Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-01-12

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

All times shown according to UTC.

Time Nick Message
00:09 jjmb joined #openstack-rally
00:13 yingjun joined #openstack-rally
00:17 jjmb1 joined #openstack-rally
00:25 anaribeiro joined #openstack-rally
00:30 jaypipes joined #openstack-rally
02:01 jjmb joined #openstack-rally
02:08 jjmb1 joined #openstack-rally
02:34 yingjun joined #openstack-rally
02:36 psd joined #openstack-rally
03:35 yingjun_ joined #openstack-rally
04:16 nkhare joined #openstack-rally
05:27 rdas joined #openstack-rally
06:00 tfreger joined #openstack-rally
06:19 openstackgerrit Li Yingjun proposed stackforge/rally: Move files to common lib(Part 4)  https://review.openstack.org/145650
06:22 rdas joined #openstack-rally
06:22 nkhare joined #openstack-rally
06:44 yfried joined #openstack-rally
06:48 yfried_ joined #openstack-rally
07:05 yfried_ boris-42: good morning. back from vacation?
07:07 neeti joined #openstack-rally
07:09 openstackgerrit Yair Fried proposed stackforge/rally: (WIP) Adds timestamps to scenarios iterations  https://review.openstack.org/146385
07:10 yingjun joined #openstack-rally
07:10 yfried_ boris-42: we've discussed this last week https://review.openstack.org/#/c/146385/1
07:11 yfried_ boris-42: could you please help me with json schema validator?
07:29 coolsvap joined #openstack-rally
07:32 prmtl joined #openstack-rally
07:36 yingjun joined #openstack-rally
07:39 msdubov_ joined #openstack-rally
07:42 yfried_ amaretskiy: andreykurilin1: I'm looking for some help re schema validation. can any of you help me?
07:43 msdubov_ yfried_: Hi! What's your problem?
07:44 yfried_ msdubov_: https://review.openstack.org/#/c/146385/1
07:45 yfried_ msdubov_: I'm replacing the scenario output from just "duration" to a dict with {"duration", "timestamp"}
07:45 yfried_ msdubov_: I meant atomic_actions output
07:46 yfried_ msdubov_: https://github.com/stackforge/rally/blob/master/rally/benchmark/runners/base.py#L123:L127
07:47 yfried_ I need to edit ^
07:47 yfried_ msdubov_: to accept {'duration': 3.7200820446014404, 'timestamp': "ts string"}
07:51 msdubov_ yfried_:I think you need something like http://paste.openstack.org/show/156407/
07:51 msdubov_ yfried_:By the way are you sure you need those "date-time" strings for timestamps?
07:51 msdubov_ yfried_: We can store just an integer timestamp
07:52 msdubov_ yfried_: (But perhaps it will be a bit harder then to use this information while generating HTML reports)
08:05 yfried_ msdubov_: we can discuss the format later.
08:08 msdubov_ yfried_: ok
08:09 yfried_ msdubov_: tnx
08:12 fhubik joined #openstack-rally
08:14 arxcruz joined #openstack-rally
08:22 yingjun joined #openstack-rally
08:23 nmagnezi joined #openstack-rally
08:26 msdubov_ joined #openstack-rally
08:59 andreykurilin joined #openstack-rally
09:00 fhubik joined #openstack-rally
09:01 oanufriev joined #openstack-rally
09:19 yingjun joined #openstack-rally
09:23 openstackgerrit Yair Fried proposed stackforge/rally: (WIP) Adds timestamps to scenarios iterations  https://review.openstack.org/146385
09:23 yfried_ msdubov_: now I'm stuck with updating the unittests
09:27 openstackgerrit Nikita Konovalov proposed stackforge/rally: [Sahara] Add Cloudera plugin support  https://review.openstack.org/146400
09:27 openstackgerrit Nikita Konovalov proposed stackforge/rally: Neutron network context support  https://review.openstack.org/143671
09:27 openstackgerrit Nikita Konovalov proposed stackforge/rally: [Sahara] Add Cloudera plugin support  https://review.openstack.org/146400
09:29 andreykurilin joined #openstack-rally
09:45 openstackgerrit Merged stackforge/rally: Add list_volumes scenario for Cinder test  https://review.openstack.org/142500
09:46 msdubov_ yfried_: Well basically it's usually just all about editing sample configs so that they pass the schema validation. Or do you have some more complicated stuff you are stuck with?
