Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-12-21

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

All times shown according to UTC.

Time Nick Message
00:47 yingjun joined #openstack-rally
01:07 yingjun joined #openstack-rally
01:28 yingjun joined #openstack-rally
02:23 baker joined #openstack-rally
02:34 yingjun_ joined #openstack-rally
02:47 chenli joined #openstack-rally
02:58 yingjun joined #openstack-rally
04:01 chenli joined #openstack-rally
04:14 zerda joined #openstack-rally
04:36 psd joined #openstack-rally
05:04 chenli joined #openstack-rally
05:11 rdas joined #openstack-rally
05:22 tfreger joined #openstack-rally
05:34 shausy joined #openstack-rally
05:58 yingjun joined #openstack-rally
06:11 openstackgerrit Yan Ning proposed openstack/rally: Add monasca tests in plugin - Part 0: metrics  https://review.openstack.org/248298
06:20 openstackgerrit Yan Ning proposed openstack/rally: Add monasca tests in plugin - Part 0: metrics  https://review.openstack.org/248298
06:27 itzikb joined #openstack-rally
06:38 anshul joined #openstack-rally
06:52 yingjun joined #openstack-rally
07:20 anshul joined #openstack-rally
07:34 lezbar joined #openstack-rally
07:37 neeti joined #openstack-rally
07:43 lezbar__ joined #openstack-rally
08:02 anshul joined #openstack-rally
08:42 bochi-michael joined #openstack-rally
08:58 psuriset_ joined #openstack-rally
09:20 amaretskiy joined #openstack-rally
09:40 yfried joined #openstack-rally
09:45 BaQs_ joined #openstack-rally
09:48 hughsaunders_ joined #openstack-rally
09:48 lewo_ joined #openstack-rally
09:51 jlk joined #openstack-rally
09:51 jlk joined #openstack-rally
09:57 rzerda joined #openstack-rally
09:58 aarefiev_ joined #openstack-rally
10:01 akuznetsova_ joined #openstack-rally
10:02 hparekh__ joined #openstack-rally
10:03 kambiz_ joined #openstack-rally
10:04 dmellado_ joined #openstack-rally
10:06 raginbaj- joined #openstack-rally
10:07 coolsvap- joined #openstack-rally
10:12 yfried_ joined #openstack-rally
10:18 openstackgerrit joined #openstack-rally
10:22 rvasilets___ joined #openstack-rally
10:37 e0ne joined #openstack-rally
10:54 openstackgerrit joined #openstack-rally
10:58 adiantum joined #openstack-rally
11:04 jaypipes joined #openstack-rally
11:25 ikhudoshyn joined #openstack-rally
11:30 MaxPC joined #openstack-rally
12:02 boris-42_ joined #openstack-rally
12:04 tfreger1 joined #openstack-rally
12:05 tfreger1 joined #openstack-rally
12:07 amaretskiy redixin ping
12:07 amaretskiy redixin we told recently about try/except in get_rows
12:07 amaretskiy i know what you are dislike
12:08 redixin cool
12:08 amaretskiy however before putting -1 please just look at old code
12:08 amaretskiy and think about problems of old code that are not a topic of this explicit patch
12:09 amaretskiy the patch is simply does old trick a bit better, and solving the trick completely is not a topic of the patch
12:10 amaretskiy let's do the work by principle ` task === 1 patch
12:10 amaretskiy *1 task === 1 patch
12:10 shausy joined #openstack-rally
12:10 amaretskiy redixin ^
12:11 redixin the problem is that some methods are raising RallyException instead of returning None?
12:11 redixin amaretskiy: ^
12:14 redixin is this possible to make all algorithms behave identically? e.g. all return None if no data. or all raise NoData exception
12:28 aix joined #openstack-rally
12:36 boris-42 amaretskiy: Btw
12:37 boris-42 amaretskiy: let's make our reports better
12:37 boris-42 amaretskiy: let's remove graphs from reports if there is only 1 iteration
12:37 boris-42 amaretskiy: it's useless information
12:37 boris-42 amaretskiy: and looks quite ugly like here https://twitter.com/samadhanam/status/678882548844482562
12:39 boris-42 amaretskiy: doesn't seem to be a hard change?
