Monday, 2014-01-27

« Sunday, 2014-01-26 Index Tuesday, 2014-01-28 »
*** matsuhashi has joined #openstack-trove00:09
*** mnaser has quit IRC00:16
*** SnowDust has quit IRC00:20
*** amcrn has joined #openstack-trove00:33
*** matsuhashi has quit IRC00:33
*** matsuhashi has joined #openstack-trove00:48
*** yogesh has joined #openstack-trove00:48
*** nosnos has joined #openstack-trove00:58
*** jcooley_ has joined #openstack-trove01:12
*** yogesh has quit IRC01:31
*** nosnos_ has joined #openstack-trove01:53
*** nosnos has quit IRC01:54
*** matsuhashi has quit IRC02:26
*** matsuhashi has joined #openstack-trove02:32
*** erkules_ has joined #openstack-trove02:46
*** erkules has quit IRC02:46
*** matsuhashi has quit IRC03:18
*** haomaiwang has joined #openstack-trove03:27
*** haomaiwang has quit IRC03:29
*** SnowDust has joined #openstack-trove03:30
*** haomaiwa_ has joined #openstack-trove03:30
*** Barker has quit IRC03:42
*** hub_cap has quit IRC03:51
*** hub_cap has joined #openstack-trove03:52
*** robertmyers has joined #openstack-trove03:56
*** hub_cap_ has joined #openstack-trove04:00
*** SnowDust has quit IRC04:13
*** matsuhashi has joined #openstack-trove04:40
*** robertmyers has quit IRC04:43
*** thedodd has joined #openstack-trove04:49
*** jcooley_ has quit IRC04:50
juicehello amcrn - saw your bug on the timestamp issue in tox - can you please review this (https://review.openstack.org/#/c/66063/).  it fixes that issue along with a few others surrounding instance/status.  thanks04:52
*** thedodd has quit IRC04:54
amcrnwhat's up juice; yeah, i'm still fairly positive there's a race that's not being accommodated for05:11
*** jcooley_ has joined #openstack-trove05:11
*** jcooley_ has quit IRC05:19
*** jcooley_ has joined #openstack-trove05:20
*** werebutt has joined #openstack-trove05:22
*** werebutt has left #openstack-trove05:22
*** jcooley_ has quit IRC05:24
amcrnjuice: can you rebase?05:30
amcrnnevermind, was simple05:31
*** erkules_ is now known as erkules05:41
*** shivamshukla has joined #openstack-trove05:54
*** jcooley_ has joined #openstack-trove06:01
*** SergeyLukjanov_ is now known as SergeyLukjanov06:04
hub_capjuice: wtf is up w/ that ascii art dude?06:05
*** jcooley_ has quit IRC06:14
amcrnlol06:17
amcrnjuice: btw, i take it back, patch works like a charm, awesome job. i had a mini-patch that was close to solving the issue, but i couldn't get around message delivery timing issues in the tests :/. glad you did all the hard work :)06:18
*** SergeyLukjanov is now known as SergeyLukjanov_06:31
hub_capjuice: why is abc introduced in your patchest? if anything, shouldtn that be a separate patchset applied before the one that amcrn is speaking of06:37
*** diakunchikov has quit IRC06:45
*** SergeyLukjanov_ is now known as SergeyLukjanov06:57
*** SergeyLukjanov is now known as SergeyLukjanov_07:20
*** flaper87|afk is now known as flaper8707:37
*** SergeyLukjanov_ is now known as SergeyLukjanov07:48
*** SergeyLukjanov is now known as SergeyLukjanov_08:03
*** SergeyLukjanov_ is now known as SergeyLukjanov08:03
*** SergeyLukjanov is now known as SergeyLukjanov_08:17
*** matsuhashi has quit IRC08:36
*** matsuhashi has joined #openstack-trove08:37
*** diakunchikov has joined #openstack-trove08:51
*** SnowDust has joined #openstack-trove09:28
openstackgerritSushil Kumar proposed a change to openstack/trove: Corrects matching of heat resources status  https://review.openstack.org/6930409:34
*** SushilKM has joined #openstack-trove09:35
openstackgerritSushil Kumar proposed a change to openstack/trove: Corrects matching of heat resources status  https://review.openstack.org/6930409:39
*** flaper87 is now known as flaper87|afk09:55
*** SnowDust has quit IRC09:55
*** amcrn has quit IRC09:58
*** SnowDust has joined #openstack-trove10:08
openstackgerritSushil Kumar proposed a change to openstack/trove: Corrects matching of heat resources status  https://review.openstack.org/6930410:16
*** haomaiwa_ has quit IRC10:29
*** haomaiwang has joined #openstack-trove10:30
*** SnowDust has quit IRC10:34
*** SnowDust has joined #openstack-trove10:52
*** SnowDust has quit IRC11:06
*** SergeyLukjanov_ is now known as SergeyLukjanov11:17
*** flaper87|afk is now known as flaper8711:18
*** SnowDust has joined #openstack-trove11:26
*** SergeyLukjanov is now known as SergeyLukjanov_11:33
*** matsuhashi has quit IRC11:40
*** matsuhashi has joined #openstack-trove11:45
*** igor has joined #openstack-trove11:55
*** SushilKM has quit IRC12:11
*** matsuhashi has quit IRC12:12
*** matsuhashi has joined #openstack-trove12:13
*** isvirido1 is now known as isviridov12:15
openstackgerritshivam shukla proposed a change to openstack/trove: Tests for heat based instance workflow  https://review.openstack.org/6649912:19
vgnbkrWhat's the status of the Fedora int-test?  I can't get redstack to work for DISTRO=fedora, either from an ubuntu host or a fedora.13:09
*** pdmars has joined #openstack-trove13:27
*** Barker has joined #openstack-trove13:31
*** SergeyLukjanov_ is now known as SergeyLukjanov13:31
*** grapex has quit IRC13:50
*** freyes has joined #openstack-trove13:56
*** rustlebee is now known as russellb14:12
*** jcru has joined #openstack-trove14:21
*** SushilKM has joined #openstack-trove14:21
*** SushilKM has quit IRC14:30
*** grapex has joined #openstack-trove14:32
*** grapex has quit IRC14:34
*** grapex has joined #openstack-trove14:35
*** radez_g0n3 is now known as radez14:38
*** kevinconway has joined #openstack-trove14:40
*** robertmyers has joined #openstack-trove14:44
*** SushilKM has joined #openstack-trove14:48
SushilKMHey Denis r u thr14:49
