Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-05-12

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

All times shown according to UTC.

Time Nick Message
00:36 botang joined #openstack-rally
00:48 dmorita joined #openstack-rally
00:56 Topic for #openstack-rally is now ☁ Rally RoadMap: http://goo.gl/JZkmwY ☁ Review Dashboard: http://goo.gl/YuCJfl ☁ IRC logs http://irclog.perlgeek.de/openstack-rally ☁ Key persons to ask:  boris-42, yfried, msdubov, rediskin, andreykurilin, amaretskiy  ☁ Documentation: https://rally.readthedocs.org/en/latest/ ☁ To publish changes to Rally:  https://rally.readthedocs.org/en/latest/contribute.html
00:56 openstackstatus NOTICE: Gerrit has been downgraded to version 2.8 due to the issues observed today. Please report further problems in #openstack-infra.
01:21 botang joined #openstack-rally
02:03 botang joined #openstack-rally
02:11 pradeep joined #openstack-rally
02:31 jjmb joined #openstack-rally
03:05 yingjun joined #openstack-rally
03:18 baker joined #openstack-rally
03:19 avico joined #openstack-rally
03:37 yingjun_ joined #openstack-rally
03:54 yingjun joined #openstack-rally
04:03 baker joined #openstack-rally
04:22 tfreger joined #openstack-rally
04:37 rdas joined #openstack-rally
04:40 anshul joined #openstack-rally
04:56 jaypipes joined #openstack-rally
05:00 yingjun joined #openstack-rally
05:41 pradeep joined #openstack-rally
06:28 coolsvap joined #openstack-rally
06:37 yingjun joined #openstack-rally
06:50 kiran-r joined #openstack-rally
06:54 agarciam joined #openstack-rally
07:01 pradeep joined #openstack-rally
07:24 neeti joined #openstack-rally
07:28 exploreshaifali joined #openstack-rally
07:40 openstackgerrit Vitaly Gusev proposed openstack/rally: [Ceilometer] Add CeilometerNovaMetrics.get_nova_notifications scenario  https://review.openstack.org/154803
07:54 openstackstatus joined #openstack-rally
07:54 openstack joined #openstack-rally
08:01 karimb joined #openstack-rally
08:02 tfreger1 joined #openstack-rally
08:03 arxcruz joined #openstack-rally
08:03 exploreshaifali joined #openstack-rally
08:05 yfried joined #openstack-rally
08:10 nkhare joined #openstack-rally
08:10 openstackgerrit Mikhail Dubov proposed openstack/rally: Release 0.0.4  https://review.openstack.org/180498
08:17 openstackgerrit maplelabs proposed openstack/rally: Loadbalancer Member and pool functionality support  https://review.openstack.org/182189
08:17 e0ne joined #openstack-rally
08:17 kiran-r joined #openstack-rally
08:19 redixin joined #openstack-rally
08:21 pradeep joined #openstack-rally
08:23 yingjun joined #openstack-rally
08:31 yfried boris-42: was there a meeting last night?
08:38 jjmb1 joined #openstack-rally
08:45 jjmb joined #openstack-rally
08:45 openstackgerrit Yuriy Nesenenko proposed openstack/rally: Add Cinder create_backup volume scenario  https://review.openstack.org/167171
08:45 jjmb joined #openstack-rally
08:46 jjmb joined #openstack-rally
08:53 arxcruz joined #openstack-rally
09:04 openstackgerrit Igor Degtiarov proposed openstack/rally: [Ceilometer] Add  scenario for events and traits  https://review.openstack.org/180514
09:09 tfreger joined #openstack-rally
09:16 aix joined #openstack-rally
09:22 fhubik joined #openstack-rally
09:31 openstackgerrit Nikita Konovalov proposed openstack/rally: Wait for status  https://review.openstack.org/172466
09:31 openstackgerrit Nikita Konovalov proposed openstack/rally: [Sahara] More debug logging  https://review.openstack.org/170491
09:32 openstackgerrit Nikita Konovalov proposed openstack/rally: [Sahara] Split EDP context  https://review.openstack.org/177682
09:37 kiranr joined #openstack-rally
09:51 aswadr joined #openstack-rally
09:54 yfried andreykurilin1: tnx
09:55 yfried boris-42: https://review.openstack.org/#/c/168849/ could you please take a look?
