Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-02-17

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

All times shown according to UTC.

Time Nick Message
00:12 meteorfox boris-42: hey, I'm trying to run Rally against Juno stable branch, and I followed the docs to install it, however when it tried to configure it, I got this error "ImportError: No module named oslo_db"
00:13 meteorfox boris-42: is that familiar?
00:14 openstackgerrit Merged stackforge/rally: Add floating-IP aware VM boot/delete utils  https://review.openstack.org/151384
00:38 meteorfox boris-42: I was able to work around it, previously I used the "enable_service rally" approach with devstack, and cause that error. I "unstacked' everything, and removed rally from the enable service, then I just did a stack of vanilla devstack, and then installed and configured rally seperately.
00:40 dmorita joined #openstack-rally
02:49 ilbot3 joined #openstack-rally
02:49 Topic for #openstack-rally is now ☁ Rally RoadMap: http://goo.gl/JZkmwY ☁ Rally IRC chat logs http://irclog.perlgeek.de/openstack-rally ☁ Key persons to ask:  boris-42, msdubov, rediskin, andreykurilin, amaretskiy  ☁ To publish changes to Rally:  https://wiki.openstack.org/wiki/Rally/Develop#How_to_contribute   ☁ Rally weekly updates:  https://wiki.openstack.org/wiki/Rally/Updates
05:08 nkhare joined #openstack-rally
05:18 exploreshaifali joined #openstack-rally
05:22 rdas joined #openstack-rally
06:06 pboldin joined #openstack-rally
06:08 anshul joined #openstack-rally
07:03 neeti joined #openstack-rally
07:09 yfried joined #openstack-rally
07:13 yfried_ joined #openstack-rally
07:54 amaretskiy joined #openstack-rally
07:59 nmagnezi joined #openstack-rally
08:04 psd joined #openstack-rally
08:05 andreykurilin_ joined #openstack-rally
08:14 openstackgerrit Alexander Maretskiy proposed stackforge/rally: (WIP) Add chunks support and chart `Parallel iterations' to report  https://review.openstack.org/146814
08:19 arxcruz joined #openstack-rally
08:44 pboldin joined #openstack-rally
08:45 oanufriev joined #openstack-rally
08:54 exploreshaifali joined #openstack-rally
08:57 boris-42 meteorfox: so I already saw that issue
08:57 boris-42 meteorfox: we will make workaround
09:27 pboldin joined #openstack-rally
09:31 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
09:38 kairat_kushaev joined #openstack-rally
09:46 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
09:46 openstackgerrit Sergey Skripnick proposed stackforge/rally: Test "Add murano benchmarks" vs mos  https://review.openstack.org/156524
09:48 e0ne joined #openstack-rally
09:48 aix joined #openstack-rally
09:51 psd_ joined #openstack-rally
09:56 exploreshaifali boris-42, around ?
10:10 pboldin joined #openstack-rally
10:12 openstackgerrit Sergey Skripnick proposed stackforge/rally: Test "Add murano benchmarks" vs mos  https://review.openstack.org/156524
10:13 pboldin1 joined #openstack-rally
10:17 aix joined #openstack-rally
10:23 Mehrnaz joined #openstack-rally
10:31 openstackgerrit Merged stackforge/rally: Add create and delete routers scenarios for Neutron  https://review.openstack.org/113177
10:32 ilyashakhat joined #openstack-rally
10:33 openstackgerrit Wataru Takase proposed stackforge/rally: Add name pattern filter for resource cleanup  https://review.openstack.org/139643
10:37 openstackgerrit Roman Vasilets proposed stackforge/rally: Adding Multiple Request Scenario  https://review.openstack.org/117705
10:38 e0ne joined #openstack-rally
10:43 openstackgerrit Roman Vasilets proposed stackforge/rally: Adding Multiple Request Scenario  https://review.openstack.org/117705
10:49 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano benchmarks  https://review.openstack.org/137661
10:49 openstackgerrit Roman Vasilets proposed stackforge/rally: Add murano base for murano benchmarks  https://review.openstack.org/137650
11:02 e0ne joined #openstack-rally
11:23 psd joined #openstack-rally
11:24 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create and check stack scenario  https://review.openstack.org/155685
11:24 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add context that pre-creates heat stacks  https://review.openstack.org/154438
11:24 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create-suspend-resume stack scenario  https://review.openstack.org/155346
11:24 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add list-stack-with-resources scenario  https://review.openstack.org/154463
11:25 e0ne joined #openstack-rally
11:26 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
11:30 msdubov joined #openstack-rally
11:32 cdent joined #openstack-rally
11:32 exploreshaifali joined #openstack-rally
11:55 yingjun joined #openstack-rally
11:55 e0ne joined #openstack-rally
11:57 andreykurilin__ joined #openstack-rally
11:59 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create and check stack scenario  https://review.openstack.org/155685
11:59 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create-suspend-resume stack scenario  https://review.openstack.org/155346
12:00 nkhare_ joined #openstack-rally
12:09 exploreshaifali joined #openstack-rally
12:17 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create and check stack scenario  https://review.openstack.org/155685
12:24 openstackgerrit Nir Magnezi proposed stackforge/rally: Adds run_bash_command function, to vm utils  https://review.openstack.org/154825
12:26 exploreshaifali hello boris-42 there?