*** robertmyers has quit IRC14:50
*** robertmyers has joined #openstack-trove14:51
SushilKMhey robertmyers14:52
SushilKMredstack rd-client is broken14:53
SushilKMcan u luk this one https://review.openstack.org/#/c/68859/14:53
*** SnowDust has quit IRC14:58
*** amytron has quit IRC15:01
SushilKMhi Robert15:02
SushilKMsaw ur comment15:02
SushilKMlet me tell u why this does not works15:02
SushilKMsince when you source openrc from devstack it takes first argument as username15:02
SushilKMso list becomes the username instead of demo15:03
robertmyersso then we should drop this command from redstack15:05
robertmyersit is not needed at all15:05
robertmyersjust source openrc and then use the trove command directly15:05
robertmyersthis is only there to show how to do it15:06
robertmyersredstack is close to being completely removed15:06
SushilKMif its there to show then it shud work correctly, or it should not be there15:06
robertmyersWell maybe we should fix it, but it is hardly holding up development15:07
robertmyersthis is a low priority, I would vote for removing it15:07
SushilKMif someone executes this one they would not see the correct output15:08
robertmyersand just add to the docs on how to use trove with devstack15:08
robertmyersI'm just saying this is not a high priority to fix right now15:08
robertmyersit is not blocking anything15:09
SushilKMhmmm, its not a blocker, agree to that15:09
*** demorris has joined #openstack-trove15:10
*** Barker has quit IRC15:15
*** Barker has joined #openstack-trove15:15
*** Barker has quit IRC15:15
*** Barker has joined #openstack-trove15:16
*** rwsu has joined #openstack-trove15:16
openstackgerritSushil Kumar proposed a change to openstack/trove: Corrects spelling typos  https://review.openstack.org/6938315:19
openstackgerritSushil Kumar proposed a change to openstack/trove: Corrects help messages in cfg.py  https://review.openstack.org/6871715:27
*** jasonb365 has joined #openstack-trove15:30
*** igor has quit IRC15:31
*** igor has joined #openstack-trove15:32
*** amytron has joined #openstack-trove15:32
*** datsun180b has joined #openstack-trove15:33
*** igor has quit IRC15:36
*** SushilKM has quit IRC15:37
*** SergeyLukjanov is now known as SergeyLukjanov_15:38
*** SushilKM has joined #openstack-trove15:39
*** freyes has quit IRC15:39
*** ashestakov has joined #openstack-trove15:39
*** plodronio has joined #openstack-trove15:42
openstackgerritSushil Kumar proposed a change to openstack/trove: Corrects spelling typos  https://review.openstack.org/6938315:42
*** tanisdl has joined #openstack-trove15:45
*** SushilKM has quit IRC15:49
*** SergeyLukjanov_ is now known as SergeyLukjanov15:58
*** rnirmal has joined #openstack-trove16:02
*** amytron_ has joined #openstack-trove16:04
*** amytron has quit IRC16:04
*** jasonb365 has quit IRC16:04
*** amytron_ is now known as amytron16:04
*** demorris has quit IRC16:06
*** jasonb365 has joined #openstack-trove16:10
*** nosnos_ has quit IRC16:14
*** matsuhashi has quit IRC16:15
hub_capvgnbkr: we dont run the int-tests for fedora16:19
hub_caphmph am i not chatting in this channel?16:20
*** hub_cap has left #openstack-trove16:20
*** hub_cap has joined #openstack-trove16:21
vgnbkrhub_cap: Great.  So it is normal practice to just run int-tests on ubuntu then submit for review?16:23
hub_capvgnbkr: yessir, thats all the gate runs16:25
hub_capbut eventually we want the gate to run fedora16:25
hub_capdo the int-tests fail if you set the vm to run fedora?16:25
hub_capid like to hope not :)16:25
*** Barker has quit IRC16:28
hub_caplol man i totally spaced and forgot how i set up pgup hahah16:29
*** thedodd has joined #openstack-trove16:29
hub_capand subsequently i mightve missed a hilight if anyone ping'd me since last night16:29
*** SergeyLukjanov is now known as SergeyLukjanov_16:30
*** SushilKM has joined #openstack-trove16:36
*** Barker has joined #openstack-trove16:37
ashestakovhub_cap: vipul SlickNik grapex pls review neutron stuff https://review.openstack.org/#/c/61481/ https://review.openstack.org/#/c/63212/16:37
vgnbkrhub_cap: DISTRO=fedora failed for me on both ubuntu and fedora hosts.  On ubuntu, I seem to recall it failed because it tried to run apt-get on fedora.  On fedora host, it couldn't find percona-xtrabackup plus other issues.16:44
hub_capok thats not good!16:44
hub_capit used to detect that it was fedora, right ashestakov16:44
hub_capashestakov: did the work to build out the fedora impl for us (horray!)16:45
vgnbkrIt does get quite far in both cases.16:45
*** SnowDust has joined #openstack-trove16:45
vgnbkrI was using fedora20, and I did wasn't able to install the VirtualBox guest additions, so I suspect there may be some issues with 20.16:46
ashestakovwhats problem with fedora?16:47
SushilKMhi hub_cap16:48
SushilKMcurrently heat code, is broken up, I have put in a fix, can u see that https://review.openstack.org/#/c/69304/16:49
SushilKMDenis suggested me to use resource.state but it seems like there is no resource.state instead they have resource.resource_status in heat16:50
SushilKMwhich is crapping the heat code, currently16:50
hub_capso they changed their resource object?16:50
hub_capthat means they changed a contract, right?16:51
