:53
*** rongze has quit IRC15:00
*** robertmyers has quit IRC15:04
*** rongze has joined #openstack-trove15:16
*** mmcdaris has quit IRC15:17
*** jcru has joined #openstack-trove15:28
*** amytron has joined #openstack-trove15:33
*** tanisdl has joined #openstack-trove15:36
*** datsun180b has joined #openstack-trove15:36
*** SushilKM has joined #openstack-trove15:38
*** plodronio has joined #openstack-trove15:48
*** jmontemayor has joined #openstack-trove15:49
*** SushilKM has quit IRC15:50
*** demorris_ has joined #openstack-trove16:01
*** demorris has quit IRC16:01
*** demorris_ is now known as demorris16:01
*** jcooley_ has joined #openstack-trove16:10
*** jasonb365 has joined #openstack-trove16:10
*** robertmy_ has joined #openstack-trove16:11
*** robertmy_ has quit IRC16:13
*** robertmy_ has joined #openstack-trove16:13
*** robertm__ has joined #openstack-trove16:17
*** robertmy_ has quit IRC16:18
*** robertm__ has quit IRC16:18
*** robertmy_ has joined #openstack-trove16:20
*** robertmy_ has quit IRC16:22
*** robertmyers has joined #openstack-trove16:22
*** robertmyers has quit IRC16:23
hub_cap_hey, austin is frozen, so it might be a while before the cro-magnon start stirring... its gotta thaw there first16:25
*** hub_cap_ is now known as hub_cap16:25
*** robertmyers has joined #openstack-trove16:25
denis_makogonhub_cap, hi =)16:26
hub_caphowdy16:26
hub_capalso, im not here to work yet, i still have to make coffee/breakfast but ill be back in ~30 minutes16:28
denis_makogonhttps://review.openstack.org/#/c/60553/ https://review.openstack.org/#/c/60532/16:30
denis_makogonit might be interesting16:30
ikhudoshynhub_cap, hey, ive got +1s from robots at last, could u pls check https://review.openstack.org/#/c/57189/16:35
hub_capcommented on the first one denis_makogon .. u might want to update the commit msg16:36
denis_makogonhub_cap, already16:36
hub_capikhudoshyn: im not really around for like ~20 more min... remind me16:36
ikhudoshynhub_cap, sure16:36
*** ashestakov has quit IRC16:39
hub_capikhudoshyn: i lied, i did it16:39
hub_capand my wife poured me coffee cuz i havent left my desk now hehe16:41
*** jcooley_ has quit IRC17:10
demorriscurious, how do y'all program? http://www.reddit.com/r/ProgrammerHumor/comments/1rx2p0/programming_in_movies_vs_programming_in_real_life/17:18
hub_capdemorris: OMG LOL17:23
hub_capim def the green screen one17:23
demorris:)17:23
*** Flying_Bond has joined #openstack-trove17:24
*** ashestakov has joined #openstack-trove17:24
*** mmcdaris has joined #openstack-trove17:25
demorrisI am more this style - http://www.youtube.com/watch?v=dxIPcbmo1_U17:28
*** SushilKM has joined #openstack-trove17:30
*** grapex has joined #openstack-trove17:31
hub_capomg17:32
hub_capif it wasnt for that stupid slow UI, she wouldve hacked that so miuch faster17:32
demorrisIt's a Unix system!17:32
hub_capi know this!17:33
hub_caplol @ that line17:33
hub_caphmmmmm i dont know if i should map shift+bkspace or super+bkspace ot delete17:34
hub_capwhat do yall recommend?17:34
hub_capim thinking shift17:34
pdmarsdemorris: "Denis, our lives are in your hands and you have butterfingers?"17:34
hub_cappdmars: its actually buttafingahs17:35
pdmarshub_cap: true17:35
hub_caphttp://www.hark.com/clips/wsmqpbjnpb-lives-are-in-your-hands-and-you-have-butter-fingers17:35
hub_capi couldnt do an inflection in the text tho17:36
pdmarsthat was pretty close17:36
hub_capya i think the a needed a umlaut or something17:36
hub_capmaybe a accent mark17:36
hub_cap:P17:36
demorrispdmars: :)17:37
openstackgerritSushil Kumar proposed a change to openstack/trove: Updates tox.ini to remove suppressed  https://review.openstack.org/6056417:38
*** esp has joined #openstack-trove17:48
openstackgerritSushil Kumar proposed a change to openstack/trove: Updates tox.ini to remove suppressed H403, H702.  https://review.openstack.org/6056417:52
*** flaper87 is now known as flaper87|afk17:54
*** mmcdaris has quit IRC17:54
hub_capbrb in a bit17:55
