Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-04-02

| Channels | #openstack-rally index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 boris-42 meteorfox: really I dislike using unit tests for performance testing
00:00 boris-42 meteorfox: this is the major reason why I started rally project=)
00:01 boris-42 meteorfox: for now it depends too much on openstack
00:01 boris-42 meteorfox: but I am working to make rally just a common tool
00:01 boris-42 meteorfox: if you need micro benchmarking write rally plugins that deals with code directly
00:02 boris-42 meteorfox: and you won't need to care about generating reports, creating load generators, doing setup, cli and other stuff
00:03 boris-42 meteorfox: am I missing something?
00:09 kun_huang working-hard man boris-42: what about the certification stuff?
00:10 boris-42 kun_huang: wanna help with that?
00:10 kun_huang boris-42: yep, wanna a little help
00:11 boris-42 kun_huang: so the idea is quite simple
00:11 boris-42 kun_huang: you should read this tutorial https://review.openstack.org/#/c/168867/
00:11 boris-42 kun_huang: about task templates
00:12 boris-42 kun_huang: so current rally input files are jinja2 templates
00:12 boris-42 kun_huang: which means that they are parametrized and flexible
00:12 boris-42 kun_huang: we can make one large task that accept few arguments and works for all clouds
00:17 kun_huang boris-42: nice, it's not implemented yet?
00:18 boris-42 kun_huang: yep even before 0.0.1
00:18 boris-42 kun_huang: that patch just adds docs
00:18 kun_huang boris-42:  good step
00:19 kun_huang boris-42:  btw, last time you said rally has big tent to become openstack project
00:19 kun_huang boris-42:  you mean ttx's new tag system?
00:19 boris-42 kun_huang: yep
00:19 boris-42 kun_huang: here is the patch https://review.openstack.org/#/c/169357/
00:19 boris-42 kun_huang: on the 7 april there will be TC meeting and we will know
00:20 kun_huang boris-42:  hah, is TC meeting open?
00:20 boris-42 kun_huang: yep
00:20 kun_huang boris-42:  will be there and say good words :))))
00:20 boris-42 kun_huang: https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee
00:21 boris-42 kun_huang: btw it will be nice if you manage to do more reviews and get back into cores faster=)
00:22 kun_huang boris-42: trying :)
00:25 dmorita joined #openstack-rally
00:41 prashantS joined #openstack-rally
01:00 yingjun joined #openstack-rally
01:04 prashantS joined #openstack-rally
01:21 logan2 joined #openstack-rally
01:35 baker joined #openstack-rally
01:45 jjmb joined #openstack-rally
02:05 panbalag joined #openstack-rally
02:25 yingjun joined #openstack-rally
02:42 prashantS joined #openstack-rally
02:53 rvcehimanshu joined #openstack-rally
02:58 baker joined #openstack-rally
03:05 boris-42 kun_huang: btw
03:05 boris-42 kun_huang: https://github.com/boris-42/profimp
03:05 kun_huang boris-42: =, = not sleeping
03:05 boris-42 kun_huang: finishing unit tests for code
03:06 boris-42 kun_huang: this is instead of that patch in Rally for tracing importans
03:06 kun_huang boris-42:  ah, watching now
03:07 boris-42 kun_huang: there is no code yet
03:07 boris-42 kun_huang: but it will be soon
03:07 boris-42 kun_huang: I spend a bit time undrestand how to setup travis and so on
03:08 boris-42 but in local env it works nice
03:11 prashantS joined #openstack-rally
03:13 kun_huang boris-42: using travis for profimp?