hub_capim confused :)16:51
SushilKMI too was, but i agreed for Denis, when he showed me something in  heat code16:52
SushilKMbut it seems they r using resource.state as their own terminology16:52
SushilKMbut with resource-show api they only publish resource.resource_status16:52
*** jasonb365 has quit IRC16:53
*** Barker has quit IRC16:54
*** SergeyLukjanov_ is now known as SergeyLukjanov16:54
*** mat-lowe_ has joined #openstack-trove16:54
hub_capso then resource.state never worked?16:55
hub_capcuz what it looks like, is that resource.state used to work, but now you are using resource.resource_status cuz something changed16:55
SushilKMyeah, it is like that only16:55
SushilKMfor me, at least16:55
hub_capor was resource.state never tested??/?16:55
*** ViswaV has joined #openstack-trove16:56
*** SushilKM has quit IRC16:56
openstackgerritAndrey Shestakov proposed a change to openstack/trove: Add mgmt API for datastores  https://review.openstack.org/6940816:57
*** SushilKM has joined #openstack-trove16:59
*** mat-lowery_ has joined #openstack-trove16:59
SushilKMIt was this place that he suggested of resource.state and its values https://github.com/openstack/heat/blob/master/heat/engine/stack_resource.py#L12116:59
*** mat-lowery has quit IRC16:59
*** thedodd has quit IRC17:00
*** SushilKM has quit IRC17:00
*** jasonb365 has joined #openstack-trove17:01
*** thedodd has joined #openstack-trove17:01
hub_capyea but lets try to run the heat path next time we commit to it (aka the big exception handling code)17:02
openstackgerritAndrey Shestakov proposed a change to openstack/python-troveclient: Add mgmt API for datastores  https://review.openstack.org/6941017:02
hub_capwhat this just points out is that if we want to consider heat a first class citizen is that it needs tests, right grapex117:02
hub_capDOH, grapex !17:02
*** SergeyLukjanov is now known as SergeyLukjanov_17:03
*** mat-lowe_ has quit IRC17:04
*** mat-lowery_ is now known as mat-lowery17:05
*** SushilKM has joined #openstack-trove17:06
SushilKMsure, i wud keep that in mind while putting any new change for heat code, and also there is going on work for unit and int test cases for heat code, soon we may have that on :)17:06
hub_capgood, im speaking on int-tests really17:07
SushilKM:)17:07
kalebpomeroyI'm starting work on the datastore capability features soon. If anyone wants to provide feedback, blueprint is here: https://blueprints.launchpad.net/trove/+spec/capabilities17:07
*** SushilKM has quit IRC17:07
kalebpomeroywiki/more docs are here: https://wiki.openstack.org/wiki/Trove/trove-capabilities17:07
hub_caphey kalebpomeroy.. i wonder, do u think capabilities might be something we talk about at the mid cycle sprint?17:09
hub_capor is it pretty stright forward?17:09
ashestakovi have question about mgmt api route17:09
kalebpomeroyI believe its really straightforward17:09
ashestakovpls look to https://review.openstack.org/#/c/69408/1/trove/extensions/routes/mgmt.py L11617:09
grapexhub_cap: Agreed17:10
ashestakovi need connect route '{tenant_id}/mgmt/datastores/versions' to MgmtDatastoreVersionController() to update action17:10
ashestakovcan i do it without create new resource?17:10
hub_caphey ashestakov im not familar w/ the :(datastore_id) for extensions either heh17:11
hub_capu can absolutely do it, if u look at the instance code there is examples of routes that are explicitly defined17:11
hub_capashestakov: ^ ^17:11
ashestakovhub_cap: why not :(datastore_id) ?17:13
hub_capive just never seen it before :)17:13
hub_capmaybe my memory just alludes me17:13
*** Barker has joined #openstack-trove17:13
hub_capbut lets not focus on that17:14
ashestakovhub_cap: for "instances" resource there are actions like "/mgmt/instances/root"17:15
hub_caplook @ instance actions17:16
hub_captahts a call to a nonstandard method iirc17:16
ashestakovi need like https://github.com/openstack/trove/blob/master/trove/common/api.py#L5217:16
hub_capso i guess im confused w/ your question17:19
hub_capis all u need to call a specific method based on a route17:19
hub_capand that method is not called a standard thing (list/show/delete..)17:19
ashestakovbut i need standard "update" call for that alias17:20
hub_capwhat is a standard update call?17:21
hub_capu never said the verb either, POST, GET, ?17:21
ashestakovPUT17:21
hub_capso its hard to understand "update call"17:21
hub_capso u just want a PUT? to call a method?17:21
hub_capthats done by default, the documentation for routes should show u what it shoudl be called17:21
hub_capGET/POST/PUT/DELETE/PATCH(maybe not patch?) are all magic and go to specifically named methods17:22
ashestakovhttps://review.openstack.org/#/c/69408/1/trove/extensions/routes/mgmt.py L100 this is route for resource, it accepts POST/PUT calls. I need add alias '{tenant_id}/mgmt/datastores/versions' for same controleed and for same "update" method17:22
hub_capand are exposed naturaly for everuy reousrce17:22
hub_capisint that the 3rd route there ashestakov ?17:23
hub_capu already have that route defined17:23
vgnbkrashestakov: the first issue with fedora is that it couldn't find percona-xtrabackup, but as I said that could be because I was using fedora 2017:23
hub_capline 11617:23
hub_capvgnbkr: u might need to check if the image builder added any 3rd party repos (if needed)17:23
ashestakovvgnbkr: do you using pre-installed image?17:23
ashestakovhub_cap: at L116 i defined different resource for same controller17:24