12:44 pboldin1 review pls https://review.openstack.org/#/c/138466/
12:44 openstackgerrit Nir Magnezi proposed stackforge/rally: Adds run_bash_command function, to vm utils  https://review.openstack.org/154825
12:51 nkhare_ joined #openstack-rally
12:53 psd_ joined #openstack-rally
13:12 openstackgerrit Nir Magnezi proposed stackforge/rally: Adds run_bash_command function, to vm utils  https://review.openstack.org/154825
13:22 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create and check stack scenario  https://review.openstack.org/155685
13:22 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add context that pre-creates heat stacks  https://review.openstack.org/154438
13:22 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create-suspend-resume stack scenario  https://review.openstack.org/155346
13:22 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add list-stack-with-resources scenario  https://review.openstack.org/154463
13:24 wtakase joined #openstack-rally
13:28 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
13:30 pboros joined #openstack-rally
13:31 psd `sh
13:34 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create and check stack scenario  https://review.openstack.org/155685
13:34 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add context that pre-creates heat stacks  https://review.openstack.org/154438
13:34 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add create-suspend-resume stack scenario  https://review.openstack.org/155346
13:34 openstackgerrit Kairat Kushaev proposed stackforge/rally: Add list-stack-with-resources scenario  https://review.openstack.org/154463
13:39 himangi joined #openstack-rally
13:51 psd_ joined #openstack-rally
14:02 openstackgerrit Yusuke Ide proposed stackforge/rally: Add user-friendly message if the task cannot be deleted.  https://review.openstack.org/156595
14:09 aarefiev joined #openstack-rally
14:57 exploreshaifali joined #openstack-rally
15:01 mwagner_lap joined #openstack-rally
15:03 cdent_ joined #openstack-rally
15:13 pboldin joined #openstack-rally
15:24 psd__ joined #openstack-rally
15:26 boris-42 exploreshaifali: hey hey
15:27 exploreshaifali boris-42, hello :)
15:27 exploreshaifali boris-42, I was thinking the project that I will work on "Generating trends of  N tasks", how will it be helpful for users?
15:28 exploreshaifali through rally anyone can check if the current envireoment of OpenStack is as per their requirement or not.... in terms of scalibility and performance
15:29 exploreshaifali for that we put different benchmarks or scenarios in a file and consider them as task
15:29 exploreshaifali so what will be the need of comparing N tasks ?
15:30 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
15:30 boris-42 exploreshaifali: let me expalin
15:30 boris-42 typical use case
15:30 boris-42 1) I am DevOps, I have deployment. And I have large task (with a lot of tests/benchmarks inside it) for validaton of it
15:31 boris-42 2) Running rally for first time to get baseline
15:31 boris-42 3) Chagning something in configuration/code
15:31 boris-42 4) Running rally one more time
15:31 boris-42 At this moment users should compare by eyes long tasks
15:32 boris-42 exploreshaifali: actually no I didn't ask you to "compare" N tasks
15:32 boris-42 exploreshaifali: I was asking to make trends no?
15:32 boris-42 these are 2 different tasks
15:32 exploreshaifali boris-42, yes
15:32 boris-42 exploreshaifali: so about comparing let's finish *
15:32 exploreshaifali boris-42, okay
15:32 boris-42 so end user would like to compare how it was before and now
15:32 boris-42 and it will be nice to have such command
15:33 boris-42 exploreshaifali: about your "trend"
15:33 boris-42 It is as well important
15:33 boris-42 For example you have CI where you are testing on daily basis
15:33 exploreshaifali boris-42, there is no such command (trend) at present......  right?