03:18 boris-42 kun_huang: yep
03:18 boris-42 kun_huang:  I just decided to try to work on github
03:18 boris-42 kun_huang: to compare with OpenStack CI
03:19 kun_huang boris-42:  I know this but never tried
03:19 kun_huang boris-42:  we could have a try
03:21 boris-42 kun_huang: so i set it and it works well
03:21 boris-42 kun_huang: but I like more OpenStack CI=)
03:24 pradeep joined #openstack-rally
03:30 prashantS joined #openstack-rally
03:38 yingjun_ joined #openstack-rally
03:50 prashantS joined #openstack-rally
04:00 boris-42 kun_huang:  https://github.com/boris-42/profimp/pull/4
04:00 boris-42 kun_huang: okay it should be soon merged
04:02 openstackgerrit Prabhjyot Singh Sodhi proposed stackforge/rally: Create 'add cloud info' spec file  https://review.openstack.org/169432
04:03 boris-42 kun_huang: ok now it works https://github.com/boris-42/profimp
04:04 prashantS joined #openstack-rally
04:05 nkhare joined #openstack-rally
04:06 kun_huang boris-42: https://travis-ci.org/boris-42/profimp/jobs/56830362 this job details page is slow at my office
04:06 boris-42 kun_huang: heh =)
04:06 boris-42 kun_huang: but in any case I merged that change and it works
04:08 boris-42 kun_huang: try to run
04:09 boris-42 kun_huang: profimp "import keystoneclient"
04:09 boris-42 LOL
04:10 rvcehimanshu joined #openstack-rally
04:16 psd joined #openstack-rally
04:20 rvcehimanshu joined #openstack-rally
04:27 prashantS joined #openstack-rally
04:36 aswadr joined #openstack-rally
04:41 rvcehimanshu joined #openstack-rally
04:54 yingjun joined #openstack-rally
04:58 yfried joined #openstack-rally
05:00 kiran-r joined #openstack-rally
05:12 openstackgerrit pradeep proposed stackforge/rally: Add "pause and unpause" scenario for Nova  https://review.openstack.org/169598
05:16 pradeep joined #openstack-rally
05:17 anshul joined #openstack-rally
05:19 rvcehimanshu joined #openstack-rally
05:40 neeti joined #openstack-rally
05:46 openstackgerrit pradeep proposed stackforge/rally: Add "suspend and resume" scenario for Nova  https://review.openstack.org/169150
06:01 yingjun joined #openstack-rally
06:02 rdas joined #openstack-rally
06:02 pradeep yingjun: ping
06:03 pradeep yingjun: made changes as you suggested
06:03 pradeep for both
06:03 yingjun pradeep, nice, will review
06:03 pradeep yingjun: is there any action item list or TODO list for rally?
06:04 yingjun pradeep: https://docs.google.com/spreadsheets/d/16DXpfbqvlzMFaqaXAcJsBzzpowb_XpymaK2aFY2gA2g/edit#gid=0
06:04 pradeep yingjun: what ever i take up, i see patches next day. Which is nice. Its just that, i dont want to duplicate efforts
06:04 pradeep yingjun: excellent.
06:04 yingjun pradeep, roadmap
06:04 pradeep yingjun: thanks
06:05 yingjun pradeep:-)
06:28 rvcehimanshu joined #openstack-rally
06:34 yingjun_ joined #openstack-rally
07:08 arxcruz joined #openstack-rally
07:11 nmagnezi joined #openstack-rally
08:09 openstackgerrit Sergey Skripnick proposed stackforge/rally: WIP: super new rally_gate.py  https://review.openstack.org/163785
08:16 amaretskiy joined #openstack-rally
08:18 e0ne joined #openstack-rally
08:21 andreykurilin_ joined #openstack-rally
08:22 yingjun joined #openstack-rally
08:26 yingjun_ joined #openstack-rally
08:27 andreykurilin_ joined #openstack-rally
08:36 aswadr joined #openstack-rally
08:36 hafe joined #openstack-rally
08:40 kun_huang yfried: ping
08:40 yfried kun_huang: pong
08:40 kun_huang yfried: http://logs.openstack.org/71/167471/3/check/gate-rally-dsvm-rally/98c7450/logs/devstack-gate-cleanup-host.txt
08:41 kun_huang yfried: search "cp: cannot"
08:41 kun_huang yfried: do you know this issue?