ashestakovhub_cap: and i want to know is it possible to avoid one more resource, and just make mapper.connect()17:25
hub_capso youre asking if u can remove the resource in 116 ?17:25
ashestakovhub_cap: yes17:25
ashestakovand attach that route to "update" methos17:25
*** SergeyLukjanov_ is now known as SergeyLukjanov17:26
SnowDusthello all from RAX / HP / IBM17:26
SnowDustcan we discuss python-troveclient/tools/install_venv_common.py17:26
hub_capoh ok, so waht u are saying is that u have /versions u want to append to /datastores/ and /datastores/{id}, and u want to know if u can do it in 1x route instead of 2, right ashestakov ?17:27
hub_capSnowDust: yea i say nuke it, i read denis_makogon 's reply, and since _I_ told u to ask the list17:27
SnowDustyeah ..17:27
SnowDust:)17:27
hub_capand i also am pretty sure we dont use it in any of trove-integratio17:27
SnowDuststill .. it was a last time call ...17:27
hub_capand honestly i dont know what the last bit of his msg was talking about17:28
SnowDustyeah .. :)17:28
hub_capdenis_makogon: around?17:28
ashestakovhub_cap: yes, i need two routes PUT '{tenant_id}/mgmt/datastores/:(datastore_id)/versions' and '{tenant_id}/mgmt/datastores/versions' in one resource17:28
ashestakovhub_cap: https://gist.github.com/andreyshestakov/8557935#file-datastores_mgmt_api-L60-L61:17:30
hub_capashestakov: ok now i understand yer question, give me a few min17:30
SnowDustok .. will wait an hour ...17:30
SnowDustthen will wipe it off ...17:31
*** freyes has joined #openstack-trove17:31
hub_capSnowDust: sure.. but i dont think anyone will mind. if denis_makogon is around, maybe he can explain the stuff at the end of his message about the other clients17:32
SnowDustyeah .. :) seriously .. his research as he says !17:32
hub_capya i didnt understand what he meant. but lets not merge the code berfore he gets a chance to explain it17:33
hub_capbut feel free to propose it17:33
*** yidclare has joined #openstack-trove17:33
SnowDustyeah thats why .. i never hurried on it ..17:33
SnowDusthey krast wassup ?17:34
*** jcooley_ has joined #openstack-trove17:34
hub_capashestakov: so i didnt see anything offhand, but you should check out the routes docs17:34
ashestakovhub_cap: i dived in routes, but i didnd found any solution17:35
hub_capive had to look @ the routes docs a few times to understand whats going on.. those are just objects that feed info to python-routes17:35
hub_capok17:35
hub_capthen id say no17:35
hub_capif youve already looked at routes :)17:35
vgnbkrashestakov: ran fedora live 20 under VirtualBox, defined DISTRO=fedora, then run ./redstack install17:35
ashestakovthat is not clear python-routes, that is trove extention for routes17:36
hub_capashestakov: k17:36
hub_capvgnbkr: im not sure that DISTRO=... is good outside of teh redstack.rc17:36
hub_capdid u define it there?17:36
hub_capashestakov: did we make that external?17:36
hub_capyea vgnbkr , https://github.com/openstack/trove-integration/blob/master/scripts/redstack.rc17:37
hub_capdid u put define it there, or as a env var?17:37
hub_capu might want to wrap that DISTRO=ubuntu17:37
hub_capto17:37
ashestakovhub_cap: oh, not trove ext, just common openstack https://github.com/openstack/trove/blob/master/trove/openstack/common/extensions.py17:37
hub_capDISTRO=${DISTRO:-ubuntu}17:38
*** igor has joined #openstack-trove17:38
hub_capyea those are broken too ashestakov17:38
*** thedodd has quit IRC17:38
hub_capi dont think they are used or updated anymore ;)17:38
hub_capour wsgi and our extensions from common are both mostly unused by other teams :D17:38
ashestakovit used for extentions https://github.com/openstack/trove/blob/master/trove/extensions/routes/mgmt.py#L3217:38
hub_capsure17:39
hub_capby our project it is17:39
hub_capcheck nova17:39
*** thedodd has joined #openstack-trove17:39
*** amcrn has joined #openstack-trove17:39
ashestakovdoes nova has extentions?17:39
hub_capyup17:39
hub_capi know nova, glance and cinder have extensions17:40
hub_capand thye used to use this common extensions code17:40
hub_capbut i think they all stopped :)17:40
hub_capand iirc, using some combination of entrypoints+stevedore17:40
*** SergeyLukjanov is now known as SergeyLukjanov_17:41
juicehub_cap: amcrn: I am around when current discussions trickle down to discussion mega-patch17:41
juiceshort answer - it wasn't intentional but it was all scaffolding to help keep me clear on what was going on while fixing the timestamp issue which turns out to involve other things17:42
juicebut more on that later17:42
juiceabc can be easily lifted to another patch set17:42
hub_capMEGAPATCH juice17:45
hub_caplets chat17:45
hub_capbreak it up17:45
hub_capdone17:45
hub_capgood chat17:45
hub_cap;)17:45
ashestakovvgnbkr: can you login to you fedora instance and find xtrabackup manually?17:45
juiceok17:46
vgnbkrI meant it couldn't find it with yum (or whatever)17:46
juiceso ascii art...17:46
juice:)17:46
vgnbkrhub_cap: Yes, I'll be delivering: DISTRO=${DISTRO:-ubuntu}17:46
juicefor maintainers or people new to the project, our objects geography (for lack of better word) is a little unclear17:46
hub_capvgnbkr: <317:46
hub_capyes do the ascii art buildings explain anything juice ??17:47
juiceso I thought it would help to have a few embedded diagrams to help as reminders17:47
hub_cap:D im all for more documentation, but it seems kinda distracting17:47