*** grapex has quit IRC17:56
*** rnirmal has quit IRC18:05
*** SushilKM has quit IRC18:05
*** grapex has joined #openstack-trove18:07
*** SnowDust has joined #openstack-trove18:11
datsun180bso what's it take to ask for some eyeballs on an itty bitty tiny baby fix that's been sort of sitting around for a while like https://review.openstack.org/#/c/56966/ and would just love to get merged18:14
*** yogesh has joined #openstack-trove18:18
*** demorris has quit IRC18:21
*** yogesh has quit IRC18:23
*** yogesh has joined #openstack-trove18:23
*** yogesh has quit IRC18:23
*** yogesh has joined #openstack-trove18:24
*** yidclare has joined #openstack-trove18:26
*** rnirmal has joined #openstack-trove18:31
*** amcrn has joined #openstack-trove18:32
*** Flying_Bond has quit IRC18:33
*** mmcdaris has joined #openstack-trove18:36
*** Flying_Bond has joined #openstack-trove18:38
*** yogeshmehra has joined #openstack-trove18:40
SlickNikhub_cap / grapex: Can you guys take a look at https://review.openstack.org/#/c/58062/ and https://review.openstack.org/#/c/58066/ when you get a chance? Thanks!18:42
*** imsplitbit has quit IRC18:49
*** Ikhudoshyn__ has joined #openstack-trove18:49
openstackgerritA change was merged to openstack/trove: Added ability to pass custom command-line options to backup runner  https://review.openstack.org/5806218:49
openstackgerritA change was merged to openstack/trove: Clean up zombie processes on backup failure  https://review.openstack.org/5806618:51
*** Ikhudoshyn__ has quit IRC18:58
*** yogeshmehra has quit IRC18:58
*** tanisdl has quit IRC18:59
*** westmaas is now known as westmau519:07
SlickNikThanks grapex!19:07
datsun180bcool, thanks for looking at that, i'll rebase it19:11
openstackgerritDenis M. proposed a change to openstack/trove: RootReport through Conductor  https://review.openstack.org/5941019:12
openstackgerritA change was merged to openstack/trove: setting fake mode to print startup errors  https://review.openstack.org/5844019:13
openstackgerritEd Cranford proposed a change to openstack/trove: Extract suffix from req URL to avoid escaping dots  https://review.openstack.org/5696619:14
*** tanisdl has joined #openstack-trove19:16
openstackgerritAndrey Shestakov proposed a change to openstack/python-troveclient: Datastore support improvements  https://review.openstack.org/6058819:16
*** mmcdaris has quit IRC19:17
*** plodronio has left #openstack-trove19:22
openstackgerritA change was merged to openstack/trove: Unittest Case for get_user  https://review.openstack.org/4820119:47
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: changes to support configuration groups  https://review.openstack.org/5844519:48
*** radez is now known as radez_g0n319:49
*** freyes has quit IRC19:49
*** yogeshmehra has joined #openstack-trove19:49
*** yogeshmehra has quit IRC19:50
grapexI've got a question for the room: Why are all of these pull requests for the guests have the agent raise NotImplementedError? Seems like Trove should be able to figure out if an agent won't be able to do some action based on the datastore type.19:50
grapexAlso, there's really no difference between raising NotImplementedError and simply not adding the method- both suggest the agent can't support the RPC call.19:50
*** amytron has quit IRC19:51
kevinconwaygrapex: what response gets sent back if the method it's looking for isn't there?19:52
grapexThe guest manager should raise an exception saying there is no such method.19:52
grapexSo to me it's the same- raising NotImplementedErrors everywhere seems unnecessary19:53
kevinconwayi guess raising NotImplemented would only be required if the guest used an ABCMeta19:53
grapexYeah- it already uses that Manager class which should inform the caller the method it wanted is missing19:54
*** mmcdaris has joined #openstack-trove19:56
*** yidclare has quit IRC19:57
ikhudoshyngrapex, SlickNik hi, guys, could u pls take a look https://review.openstack.org/#/c/57189/19:57