08:41 aswadr joined #openstack-rally
08:41 yfried kun_huang: nope
08:42 kun_huang yfried: okay, let me ask tempest guys
08:42 hafe using scenario VMTasks.boot_runcommand_delete I don't understand if the script should exist on the test host or in the VM
08:43 yfried hafe: the script should be in a path accessible to the python interpreter executing the test code
08:43 yfried hafe: so yes- on the rally machine
08:44 hafe but if it is not on test host it fails
08:44 yfried hafe: what's "test host"?
08:44 hafe so you mean it is copied to the server
08:44 yfried hafe: it's not even copied
08:44 yfried hafe: the test opens the file and executes in line-by-line via ssh to the vm
08:45 hafe OK will try a bit more
08:45 hafe and it must return json?
08:49 yfried hafe: only if you want the output displayed, I think
08:49 yfried hafe: I mean that it will execute whatever code you want but it expects json output to be able to parse it
08:52 _kiran_ joined #openstack-rally
08:53 andreykurilin_ joined #openstack-rally
08:54 hafe yfried: I got it to work, thx
08:58 oanufriev joined #openstack-rally
08:59 andreykurilin_ joined #openstack-rally
09:13 e0ne joined #openstack-rally
09:15 openstackgerrit Alexander Maretskiy proposed stackforge/rally: [Spec] Rally upgrades  https://review.openstack.org/167319
09:21 yfried andreykurilin: re https://review.openstack.org/#/c/168849/
09:21 yfried andreykurilin: could you please explain your comment?
09:29 openstackgerrit Merged stackforge/rally: Add "suspend and resume" scenario for Nova  https://review.openstack.org/169150
09:43 openstackgerrit Alexander Maretskiy proposed stackforge/rally: [Spec] Rally upgrades  https://review.openstack.org/167319
09:49 tosky joined #openstack-rally
09:52 andreykurilin yfried: hi, sure
09:54 andreykurilin yfried: auto_generated config file is stored as ~/.rally/tempest/for-deployment-<uuid>/tempest.conf , so copy only path to this file will not save config while re-install.  you should check location of config file and copy it to temp dir, uninstall&install tempest and than, return it back:)
09:55 psd joined #openstack-rally
10:02 rvcehimanshu joined #openstack-rally
10:09 psd joined #openstack-rally
10:10 openstackgerrit Alexander Maretskiy proposed stackforge/rally: [Spec] Rally upgrades  https://review.openstack.org/167319
10:28 olkonami joined #openstack-rally
10:31 cdent joined #openstack-rally
10:44 nkhare joined #openstack-rally
10:48 msdubov joined #openstack-rally
11:00 rook__ joined #openstack-rally
11:09 _kiran_ joined #openstack-rally
11:19 psd joined #openstack-rally
11:41 yfried|afk joined #openstack-rally
11:53 nkhare joined #openstack-rally
11:59 panbalag joined #openstack-rally
12:00 openstackgerrit Mikhail Dubov proposed stackforge/rally: Add streaming algorithms and SLA to check for outliers  https://review.openstack.org/148079
12:00 boris-42 joined #openstack-rally
12:00 pboros joined #openstack-rally
12:02 openstackgerrit Oleh Anufriiev proposed stackforge/rally: Removed task validation from api.Task.start  https://review.openstack.org/158899
12:02 boris-42 yfried|afk: ping
12:18 rook__ joined #openstack-rally
12:41 e0ne joined #openstack-rally
12:47 rvcehimanshu joined #openstack-rally
12:54 yfried|afk andreykurilin gotcha. nice catch
12:54 yfried|afk boris-42: pong
12:55 boris-42 yfried|afk: I have small question
12:55 yfried|afk boris-42: I might have a small answer .. :)
12:55 boris-42 yfried|afk: regarding OS-SERVICE-ENDPOINT vs OS_ENDPOINT
12:55 boris-42 yfried|afk: what is the different?)
12:55 yfried|afk boris-42: oh - no idea
12:55 boris-42 yfried|afk: so I believe that we need OS_ENDPOINT
12:55 pradeep joined #openstack-rally
12:55 boris-42 yfried|afk: at least users want it
12:56 yfried|afk boris-42: I could easily change that
12:56 yfried|afk boris-42: assuming you know which
12:56 boris-42 yfried|afk: could you please ?