09:55 cdent joined #openstack-rally
09:56 kiranr joined #openstack-rally
09:57 andreykurilin1 yfried: np:)
10:07 openstackgerrit maplelabs proposed openstack/rally: corrected pep8 errors  https://review.openstack.org/182221
10:21 openstackgerrit maplelabs proposed openstack/rally: LoadBalancer member and pool functionalities  https://review.openstack.org/182223
10:22 openstackgerrit maplelabs proposed openstack/rally: corrected pep8 errors  https://review.openstack.org/182221
10:27 kiran-r joined #openstack-rally
10:40 openstackgerrit Kairat Kushaev proposed openstack/rally: Add scenario for heat resource group  https://review.openstack.org/178751
10:40 openstackgerrit Kairat Kushaev proposed openstack/rally: Pass parameters, files and env  for heat stacks  https://review.openstack.org/178631
10:40 openstackgerrit Kairat Kushaev proposed openstack/rally: Add create-snapshot-restore-delete stack scenario  https://review.openstack.org/179037
10:41 kiran-r joined #openstack-rally
10:42 kairat_kushaev pradeep: Hello, could you please explain in more details why do I need to update fakes here https://review.openstack.org/#/c/179037/?
11:11 tosky joined #openstack-rally
11:13 openstackgerrit Igor Degtiarov proposed openstack/rally: [Ceilometer] Add  scenario for events and traits  https://review.openstack.org/180514
11:14 nkhare joined #openstack-rally
11:16 boris-42 yfried: one question https://review.openstack.org/#/c/168849/
11:16 boris-42 yfried: why del verifier
11:18 yfried boris-42: you are actually right, since I'm assigning a new value
11:18 yfried boris-42: I'll remove it
11:20 boris-42 yfried: thanks
11:21 boris-42 yfried: otherwise lgtm
11:21 openstackgerrit Kairat Kushaev proposed openstack/rally: Add create-snapshot-restore-delete stack scenario  https://review.openstack.org/179037
11:21 openstackgerrit Yair Fried proposed openstack/rally: Adds tempest re-install to API/CLI  https://review.openstack.org/168849
11:21 yfried boris-42: ^ new version
11:21 yfried andreykurilin1: ^
11:23 openstackgerrit Sergey Skripnick proposed openstack/rally: Improve installation script  https://review.openstack.org/162418
11:24 kairat_kushaev boris-42: Hello, do you propose to update rally-tests for all scenarios here: https://review.openstack.org/#/c/178631/?
11:30 anshul joined #openstack-rally
11:30 openstackgerrit Mikhail Dubov proposed openstack/rally: [Spec] Enhance Rally API  https://review.openstack.org/182245
11:33 openstackgerrit Vitaly Gusev proposed openstack/rally: [Ceilometer] Add CeilometerNovaMetrics.get_nova_pollsters scenario  https://review.openstack.org/160387
11:36 pradeep kairat_kushaev:  under FakeHeatClient , stacks doesnt have snapshot defined
11:37 pradeep kiran-r:  you might want to take a look at it tests/unit/fakes.py
11:38 pradeep kairat_kushaev:  Did u run tox?
11:38 kairat_kushaev kairat_kushaev: yep, it is ok with my patch. I am not using any Fakes. so Why should I update FakeHeatClient?
11:43 stpierre joined #openstack-rally
11:46 karimb joined #openstack-rally
11:57 exploreshaifali joined #openstack-rally
12:01 nkhare joined #openstack-rally
12:02 openstackgerrit Jens Rosenboom proposed openstack/rally: Fix security groups cleanup  https://review.openstack.org/179428
12:02 pradeep kairat_kushaev: if you dont see any failuers with tox, then fine.