10:20 boris-42 yfried_: sorry I was sleeping)
10:20 boris-42 yfried_: I'll take a look
10:21 yfried_ boris-42: msdubov_: I'm stuck with the unittest. don't know how to get mock to generate the TS the same way it does "duration"
10:22 msdubov_ yfried_: Can you please provide us with a link to that test?
10:27 yingjun joined #openstack-rally
10:39 arxcruz andreykurilin1, hey :)
10:56 andreykurilin joined #openstack-rally
10:59 pcaruana joined #openstack-rally
11:03 yfried_ msdubov_: http://logs.openstack.org/85/146385/2/check/gate-rally-python27/2e2943f/console.html
11:20 prmtl joined #openstack-rally
11:35 yfried_ msdubov_: ^ is the jenkins log enough?
11:36 msdubov_ yfried_: Yep, I just wanted to see which test fails, will look into it a bit later
11:36 boris-42 yfried_: it's better to give this http://logs.openstack.org/85/146385/2/check/gate-rally-python27/2e2943f/testr_results.html.gz
11:36 boris-42 yfried_: cause it's pretty HTML page=)
11:40 rdas joined #openstack-rally
11:43 yfried_ boris-42: tnx. didn't know that (I'm used to tempest)
11:43 yfried_ boris-42: any pointers on the mocks?
11:43 boris-42 yfried_: looking now on your patch
11:43 yfried_ boris-42: tnx
11:45 boris-42 yfried_:  one more time what patch?
11:46 boris-42 https://review.openstack.org/#/c/146385/
11:46 boris-42 ah this one?
11:47 boris-42 yfried_: why do you touch atomic actions?
11:49 yfried_ boris-42: I'm adding timestamp to each atomic action, so now we can know when it happened and not only how long
11:49 boris-42 yfried_: okay
11:49 boris-42 maybe it make sense
11:53 cdent joined #openstack-rally
11:54 andreykurilin1 arxcruz: hi!
11:54 arxcruz andreykurilin1, hey :)_
11:55 arxcruz andreykurilin1, did you saw my comment on https://review.openstack.org/#/c/145458/ ?
11:55 andreykurilin1 arxcruz: one moment
11:59 boris-42 yfried_: remove those changes
11:59 boris-42 yfried_: https://review.openstack.org/#/c/146385/2/rally/benchmark/processing/utils.py
11:59 boris-42 yfried_: it's related to rally task report command
11:59 boris-42 yfried_: you'll need to change our mako templates if you add this
12:01 yfried_ boris-42: well, following our discussion, we do need to add timestamp to atomic action
12:01 boris-42 yfried_: we need but don't touch processing code
12:02 boris-42 yfried_: you can't touch it without changing rally task report templates
12:02 boris-42 yfried_: and it will be quite huge change
12:02 yfried_ boris-42: then how would you suggest we do that>
12:02 yfried_ ?
12:02 boris-42 yfried_: I mean step by step approach insted all in one patch
12:03 yfried_ boris-42: so only iterations for now?
12:03 boris-42 yfried_: if you are touching processing -> you must fix templates
12:03 yfried_ boris-42: ok
12:03 boris-42 yfried_: chose what you would like
12:03 boris-42 fixing templates and iterations times in one patch
12:03 yfried_ boris-42: the step-by-step makes sense
12:03 boris-42 or 1 by 1
12:03 boris-42 you can't do 1.5 step in one patch=)
12:04 boris-42 so I see something like
12:04 boris-42 1) collect timestampts
12:05 boris-42 2) add them to rally task detailed and rally task results
12:05 boris-42 3) add them to rally task report
12:07 andreykurilin1 arxcruz: ok. i'm confused. in the one hand, you are right and I can't imagine where "items = dict1.items() + dict2.items()" can be useful in Rally, so we can check "dict(dict1.items() + dict2.items())" constuction:) but in other hand, we have not objects with "items" attribute and since "dict1.items() + dict2.items()" is incompatible issue, it would be better to check exactly it.