12:44 amaretskiy boris-42 sure, will do this
12:45 amaretskiy redixin this change (returning None instead of raise) was discussed long time ago
12:46 amaretskiy redixin and this change of streaming algorithms is not related to spec implementation - it can be done any time before or after output-related patches are merged
12:47 boris-42 ikhudoshyn: ping
12:47 ikhudoshyn boris-42, sup
12:47 boris-42 amaretskiy: btw we should renmate streaming algroitms file
12:47 boris-42 amaretskiy: to just streaming
12:47 boris-42 ikhudoshyn: so do you have any progress related to tasks for newbies?
12:48 adiantum joined #openstack-rally
12:48 amaretskiy boris-42 I will implement this change (do not display chart for 1 iteration) with changes related to output - this is actually a very small change
12:48 ikhudoshyn boris-42, not exactly
12:48 boris-42 amaretskiy: why not just separated change hm?
12:48 boris-42 ikhudoshyn:  ?)
12:48 redixin fix master, and then write simple code without any workarounds or write new code with workarounds and then fix new code and then remove workarounds
12:49 ikhudoshyn boris-42, I started diggin thru the code base
12:49 ikhudoshyn boris-42, now I only have a vision about parts I've been tinkering with
12:50 ikhudoshyn boris-42, unfortunately there are not many parts that could be easily given to newbies
12:50 ikhudoshyn boris-42, but i'm working on it)
12:50 boris-42 ikhudoshyn: ok
12:50 boris-42 ikhudoshyn: there is bunch of bad places
12:50 redixin amaretskiy: why not fix small nits before writing your new code?
12:50 boris-42 ikhudoshyn: with bad tests / not covered
12:50 boris-42 ikhudoshyn: or bad names
12:51 boris-42 ikhudoshyn: or just ugly hacks that should be refactored
12:51 boris-42 ikhudoshyn: one of task can be like remove reports if there is only 1 iteration
12:51 boris-42 ikhudoshyn: it can be done by newbie instead of amaretskiy ^
12:51 boris-42 ikhudoshyn: you can ask everybody from team about tasks that they know about and just collect them
12:51 amaretskiy redixin: because this is not a samll nit, this change (None or raise) required discussion
12:52 ikhudoshyn boris-42, remove reports?
12:52 amaretskiy because that was discussed before
12:52 redixin amaretskiy: so why not just discuss and fix that before writing shit over existing shit?
12:52 redixin openstack style
12:53 amaretskiy boris-42 redixin: we have a discusstion with redixin regarding Nobe vs raise in streaming algorithms when no data is processed
12:53 ikhudoshyn boris-42, will do. I guess I'll collect them as lo-prio tasks in launchpad
12:53 lane_kong left #openstack-rally
12:53 amaretskiy redixin as far as I remember this was discussed with boris-42
12:54 amaretskiy redixin: this would be pitty for me to solve some unrelated stuff while doing something concrete
12:54 akrzos joined #openstack-rally
12:54 amaretskiy redixin: None vs rais ein streaming algorithms - this is something completely nor related
12:55 amaretskiy redixin - this change is simple - you can submit a patch
12:55 amaretskiy redixin and the problem will be solved while the patch is merged - no matter before or after my patches
12:55 redixin amaretskiy: it will produce a merge conflicts
12:56 amaretskiy redixin is that a problem? and will this conflict will be problematic?
12:56 amaretskiy redixin merge conflict will be a quite simple
12:57 amaretskiy redixin i just want to explain that i do not see advantages to do something unrelated while implementing concrete tasks
12:58 amaretskiy redixin this would be 1) extra patch set in long patches chain 2) this requires discussion because this *already* was discussed nad 3) id discussion will throw this change away - it will affect my concrete task and make a delay
12:58 redixin amaretskiy: i just want to explain, that is better to fix nits before writing code with workarounds. later you may write simple nice code without any workarounds
12:58 amaretskiy redixin please see this resons
12:59 amaretskiy redixin let's move this to out today meeting, this is a good topic
13:00 redixin amaretskiy: if you upload patchset with obvious nits, you may wait a long time before it get merged
13:00 redixin instead you may fix nits before. it will merge faster overall
13:01 redixin amaretskiy: but it is just my opinion. you may just ignore it, and continue to wait for +2 from someone else
13:01 amaretskiy redixin: i say this again - using `raise' *was discussed* long time ago
13:01 amaretskiy redixin so I can not just make a patch and switch code to None
13:02 redixin amaretskiy: so ask people who was discussing it for +2
13:02 amaretskiy redixin - what does stop you from making a patch? merge confloct will be quite simple
13:03 amaretskiy redixin: this is your idea (using None) so this looks like your task
13:03 amaretskiy redixin: not mine
13:03 amaretskiy redixin: however I'm not thinking that using None is bad idea, just do not want to waste time
13:04 amaretskiy redixin: on something that delays a huge improvement
13:04 redixin amaretskiy: you just want us to +2 code with nits?