12:02 kairat_kushaev pradeep: ok
12:03 openstackgerrit Yair Fried proposed openstack/rally: Split SLA to base and plugins  https://review.openstack.org/180509
12:03 openstackgerrit Yair Fried proposed openstack/rally: Split scenarios under Plugins  https://review.openstack.org/181738
12:03 openstackgerrit Yair Fried proposed openstack/rally: Split runners to base and plugins  https://review.openstack.org/180904
12:03 yfried boris-42: ^ please take a look when you have the time
12:04 openstackgerrit maplelabs proposed openstack/rally: LoadBalancer member and pool functionalities(corrected)  https://review.openstack.org/182223
12:06 boris-42 yfried: thank you fro working on that I will
12:09 yfried boris-42: sure
12:12 openstackgerrit Roman Vasilets proposed openstack/rally: Add deploy environment murano benchmark  https://review.openstack.org/171625
12:15 openstackgerrit Mikhail Dubov proposed openstack/rally: [Spec] Enhance Rally API  https://review.openstack.org/182245
12:15 openstackgerrit maplelabs proposed openstack/rally: LoadBalancer member and pool functionalities(corrected)  https://review.openstack.org/182223
12:18 openstackgerrit maplelabs proposed openstack/rally: LoadBalancer member and pool functionalities(corrected)  https://review.openstack.org/182223
12:24 MaxPC joined #openstack-rally
12:28 aix joined #openstack-rally
12:43 boris-42 yfried: one nit here https://review.openstack.org/#/c/180509/
12:44 boris-42 yfried: in docs
12:44 boris-42 yfried: rest seems ok
12:44 rdas joined #openstack-rally
12:45 yfried boris-42: tnx
12:45 openstackgerrit Roman Vasilets proposed openstack/rally: Add functional test for task samples  https://review.openstack.org/181400
12:57 boris-42 yfried: so otherwise lgtm
12:57 openstackgerrit Mikhail Dubov proposed openstack/rally: [Spec] Enhance Rally API  https://review.openstack.org/182245
13:04 yfried boris-42: I'm working on context now and I'm not sure how to handle cleanup - is it common or openstack?
13:04 yfried boris-42: or maybe it should remain in base
13:13 openstackgerrit Sergey Skripnick proposed openstack/rally: Improve installation script  https://review.openstack.org/162418
13:13 openstackgerrit Yair Fried proposed openstack/rally: Split SLA to base and plugins  https://review.openstack.org/180509
13:13 openstackgerrit Yair Fried proposed openstack/rally: Split scenarios under Plugins  https://review.openstack.org/181738
13:13 openstackgerrit Yair Fried proposed openstack/rally: Split runners to base and plugins  https://review.openstack.org/180904
13:13 openstackgerrit Yair Fried proposed openstack/rally: [Context] Split context into plugins  https://review.openstack.org/182284
13:16 openstackgerrit Jens Rosenboom proposed openstack/rally: Fix security groups cleanup  https://review.openstack.org/179428
13:18 openstackgerrit Yair Fried proposed openstack/rally: Split SLA to base and plugins  https://review.openstack.org/180509
13:18 openstackgerrit Yair Fried proposed openstack/rally: Split scenarios under Plugins  https://review.openstack.org/181738
13:18 openstackgerrit Yair Fried proposed openstack/rally: [Context] Split context into plugins  https://review.openstack.org/182284
13:18 openstackgerrit Yair Fried proposed openstack/rally: Split runners to base and plugins  https://review.openstack.org/180904
13:31 openstackgerrit Merged openstack/rally: Adds tempest re-install to API/CLI  https://review.openstack.org/168849
13:45 georgem1 joined #openstack-rally
13:47 asrangne joined #openstack-rally
13:48 openstackgerrit Nikita Konovalov proposed openstack/rally: [Sahara] Improve Image context  https://review.openstack.org/166859
13:48 georgem1 If I run a rally task (NovaServers.boot_and_delete_server) with concurrency: 50 and times: 10, does it mean that it will start 50 instances at a time, delete them and repeat 10 times?
13:49 georgem1 because the task finished successfully but I don't see 500 deleted instances when I run: nova list --all-ten --deleted | grep rally | wc -l
13:49 yfried georgem1: it will start 10 instances at a time. repeat 5 times
13:49 yfried georgem1: ^ ideally
13:50 yfried georgem1: afaik, in practice what you are saying is, "do this task 50 times" and "dedicate 10 workers to it"
13:50 openstackgerrit Merged openstack/rally: Setup volume detach timings separately  https://review.openstack.org/179087
13:51 yfried georgem1: so if the first op takes inite time due to some error, the 9 remaining threads will have to devide 49 ops as they see fit
13:51 georgem1 yfried: this is the task http://pastebin.com/bhfCnxt6
13:51 yfried georgem1: am I clear?
13:51 yfried the total number of ops (VMs created) will be 50
13:52 yfried georgem1: ^ and since the scenario is "_and_delete" - each op will delete its own VM.