12:07 arxcruz andreykurilin1, okay, I'll fix the regex then :)
12:08 andreykurilin1 arxcruz: thanks
12:08 arxcruz np
12:12 redixin joined #openstack-rally
12:16 andreykurilin1 boris-42: https://review.openstack.org/#/c/137716/11/rally/benchmark/runners/base.py . It seems to me, we should not ignore E126 rule( https://github.com/stackforge/rally/blob/master/tox.ini#L41 ), so such ugly code would prevented by pep8:)
12:22 msdubov_ joined #openstack-rally
12:38 openstackgerrit Anastasia Kuznetsova proposed stackforge/rally: Add Mistral benchmark  https://review.openstack.org/144320
12:42 openstackgerrit Yair Fried proposed stackforge/rally: (WIP) Adds timestamps to scenarios iterations  https://review.openstack.org/146385
12:47 openstackgerrit Merged stackforge/rally: Move files to common lib(Part 4)  https://review.openstack.org/145650
12:52 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
12:52 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
12:53 nkhare joined #openstack-rally
12:58 tfreger joined #openstack-rally
13:00 openstackgerrit Yair Fried proposed stackforge/rally: (WIP) Adds timestamps to scenarios iterations  https://review.openstack.org/146385
13:00 yfried_ boris-42: https://review.openstack.org/#/c/146385/
13:00 yfried_ how is that?
13:01 yfried_ though I don't see where we log the added data
13:01 pboros joined #openstack-rally
13:16 openstackgerrit Yair Fried proposed stackforge/rally: (WIP) Adds timestamps to scenarios iterations  https://review.openstack.org/146385
13:21 boris-42 yfried_: ? not sure
13:21 boris-42 yfried_: what?)
13:22 yfried_ boris-42: I'm saying that I've added the timestamps to the scenario iterations but they aren't logged or displayed in the html files. I wonder how to do that
13:22 nkhare joined #openstack-rally
13:26 fhubik_afk joined #openstack-rally
13:27 mwagner_lap joined #openstack-rally
13:27 fhubik_afk joined #openstack-rally
13:27 boris-42 yfried_: you should make another patch
13:27 boris-42 yfried_: to display them
13:28 yfried_ boris-42: yeah. I guess. could you point me to the relevant modules?
13:30 exploreshaifali joined #openstack-rally
13:32 boris-42 yfried_: one of them you already found
13:32 boris-42 yfried_: it's processing
13:32 boris-42 yfried_: this is template for HTML report https://github.com/stackforge/rally/blob/master/rally/ui/templates/task/report.mako
13:34 boris-42 yfried_: thsese lines generate HTML (passing args to that template)
13:34 boris-42 yfried_: https://github.com/stackforge/rally/blob/master/rally/benchmark/processing/plot.py#L308-L311
13:34 boris-42 yfried_: so you need to change process_reults
13:34 boris-42 yfried_: https://github.com/stackforge/rally/blob/master/rally/benchmark/processing/plot.py#L258
13:34 boris-42 to add this data
13:34 boris-42 and template
13:45 wtakase joined #openstack-rally
13:46 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
13:46 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
13:48 boris-42 arxcruz: ping
13:48 arxcruz boris-42, pong
13:48 boris-42 arxcruz: how are you?
13:48 arxcruz I'm doing great as usual :) what about you? 100% or still in vacations ?
13:49 boris-42 arxcruz: 200%)
13:49 arxcruz :D
13:49 boris-42 arxcruz: are you working on functional framwork?
13:49 arxcruz boris-42, I'll start this week
13:49 boris-42 great
13:51 wtakase boris-42: Hi!
13:51 boris-42 wtakase: oh HI=)
13:51 boris-42 wtakase: glad to see you in IRC=)
13:53 wtakase boris-42: I'd like to know current status about support-existing-users.
13:54 tfreger joined #openstack-rally
13:56 boris-42 wtakase: i have hard porn with it
13:57 tfreger joined #openstack-rally
13:58 wtakase boris-42: ?? what do you mean?
13:59 boris-42 wtakase: I am working on it as hard as I can
14:02 wtakase boris-42: I see. Thanks!! I am really interested in the feature. If you need any help, I can help you!!
14:03 boris-42 wtakase: unfortunatelly I don't know how to split this task to work in parallel on it
14:05 wtakase boris-42: OK, I am waiting for your patch.
14:09 boris-42 wtakase: sure
14:13 boris-42 wtakase: btw do you have some CI in Cern?
14:13 boris-42 wtakase: for OpenStack testing
14:14 boris-42 wtakase: Rally team is looking for ThridPart testing for other OpenStack distrubutions
14:15 openstackgerrit Anastasia Kuznetsova proposed stackforge/rally: Add Mistral benchmark  https://review.openstack.org/144320
14:17 fhubik Hi, I was just wondering, are there some basic unfinished tasks suitable for Rally-beginners? I found some of them on Trello, but they seems to be already finished and merged. Thanks.