13:05 amaretskiy I do not understand what is a nit we are discussing - is there using raise RallyException from streaming algorithms?
13:06 amaretskiy redixin ^
13:06 redixin nit is streaming algorithms behave different
13:06 amaretskiy redixin: I do not understand
13:06 redixin if isinstance(algo1) then do something
13:07 redixin thing for collecting data from batch of algo objets should know details
13:07 amaretskiy redixin - I partially agree with comment on code - there will be another patch set
13:08 amaretskiy redixin: but I will relay on `except RallyException' while calling algo.result() untill streaming algorithms raise this exceptions
13:09 amaretskiy redixin: streaming algorithms are not changed by my patches
13:09 amaretskiy redixin: thsi is complete different part of code that I do not touch in my improvement of output
13:10 redixin amaretskiy: nit in streaming algorithms makes your code more complex. why not just fix streaming algorithms before writing new code?
13:10 amaretskiy redixin: give examples
13:11 redixin amaretskiy: or you may even fix streaming algorithms in the same patchset
13:11 amaretskiy redixin: fix what? please explain
13:11 redixin oh forget it
13:11 amaretskiy redixin: ok
13:11 amaretskiy redixin: I will move a part of get_rows() to new method
13:12 redixin i dont care
13:12 amaretskiy redixin: to make class more flexible
13:12 redixin why shoud i review this? this is not my idea
13:12 amaretskiy redixin: I hope you are open for dialog
13:14 amaretskiy redixin: do you insist on returning None from algo.result() (instead of raising RallyException) if there is no data? If yes I will create a topic on our meeting
13:14 redixin I understand your position. You are changing only reports stuff and not interested in any other parts of code.
13:16 openstackgerrit Lee Yarwood proposed openstack/rally: install_rally.sh: Add redhat-rpm-config to package list for EL distros  https://review.openstack.org/260000
13:18 tosky joined #openstack-rally
13:19 openstackgerrit Illia Khudoshyn proposed openstack/rally: Prepare for distribured runner, part 2  https://review.openstack.org/254780
13:38 gubouvier joined #openstack-rally
13:42 openstackgerrit Yaroslav Lobankov proposed openstack/rally: [Verify] Adding "xfail" mechanism for Tempest tests  https://review.openstack.org/260013
13:50 redixin rvasilets ping
13:51 redixin rvasilets___: hiyo
13:51 rvasilets hi
13:51 rvasilets redixin, ^
13:52 redixin rvasilets: wee also need to move rally sqlite file to /opt/rally in dockerfile
13:52 redixin rvasilets: do you want to to it in your patch?
13:52 redixin rvasilets: or I can make another one. there will be conflicts
13:53 rvasilets redixin, Ok, I will do this
13:53 redixin or wait
13:54 redixin hmm im not sure if this file should be at /home/rally or not
13:54 redixin it should be either at /opt/rally or /home/rally/.rally; the problem is it shoud not be directly in /home/rally
13:55 redixin rvasilets: ^
13:55 rvasilets then  as for me better in /opt/rally
13:56 rvasilets if it would be all ok with permitions  to db
13:56 redixin rvasilets: it is not too useful. user should commit all changes after exited from container
13:56 redixin or keep database in volume (-v /home/rally)
13:56 rvasilets Does user need to commit db?
13:56 redixin I mean commit changes to docker container
13:57 redixin docker will discard all changes in filesystem
13:57 rvasilets Then there is the same problem for samples ant all that locateed in /opt/rally/
13:57 redixin samples are read-only
13:58 redixin it just samples. nobody want to edit samples
13:58 rvasilets oO I thought if I biuld my own container then I can do everything with it
13:59 redixin you can edit /bin/sh on your system, but you dont want to do it
13:59 redixin or edit /usr/share/docs/whatever
13:59 redixin but you dont want to do it
13:59 redixin the same for samples. you can copy sample and edit it. but not to edit in place
14:00 rvasilets Ok. To summarize all this) What I need to do? - To move db to proper place
14:00 rvasilets Yes?