13:52 baker joined #openstack-rally
13:53 georgem1 yfried: I thought that it means it will start 50 instances and delete them 10 times in a row: http://en.wikipedia.org/wiki/Concurrency_%28computer_science%29
13:53 yfried georgem1: if you run "boot_server" server instead, 50 VMs will be created and the scenario will finish. THEN, cleanup will delete all 50
13:53 baker joined #openstack-rally
13:53 yfried georgem1: that's a very common misunderstanding
13:54 yfried georgem1: take a look at https://rally.readthedocs.org/en/latest/ and if it seems unclear, you are welcom to open a bug and/or edit the doc
13:54 stpierre joined #openstack-rally
13:54 yfried georgem1: btw, if you take a look at the results.html file you'll things more clearly
13:57 georgem1 yfried: the report shows a count of 10 for the iteration, so it did 10 iterations of 5 instances (I assume), which is different than what you said
13:57 rook joined #openstack-rally
13:59 georgem1 ok, so if I want to have 50 instances started once and then deleted and repeat the process 10 times, I should change to concurrency: 500 and times: 10
14:00 yfried georgem1: could you please share your HTML?
14:00 yfried georgem1: times:500, concurrency:10
14:01 yfried georgem1: but maby RPS runner might suite you better?
14:02 tosky joined #openstack-rally
14:02 georgem1 http://picpaste.com/Screen_Shot_2015-05-12_at_10.02.12_AM-DwVcNSlU.png
14:05 baker joined #openstack-rally
14:06 georgem1 yfried: yeah, I definitely don't understand how this works… I tried again with   "concurrency": 500,  "times": 10 and the count of deleted instances containing the word rally went up by only 10
14:06 yfried boris-42: andreykurilin1: redixin: amaretskiy: where's the design summit agenda/etherpads re Rally?
14:08 yfried georgem1: I can't help you with screen captures. I need to see the actual HTML file
14:08 yfried georgem1: what do you mean "count of deleted instances containing the word rally"
14:09 georgem1 yfried: this is how I see how many instances were created by rally
14:09 yfried georgem1: are you doing "nova list | grep rally | grep delete" | wc -l
14:09 yfried georgem1: or something like that?
14:09 georgem1 nova list --all-ten --deleted | grep rally | wc -l
14:09 georgem1 yes
14:09 yfried georgem1: that wouldn't work
14:09 georgem1 why?
14:09 yfried georgem1: please consider that you are actually looking at the VMs *currently being deleted*
14:10 yfried georgem1: some VMs might be deleted faster, prior or after you are executing teh query
14:10 georgem1 yfried: no, I'm pretty sure I see instances deleted long time ago
14:13 boris-42 yfried: it's openstack
14:13 boris-42 yfried: for now
14:13 yfried boris-42: what do you mean?
14:13 yfried georgem1: ok. I was unfamiliar with the "--deleted" option
14:13 * yfried rechecking georgem1 command
14:14 andreykurilin1 yfried: https://libertydesignsummit.sched.org/type/design+summit/rally#.VVIKmeRZ5Qs
14:14 boris-42 yfried: rally cleanup context
14:14 boris-42 yfried: is for now openstack
14:15 yfried boris-42: you want me to move the entire cleanup dir under plugins.openstack? ok but it seems that it should have something in common, for WHEN we have non-openstack scenario/resources
14:15 openstackgerrit Chris St. Pierre proposed openstack/rally: Add Heat scenario to test scaling policies  https://review.openstack.org/178436
14:15 boris-42 yfried: it's not common at all
14:15 boris-42 yfried: there is nothing common inside it for now
14:16 yfried boris-42: I understand. just thinking ahead
14:16 yfried boris-42: some base/abstarct
14:16 boris-42 yfried: someday when somebody decide to do something abstract based on what we have
14:16 boris-42 yfried: we will handle it properly
14:16 yfried andreykurilin1: any rally talks in the main summit?
14:18 yfried [17:06] <georgem1> yfried: yeah, I definitely don't understand how this works… I tried again with   "concurrency": 500,  "times": 10 and the count of deleted instances containing the word rally went up by only 10
14:19 yfried georgem1: ^ so this is in line with what I've said - "times:10" means that you will be launching 10 VMs total.