14:17 boris-42 fhubik: we removed trello
14:18 boris-42 fhubik: it was not suitable...
14:18 boris-42 fhubik: https://bugs.launchpad.net/rally
14:18 fhubik boris-42: ok ;)
14:18 boris-42 fhubik: this seems very good
14:19 boris-42 fhubik: https://bugs.launchpad.net/rally/+bug/1409213
14:19 boris-42 fhubik: for the first patch
14:19 boris-42 fhubik: you can assign it to you and start working on it
14:19 wtakase boris-42: No, we used own python scripts which are simple for service functional testing.
14:20 wtakase boris-42: And we plan to use Rally instead of them.
14:20 fhubik boris-42: thanks, that's just what I was looking for
14:20 boris-42 wtakase: I mean how do you deploy OpenStack?
14:21 boris-42 wtakase: if you have some automation, we can make CI for Rally (e.g. somebody push patch, job is run against your deployment)
14:22 boris-42 wtakase: so patches in rally won't break rally for your OpenStack
14:23 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
14:23 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the context benchmark_image  https://review.openstack.org/138466
14:23 paboldin tadam!
14:23 boris-42 paboldin: lol
14:24 boris-42 paboldin: those patches always scares me
14:24 boris-42 =)
14:24 paboldin :D
14:24 openstackgerrit Alexander Maretskiy proposed stackforge/rally: Improvements for scenario VMTasks.boot_runcommand_delete  https://review.openstack.org/144337
14:25 wtakase boris-42: sorry, I have a meeting so I'll response later.
14:46 gtt116 joined #openstack-rally
14:47 anaribeiro joined #openstack-rally
14:48 openstackgerrit Nobuto MURATA proposed stackforge/rally: fix devstack URL  https://review.openstack.org/146480
14:57 openstackgerrit Yair Fried proposed stackforge/rally: Adds timestamps to scenarios iterations  https://review.openstack.org/146385
15:02 yfried_ boris-42: I wonder where and how should we display iteration timestamp data
15:02 yfried_ boris-42: http://logs.openstack.org/85/146385/5/check/gate-rally-dsvm-rally/e4b81e4/rally-plot/results.html.gz#/Authenticate.keystone
15:04 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
15:04 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
15:05 yfried_ boris-42: It doesn't add anything unless we have parallel runs
15:12 boris-42 yfried_: why so we can have special graphs
15:12 boris-42 yfried_: with such info
15:13 boris-42 yfried_: instead of iteration number use this time somehow
15:13 yfried_ boris-42: oh, good idea
15:13 boris-42 yfried_: and have some switcher on top
15:13 yfried_ boris-42: or - we can somehow toggle between TS and #
15:13 boris-42 yfried_: or we can draw one line (that shows amount of parallel running threads)
15:14 yfried_ boris-42: or is that what you meant by switcher?
15:14 boris-42 yfried_: have on page switching mode
15:14 boris-42 between # and TS graphs
15:14 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
15:14 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the context benchmark_image  https://review.openstack.org/138466
15:14 openstackgerrit Alexander Maretskiy proposed stackforge/rally: Optional parameter "endpoint" is added to deployment configuration  https://review.openstack.org/146490
15:15 yfried_ boris-42: so, I've really no idea how to do that. I'll dive into it and see what's what, but I've never messed with html and or xml displays so don't hold your breath waiting for this
15:15 yfried_ boris-42: on the other hand, we can merge the first patch since it's already passing gate
15:15 boris-42 yfried_: yep
15:16 boris-42 yfried_: amaretskiy can work on this
15:16 boris-42 yfried_: after he finish network patches
15:16 boris-42 yfried_: for now we can just add extra info
15:16 boris-42 and use it a bit later
15:17 openstackgerrit Alexander Maretskiy proposed stackforge/rally: Optional parameter "endpoint" is added to deployment configuration  https://review.openstack.org/146490
15:21 msdubov_ joined #openstack-rally
15:25 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
15:32 msdubov_ joined #openstack-rally
15:35 msdubov_ joined #openstack-rally
15:35 boris-42 jlk: around?
15:37 msdubov_ joined #openstack-rally
15:38 msdubov_ joined #openstack-rally
15:44 msdubov_ joined #openstack-rally
15:46 openstackgerrit Joe Talerico proposed stackforge/rally: Created new single vm workload sample script using linpack  https://review.openstack.org/115439
15:48 msdubov_ joined #openstack-rally
15:51 wtakase joined #openstack-rally
15:52 wtakase boris-42: We don't use any CI tools.