12:56 yfried|afk boris-42: pushing now
12:56 boris-42 yfried|afk: thank you
12:56 pradeep yfried|afk: boris-42: ping
12:57 yfried|afk pradeep: ?
12:57 pradeep When there is "Patch in Merge Conflict" . I will have to rebase patch with upstream and resend with commit id?
12:57 pradeep yfried|afk: ^
12:57 yfried|afk pradeep: no need to ping, if you see as talking right now :)
12:57 pradeep yfried|afk: :)
12:57 yfried|afk panbalag: yes
12:58 yfried|afk pradeep: it means that auto rebase didn't work, and you need to manually resolve the conflicts
12:58 yfried|afk pradeep: as long as you keep the change-id (automatically generated at the end of the commit message)
12:59 panbalag yfried|afk, what is the "yes" for? I forgot what I asked :)
12:59 yfried|afk pradeep: you'd be just applying a new patchset to you patch
12:59 yfried|afk panbalag: sorry
12:59 pradeep yfried|afk:  I was sending my patchset as older commit as base.  Now i will rebase with upstream and resend
12:59 pradeep Is that what you mean?
12:59 yfried|afk panbalag: "p<TAB>" auto completes your nick instead of pradeep
13:00 panbalag yfried|afk, ok :)
13:00 yfried|afk pradeep: yes
13:00 pradeep yfried|afk: perfect. I will do that.
13:00 pradeep Thanks
13:00 yfried|afk pradeep: make sure you are always working on top of an up-to-date master branch
13:00 msdubov joined #openstack-rally
13:01 psd joined #openstack-rally
13:02 pradeep yfried|afk: sure.  2 of my different patche sets  were altering same file. One accepted. 2nd one has issue Thats the culprit :)
13:02 pradeep yfried|afk: i will make sure next time.
13:02 pradeep onwards
13:03 openstackgerrit Yair Fried proposed stackforge/rally: Adds endpoint to --fromenv deployment creation  https://review.openstack.org/166607
13:03 yfried|afk boris-42: ^
13:05 boris-42 yfried|afk: +2
13:09 boris-42 yfried|afk: did you have a chance to think about proposal regarding to dual stuff?
13:09 yfried|afk boris-42: no.
13:09 yfried|afk boris-42: sorry
13:10 yfried|afk boris-42: and I probably won't have time next week as well (holiday)
13:10 yfried|afk boris-42: I'm still in my previous state of mind
13:11 yfried|afk boris-42: that current patch and structure are fine (with the possiblity of a not-implemented exception added to novawrapper)
13:11 yfried|afk boris-42: and that any new idea you have, which so-far sounds good, shouldn't block this patch series
13:12 boris-42 yfried|afk: it is not fine
13:12 yfried|afk boris-42: but I might be too-close the issue, as nmagnezi is my colleague and I talked with him a lot about this topic
13:12 boris-42 yfried|afk: it breaks to big things
13:12 boris-42 2 big things*
13:12 yfried|afk boris-42: sorry, I don't feel the same way.
13:13 boris-42 yfried|afk: but you can't use anymore wrappers as they were designed => they become useles
13:13 yfried|afk boris-42: I suggest you bring another core member on board for this
13:13 boris-42 yfried|afk: you can't use network context for nova -> it becomes useless
13:13 boris-42 yfried|afk: this 2 things are broken and it is very bad
13:14 boris-42 yfried|afk: I can make more detailed comment in patch with my proposal how to fix things
13:14 boris-42 yfried|afk: amaretskiy ping
13:14 amaretskiy hi
13:14 boris-42 yfried|afk: amaretskiy so we have here debate=)
13:14 yfried|afk boris-42: you are correct in general. it's a fundamental problem that needs to be addressed regardless of this patch
13:14 boris-42 yfried|afk: nope without this patch -> we don't have fundamental problem
13:15 boris-42 yfried|afk: this patch introduce issue
13:15 boris-42 yfried|afk: I really dislike merging patches that introduce issues instead of finding the way to do the same without issues
13:15 nmagnezi boris-42, without this patch we don't have anything basically :-)
13:15 yfried|afk boris-42: you say introduce, I say expose :)
13:15 boris-42 yfried|afk: really we can just create 1 separated context that will work together with netowrk context
13:15 boris-42 yfried|afk: for dual api
13:16 boris-42 yfried|afk: why we need to do any changes in network context ?