*** mmcdaris has quit IRC19:57
*** mmcdaris has joined #openstack-trove19:58
*** vipul is now known as vipul-away19:59
*** vipul-away is now known as vipul19:59
cp16netgrapex: yeah i think that makes sense but will that raise a different exception if the call went to the wrong manager?19:59
grapexcp16net: Wrong manager?20:01
*** datsun_F40PH has joined #openstack-trove20:06
datsun_F40PHoh neat, my vpn decided to take a nap20:06
datsun_F40PHwell at least i'm on public20:06
cp16netgrapex: well if the manger didnt have the method20:07
cp16netthats what i was thinking20:07
grapexcp16net: Then the RPC call simply fails with some exception saying the method wasn't there.20:08
*** datsun180b has quit IRC20:08
*** datsun_F40PH is now known as datsun180b20:08
kevinconwaygrapex: does it return a 501 NotImplemented as the final status when the RPC fails to find the method?20:10
*** mmcdaris has left #openstack-trove20:10
kevinconwayI haven't looked into the mechanism that does it, but there are 400 BadRequest exceptions thrown in the mysql guest that bubble up to the client that called the API20:10
kevinconwayso should they maybe raise HTTPNotImplmented of some kind? like a trove exception?20:11
*** mwagner_lap has joined #openstack-trove20:11
*** yogesh has quit IRC20:12
*** yogesh has joined #openstack-trove20:12
*** Flying_Bond has quit IRC20:14
grapexikhudoshyn: https://review.openstack.org/#/c/57189/ I responded20:14
grapexOne qualm: please rename "mount_point" to "data_dir" where it appears in the backup code20:15
grapexSomeone reading the code might infer that we're doing backups of the entire volume which may or may not be true for all datastores.20:15
*** rongze has quit IRC20:15
*** yogesh has quit IRC20:17
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: fake mode update after changes to trove  https://review.openstack.org/5844120:21
*** harlowja has quit IRC20:23
*** imsplitbit has joined #openstack-trove20:23
*** harlowja has joined #openstack-trove20:23
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: fake mode update after changes to trove  https://review.openstack.org/5844120:30
openstackgerritCraig Vyvial proposed a change to openstack/trove: revert change to the trove.test.conf  https://review.openstack.org/6061220:33
*** tanisdl has quit IRC20:52
*** denis_makogon_ has joined #openstack-trove20:52
*** denis_makogon has quit IRC20:58
*** denis_makogon_ is now known as denis_makogon20:58
*** tanisdl has joined #openstack-trove20:58
*** dmakogon_ has joined #openstack-trove20:58
denis_makogonguys, could you please review https://review.openstack.org/#/c/59410/20:59
*** SnowDust has quit IRC21:00
*** yogesh has joined #openstack-trove21:06
*** demorris has joined #openstack-trove21:11
*** SnowDust has joined #openstack-trove21:12
*** yidclare has joined #openstack-trove21:13
*** radez_g0n3 is now known as radez21:13
hub_capugh... stdby keeps making my laptop go crazy21:14
*** yogesh has quit IRC21:34
denis_makogonhub_cap, could you please re-review https://review.openstack.org/#/c/59410/ ? we already have two ready reviews, ikhudoshyn 's review(already approved) depends on mine,21:38
hub_capcan u link ikhudoshyn's review?21:40
denis_makogonhub_cap, https://review.openstack.org/#/c/59078/21:42
hub_capikhudoshyn: your dependent review is outdated... im not sure if itll auto merge (But its a good way to test it out!)21:43
hub_capdenis_makogon: he may have to rebase after yours gets in :P21:44
hub_capim not sure how gerrit works in this case21:44
denis_makogonmine already rebased21:44
denis_makogoncould mine land first ?21:44
denis_makogonhub_cap, he might rebase anyway21:47
denis_makogonlol, B. Singer anounced X-Men: Apocalypse movie =))21:48
cweidtotally trying this when I get home... http://i.imgur.com/eo999FM.gif21:51
hub_capwowowwowowo cweid21:52
openstackgerritRobert Myers proposed a change to openstack/python-troveclient: Adding pagination support for backups  https://review.openstack.org/6062621:53