14:19 yfried georgem1: you've dedicated 500 threads to do so (which is, obviously, an overkill)
14:20 yfried georgem1: so are you good? have you looked at RPS runner, seems like it's exactly what you are looking for
14:25 openstackgerrit Sergey Skripnick proposed openstack/rally: Improve installation script  https://review.openstack.org/162418
14:29 openstackgerrit Chris St. Pierre proposed openstack/rally: Fix security groups cleanup  https://review.openstack.org/179428
14:31 georgem1 yfried: thanks for help, I'll do some reading on the RPS runner
14:31 yfried georgem1: sure. but does my explanation make sense to you?
14:32 georgem1 yfried: yes, it does now
14:34 yfried georgem1: great. glad I could help
14:45 yfried boris-42: now I need another core with the patience to go over these patches...
14:46 yfried boris-42: msdubov: https://review.openstack.org/#/q/topic:bp/split-plugins,n,z how do I add these to the next release?
14:57 openstackgerrit Alexander Maretskiy proposed openstack/rally: [Fuel] Add Fuel scenario `list_environments'  https://review.openstack.org/178731
15:03 baker joined #openstack-rally
15:04 openstackgerrit Nikita Konovalov proposed openstack/rally: Wait for status  https://review.openstack.org/172466
15:04 baker joined #openstack-rally
15:08 kiran-r joined #openstack-rally
15:11 kiranr joined #openstack-rally
15:11 boris-42 yfried: OH https://review.openstack.org/#/c/182284/2
15:11 boris-42 yfried: a lot of files=)
15:14 aswadr joined #openstack-rally
15:16 yfried boris-42: every file that uses "context"
15:17 yfried boris-42: https://review.openstack.org/#/c/181738/4 a lot more
15:17 boris-42 yfried: split it please
15:17 boris-42 yfried: to few patches
15:17 boris-42 yfried: you don't need to move everything in one step *
15:19 pradeep joined #openstack-rally
15:20 yfried boris-42: considering that some tests use "plugins.openstack.scenario" stuff it would make the patches unnecessary complex
15:20 yfried boris-42: better handle each plugin in a single patch. it's hell to sort as ti is
15:21 boris-42 yfried: I am asking another thing
15:21 boris-42 yfried: directories are mostly indpenedent
15:21 boris-42 yfried: except nova/vm
15:21 boris-42 yfried: so you can move few directories of scenarios in each patch
15:22 boris-42 yfried: e.g. zaqar, murano, mistral, designate can be move in one patch
15:22 boris-42 yfried: after that nova, vm in another
15:22 boris-42 yfried: glance, cinder, sahara and rest in third
15:22 boris-42 and so on
15:26 yfried boris-42: I'd rather not. it was a hard work to do this as it was. splitting it would be even harder
15:26 boris-42 yfried: why do you think so?
15:27 boris-42 yfried: it's a bit hell for reviewer https://review.openstack.org/#/c/181738/
15:27 klindgren joined #openstack-rally
15:28 yfried boris-42: because it means that's i'll have to revert PARTS of this patch in order to split it
15:29 yfried boris-42: I know these are a lot of files, but the biggest change to a single is 12 lines.
15:30 boris-42 yfried: yep but I can't even check that you moved everything properly..=( with proper names
15:30 boris-42 and there is no typos in files
15:30 boris-42 filenames..
15:30 yfried boris-42: the move was done using pycharm's refactor, so the filenames should be intact
15:31 yfried boris-42: but I understand what you are saying
15:31 yfried boris-42: I'll try
15:38 openstackgerrit Roman Vasilets proposed openstack/rally: Add deploy environment murano benchmark  https://review.openstack.org/171625
15:45 jaypipes joined #openstack-rally
16:01 gema_ joined #openstack-rally
16:02 klindgren joined #openstack-rally
16:08 gema_ joined #openstack-rally
16:10 yfried boris-42: it's too hard to split this retroactively https://review.openstack.org/#/c/181738/4/rally/common/opts.py,cm
16:10 openstackgerrit Sergey Skripnick proposed openstack/rally: Improve installation script  https://review.openstack.org/162418
16:11 yfried boris-42: btw, if the names were wrong, the tests would have failed, as some imports are strings and some are python syntax
16:11 exploreshaifali joined #openstack-rally
16:12 yfried boris-42: so it's enough to look at the name changes in the test files and you don't need to look that the diff list in gerrit
16:12 yfried boris-42: sorry that's it's too long. I'll try to split the work better in the furure
16:19 davideagnello joined #openstack-rally
16:26 davideagnello joined #openstack-rally
16:34 yfried boris-42: when is the rally mtg?