juiceha ha ... well a class diagram would be better :)17:47
* hub_cap pukes17:47
hub_capbut srsly, its all good work17:47
ashestakovvgnbkr: do you using "$DATASTORE_TYPE" == "percona" with fedora?17:47
hub_capjust make it into s few small works of art ;)17:47
vgnbkrashestakov: no17:48
hub_capashestakov: i think perc bkup is default for non perc installs too17:48
juicewhen you have this pinwheel of instance, server and compute instance, it gets a little dizzyying when you have logic bouncing around between those classes17:48
*** Barker has quit IRC17:48
juiceI am not married to the art work and it's easy to put abc in another patchset17:49
ashestakovhub_cap: frdora-percona has script for repos https://github.com/openstack/trove-integration/blob/master/scripts/files/elements/fedora-percona/install.d/05-percona-server17:49
juicewould that ease your indigestion?17:49
hub_capjuice: i think there was 3 things, right?17:49
ashestakovbut fedora-mysql hasnt17:49
juicelike you, I just want this chapter closed17:49
juicelet me look at what you wrote17:49
hub_capashestakov: u dont need percona server for xtrabackup17:49
hub_capjuice: so do i17:49
hub_capits my highest prio, to get the bug merged17:49
ashestakovbut you need repo to retrieve percona-xtrabackup17:50
juiceoh actually 217:50
hub_capi dont want to see our gate failing17:50
hub_capashestakov: oh the repo is only put to percona on fedora??17:50
hub_capi think thats in the base in ubuntu17:50
ashestakovyes17:50
juiceum the using iss I did not change someone else must have before me17:50
hub_capboom vgnbkr bug found, fix it up17:50
hub_capthx ashestakov17:50
juicebut it looks there were a few missing pieces17:50
juiceso I can pull out abc into another patch and then there will be just two17:51
*** Barker has joined #openstack-trove17:53
hub_capjuice: link me the review plz17:56
*** demorris has joined #openstack-trove17:57
juicethis one -> https://review.openstack.org/#/c/66063/ or something else?17:57
*** ashestakov has quit IRC18:00
*** Ranjitha has joined #openstack-trove18:00
hub_capahh the _other_ one juice18:00
hub_capwas that you fixed the service status to use a object everywehre instead of just a string18:01
hub_capusing the ISS object18:01
hub_capthat seems like it cuold be patch #118:01
hub_caperr18:01
hub_cap#218:01
SnowDusthub_cap : https://review.openstack.org/#/c/6443918:01
SnowDustcan we get this in now ?18:01
hub_cap#1 = abc meta (doesnt touch anything) --- also somewhat contriversial(sp*)18:01
hub_capSnowDust: yea im not sure why no one has approved yet18:02
SnowDust:)18:02
vgnbkrhub_cap: happy to take a look.  Can I submit my change for first-boot.d independently (without fedora working)?18:04
hub_capvgnbkr: yes18:04
hub_capwell18:04
vipulSnowDust: approved18:04
hub_capdo u plan on doing the other one too?18:04
hub_capmaybe make it the first review in a 2 review dependency (its just 2 separate commits in the same branch, wo diff gerrit ids)18:05
juiceis abc controversial18:05
hub_capHAHAHAHHAHHHAAHH18:05
vgnbkrhub_cap: didn't you ask me to look at it?  (did I misunderstand your comment above?)18:05
hub_capjuice: my hilight highlighted controversial18:05
*** harlowja_away is now known as harlowja18:05
hub_capcuz nTROVEr18:05
hub_capvgnbkr: not at all18:05
hub_capso push your first one18:05
juicewell I like it but I can remove it18:06
hub_capand if u get finished w/ #2 first, then make #2 the first in the patch set "Chain"18:06
hub_capill help u w/ that18:06
juicesounds good18:06
hub_capnot u juice18:06
hub_caplol18:06
hub_capthat was for vgnbkr18:06
hub_capthere is no hope for u juice  ;)18:06
hub_capso heres the deal w/ abc juice18:06
juicethere is a just enough to keep me going18:06
hub_capif u ask a non java python dev about it, they will puke18:07
hub_capid say make the abc completely independent18:07
juiceunderstood18:07
hub_capand make the patchset #2(string to ISS object)->#3(fix of the stupid timestamps)18:07
hub_capand feel free to submit the abc one18:07
juicei know the pythonic way is to keep folks on their toes18:07
hub_capbut u might also want to bring it up18:07
juiceshould have been called rattlesnake18:07
hub_capjuice: well done ;)18:07
hub_caplol, u get bit once and u DIE18:08
hub_capvipul: thx for the approve on that.. lol18:08
juicein that case a slow and steady strangulation is more appropropriate :)18:08
hub_cap:)18:11
juicelet me actually explain myself a bit further...I really do like the ease and lightness of python...I really do.  On the other hand as a maintainer of someone else's python code, I find there are some conventions that cause a noticeable amount of mental overhead keeping track of usage vs. declaration18:15
juiceand thus anything that helps us as maintainers keep things clear and concise helps us to deliver higher quality code and in a shorter amount of time with greater ease18:15
hub_capim all about the abc's yo18:16
hub_capfwiw18:16
hub_capso is cweid18:16
juice:)18:16
juiceI just wanted to put it out there so it didn't seem random18:17
hub_capjuice: the fact that u put it out there was random18:17
hub_capHAHAH18:17
juiceoh man :)18:17
grapexjuice: I came close to introducing ABCs to Trove on a few features I did, but by the time I was done working I often found they weren't as good as what I wanted.18:18
grapexI'm ok with them if you find a good use case. Not sure what the OpenStack Gods will say though.18:18