hub_capthats crazy21:53
cweidYeah wish you would have saved all them AOL cds huh?21:54
grapexdenis_makogon hub_cap: Sorry, I just -2'd this: https://review.openstack.org/#/c/59410/21:54
denis_makogongrapex, for what reason ?21:55
grapexI think the correct fix is to change the root history from the Trove side- the guest shouldn't have to tell Trove root was enabled when Trove asked it to in the prepare call21:55
hub_capi think SlickNik asked this before grapex21:56
grapexhub_cap: I don't see his comment in gerrit21:57
denis_makogongrapex, i see what you mean, but there's no wat to know if guest successfully finished this task21:57
grapexdenis_makogon: If the prepare call fails, it never enters ACTIVE state.21:57
denis_makogongrapex, robertmyers asked : Looks good, however it seems like this doesn't need to be done on the guest side. Can't the taskmanager api update the root record when we make the call? Either way I'm fine with this.21:57
grapexI am curious21:58
grapexif a user asks for a new instance, and wants a root user, and then gets a new instance without, how will they know that it failed?21:59
datsun180bthe thinking was the user create calls are async just in case it takes nonzero time to execute the command in mysql21:59
datsun180bso then if it's async then only the guest would know if/when it completed successfully without the taskman polling for it or something21:59
denis_makogongrapex, problem, that guest cannot report, all cals are async22:00
datsun180bat least i think that was the rationale22:00
grapexIt's like the instance is in a semi-initialized state. Also, if it doesn't already, shouldn't enable_root raise an exception (and thus end the prepare call and cause a failure) if it fails anyway?22:00
grapexdeni_makogon: No, it can.22:00
denis_makogonif you want to know if root unabled, prepare call should be "call istead of cast"22:00
grapexIt does so by updating the service_status, the way it always does22:00
SlickNikdenis_makogon: why can't we do it in the taskmanager after the call to the guestagent. It's a _call_ and not a _cast_ so we can do it only if the guest agent has succeeded.22:00
grapexSo if I want a new instance which is restored from some backup, if the restore process fails we don't just say the instance is ACTIVE and leave it at that.22:01
grapexThe prepare call should always do what it was told or leave the guest in a failed state.22:01
denis_makogongrapex, so you suggesting to move report call from guest22:02
grapexdenis_makogon: yes. It could live along side the prepare call in the taskmanager code.22:02
*** yogesh has joined #openstack-trove22:02
*** jcooley_ has joined #openstack-trove22:03
denis_makogongrapex, https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L231 after last call, before return22:03
denis_makogongrapex, ok, i'm fine with that22:03
grapexdenis_makogon: Why not put it even closer to prepare though22:03
grapexlike right before or after the call22:04
grapexSo the instance is basically "born" with root history22:04
denis_makogongrapex, becasuse if prepare would fail we would have another message in MQ22:04
denis_makogonwithout consumer22:04
denis_makogongrapex, not so good case22:05
grapexdenis_makogon: I mean, don't even call conductor- set the root history in the database without talking to conductor.22:05
denis_makogonwhat if instance failed, but root enabled because it was written to db22:05
denis_makogongrapex, i understood it22:05
*** ashestakov has quit IRC22:06
denis_makogongrapex, and i will be silly if taskmanager would call conductor while it has direct access to db22:06
grapexdenis_makogon: If the instance fails, most likely ops or someone will have to fix it, or the user will just delete it. Either way at that point the experience is already subpar, no one would care that the root history was enabled22:06