15:33 boris-42 exploreshaifali: yep that is what you would like to implement
15:34 boris-42 exploreshaifali: so the idea is to have N similar tasks that were run over time and build one report where we can see how it is changed over time
15:35 exploreshaifali boris-42, okay!.... I thought the all tasks will run at same time(in parallel) and we need to generate trend of them
15:36 exploreshaifali boris-42, now its pretty clear....... its like "x was happening 1 hour ago and now it is like y"
15:38 exploreshaifali boris-42, for this the tasks will be "same"
15:38 exploreshaifali ?
15:40 exploreshaifali or we can generate trend from similar tasks also..... if so, how? I am unable to create a picture in my mind
15:45 boris-42 exploreshaifali: it depends on case
15:46 boris-42 exploreshaifali: like you can run peridocially tasks against cloud
15:46 boris-42 exploreshaifali: like every hour
15:46 boris-42 exploreshaifali: and track how performance is changed druing thime
15:46 boris-42 exploreshaifali: or it can be integrated in process
15:46 boris-42 of development
15:46 boris-42 where you would like to track how performance is changed over time
15:47 boris-42 exploreshaifali: these are 2 most often cases for trend functionallity
15:49 exploreshaifali boris-42, okay, but we will able to generate treds of "same " tasks only....... how we generate trends from "similar" tasks?
15:50 pboros_ joined #openstack-rally
15:57 pboros joined #openstack-rally
15:59 boris-42 exploreshaifali: with some hacks
16:00 boris-42 exploreshaifali: but first step is to make for "same" tasks
16:00 boris-42 exploreshaifali: and then start extending
16:00 exploreshaifali boris-42, okay :D
16:01 exploreshaifali boris-42, Thanks :)
16:07 exploreshaifali boris-42, can I add your name https://wiki.openstack.org/wiki/GSoC2015 here under Mentors section ?
16:13 yingjun_ joined #openstack-rally
16:14 boris-42 exploreshaifali yep sure=)
16:15 exploreshaifali boris-42, okay :)
16:15 boris-42 exploreshaifali: so we will move with quite small steps
16:15 boris-42 exploreshaifali: for example we will have at first point only one graph
16:15 boris-42 exploreshaifali: based on duration
16:15 boris-42 exploreshaifali: then add errors
16:15 boris-42 exploreshaifali: and so on=)
16:16 exploreshaifali boris-42, okay, that seems to be good approach..... but please can you hint me a bit for hacks to do for generating trends for similar tasks?
16:16 exploreshaifali boris-42, I want to make a picture in my mind
16:29 pboros_ joined #openstack-rally
16:32 exploreshaifali brb
16:32 boris-42 ekarlso: heh
16:33 boris-42 ekarlso: sry
16:34 andreykurilin_ joined #openstack-rally
16:47 msdubov joined #openstack-rally
16:48 ekarlso boris-42: ...
16:49 andreykurilin_ joined #openstack-rally
16:52 Topic for #openstack-rally is now ☁ Rally RoadMap: http://goo.gl/JZkmwY ☁ Rally IRC chat logs http://irclog.perlgeek.de/openstack-rally ☁ Key persons to ask:  boris-42, msdubov, rediskin, andreykurilin, amaretskiy  ☁ Documentation: https://rally.readthedocs.org/en/latest/ ☁ To publish changes to Rally:  https://rally.readthedocs.org/en/latest/contribute.html
16:59 oanufriev joined #openstack-rally
17:12 psd_ joined #openstack-rally
17:26 jaypipes joined #openstack-rally
17:36 exploreshaifali joined #openstack-rally
17:38 e0ne joined #openstack-rally
18:01 openstackgerrit Alexander Maretskiy proposed stackforge/rally: (WIP) Add chunks support and chart `Parallel iterations' to report  https://review.openstack.org/146814
18:05 abhirc joined #openstack-rally
18:14 andreykurilin_ joined #openstack-rally
18:17 himangi joined #openstack-rally
18:32 psd__ joined #openstack-rally
18:51 meteorfox boris-42: quick question, is there a scenario where I can build, say 10 concurrent instances (with different users), but get the metrics for the individual boot times, and not for the whole group
18:51 meteorfox ?
18:52 meteorfox it seems boot-and-delete.json, only gives metrics for the whole time of building 2 instances, and not for each build seperately
18:53 boris-42 meteorfox: h
18:53 boris-42 meteorfox: actually output for rally is much richer
18:54 boris-42 meteorfox: did you run "rally task report" command?