13:16 boris-42 nmagnezi: ^
13:16 nmagnezi boris-42, for dual-stack?
13:16 boris-42 nmagnezi: yep
13:16 boris-42 nmagnezi: so network context will setup networks as now
13:17 boris-42 nmagnezi: after that dual-stack will do magic with IPv6 if it specified
13:17 boris-42 nmagnezi: so netowrk context will still work in the same way for nova netowrk and neutron (which is what I am asking)
13:17 nmagnezi boris-42, so first, if ipv6 was not specified, the existing behavior remains intact
13:17 boris-42 nmagnezi: and you will get support of IPv6 where you need it
13:18 boris-42 nmagnezi: yep but I don't know why did you put that code in netowrk context and not make separated one?
13:18 nmagnezi boris-42, let me explain
13:18 boris-42 nmagnezi: ok
13:18 amaretskiy boris-42, yfried|afk: regarding OS_ENDPOINT vs. OS_SERVICE_ENDPOINT - do *not* use OS_SERVICE_ENDPOINT!!!
13:18 amaretskiy because in this case keystone cli will work in other way
13:18 nmagnezi boris-42, the current state is: you have a network setup that creates an IPv4 network + router for you, right?
13:19 boris-42 amaretskiy: lol
13:19 boris-42 amaretskiy: ok=)
13:19 amaretskiy it will start asking for OS_SERVICE_TOKEN
13:19 boris-42 amaretskiy: btw how?
13:19 boris-42 amaretskiy: ah
13:19 boris-42 nmagnezi: yep
13:19 nmagnezi boris-42, now.. if you want your tests to involve IPv6.. there 2 two approaches:
13:20 nmagnezi boris-42, 1. if you have an IPv6 Infra (hence, your tester has IPv6 address) you may use IPv6 in a similar way you use IPv4 plus some changes related to the neutron subnet creation (additional parameters)
13:21 nmagnezi boris-42, 2. if you do not have such infra, you will have to communicate with your instance in IPv4 and that instance will have additional IPv6 NIC (hence, dualstack)
13:21 boris-42 nmagnezi: okay but probably you didn't get what I said
13:22 nmagnezi wait :-)
13:22 boris-42 nmagnezi: ok
13:22 nmagnezi boris-42, now, i examined the NeutronWrapper and there are some assumptions and couplings there that made it not so simple todo.
13:22 nmagnezi boris-42, first the create_network function also creates a neutron router
13:22 nmagnezi boris-42, meaning it's not that east to have two neutron networks with the same router
13:23 nmagnezi boris-42, so i had to modify that..
13:23 nmagnezi boris-42, second.. if you want to enable that dual-stack mode, the network context does not know how to get IPv6 related details such as ra_mode and address_mode
13:24 nmagnezi boris-42, not to mention setting up both IPv4 and IPv6 nets with the same router
13:24 yfried|afk nmagnezi: ^ don't forget clean up and ports ... :)
13:24 nmagnezi boris-42, plus the cleanup of that context is not built for deleting routers that were not created as part of a neutron network creation
13:24 boris-42 nmagnezi: yfried|afk guys
13:24 boris-42 one more time
13:24 yfried|afk nmagnezi: and the coupling magic hihi
13:25 boris-42 WHY all these changes can't be done in separated context?
13:25 boris-42 like network context will do stuff as usually
13:25 yfried|afk boris-42: you mean replace network context with a different context?
13:25 nmagnezi boris-42, you mean create another context name  "dual-stack-networks" ?