*** radez is now known as radez_g0n322:06
*** mmcdaris has joined #openstack-trove22:06
grapexdenis_makogon: Sorry, I guess I misunderstood what you meant about having another message in MQ.22:06
*** mmcdaris has quit IRC22:07
grapexI thought you understood it, but I just wanted to restate just in case.22:07
*** yogeshmehra has joined #openstack-trove22:07
*** mmcdaris has joined #openstack-trove22:07
denis_makogongrapex, no proble,22:07
*** mmcdaris has left #openstack-trove22:07
denis_makogongrapex, i mean that i would like to avoid such things22:07
denis_makogonif instance fail - we should not get any other reports, except status22:08
denis_makogonif instance failed, root report should not even being reported22:08
openstackgerritRobert Myers proposed a change to openstack/trove: Paginate backup list api  https://review.openstack.org/6035822:09
denis_makogongrapex, i'm definitely ok with simplifying prepare call22:09
grapexdenis_makogon: If the instance fails and you have an entry in root history enabled, I don't see what the problem is with that.22:10
denis_makogongrapex, incorrect reporting to end-user22:10
denis_makogongrapex, it would looks like incosistent state22:11
grapexThe end user's instance already says FAILED. I don't think they'd care about it22:11
grapexlike if I'm waiting for food at a restaurant for hours, and then I find someone else took my order when they called my name, I don't care if they did or didn't include pickles and onions.22:11
denis_makogongrapex, anyway, let's avoid such cases, we can do that22:12
grapexdenis_makogon: Is your suggestion to put it around where the usage event is sent?22:12
denis_makogonyes22:12
grapexdenis_makogon: So, let's say an operator shuts off the task manager daemon22:13
denis_makogoni think it's perfect place22:13
grapexa usage event not being sent means the operator probably loses a bit of money or something22:13
*** robertmy_ has joined #openstack-trove22:13
grapexBut if we send the event earlier, its possible the customer is overbilled22:13
grapexSo, if we set root history earlier, what is the danger to the user?22:14
grapexThe only way in which it becomes misleading is if there is an error provisioning their instance22:14
grapexso the instance goes to FAILED status22:14
grapexand then an operator steps in and manually fixes it somehow22:14
grapexand then the operator doesn't enable root during the manual process, plus the user decides to change their mind and not enable root either22:15
grapexand then later the user gets upset when they see root is enabled- to me that's a hypothetical22:15
grapexthe risk on the other hand is that there is a time out in the taskmanager, so the root history is *never* set22:15
grapexbut then the instance ends up eventually provisioning - maybe it just took a long time- and now on top of the usage event issue, the root history is not set.22:16
denis_makogonyes, and _never_set_ is ok22:16
grapexI disagree.22:16
*** robertmyers has quit IRC22:16
grapexIf never set is ok for users who have root, why even have a root history?22:16
openstackgerritCraig Vyvial proposed a change to openstack/trove: making the bin scripts called with main()  https://review.openstack.org/5422522:16
denis_makogongrapex, you suggesting to put it after prepare call &22:17
denis_makogon?22:17
*** robertmy_ has quit IRC22:17
grapexOr right before- just don't put it after waiting to see the status changed.22:17
denis_makogonhttps://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L20522:18
*** yogeshmehra has quit IRC22:18
openstackgerritCraig Vyvial proposed a change to openstack/trove: changing the routes for the api to be clear  https://review.openstack.org/5422422:18
grapexdenis_makogon: inserting it on line 203, or right after 205 would make sense to me.22:18
denis_makogongrapex, ok, i'll do that22:19
denis_makogongrapex, thanks for brainstorm =)22:19
denis_makogongrapex, anyway, we need to break connection between guest and back-end22:20