16:36 yfried redixin: amaretskiy: andreykurilin1: ^
16:36 yfried meteorfox: ^^
16:39 amaretskiy yfried: we have rescheduled meetings to wednesday
16:39 amaretskiy in mailing list
16:39 yfried amaretskiy: so tomorrow?
16:39 amaretskiy unfortunately wiki is not updated https://wiki.openstack.org/wiki/Rally#Where_can_I_discuss_.26_propose_changes.3F
16:39 amaretskiy yfried I guess - yes :)
16:41 rook amaretskiy yfried is there any place I can get more information on rps?
16:41 rook comparing rps / constant ?
16:41 gema_ joined #openstack-rally
16:43 boris-42 yfried: okay
16:43 boris-42 rook: so what you are interested in?
16:44 boris-42 rook: contant - has contant amount of workers per whole test
16:44 boris-42 rook: rps - has not constant amount of workers per whole test
16:44 boris-42 rook: rps - will start new iteration each (1 / rps) second
16:45 boris-42 rook: constant - will start new iteration when some of previous iteration get finished*
16:45 boris-42 rook: ^ that's the whole difference
16:45 rook boris-42 thanks!
16:46 rook boris-42 with SLA do we have a mechanism to increase load on successful iterations? ( i had this scripted previously )
16:47 rook for example : if I ran the VMTask scenario, with concurrency of 100 and times 100.. if that returns success, i would want to increase concurrency and times, to 150/150 for example.
16:47 rook to simulate a bootstorm.
16:47 rook the desired outcome would be, it would keep increasing N instances, until i could no longer sustain the SLA
16:48 boris-42 rook: so for now we are supporting only aborting load on SLA failure
16:49 boris-42 rook: to do what you want you can implement own runner
16:50 rook boris-42 hm ok
16:51 boris-42 rook: basically you can extend rps runner to do the ramp stuff
16:51 boris-42 rook: or we can extend it
16:52 rook boris-42: i have scripted this outside of Rally in the past.
16:52 rook might make sense to bring the logic into Rally at some point.
16:52 boris-42 rook: https://github.com/openstack/rally/blob/master/rally/benchmark/runners/rps.py#L142-L214
16:53 boris-42 rook: so I am not sure that we will be able to do the next steps based on SLA results
16:53 boris-42 rook: because it affects engine
16:53 rook okay
16:53 openstackgerrit Nikita Konovalov proposed openstack/rally: [Sahara] More debug logging  https://review.openstack.org/170491
16:54 boris-42 rook: at least you can push feature requres
16:54 boris-42 rook: so we will implement this at some point of time https://github.com/openstack/rally/tree/master/doc/feature_request
16:55 openstackgerrit Alexander Maretskiy proposed openstack/rally: [Fuel] Add Fuel scenario `list_environments'  https://review.openstack.org/178731
16:55 PerfBeing joined #openstack-rally
17:02 ptoohill joined #openstack-rally
17:03 dmorita joined #openstack-rally
17:04 openstackgerrit Roman Vasilets proposed openstack/rally: Add deploy environment murano benchmark  https://review.openstack.org/171625
17:05 vkmc joined #openstack-rally
17:08 openstackgerrit Roman Vasilets proposed openstack/rally: Add deploy environment murano benchmark  https://review.openstack.org/171625
17:15 rook boris-42: is vmtask scenario back to defaulting to floating ip?
17:18 rook NM - i was looking at a old repo.
17:18 rook good to go.
17:22 harlowja joined #openstack-rally
17:22 boris-42 rook: it works with both
17:23 boris-42 rook: fixed/floating
17:23 rook yup
17:23 rook boris-42: i had to change branches ;)
17:24 boris-42 rook: heh
17:25 boris-42 rook: we have a lot of changes recently
17:31 baker joined #openstack-rally
17:32 baker joined #openstack-rally
17:32 openstackgerrit Roman Vasilets proposed openstack/rally: Add deploy environment murano benchmark  https://review.openstack.org/171625
17:37 baker joined #openstack-rally
18:01 georgem1 left #openstack-rally
18:08 pradeep joined #openstack-rally
18:26 boris-42 stpierre: so what do you think about the very last version https://review.openstack.org/#/c/162418/ ?