15:52 wtakase boris-42: We build RPMs based on RDO packages and deploy them using Puppet.
15:53 wtakase boris-42: We have two environments QA and production.
15:53 wtakase boris-42: We plan to run Rally against both environments and compare the result for checking status.
16:03 Soumit joined #openstack-rally
16:24 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
16:24 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
16:35 fhubik boris-42: sorry I am bothering again, but I cannot repreduce that ticket you suggested (https://bugs.launchpad.net/rally/+bug/1409213), cau you direct me how to reproduce?
16:39 paboldin review pls: https://review.openstack.org/#/c/138466/
16:48 msdubov_ joined #openstack-rally
16:49 boris-42 fhubik: run "tox"
16:52 coolsvap joined #openstack-rally
16:52 fhubik boris-42: unfortunately, no such error there
16:57 boris-42 fhubik: so they disappeared =(
16:58 boris-42 fhubik: in any case bug is valid
16:58 boris-42 fhubik: we shouldn't import oslo.foo we should import oslo_foo
16:58 boris-42 fhubik: for all oslo libs
16:59 boris-42 fhubik: according to this BP https://blueprints.launchpad.net/oslo-incubator/+spec/drop-namespace-packages
17:01 openstackgerrit Alexander Maretskiy proposed stackforge/rally: Optional parameter "endpoint" is added to deployment configuration  https://review.openstack.org/146490
17:02 fhubik boris-42: ok, so should I repair this one and review or wait for someone's more skilled extensive refactoring?
17:02 boris-42 fhubik: it's imho simple task
17:03 boris-42 fhubik: just find all occurences of "oslo.<some_lib>" and change them to "oslo_<some_lib>"
17:03 boris-42 fhubik: so it's still simple task
17:03 fhubik boris-42: Ok, Il'l try, thanks
17:12 boris-42 fhubik: you can fix all oslo.config
17:12 boris-42 fhubik: as a first patch
17:12 fhubik boris-42: and oslo.utils as second?
17:13 openstackgerrit Pavel Boldin proposed stackforge/rally: Add the generic script-based benchmark for the VMs  https://review.openstack.org/141671
17:13 fhubik boris-42: and so on?
17:17 boris-42 fhubik: yep
17:21 openstackgerrit Yuriy Nesenenko proposed stackforge/rally: Added list_snapshots scenario for Cinder test  https://review.openstack.org/146580
17:34 andreykurilin joined #openstack-rally
18:02 jlk boris-42: I am now. Just doing Monday stand up
18:03 leeantho joined #openstack-rally
18:04 leeantho_ joined #openstack-rally
18:09 boris-42 jlk: this is fix for MOS https://review.openstack.org/#/c/146490/
18:09 boris-42 jlk: after your patch=)
18:09 boris-42 jlk: if we pass to pulbic enpoint with with 35357 port everything work
18:10 boris-42 jlk: I really don't know wtf is that but it works=)
18:11 jlk hah, okay. That might make Rackspace clouds work too.
18:13 jlk boris-42: if end point is set, seems like keystone will always use that end point, and would never use auth_url. Interesting.
18:14 jlk Did they try just setting auth_url to what they are setting endpoint to?
18:24 boris-42 jlk: but they should already work? no?
18:24 boris-42 jlk: I mean rack space after we merge your patch?
18:33 jlk boris-42: Rackspace does the same thing Fuel did. A wholly separate URL for admin interface where admin users can auth
18:34 jlk I don't think the service catalog would work, because the admin user wouldn't be able to auth at the normal user auth_url
18:37 boris-42 jlk: as far as I know endpoint is used instead of managment url
18:37 boris-42 jlk: when admin wants to go to it*
18:37 boris-42 it's some kind of dark magic
18:38 jlk from my memory of looking at the keystone code, if endpoint was passed in, keystone would blindly use it. So in those scenarios, I don't see that auth_url is used at all.
18:38 jlk so I'd be curious if they would test by setting auth_url to their endpoint url and see what happens.