14:00 redixin I was wrong when decided to move db to /opt/rally
14:00 rvasilets But place isn't choosed yet
14:01 redixin it should be in /home/rally/.rally/
14:01 redixin or even deeper
14:01 redixin /home/rally/.rally/db/
14:02 rvasilets https://allisonmaruska.files.wordpress.com/2015/10/we-need-to-go-deeper.jpg
14:02 andreykurilin boris-42: meeting?
14:02 rvasilets Ok where db would be located now?
14:02 redixin rvasilets: but there is another problem -_-
14:03 redixin we need to go deeper -_-
14:03 rvasilets )
14:04 redixin rvasilets: rally should create directory when recreating database
14:04 rvasilets redixin, Now db is here yes? https://github.com/openstack/rally/blob/master/install_rally.sh#L40
14:04 redixin this is default value
14:05 redixin rvasilets: Ill try to fix creating underlying directory and then ping you again
14:05 rvasilets or here https://github.com/openstack/rally/blob/master/install_rally.sh#L619 ...
14:05 rvasilets redixin, ok
14:40 openstackgerrit Sergey Skripnick proposed openstack/rally: Do not put sqlite file in /home/rally  https://review.openstack.org/260033
14:45 gsagie Hello all, is there any good report that can show me what failed? (http://logs.openstack.org/19/260019/2/check/gate-dragonflow-dsvm-rally-nv/01b7ef2/logs/)
14:46 gsagie or something graphical like testr for tempest
14:47 boris-42 gsagie: hmmm
14:47 boris-42 gsagie: is it your implementation of rally job?
14:48 boris-42 gsagie: this is how it should look like http://logs.openstack.org/80/254780/15/check/gate-rally-dsvm-rally/2dd0cfa/
14:48 redixin gsagie: http://logs.openstack.org/19/260019/2/check/gate-dragonflow-dsvm-rally-nv/01b7ef2/console.html#_2015-12-21_14_29_55_108
14:48 boris-42 redixin: however for some reason I don't see reports
14:49 redixin gsagie: also it seems report is generated http://logs.openstack.org/19/260019/2/check/gate-dragonflow-dsvm-rally-nv/01b7ef2/console.html#_2015-12-21_14_29_53_213
14:49 redixin but not published for some reason
14:49 boris-42 redixin: no reports here http://logs.openstack.org/19/260019/2/check/gate-dragonflow-dsvm-rally-nv/01b7ef2/
14:49 boris-42 redixin: seems like they forgot to put publishing bot
14:50 redixin > (16:49:14) redixin: but not published
14:53 redixin gsagie: you need something like this https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/rally.yaml#L29-L42
14:53 redixin gsagie: to add here https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/dragonflow.yaml#L90
14:58 redixin boris-42: 'rally-manage db recreate' should create underlying directory for sqlite file
14:58 redixin if it not exists
15:00 gsagie redixin: ok, thanks
15:05 adiantum joined #openstack-rally
15:09 stpierre joined #openstack-rally
15:15 baker joined #openstack-rally
15:16 psd How do I run a particular test with tox?
15:18 stpierre tox -e py27 -- full.name.of.test.module.and.class.and.function
15:19 stpierre if you just specify a class it'll run all tests in that class, and if you just specify a module it'll run all tests in that module, and so on
15:19 stpierre you might also find 'tox -e py27 -- --failing' useful
15:19 stpierre obv. s/py27/py34/ to run the tests with py3k instead of python2
15:20 psd Thanks stpierre!
15:23 stpierre joined #openstack-rally
15:33 boris-42 stpierre: so we discussed today during the meeting
15:33 boris-42 stpierre: about wait_for
15:35 openstack joined #openstack-rally
15:36 openstackstatus joined #openstack-rally
15:42 openstackgerrit Yan Ning proposed openstack/rally: Add monasca tests in plugin - Part 0: metrics  https://review.openstack.org/248298
15:48 stpierre boris-42: yeah, sorry i missed it. it's a tough time
15:49 boris-42 stpierre: it is
15:49 boris-42 stpierre: I know
15:49 boris-42 stpierre: monday 6.a.m is bad time for meeting
15:49 stpierre yeah, i know it's worse for you
15:49 stpierre looks like the logs aren't up yet -- what'd you decide?
16:12 jose_lausuch joined #openstack-rally
16:13 jose_lausuch boris-42: is it possible to install tempest from rally (rally verify install) without having actual connectivity to the openstack deployment. neither rc file sourced?