18:29 cdent joined #openstack-rally
18:29 exploreshaifali joined #openstack-rally
18:31 stpierre lemme take a look. i was waiting for it to pass gate first :)
18:32 jlk boris-42: ping; might have found a comment bug, not positive.  https://github.com/openstack/rally/blob/master/rally/benchmark/scenarios/nova/servers.py#L327  says it'll make a volume, but I don't see anything in that test about a volume
18:35 boris-42 jlk: can you fix it please?)
18:35 jlk sure, I just wanted to confirm that the test doesn't actually do a volume
18:35 boris-42 jlk: seems like we missed copy paste bug=(
18:35 boris-42 jlk: yep docstring is wired
18:36 boris-42 jlk: so we should fix docstring
18:36 jlk fixing
18:36 boris-42 jlk: btw comment regarding to cleanup is bad as well
18:37 jlk I just noticed that too :)
18:38 boris-42 jlk: actually we should sit down someday
18:38 jlk yes
18:38 boris-42 jlk: and just do the docstring cleanup
18:39 boris-42 jlk: we are doing it periodically
18:40 openstackgerrit Jesse Keating proposed openstack/rally: Fixup resize_server docstring to match reality  https://review.openstack.org/182405
18:42 boris-42 jlk: ^ +2+1ed
18:50 baker joined #openstack-rally
18:50 baker joined #openstack-rally
18:51 e0ne joined #openstack-rally
18:54 boris-42 stpierre: thanks for review
18:54 stpierre np. sorry i keep -1'ing this :(
18:54 boris-42 redixin: so seems like your patch is almost ready for merging
18:54 boris-42 redixin: there are few comments from me and stpierre
18:55 redixin boris-42, nice
18:56 boris-42 redixin: nothing serious just nitpicks mostly
18:57 yfried joined #openstack-rally
19:10 jlk boris-42: hi there.
19:11 jlk I want to run a resize test with confirm, and then another without confirm (to ensure resize revert works)
19:11 jlk but if I make two tests in the same file and that's the only difference, rally seems to only do one test
19:12 boris-42 jlk: nope
19:12 jlk maybe it's my older version of rally?
19:12 boris-42 jlk: nope
19:12 boris-42 jlk: seems like you are putting bad input file
19:12 jlk maybe
19:12 boris-42 jlk: paste it
19:12 jlk just a tic
19:12 jlk oooh
19:13 jlk if i want two runs under the same name, do I just make two list entries under the name block?
19:13 * jlk tries that first
19:15 boris-42 jlk: https://rally.readthedocs.org/en/latest/tutorial/step_2_running_multple_benchmarks_in_a_single_task.html
19:15 boris-42 jlk: ^ this is what you are looking for
19:15 boris-42 jlk: I agree that current input format is odd
19:15 jlk boris-42: yeah I had the yaml wrong
19:15 jlk is yaml or json preferred?
19:15 boris-42 jlk: any
19:16 openstackgerrit_ joined #openstack-rally
19:16 jlk bingo
19:16 jlk got it to work
19:17 boris-42 jlk: heh new format is much better...
19:17 boris-42 jlk: https://github.com/openstack/rally/blob/master/doc/specs/in-progress/new_rally_input_task_format.rst
19:17 jlk yeah, one of my epics this sprint is to upgrade our rally
19:18 jlk neat
19:18 boris-42 jlk: what version do you use of rally?
19:19 jlk hrm, let me see if I can discover that
19:19 boris-42 jlk: there should be any issues to move from 0.0.1 -> 0.0.4
19:19 boris-42 shoulndn't*
19:19 jlk 0.0.1
19:19 boris-42 jlk: not sure that it's the first version..
19:19 jlk it might not be
19:19 boris-42 jlk: how did you install it? git clone && install_rally.sh?
19:19 jlk yeah
19:20 boris-42 jlk: can you show me the git log -1 ?
19:20 jlk just a sec, running a test from that cli
19:20 jlk our install is somewhat janky because it was a "spike", a trial. We never did the work to fully SCM it
19:21 openstackgerrit Merged openstack/rally: [Sahara] Improve Image context  https://review.openstack.org/166859
19:21 boris-42 jlk: how old installation is?