18:38 boris-42 jlk: https://github.com/openstack/python-keystoneclient/blob/master/keystoneclient/httpclient.py#L610-L611
18:38 boris-42 jlk: take a look here^
18:39 jlk yes
18:39 jlk management_url comes from catealog
18:39 boris-42 jlk: so only one place
18:39 jlk catalog
18:39 boris-42 jlk: where it is used
18:39 jlk oh right
18:39 jlk I see
18:39 jlk _endpoint is only grabbed when requesting the management url
18:39 boris-42 jlk: yep
18:39 jlk huh, yeah, this is a better fix than mine of just ripping it out
18:39 boris-42 jlk: so very very very dark magic=)
18:40 boris-42 jlk: your is good as well
18:40 boris-42 jlk: both makes sense=)
18:40 jlk might want to toss some tests around this to see if keystone code every changes here
18:48 openstackgerrit Merged stackforge/rally: fix devstack URL  https://review.openstack.org/146480
19:00 harlowja joined #openstack-rally
19:11 mwagner_lap joined #openstack-rally
19:28 boris-42 cdent: around?
19:29 cdent yessir
19:29 boris-42 cdent: just saw your email
19:29 boris-42 cdent: about gabbi
19:29 * cdent nods
19:30 boris-42 cdent: hehe=)
19:30 boris-42 cdent: nice idea
19:30 boris-42 cdent: I think I'll add such benchmark to rally=)
19:30 cdent I had someone ask me a while ago "how is this different from rally"
19:30 boris-42 cdent: haha
19:31 boris-42 cdent: and before they asked
19:31 boris-42 "how is this different from tempest"
19:31 boris-42 =)
19:31 cdent and I pointed out the difference between raw http as tests, and scenarios and benchmakrs
19:31 cdent :)
19:31 boris-42 cdent: so what I think is that this can be benchmark*
19:32 cdent oh yeah?
19:32 boris-42 cdent: yep
19:32 boris-42 cdent: just repeat it N times
19:32 boris-42 cdent: create-get-delete-get
19:32 boris-42 cdent: with proper assertations
19:32 cdent true
19:32 boris-42 cdent: btw one thing that I am interested
19:33 boris-42 cdent: how about preparing env
19:33 boris-42 cdent: like to create a volume snapshot you need a snapshot
19:33 boris-42 cdent: is there way to do that with gabbi?
19:33 cdent If you look at the linked ceilo review, there's an example in there of how a fixture is used, and also how the mongo database server is set up before the test run
19:34 cdent the fixture creates a few samples which are then retrieved in a test
19:34 boris-42 cdent: hm need to take a look better on that patch
19:34 cdent so it would be possible to have a fixture that established the snapshot and then cleaned it up when done
19:34 cdent but the fixtures themselves are left as an exercise for the impelementor
19:35 boris-42 cdent: hm this can be really integrated in rally
19:35 boris-42 cdent: I really like this idea=)
19:35 boris-42 so simple and stright
19:35 boris-42 maybe just passing this samples to your code
19:35 boris-42 like we have integrated tempest
19:35 cdent I like things simple.
19:35 boris-42 actually a bit differnt*
19:36 cdent It's far more complicated under the covers than I wanted it to be
19:36 boris-42 but Idea is nice
19:36 cdent but all the complexity is to deal with what I consider "bad form" in existing apis.
19:37 boris-42 cdent: heh if we have this enigne inside rally we could make a benchmarks based on this
19:37 cdent If you've got more questions, post the on the thread so we can get some visible dialog going
19:37 cdent As apparently in order for it to get into global-requirements it needs to be well discussed
19:37 boris-42 first of all you shoud move it
19:37 boris-42 to stackforge
19:37 boris-42 put base testing
19:37 boris-42 at least unit
19:38 boris-42 and pep
19:38 cdent I'm waiting to address the stackforge issue until a bit later
19:38 cdent it already has automated per-commit testing via travis, including pep
19:39 cdent it's time for dinner, I'll be back a bit later to check in on things
19:39 boris-42 cdent: so one advice
19:39 boris-42 cdent: don't use for base unittest
19:39 boris-42 =)
19:39 boris-42 it makes thing painfull
19:39 boris-42 =)
19:40 cdent don't use what for what?