13:25 boris-42 yfried|afk: nope
13:25 yfried|afk boris-42: or add a new context?
13:25 boris-42 USING 2 context
13:25 boris-42 network_context after that dual_stack_context
13:26 boris-42 so netowrk_context does part of owrk dual_stack finish work
13:26 boris-42 and you are specifing both of them
13:26 boris-42 in input task
13:26 nmagnezi boris-42, what's the motivation for that? if thet both handle networking why have 2 contexts?
13:26 jjmb joined #openstack-rally
13:27 boris-42 nmagnezi: one more time I don't want in code that works for nova network and neutron part of code that doesn't work for nova netowrk
13:27 boris-42 nmagnezi: the second context will handle just IPv6 setup
13:27 boris-42 that works only for neutron
13:27 boris-42 and I am perfectly ok to have context that works only for Neutron and can be used with existing context that works for both
13:28 nmagnezi boris-42, so how will you handle non IPv6 related stuff that neutron supports and nova network does not? such as Loadbalancer as a service for example
13:28 boris-42 nmagnezi: I don't want to handel that in NetwrokContext
13:28 boris-42 nmagnezi: it will be handled in other context classes that works together with network context
13:29 boris-42 nmagnezi: take a look how quotas/roles are done
13:29 boris-42 nmagnezi: they are not in users context, they are in separated context
13:29 boris-42 nmagnezi: that is the idea of context, to not be fat and do just small parts of works that can be reused
13:30 boris-42 nmagnezi: so LoadBalancer stuff will be separated context
13:30 boris-42 nmagnezi: that knows how to work with network context and maybe dual stack
13:31 nmagnezi boris-42, so if i understand you correctly, your guild-line for context (and in our case, network context) is to have the very basic common ground that is supported for a specific subject? in our case, the common ground between nova-network and neutron
13:31 nmagnezi right?
13:32 e0ne joined #openstack-rally
13:32 openstackgerrit Mikhail Dubov proposed stackforge/rally: Add streaming algorithms and SLA to check for outliers  https://review.openstack.org/148079
13:32 boris-42 nmagnezi: yep close to true. What I am talking about is to split every action like load balanicing, dual stack, other stuff in separated context (that requires netowrk context0
13:32 boris-42 nmagnezi: in network context keep stuff that can work in both neutron and nova network and can be reused
13:33 boris-42 nmagnezi: so we won't have any design issues
13:33 jjmb1 joined #openstack-rally
13:34 boris-42 nmagnezi: so in your case in rally input task you will specify something like context: {"users": ...., "network": ...., "dual_stack"}
13:34 boris-42 so users will create tenants/users after that network will create networks routers and rest of stuff dual_stack will add IPv6
13:35 boris-42 nmagnezi: and network will still work for both nova network and neutron, but dual_stack will be neutron specific
13:35 nmagnezi i understand your point
13:35 nmagnezi boris-42, regardless of that
13:36 nmagnezi boris-42, I really think there are some function from that code that will be useful to NeutronWrapper as improvements
13:36 nmagnezi boris-42, not talking about the code in the context
13:37 nmagnezi boris-42, stuff like "list_ports" "delete_router" etc
13:37 boris-42 nmagnezi: so are you agree with splitting functionality to 2 context?
13:37 boris-42 do you agree*
13:37 jjmb joined #openstack-rally
13:37 aswadr joined #openstack-rally
13:38 boris-42 nmagnezi: because NeutronWrapper is separated topic
13:38 nmagnezi boris-42, well.. while I understand your point i must admit that I think it can work well in the existing manner..