juicethe other part of it to is that i am in a relatively monogamous relationship with my IDE18:18
hub_capjuice: arent we all18:19
juiceand it helps in many many cases catch shit that we leave laying around18:19
grapexjuice: Which one are you using?18:19
hub_capthe real question, is, of course, is your IDE faithful to u?18:19
juicePyCharms18:19
hub_capwho wears the pants in the relationship?18:19
openstackgerritTim Simpson proposed a change to openstack/trove: Makes the backup tests less onerous  https://review.openstack.org/6942918:19
grapexjuice: That's kind of cool if it helps identify tiny mistakes in PyCharm18:19
grapexthough I don't think we can do something that drastic just for IDE support. I like PyCharm a lot but still prefer Sublime18:20
juicelike most relationships it is a compromise but I keep coming back for the sugar18:20
grapexjuice: Disturbing!18:20
grapexBy the way guys18:20
juiceahh grapex :)18:20
hub_capLOL juice grapex18:20
grapexthat pull review I just pushed out is almost nothing but it takes away a lot of pain of running these backup tests18:20
cp16netyou need some sugar for your juice?18:20
cp16netlol18:21
grapexI know this is the pot calling the kettle a tool for cooking food via heat but if you could look at it it wouldn't take more than a second. :D18:21
juiceweak cp16net :)18:21
*** jcooley_ has quit IRC18:22
juicebut damn you are still the beast on guitar in rock band cp16net18:22
*** jcooley_ has joined #openstack-trove18:23
cp16nethaha18:24
cp16netonly when i'm inebriated18:24
*** amcrn has quit IRC18:24
*** amcrn_ has joined #openstack-trove18:25
*** amcrn_ is now known as amcrn18:25
*** thedodd has quit IRC18:26
*** amcrn has quit IRC18:26
*** ViswaV has quit IRC18:30
*** ViswaV has joined #openstack-trove18:31
SnowDustanyone who can look into the zuul failure : https://jenkins02.openstack.org/job/gate-python-troveclient-pypy/156/console18:33
*** ViswaV has quit IRC18:33
*** khyati has joined #openstack-trove18:36
*** ViswaV has joined #openstack-trove18:37
*** yogesh has joined #openstack-trove18:40
*** ViswaV has quit IRC18:42
*** igor has quit IRC18:46
*** igor has joined #openstack-trove18:46
*** ashestakov has joined #openstack-trove18:48
*** amcrn has joined #openstack-trove18:49
esp@SnowDust you might try to ask someone in #openstack-infra18:50
SnowDustesp: sure !18:51
mat-loweryI was looking into adding "secgroup per datastore" functionality and found this blueprint (https://blueprints.launchpad.net/trove/+spec/security-groups-workflow-update) and this patch set (https://review.openstack.org/#/c/50944/3/trove/taskmanager/models.py). Patch set is abandoned with dmakogon stating in inline comment that change will come with another patch set. Can anyone help me find the patch set to which he is referring?18:51
esp@SnowDust were you able to run the tests locally?18:51
*** igor has quit IRC18:51
*** yogesh has quit IRC18:51
SnowDustno didnt try ..18:51
SnowDusti tried reasoning the console output itself :)  but cryptic18:52
*** yogesh has joined #openstack-trove18:52
*** ViswaV has joined #openstack-trove18:53
*** SergeyLukjanov_ is now known as SergeyLukjanov18:53
*** jmontemayor has joined #openstack-trove18:55
*** yogesh has quit IRC18:55
*** yogesh has joined #openstack-trove18:56
*** thedodd has joined #openstack-trove19:00
*** yogesh has quit IRC19:02
*** yogesh has joined #openstack-trove19:11
hub_capmat-lowery: hmph i can barely remember hehe19:12
hub_capdenis_makogon: would know19:12
*** igor has joined #openstack-trove19:16
mat-loweryhub_cap: Thanks. Regarding the approach taken in the patch set (i.e. ports per datastore specified in conf), what's the reason to put it there vs. datastore_versions table? And does this patch need to exist at all with Heat, which can handle secgroups per datastore now? Or is this patch a stop-gap until Heat becomes the only way? Or will Heat and Nova-based provisioning co-exist indefinitely? Sorry, lots of questions.19:17
openstackgerritTim Simpson proposed a change to openstack/trove: Makes the backup tests less onerous  https://review.openstack.org/6942919:24
openstackgerritSushil Kumar proposed a change to openstack/trove: Adds _delete_resources to FreshInstanceTasks  https://review.openstack.org/6945019:37
*** SushilKM has joined #openstack-trove19:37
*** SnowDust has quit IRC19:44
*** mat-lowery has quit IRC19:46
*** mat-lowery has joined #openstack-trove19:47
*** SushilKM has quit IRC19:48
*** SushilKM has joined #openstack-trove19:48
*** kevinconway_ has joined #openstack-trove19:51
*** kevinconway has quit IRC19:52
*** kevinconway_ is now known as kevinconway19:52
*** ViswaV has quit IRC19:54
*** amcrn has quit IRC19:54
*** SnowDust has joined #openstack-trove19:56
hub_caphey mat-lowery sry was in a meeting19:57
hub_capso, they _will_ Coexist19:57
hub_capbut we will deprecate the old way19:57
hub_capand eventualy remove it19:57
hub_capSWEET SEASON ONE OF THE MUPPET ShOW JUST GOT HeRE!!!19:58
hub_capso ther eis a really odd place where there is a disconnect between the 2 impls19:58
*** SushilKM has quit IRC19:58
hub_capaka _in_ the yaml vs _in_ a config to turn on off19:58
*** yogesh has quit IRC19:59
hub_capso ive tried to make us adhere to if its configurable from the operator perspective via a config19:59
hub_capthat we keep that in the default heat template as well (for instance jinja if stmts to turh things like dns on/off)19:59