19:42 boris-42 unittest as a base for your framework
19:43 boris-42 it really sux for functional testing
19:48 boris-42 so python -m subunit.run discover -f gabbi | subunit2pyunit
19:48 boris-42 cdent: ^
19:48 boris-42 cdent: this is blocker for any kind of load testing with your lib
19:49 boris-42 cdent: cause it takes forever do do load
19:49 boris-42 I mean discover
19:52 harlowja joined #openstack-rally
19:55 nmagnezi joined #openstack-rally
20:00 openstackgerrit Boris Pavlovic proposed stackforge/rally: Input task templates and task cmd cleanup  https://review.openstack.org/137716
20:00 boris-42 andreykurilin1: ^
20:13 cdent boris-42: the use of subunit is just for the sake of integration with testrepository
20:13 cdent it's not required
20:14 cdent the use of unittest as the base is again, for integration with existings stuff. If I had my way, it would all be py.test based simple assertions and no TestCase, TestSuite, subunit, testrepository anywhere
20:18 cdent It might make sense, boris-42, to abstract the TestCase related stuff out as a facade over something more basic
20:19 cdent And then other things can use the tests too
20:19 cdent a friend of mine wants to make it work with py.test if he ever finds the time
20:19 boris-42 cdent: what I know is that this is evil =)
20:19 boris-42 cdent: from Rally expirince=)
20:20 boris-42 cdent: I wrote some ideas to ML
20:20 boris-42 cdent: could you please reply (so we will get more interest on that topic)
20:20 boris-42 cdent: py.test is not for functional testing...
20:21 cdent why do you say that?
20:21 cdent (the py.test bit)
20:23 cdent but yeah will get on the mailing list, will switch puters to do that, biab
20:28 barra204 joined #openstack-rally
20:30 flwang joined #openstack-rally
20:30 boris-42 joined #openstack-rally
20:32 barra204_ joined #openstack-rally
20:36 mwagner_lap joined #openstack-rally
20:39 exploreshaifali joined #openstack-rally
20:39 vaidy_ril joined #openstack-rally
20:56 meteorfox joined #openstack-rally
20:56 meteorfox left #openstack-rally
21:02 andreykurilin joined #openstack-rally
21:39 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
21:39 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
21:45 klindgren_ joined #openstack-rally
21:54 nmagnezi_ joined #openstack-rally
22:07 andreykurilin joined #openstack-rally
22:23 jlk boris-42: rally-manage db recreate could use some progress  indication, or the install script should say something about it happening :/
22:24 boris-42 jlk: hm
22:24 boris-42 jlk: it should be fast lol
22:24 jlk on a 2.7G sqlite db it's taking  a while
22:24 boris-42 jlk: lol
22:24 boris-42 never tested that
22:24 boris-42 =)
22:24 jlk we were trying to break hardware, so I had a while true; do; rally loop setup
22:24 jlk for weeks
22:26 boris-42 jlk: heh=) I wish we will finish someday rally as a service
22:26 boris-42 jlk:  so no bash=)
22:27 jlk nod
22:27 boris-42 for such purpose
22:27 boris-42 and periodic runs with sla checks
22:27 boris-42 that is some kind of active monitoring
22:28 jlk question
22:28 jlk on a shared host, with a single rally install on it
22:28 jlk can the active deployment be per-user(session) ?
22:29 jlk so userA logged in can run rally against cloud A and userB logged in can run rally against cloudB?
22:29 boris-42 yep why not
22:29 boris-42 rally stores all files to ~/.rally/*
22:29 jlk just wasn't sure what the "activate" deployment stuff does
22:29 jlk ah ok.
22:29 jlk so for sure, have to use different user accounts
22:29 boris-42 let me show you the secret=)
22:29 jlk can't be different rallys per login session of the same user account
22:30 boris-42 jlk: run "cat ~/.rally/globals"
22:30 boris-42 jlk: you can
22:30 boris-42 jlk: but you can't use "default deployment feature"
22:30 boris-42 you need to run task with specifiing deployment
22:30 jlk gotcha
22:30 jlk okay
22:31 boris-42 so it can work even from one user=)
22:31 boris-42 in one session*
22:31 boris-42 if you put &
22:31 boris-42 you should be able even to run few tasks against single deployment
22:32 boris-42 at least we are trying to have FULL isoliation between 2 tasks
22:49 jlk nod
23:01 barra204__ joined #openstack-rally
23:15 boris-42 jlk: LOL
23:15 boris-42 jlk: Don't miss it https://bugs.launchpad.net/ossa/+bug/1408663
23:15 boris-42 jlk: glance strikes again=)
23:29 mwagner_lap joined #openstack-rally
23:38 shakamunyi joined #openstack-rally
23:47 jlk boris-42: yeah we caught that one last week. My havana version of the patch was safe :)

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