13:39 baker joined #openstack-rally
13:39 nmagnezi boris-42, neutron and nova-network are different. maybe the idea of the unified wrapper for both of them is problematic to begin with
13:40 boris-42 nmagnezi: so but for now it works well in case of IPv4
13:40 nmagnezi boris-42, but this is the situation at the moment. so if this is the guidline
13:40 nmagnezi this is the guildline.. :)
13:40 nmagnezi boris-42, yes. but not for all features.. (LBaaS etc)
13:41 boris-42 nmagnezi: but it shouldn't do anything related to LBaaS
13:41 boris-42 nmagnezi: as I said it is not Rally apporach to put everything in single context
13:41 nmagnezi boris-42, just pointed an example for a feature which is supported only in Neutron :)
13:41 boris-42 nmagnezi: we need to keep context classes as simple as possible and to make the do only 1 thing
13:41 nmagnezi same as IPv6
13:42 nmagnezi boris-42, yup.. i got it
13:42 boris-42 nmagnezi: why you want to put it in netowrk context?) that is unclear for me=)
13:42 nmagnezi boris-42, because it worked :D
13:42 boris-42 nmagnezi: and why it won't work in separated context?
13:43 boris-42 nmagnezi: can I make a patch set on top of your work
13:43 nmagnezi i never said it cannot
13:43 boris-42 nmagnezi: just to explain via code what I mean
13:43 boris-42 nmagnezi: ok so splitting into a lot of pieces this stuff seems for me reasonable
13:44 boris-42 nmagnezi: at least neutron specific stuff shouldn't be together with code that should work with nova network
13:44 boris-42 nmagnezi: so back to question regarding to NeutronWrapper
13:44 boris-42 nmagnezi: the idea of wrappers is to get instance of one of wrappers and work with it
13:45 boris-42 so you don't care are you working with nova network or neutron, in your case you are introducing things that works only for neutron
13:45 boris-42 and interface of NeutronWrapper and NovaWrapper are different
13:46 boris-42 if you would like to interoduce in NeutronWrapper such methods (for some reason) I would at least recomend to make them "_method" have _
13:46 boris-42 so like they are not interface
13:47 nmagnezi boris-42, and what would happen if we have things which are not implemented in nova-network? for example routers
13:47 nmagnezi boris-42, you cannot "delete-router" in nova-network
13:48 boris-42 nmagnezi: the idea of wrapper is to hide common things between nova network and neutron
13:48 boris-42 nmagnezi: if thing is not common it is better not to keep it in wrapper
13:49 nmagnezi boris-42, i understand.. but for that reason it lacks many things :(
13:49 boris-42 nmagnezi: because what is the purpose of wrapper in such case?
13:49 boris-42 nmagnezi: I would prefer to make some module called utils
13:49 boris-42 nmagnezi: and put functionallity that you need
13:49 boris-42 nmagnezi: and reuse part of this functionallity in NeutronWrapper
13:51 boris-42 nmagnezi: if you would like I can propose a change based on your patch
13:51 boris-42 nmagnezi: to show what I mean
13:56 dpaterson joined #openstack-rally
14:06 mwagner_lap joined #openstack-rally
14:12 exploreshaifali joined #openstack-rally
14:18 aix joined #openstack-rally
14:19 rook__ joined #openstack-rally
14:20 aix joined #openstack-rally
14:25 psd joined #openstack-rally
14:28 openstackgerrit pradeep proposed stackforge/rally: Add "pause and unpause" scenario for Nova  https://review.openstack.org/169598
14:30 pradeep yfried|afk:  Thanks rebased my patch after merge conflict. kool
14:37 yfried pradeep: pong
15:02 psd joined #openstack-rally
15:10 openstackgerrit Vadim Rovachev proposed stackforge/rally: Fix validation for rps parameter  https://review.openstack.org/170160
15:16 jaypipes joined #openstack-rally
15:16 openstackgerrit Roman Vasilets proposed stackforge/rally: Adding Multiple Request Scenario  https://review.openstack.org/117705
15:20 rook__ joined #openstack-rally
15:35 openstackgerrit Merged stackforge/rally: Adds endpoint to --fromenv deployment creation  https://review.openstack.org/166607
15:37 prashantS joined #openstack-rally
15:41 prashantS joined #openstack-rally
15:48 openstackgerrit Vadim Rovachev proposed stackforge/rally: Fix validation for rps parameter  https://review.openstack.org/170160
16:21 rook__ joined #openstack-rally
16:21 aswadr joined #openstack-rally
16:22 yfried|afk joined #openstack-rally
16:36 openstackgerrit pradeep proposed stackforge/rally: Add "pause and unpause" scenario for Nova  https://review.openstack.org/169598
16:39 rvcehimanshu joined #openstack-rally
16:45 anshul joined #openstack-rally
16:52 psd joined #openstack-rally
17:07 pradeep1 joined #openstack-rally
17:22 rook__ joined #openstack-rally
17:35 andreykurilin_ joined #openstack-rally
17:49 davideagnello hello, we are looking to use Rally as a test framework in stack forge project Cue for our integration and scenario testing on devstack and other OpenStack installations.  Which projects in DevStack currently use the Rally test framework?