*** michael-yu has joined #openstack-trove20:00
*** Barker has quit IRC20:02
*** datsun180b has quit IRC20:03
*** yogesh has joined #openstack-trove20:03
*** jasonb365 has quit IRC20:05
*** datsun180b has joined #openstack-trove20:05
*** kevinconway has quit IRC20:05
*** openstackgerrit has quit IRC20:06
*** pdmars_ has joined #openstack-trove20:06
*** openstackgerrit has joined #openstack-trove20:06
*** pdmars_ has quit IRC20:06
*** jcru_ has joined #openstack-trove20:06
*** pdmars_ has joined #openstack-trove20:07
*** kevinconway has joined #openstack-trove20:08
*** jcru has quit IRC20:08
*** pdmars has quit IRC20:09
*** ViswaV has joined #openstack-trove20:12
*** ViswaV_ has joined #openstack-trove20:13
*** ViswaV has quit IRC20:13
*** radez is now known as radez_g0n320:14
*** SnowDust has quit IRC20:16
*** Barker has joined #openstack-trove20:16
*** datsun180b has quit IRC20:18
*** jcooley_ has quit IRC20:22
*** jcooley_ has joined #openstack-trove20:22
mat-loweryhub_cap: Thanks. The line between what goes in datastore_versions vs. conf is blurry to me. What is wrong with modifying trove-manage call to "trove-manage datastore_version_update myds myversion mymanager myimage mypackages myactive 3306". My goal is to understand the line between table and conf since, in this case, as raw nova-based provisioning will ultimately go away (and along with it any need for secgroup columns in a table) it makes20:27
mat-lowery sense for this to be purely conf.20:27
*** jasonb365 has joined #openstack-trove20:28
hub_capsure20:28
hub_capi dont think we have a good answer to whether its in one or the other20:28
hub_capthe reason it wasnt in the db before was cuz we didnt have datastores yet :)20:28
mat-loweryhub_cap: Gotcha. Thanks.20:29
hub_capso its possible it belongs here20:30
hub_capbut we also want to make it user serviceable in the future20:30
*** datsun180b has joined #openstack-trove20:31
mat-loweryhub_cap: "user serviceable" as in the user that issues "trove create"? meaning they can specify secgroup during create?20:35
*** amcrn has joined #openstack-trove20:37
*** datsun180b has quit IRC20:40
*** jasonb365 has quit IRC20:40
*** jasonb365 has joined #openstack-trove20:42
*** SergeyLukjanov is now known as SergeyLukjanov_20:44
*** datsun180b has joined #openstack-trove20:45
*** michael-yu has quit IRC20:48
*** jcru_ has quit IRC21:01
*** pdmars has joined #openstack-trove21:01
*** jcru has joined #openstack-trove21:01
*** datsun180b_ has joined #openstack-trove21:02
*** tanisdl has quit IRC21:03
*** datsun180b has quit IRC21:03
*** pdmars_ has quit IRC21:03
*** datsun180b_ is now known as datsun180b21:03
hub_capyea im thinking cases wher internal policy dictate not running mysql on 330621:15
*** michael-yu has joined #openstack-trove21:17
*** radez_g0n3 is now known as radez21:19
*** ashestakov has quit IRC21:19
mat-loweryIs there a policy for new conf options (naming convention, default values, milestone placement)? On the flip side, when can options be modified (e.g., default value changed)?21:22
*** ViswaV_ has quit IRC21:22
*** ViswaV has joined #openstack-trove21:24
datsun180bif you're adding a new one i don't know that there are any hard rules besides you'd better put a default in cfg.py21:27
datsun180bmost of it seems to be more convention than compulsion21:28
mat-loweryThanks datsun180b.21:34
amcrna few other i can think of: don't prefix it with "trove_" unless there's a good reason (because it's redundant), if you're defaulting in cfg.py don't put the same parameter + value in the sample.conf (it's redundant), and the name shouldn't use datastore-specific terminology21:39
mat-loweryamcrn: Got it. Thanks.21:40
kevinconwayalso, if it's related to a db connection you should find a unique name for it like everyone else did...21:40
kevinconwaydb_connection, sql_connection, connect_db, db_db_db_go, are all taken21:40
mat-lowerykevinconway: lol21:41
cp16netamcrn: i think i got conf stuff fixed up finally21:50
amcrnnoiiiice.21:50
* cp16net hopes i addressed every issue21:51
cp16neti wrote all them down and slowly crossed them off my list.21:51
cp16neti'm running int tests now21:51
*** freyes has quit IRC21:51
cp16netpassed with just the subset of tests and i'm running blackbox now21:51
*** jcooley_ has quit IRC21:53
*** jcooley_ has joined #openstack-trove21:53
kevinconwaycp16net: i have an issue with config changes21:55
kevinconwayi need you to change everything21:55
amcrnseem reasonable, i agree.21:56
datsun180bcp16net: i don't have anything constructive but here's a way to replace two unrelated sections of eight lines apiece with a six-line lambda21:57
openstackgerritTim Simpson proposed a change to openstack/trove: Changing DNS to pass string to driver  https://review.openstack.org/6948021:58
kevinconwaydatsun180b: reminds me of some fun stack overflow answers related to "How do i make a multi-line lambda and assign it to a variable?"21:59
openstackgerritTim Simpson proposed a change to openstack/trove: Changing DNS to pass string to driver  https://review.openstack.org/6948022:00
cp16netkevinconway: i will kill you...22:02
*** jasonb365 has quit IRC22:03
*** rnirmal has quit IRC22:05
*** pdmars has quit IRC22:06
hub_capgrapex: ive got 1h 20m left on my battery.. that woudl put it at about 7h unplugged today22:09
grapexnice22:09
*** ViswaV has quit IRC22:18
*** ViswaV has joined #openstack-trove22:18
*** datsun180b has quit IRC22:26
esmuteHey grapex, got a question regarding https://bugs.launchpad.net/trove/+bug/127344622:33