grapexdenis_makogon: Thanks for the pull request- sorry there's this issue holding it up due to pre-existing code.22:20
denis_makogongrapex, asap22:20
denis_makogongrapex, no problem, i22:21
denis_makogongrapex, as i said, i'm totally ok with making guest prepare faster22:21
*** harlowja has quit IRC22:21
denis_makogongrapex, hub_cap, btw, have you seen ML thread about unified VM agents ?22:21
*** SnowDust has quit IRC22:22
grapexdenis_makogon: yes22:22
grapexI've been wondering when / if to respond.22:22
denis_makogongrapex, very interesting22:22
grapexI don't believe a unified agent across OpenStack projects makes any sense.22:22
denis_makogongrapex, it has a lot responses22:22
denis_makogongrapex, i think it would be nice to have framework for agents, with pluggable transport impl22:23
grapexI'm worried the end result could only be either a very bloated agent that has to do a bit of everything, or enforcing that an agent can only do several things which will require changing Trove22:23
denis_makogongrapex, and API22:23
grapexI think Trove should support an interface for agents. If some new framework is created and works well, I'd always be for using that.22:23
kevinconwayi should respond to the ML with my bash agent22:24
denis_makogongrapex, when i was playing with extracting guest code, i was common skeleton22:24
kevinconwayit listens to rabbit. that means it s cales22:24
denis_makogon*i saw22:24
*** jcooley_ has quit IRC22:24
grapexI just don't want at the next OS summit prominent people in OS who don't work on Trove to start asking why we aren't using some new framework that doesn't do things we may need.22:24
denis_makogongrapex, true22:25
grapexSo at Rackspace we use Sneaky Pete which is in C++. We also contribute to the reference agent. So I'm predisposed against the idea of creating a third agent since I've already written several trove features into two agents already.22:26
denis_makogongrapex, for now it will be nice to send into ML Trove requirements for such agent22:26
denis_makogoni think solum soon would need that22:26
*** yogeshmehra has joined #openstack-trove22:28
kevinconwayi'm still not sure what a common guest agent could even be except a framework on which to build agents.22:28
kevinconwayeven then what would the framework be? a base class on which to build your system daemon that listens to rabbit for messages?22:29
kevinconwaythat's what managers already do22:29
denis_makogonkevinconway, different transport22:29
denis_makogonkevinconway, different ways of executing tasks22:29
denis_makogonkevinconway, for now i know 2 ways22:30
denis_makogonkevinconway, API and scenarious22:30
kevinconwaybut how is that different from what a manager already does?22:30
denis_makogonkevinconway, no difference, that is why it would be easy to collect common needs and requirements22:31
kevinconwayif it is not different then what does it add?22:31
denis_makogonkevinconway, unified agents concept is only under design22:31
denis_makogonkevinconway, almost all code will be in oslo22:31
denis_makogonkevinconway, so other projects will just sync wth oslo and will develop their own agent on top of oslo lib22:33
kevinconwaydenis_makogon: manager and transport code is already in oslo22:33
denis_makogonkevinconway, yes22:33
denis_makogonkevinconway, you are right22:33
*** yogeshmehra has quit IRC22:33
denis_makogonkevinconway, but there's no was to combine them into one entity22:33
kevinconwayyou combine them as needed for your agent though22:34
denis_makogonkevinconway, use, but oslo will dictate you rules how to build it22:34
denis_makogonkevinconway, as i said - framework22:34
kevinconwaydenis_makogon: https://github.com/kevinconway/scriptpool22:36
kevinconwayit's apache licensed22:36
kevinconwaygreat agent framework22:36
denis_makogonlol22:36
kevinconwayyou can impl new transports easy22:36
kevinconwayit listens to rabbit already22:36
kevinconwayi' m working on a socket/network transport too so you can cURL to it22:37
denis_makogonwe need REST22:37