17:49 andreykurilin_ joined #openstack-rally
17:53 boris-42 davideagnello: hi there
17:54 davideagnello hello
17:54 boris-42 davideagnello: cinder/glance/keystone/neutron
17:54 boris-42 davideagnello: zaqar/designate
17:54 boris-42 davideagnello: ceilometer
17:54 boris-42 =
17:54 boris-42 =)
17:54 davideagnello great, thank you :)
17:57 tosky joined #openstack-rally
18:00 vipul davideagnello: https://github.com/stackforge/rally/tree/master/rally/benchmark/scenarios
18:00 vipul boris-42: pointed me at that a few days ago.. a good place to start to look into the different tests being run
18:01 davideagnello vipul: cool, thanks
18:02 boris-42 davideagnello: vipul https://rally.readthedocs.org/en/latest/gates.html
18:02 boris-42 davideagnello: vipul this may help as well
18:03 nmagnezi joined #openstack-rally
18:03 davideagnello boris-42: this is very helpful, thanks
18:14 e0ne joined #openstack-rally
18:15 davideag_ joined #openstack-rally
18:23 rook__ joined #openstack-rally
18:40 psd_ joined #openstack-rally
18:43 baker joined #openstack-rally
19:09 cdent joined #openstack-rally
19:29 baker joined #openstack-rally
19:45 openstackgerrit Sergey Skripnick proposed stackforge/rally: WIP: super new rally_gate.py  https://review.openstack.org/163785
20:05 baker_ joined #openstack-rally
20:09 exploreshaifali joined #openstack-rally
20:12 panbalag joined #openstack-rally
20:25 rook__ joined #openstack-rally
20:31 rook__ joined #openstack-rally
20:45 baker joined #openstack-rally
20:51 psd joined #openstack-rally
21:16 panbalag joined #openstack-rally
21:18 davideagnello joined #openstack-rally
21:32 davideagnello joined #openstack-rally
21:40 vaidy_ril joined #openstack-rally
21:54 boris-42 davideagnello: vipul btw
21:54 boris-42 https://github.com/boris-42/profimp
21:54 boris-42 ^ tool to improve your imports
22:18 andreykurilin_ joined #openstack-rally
22:45 prashantS joined #openstack-rally
22:46 tosky left #openstack-rally
23:16 mwagner_lap joined #openstack-rally
23:32 davideagnello boris-42: when deploying the Rally test framework, it looks like it also brings in scenario tests for other OpenStack services with it like Neutron, cinder, etc.. is there a way to install just the framework without other tests and clients?
23:34 boris-42 davideagnello: nope
23:34 boris-42 davideagnello: why do you need that?)
23:37 davideagnello we would be writing integration/scenario/stress tests that target only our service (Cue) without testing other services (Cinder, Neutron, etc..)
23:39 davideagnello boris-42: that would mean every time we are deploying Rally to test Cue, we would be pulling in all other tests with it?
23:50 rook joined #openstack-rally
23:53 boris-42 davideagnello: I really don't no what is the issue...
23:53 dmorita joined #openstack-rally
23:53 boris-42 davideagnello: okay you have installed plugins and ?
23:53 boris-42 davideagnello: they own't be run
23:53 boris-42 davideagnello: and it doesn't take too much time to install all of them and all requriments

| Channels | #openstack-rally index | Today | | Search | Google Search | Plain-Text | summary