esmutewhere do you see a function being passed? What i am seeing is a string being passed as the content. The string is the IP.22:34
esmutehttps://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L59322:34
esmutehttps://github.com/openstack/trove/blob/master/trove/dns/manager.py#L5522:34
esmutehttps://github.com/openstack/trove/blob/master/trove/instance/views.py#L2922:35
*** datsun180b has joined #openstack-trove22:37
grapexYour first line, line 59322:40
grapexip.pop is the built in list pop function22:40
grapexSo my experience is, I have to call content() to get to the string.22:40
grapexesmute:^^22:40
grapexSo, as it is today, get_ip_address returns a list22:41
grapexTh taskmanager code then passes the "pop" function of that list to the DnsManager22:41
grapexwhich in turn puts it onto the content attribute of DnsEntry22:41
grapexSo at Rax I had to modify the code to call the content field first and use the result of that call, which is a string.22:41
esmuteahh its a bug..22:41
grapexI think it would make more sense to just pass a string in the first place22:42
esmutei meant to pop the first element in the array22:42
esmuteip[0]22:42
grapexesmute: I figured. Good ole dynamic languages. :) I once spent two days because I used the old style exception syntax when I meant to capture one of two types of exceptions.22:42
esmutebut i was told that was not too pythonic :P... so i meant to do ip.pop()22:42
grapexSo looking at the code, I gave the instance a "get_dns_address" function. I figured it was too important not to have its own function.22:43
esmuteyeah. i like that... i also like the decoupling of get_ip_address() from the taskmanager22:45
esmuteThanks for looking into it grapex22:48
grapexesmute: Thank you esmute.22:48
grapexAt the risk of delaying my own pull request, I think I'm going to add an integration test and some docs as well.22:48
*** rnirmal has joined #openstack-trove22:50
*** flaper87 is now known as flaper87|afk22:50
*** igor__ has joined #openstack-trove22:56
*** kevinconway has quit IRC22:57
*** igor has quit IRC22:59
openstackgerritCraig Vyvial proposed a change to openstack/trove: adding configuration group support  https://review.openstack.org/5316823:02
openstackgerritCraig Vyvial proposed a change to openstack/trove-integration: changes to support configuration groups  https://review.openstack.org/5844523:03
openstackgerritCraig Vyvial proposed a change to openstack/python-troveclient: adds support for configurations management  https://review.openstack.org/5316923:04
cp16netamcrn: so i think there was one issue with the migrations23:04
amcrn?23:05
cp16neti noticed that 019_ failed until i removed the session stuff that modifies the database23:05
cp16neti think that might need to be changed like i did with the datastore version migration23:05
vipulcp16net: you mean the tox run failures?23:06
cp16netye23:06
vipuli seem to be getting those too.. what's the fix?23:06
vipul    session.configure_db(CONF)?23:06
cp16neti fixed it by commenting these lines out23:07
cp16nethttps://github.com/openstack/trove/blob/master/trove/db/sqlalchemy/migrate_repo/versions/019_datastore_fix.py#L33-L6323:07
cp16netlol23:07
cp16netbut i think it needs to be setup more like what i did here https://github.com/openstack/trove/blob/master/trove/db/sqlalchemy/migrate_repo/versions/017_update_datastores.py#L2323:07
cp16netbecause it should not use a session23:08
amcrnthat lets you avoid the session.configure_db, yes23:08
cp16neti think thats the problem23:08
cp16netyea amcrn and i ran into something similar when we were testing the migration paths23:08
vipulthat one line shoudl just be removed, that should fix it no?23:09
cp16netno23:09
cp16netneed to remove the DBInstance.find_all23:09
vipulah i see23:09
cp16netor any of the DBDatastore calls23:09
cp16netust call directly to the database table names23:09
cp16neti *think* that will fix it23:10
vipulYea it's a circular dependency if you use the actual model classes23:10
cp16netyeah ok23:11
cp16netthat makes sense23:11
cp16netvipul: so are you making a fix for it?23:20
*** yogesh has quit IRC23:21
cp16neti could work on it if not23:24
amcrnesp: ping23:24
cp16netbecause i think i will have to fix it for my review anyway23:24
espyo23:24
amcrnwow, like lightning23:25
esplol23:25
espI'm not always sleeping you know...23:25
cp16neti'm headed out23:25
amcrnesp: regarding https://blueprints.launchpad.net/trove/+spec/create-route-for-list-availability-zones23:25
espcp16net: peace23:25
espah, right..23:25
* esp reading23:25
espamcrn: so that bp reflects the correct status.23:26
espnot started23:26
espthe AZ is probably not important to RAX and other folks but we have 3 AZ's in our deployments23:27
amcrngotcha23:27
espthe thought was that we should allow end users to choose which of the 3 AZ's to launch their instance in.23:27
espand since you can pretty name an AZ whatever you like, the customer wouldn't know what to specify unless we told them (via documentation or something)23:28
*** plodronio has quit IRC23:38
*** esp has left #openstack-trove23:43
*** thedodd has quit IRC23:45
*** dmakogon_ has joined #openstack-trove23:46
*** denis_makogon has quit IRC23:48
*** radez is now known as radez_g0n323:53
« Sunday, 2014-01-26 Index Tuesday, 2014-01-28 »

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