19:22 boris-42 jlk: if it is about 1/2 year then upgrade should be smooth
19:22 jlk hrm. less than 6 months old?
19:22 boris-42 jlk: download new code and reinstall it
19:22 boris-42 jlk: yep
19:22 jlk much less because I've only been here for 6~ months
19:23 boris-42 jlk: so you should do next thing (remember the hash of commit where you are now)
19:23 jlk yeah my card is to do that, install the new version, change our templates a bit, and show people how to do one-shot executions on a specific cluster with run-time template definitions
19:23 boris-42 jlk: backup DB
19:23 boris-42 jlk: download new code and try to run it with the same DB
19:23 boris-42 jlk: it should work
19:23 jlk I don't even think we need the old db around
19:24 boris-42 jlk: ah then it's super simple
19:24 boris-42 jlk: you can do it even now
19:24 jlk oh, question. If I wanted to re-use parts of rally to render html from result json, is that easy to do? We're thinking about embedding rally results into our internal CMDB
19:24 jlk but instead of storing pre-rendered html we want to store the json
19:24 boris-42 jlk: so rally task report
19:25 boris-42 jlk: accept both N json files and UUIDs of task
19:25 boris-42 jlk: you can even mix them
19:25 jlk ah good.
19:25 boris-42 jlk: it will create single HTML page
19:26 jlk yeah, we dump out the html right now and just toss it in a directory, which is not great
19:26 jlk we'd rather associate the runs with the customer/cluster, so we can look back and show results on-demand
19:27 jlk but without either storing raw html or writing our own renderer. Seems  using report would be good enough
19:30 jjmb joined #openstack-rally
19:33 boris-42 jlk: be back soon
19:33 boris-42 jlk: why not storing to Rally db everytyhing?
19:33 boris-42 jlk: btw?)
19:34 jlk Because I'm not sure I want to introduce another API/DB that we have to keep running/monitored/backedup/etc..
19:34 jlk I'd rather our rally be as stateless as possible and store the data in our existing system
19:36 jlk boris-42: will there be structure in the future to do things like "ensure I can still ping/ssh to the host during the test" type stuff? Like if I wanted to boot, ping, resize, ping, resize confirm/revert, ping ?
19:39 jlk boris-42: our install is from 930394d3e9afcbef20976eefefb9e72bd7fa498e, jan 20th
19:46 turul_ joined #openstack-rally
19:48 botang joined #openstack-rally
20:12 PerfBeing joined #openstack-rally
20:13 karimb joined #openstack-rally
20:25 boris-42 jlk: ok
20:26 boris-42 jlk: so what you want is actually boot_and_bounce_server
20:26 boris-42 jlk: but a bit extended wiht ping/ssh stuff
20:26 jlk perhaps, but I was specifically looking to test w/ resize
20:27 jlk since resize or migrate will send the instance elsewhere
20:27 jlk not necessary for the minimum viable product though.
20:28 boris-42 jlk: so you can extend the boot_and_bounce server
20:28 jlk ah
20:28 boris-42 jlk: to add resize action
20:29 boris-42 jlk: or migrate
20:29 jlk makes sense
20:29 boris-42 jlk: it was actually already attempt
20:29 boris-42 jlk: for livemigrate
20:37 openstackgerrit joined #openstack-rally
20:53 e0ne joined #openstack-rally
20:57 openstackgerrit Chris St. Pierre proposed openstack/rally: Add Nova scenario to boot multiple servers in one request  https://review.openstack.org/176858
21:04 openstackgerrit Merged openstack/rally: Fixup resize_server docstring to match reality  https://review.openstack.org/182405
21:07 PerfBeing joined #openstack-rally
21:38 openstackgerrit Merged openstack/rally: Split SLA to base and plugins  https://review.openstack.org/180509
21:52 openstackgerrit Roman Vasilets proposed openstack/rally: Add deploy environment murano benchmark  https://review.openstack.org/171625
22:03 tosky joined #openstack-rally
22:18 logan2 joined #openstack-rally
22:26 PerfBeing joined #openstack-rally
22:31 botang joined #openstack-rally
22:56 davideagnello joined #openstack-rally
23:40 PerfBeing joined #openstack-rally
23:59 jjmb joined #openstack-rally
23:59 PerfBeing joined #openstack-rally

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