18:54 meteorfox yeas
18:54 meteorfox yes
18:54 boris-42 every iteration is single boot
18:54 meteorfox I ran the default boot-and-delete.json
18:54 meteorfox and each iterations seems to report the build for the 2 servers
18:55 meteorfox I understand that scenario, is like a cluster boot
18:55 meteorfox but I also want to be able to get the metrics of each individual boot
19:04 meteorfox boris-42: I ran this scenario https://github.com/stackforge/rally/blob/master/samples/tasks/scenarios/nova/boot-and-delete.json when I do rally task results I get  these samples https://gist.github.com/meteorfox/8057af6eea14dd8d672c which seems to be the aggregated concurrent boot times
19:04 meteorfox is that correct?
19:08 psd boris-42, Hey!
19:09 boris-42 meteorfox: there is no aggreagtion
19:09 boris-42 meteorfox: in "rally task results"
19:09 boris-42 meteorfox: it shows full information for every iteration
19:10 boris-42 meteorfox: iterations contains from boot and delete server
19:10 boris-42 meteorfox: this was performed for one user
19:10 boris-42 meteorfox: from*
19:11 meteorfox boris-42: I would have expected to see 20 samples, since we are running 10 with 2 concurrent boots, but the count on the reports shows only 10
19:12 boris-42 meteorfox: you can't expect 20
19:12 boris-42 meteorfox: if you put to run scenario 10 times
19:12 boris-42 meteorfox: concurency is the amount of parallel running not amount of total
19:12 boris-42 psd: hey hey
19:13 meteorfox boris-42: so the boot  and delete times shown is the whole time for the 2 concurrent boots?
19:13 boris-42 meteorfox: nope
19:13 boris-42 meteorfox: there is NO aggregation
19:13 boris-42 meteorfox: it just runs 10 times python method NovaServers.boot_and_delete
19:14 psd boris-42, got that create and delete benchmark merged today, you got something else for me?
19:14 boris-42 psd: thank you for work on it!
19:14 boris-42 psd: yep we can find a lot of tasks
19:14 psd boris-42, About the adding cloud info task, how should I proceed?
19:14 boris-42 psd: ah I need to find some time to see the full picture
19:15 boris-42 psd: and then I will split that work on smaller steps
19:15 psd boris-42, that would be great
19:15 boris-42 meteorfox: so you are running 10 times that python method
19:15 boris-42 meteorfox: (and at any moment you have 2 running of them)
19:15 meteorfox boris-42: ok I see
19:16 boris-42 meteorfox: you are collecting all data that is possible (atomic actions, scenario duration, errors and so on)
19:16 boris-42 meteorfox: and putting them to arrya
19:16 boris-42 meteorfox: "constant" is just one of type of load that is supported out of box of rally
19:16 meteorfox boris-42: now I get it
19:16 meteorfox I thought it was 10 x 2, each iteration builds runs two boots
19:17 boris-42 meteorfox: nope
19:17 meteorfox but is actually 10 / 2
19:17 meteorfox or times / concurrency runs
19:17 boris-42 meteorfox: nope
19:17 boris-42 times is amount of runs
19:17 boris-42 it is independent from concurrency at all
19:18 boris-42 concurrency is just load level *
19:18 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
19:19 meteorfox boris-42: ok like a threshold in max concurrent ?
19:19 boris-42 meteorfox: ?
19:20 boris-42 meteorfox: it's much simpler
19:20 openstackgerrit joined #openstack-rally
19:20 boris-42 you have times - amount of calls of some method
19:20 boris-42 it's total
19:20 boris-42 so it will be exactly that number
19:20 meteorfox boris-42: Like, I have 10 jobs to run, but there can only be 2 of those running maximum at any given time
19:20 boris-42 meteorfox: yep
19:21 boris-42 meteorfox: every job is boot and delete vm
19:21 boris-42 meteorfox: from one randomly chosen user
19:21 meteorfox boris-42: right
19:21 meteorfox boris-42: makes sense
19:21 boris-42 meteorfox: so there are 3 more runners
19:21 psd boris-42, do you suggest some task meanwhile?