16:14 boris-42 ylobankov: andreykurilin ^
16:14 boris-42 jose_lausuch: so actually it is not
16:14 boris-42 jose_lausuch: the idea of rally verify install was to create seperated tempest installation for each deployment that you have in rally
16:14 boris-42 jose_lausuch: so we will need to change rally to cover this use case
16:15 jose_lausuch ok
16:15 jose_lausuch I was thinking of having a preinstalled virtual environment for tempest, and a command to tweak it for the deployment X
16:15 jose_lausuch but thats fine
16:16 jose_lausuch and can I run "install_rally.sh --yes" in the same situation?
16:16 openstackgerrit Yaroslav Lobankov proposed openstack/rally: [Verify] Adding "xfail" mechanism for Tempest tests  https://review.openstack.org/260013
16:17 boris-42 jose_lausuch: so
16:17 boris-42 jose_lausuch: why it's hard to run 2 commands instead of one?
16:17 boris-42 jose_lausuch: rally deployment create, rally verify start ?
16:18 jose_lausuch I would like to have as many preinstalled stuff as possible to reduce test time later on
16:18 jose_lausuch I know that for rally deployment create, I need the deployment
16:18 jose_lausuch but for install_rally?
16:20 boris-42 jose_lausuch: ?
16:20 boris-42 jose_lausuch: install_rally is just a bash script that does things that can't be done by pip
16:20 boris-42 jose_lausuch: but overall it's just get rally installe
16:20 jose_lausuch ok
16:21 jose_lausuch so, it just installs all the needed libraries
16:21 jose_lausuch ok, thanks boris-42
16:21 lutostag joined #openstack-rally
16:22 boris-42 jose_lausuch: and system packages
16:22 boris-42 jose_lausuch: and bash completition
16:22 boris-42 jose_lausuch: and set's up DB
16:22 boris-42 jose_lausuch: so what is your full use case?
16:23 boris-42 jose_lausuch: maybe we will be able to do something for it
16:24 jose_lausuch I want to include as many preinstalled stuff in a docker image (portable) for OPNFV
16:24 jose_lausuch so, for now, I will include the installation of rally in the image
16:25 jose_lausuch instead of running the command every time we run our tests
16:31 boris-42 jose_lausuch: so the most time consuming is install_rally.sh
16:31 boris-42 jose_lausuch: we can make a command that will install tempest
16:31 boris-42 jose_lausuch: then you won't have time comusing operations
16:31 boris-42 jose_lausuch: e.g. rally verify start will just copy paste already installed tempest whcih will work quite fast
16:32 jose_lausuch mmm, would that be a big issue? :)
16:43 openstackgerrit Merged openstack/rally: Designate V2 support - part 2  https://review.openstack.org/249313
16:47 boris-42 jose_lausuch: nope it won't
16:56 jose_lausuch boris-42: that would be great then :)
16:57 jose_lausuch but its not blocking though..
17:02 madhu_ak joined #openstack-rally
17:07 openstackgerrit Alexander Maretskiy proposed openstack/rally: [StreamingAlgorithms] result() returns None if data is missing  https://review.openstack.org/260114
17:13 openstackgerrit Alexander Maretskiy proposed openstack/rally: [StreamingAlgorithms] result() returns None if data is missing  https://review.openstack.org/260114
17:18 stpierre_ joined #openstack-rally
17:21 openstackgerrit Merged openstack/rally: Fixed Glance Certification part  https://review.openstack.org/259145
17:31 lgreg1 left #openstack-rally
17:40 stpierre__ joined #openstack-rally
17:41 redixin joined #openstack-rally
17:43 stpierre joined #openstack-rally
17:58 ningy joined #openstack-rally
18:14 openstackgerrit OpenStack Proposal Bot proposed openstack/rally: Updated from global requirements  https://review.openstack.org/253000
18:34 e0ne joined #openstack-rally
18:46 psuriset_ joined #openstack-rally
18:52 rvasilets___ joined #openstack-rally
19:01 pvaneck joined #openstack-rally
19:56 lezbar joined #openstack-rally
20:29 e0ne joined #openstack-rally
20:38 e0ne joined #openstack-rally
20:50 e0ne joined #openstack-rally
21:07 Aish joined #openstack-rally
21:23 redixin joined #openstack-rally
21:53 Aish joined #openstack-rally
22:52 Aish joined #openstack-rally
23:19 openstackgerrit Merged openstack/rally: install_rally.sh: Add redhat-rpm-config to package list for EL distros  https://review.openstack.org/260000
23:30 rook joined #openstack-rally
23:58 baker joined #openstack-rally

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