denis_makogoninstead of rpc/rebbit22:37
denis_makogonit's very hard to proxy rpc calls22:38
kevinconwaydenis_makogon: i plan to integrate my bash http server into it once complete. it will allow REST service development in bash that can drive the agent.22:38
denis_makogonso, ok, thanks to all for chat22:38
denis_makogonkevinconway, <---- maniac22:39
* denis_makogon gone until Mon22:39
*** denis_makogon is now known as bot_bot22:40
datsun180bdoes that mean that denis is really a robot?22:41
*** bot_bot has quit IRC22:45
*** datsun180b has quit IRC22:48
grapexOk, I've got to go. Talk to you all on Monday!22:48
hub_capcu guys22:51
hub_capcu bot_bot22:52
cp16netpeace out grapex22:58
cp16nethub_cap: did we decide to hold off on adding new datastores for a while?22:59
cp16neti think i recall this being said but wasnt sure22:59
cp16nethub_cap: at least until tempest ?22:59
hub_capcp16net: i had said it for ones that hadnt already been in review status23:00
hub_capbut i think at this point we should allow them23:00
*** yidclare has quit IRC23:00
hub_capbut we should discuss23:00
cp16netok23:00
hub_capwhat do u think cp16net ?23:00
cp16netbecause i've skipped looking at those23:00
hub_capheh23:00
hub_capid say give em a look-see23:00
hub_capim running redis right now23:00
cp16netcool23:00
jmontemayor+123:01
cp16neti've been working on getting a public image working for me23:01
hub_caphah of course youd say that jmontemayor ;)23:02
hub_capya i think bot_bot would like to know that too23:03
hub_capill talk to them on monday23:03
cp16netyeah i'm thinking is going to be difficult to test all the datastores23:03
cp16netin the public jenkins job23:03
hub_capwell.. with enough cloud vms u can test em ;)23:03
hub_capits just creating them is via a tempest fixture23:04
cp16nettrue...23:04
cp16netit would require basically a job for each datastore23:04
hub_capand we can reuse that to prov diff datastores and even clusters eventually23:04
*** jcru has quit IRC23:04
cp16netwhich would make it pretty easy i think23:04
*** yogesh has quit IRC23:05
cp16netthis might even be able to be helped out with using the fake daemon mode to run through the tests23:05
*** kevinconway has quit IRC23:05
*** yogesh has joined #openstack-trove23:05
cp16netit allowed me to test the api for configs23:05
cp16netwithout waiting 2 min for a vm to spin up23:06
cp16nethub_cap: another point is that i am sure the public wants something more than mysql by now as well23:06
cp16neti think if we merged them in we are early enough in the Icehouse cycle to be able to fix issues that come up23:07
cp16netit would not make sense to merge a new datastore in I4 time frame23:08
cp16netI3 at the latest i think23:08
cp16netwhat do you think hub_cap23:08
hub_capyes, there is no I4 :(23:09
hub_capbut we should merge them sooner rathe than later hehe23:09
*** yogesh has quit IRC23:09
*** jasonb365 has quit IRC23:10
*** pdmars has quit IRC23:12
*** pdmars has joined #openstack-trove23:13
*** pdmars has quit IRC23:14
*** yidclare has joined #openstack-trove23:15
*** rnirmal has quit IRC23:18
demorris+1 for merging them :)23:20
*** mmcdaris has joined #openstack-trove23:22
*** tanisdl has quit IRC23:22
cp16neti'll go through them some time soon23:26
cp16neti really want to get my image setup to fire up a review and test it23:26
cp16netso i can verify much mo betta23:27
*** mmcdaris has quit IRC23:33
*** tanisdl has joined #openstack-trove23:33
*** rnirmal has joined #openstack-trove23:33
hub_cap++23:36
hub_capim doin that w/ the redis review now23:36
hub_capbut for some reason rabbit wouldnt start23:36
hub_capso i killed the vm and brought up a new one heh23:36
hub_capnow its fine23:36
cp16netncie23:39
cp16netwhat all did you bake into the image?23:40
*** jmontemayor has quit IRC23:42
amcrn+1 for merging them :)23:48
hub_capumm it was kinda iterative... i can give u my root history after i sanitize it23:54

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!