19:21 abhirc joined #openstack-rally
19:22 boris-42 meteorfox: constant_for_duration (that accept duration instead of times)
19:22 boris-42 meteorfox: rps (request per second - in this case scenarios per sec)
19:22 boris-42 meteorfox: that accept times (and how much scenarios to start every second )
19:22 boris-42 meteorfox: and serial (it's just for loop)
19:23 meteorfox boris-42: cool, like open model workloads (rps)
19:23 boris-42 meteorfox: so you can take a look at code here https://github.com/stackforge/rally/blob/master/rally/benchmark/runners/serial.py#L52-L58
19:23 boris-42 meteorfox: actually load generators are plugins
19:23 boris-42 meteorfox: so if you need to generate some specific load you can do it=)
19:24 meteorfox boris-42: cool, cool
19:24 meteorfox boris-42: another question, so say, I want the other scenario, where I'm measuring like a cluster boot, where I want the end-to-end time of building N servers
19:26 psd_ joined #openstack-rally
19:29 meteorfox boris-42: the use case is, that most people build a cluster of servers not just 1 (e.g. web app, and DB, etc). We want to measure how long it take to build such sets of servers, which could possibly have different images, and/or flavors
19:31 boris-42 meteorfox: so you can do it with heat template
19:31 boris-42 meteorfox: even now
19:31 meteorfox boris-42: right, but I want to skip heat
19:31 meteorfox I might not have heat
19:32 meteorfox I don't need the orchestration, just issues the boot command with different parameters, and wait until they are all ready
19:32 meteorfox maybe, something like NovaServers.cluster_boot_and_delete
19:36 meteorfox boris-42: what do you think?
19:39 exploreshaifali joined #openstack-rally
19:40 boris-42 meteorfox: soo
19:41 boris-42 meteorfox: you can make any kind of scenarios
19:41 boris-42 meteorfox: rally is framework
19:41 boris-42 meteorfox: so it can accept few images
19:41 boris-42 meteorfox: boot some VM with those
19:41 boris-42 meteorfox: run something stuff
19:41 boris-42 meteorfox: collect something
19:41 boris-42 meteorfox: and do the things that you need in parallel
19:42 boris-42 meteorfox: how that sound?
19:42 meteorfox boris-42: heh, right. Sounds cool, before I start implementing this, I just wanted to make sure there isn't something out there I could use.
19:43 meteorfox boris-42: how would you go around to implement this? Would you add a method under NovaServers scenarios?
19:44 meteorfox boris-42: or is there a different route?
19:49 pboldin joined #openstack-rally
19:59 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
19:59 openstackgerrit Sergey Skripnick proposed stackforge/rally: More verbose WrongStatusException  https://review.openstack.org/156737
20:19 boris-42 meteorfox: nope I will chose some other class
20:19 boris-42 meteorfox: maybe call it RealLifeScenarios ?)
20:20 boris-42 meteorfox: https://github.com/stackforge/rally/tree/master/rally/benchmark/scenarios adding here new directory
20:20 boris-42 meteorfox: reallife
20:20 boris-42 meteorfox: or something like that
20:40 andreykurilin_ joined #openstack-rally
20:43 openstackgerrit OpenStack Proposal Bot proposed stackforge/rally: Updated from global requirements  https://review.openstack.org/156761
20:48 openstackgerrit Sergey Skripnick proposed stackforge/rally: More verbose WrongStatusException  https://review.openstack.org/156737
21:10 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
21:39 himangi hey! I wish to start contributing to rally.. I would appreciate any help!
22:08 openstackgerrit Sergey Skripnick proposed stackforge/rally: Add new script for gate jobs  https://review.openstack.org/144633
22:36 boris-42 himangi: hey hey
22:39 himangi hi boris-42
22:39 boris-42 himangi: so what you would like to contribute?
22:43 himangi boris-42, I am a developer.. I have some previous experience working on the Linux kernel and am looking forward to work with openstack as a GSOC intern
22:50 boris-42 himangi: so we can start from some simple tasks (bugs)
22:50 himangi boris-42, sure
22:50 boris-42 himangi: do you have any ideas related to rally ? topics that you would like to work?
22:50 himangi boris-42: not right now
22:51 himangi boris-42: I have been to a lot of conferences and heard talks on openstack and found it really interesting
23:04 boris-42 himangi: and you chose to work on rally?)
23:05 openstackgerrit Merged stackforge/rally: Updated from global requirements  https://review.openstack.org/156761
23:07 himangi boris-42:  yes, It looked really cool and benchmarking is quite an important part of any project.. It so cool that openstack has a well-defined system for the same..
23:07 himangi It will help me learn about how openstack works in a better way
23:08 himangi I hope to work with openstack for a long term and rally seems to be a good place to start!
23:48 boris-42 himangi: ok great
23:49 boris-42 himangi: you can start from readthedocs and plaing with rally https://rally.readthedocs.org/en/latest/
23:49 boris-42 himangi: there is step by step guide
23:49 boris-42 himangi: after finishing you should get familiar with contribution process so we will find some simple bug to fix
23:49 boris-42 himangi: and then we can think about future stuff
23:49 boris-42 himangi: sorry but I ahve to go and sleep now=)
23:49 boris-42 it's too late here
23:50 himangi sure boris-42 .. good night :)
23:50 boris-42 himangi: see you

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