Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2016-03-31

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

All times shown according to UTC.

Time Nick Message
00:22 diogogmt joined #openstack-rally
00:35 rpromyshlennikov joined #openstack-rally
00:37 mvk_ joined #openstack-rally
00:39 omar_rally_noob joined #openstack-rally
00:43 omar_rally_noob how do you know exactly what a task is doing for example https://github.com/openstack/rally/blob/master/samples/tasks/scenarios/authenticate/keystone.json I understand the task is going to create 3 tenants with 50 users in each tenant and auth 100 times with 5db transactions at a time.
00:44 omar_rally_noob does the task produce logs or detailed output, all i get are results from running these tasks
00:49 yingjun joined #openstack-rally
00:54 stpierre joined #openstack-rally
01:06 boris-42 omar_rally_noob: hi
01:07 baohua joined #openstack-rally
01:07 boris-42 omar_rally_noob: you can run "rally task results" to get all data in JSON or "rally task report" to get HTML reports
01:07 boris-42 omar_rally_noob: as well you can run rally -v task start ... to get logs of running rally
01:07 boris-42 or -vd to get debug
01:48 ilbot3 joined #openstack-rally
01:48 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
01:58 yanyanhu joined #openstack-rally
02:07 akrzos joined #openstack-rally
02:32 openstackgerrit Yanyan Hu proposed openstack/rally: Add basic support for Senlin service  https://review.openstack.org/298109
02:57 openstackgerrit Merged openstack/rally: Change deployment model schema  https://review.openstack.org/276821
03:06 yanyanhu joined #openstack-rally
03:27 ajo_ joined #openstack-rally
03:29 rpromysh1 joined #openstack-rally
03:31 dmellado_ joined #openstack-rally
03:39 serverascode joined #openstack-rally
04:15 dratushnyy joined #openstack-rally
04:38 rdas joined #openstack-rally
04:59 shausy joined #openstack-rally
04:59 rcernin joined #openstack-rally
05:09 dratushnyy joined #openstack-rally
05:28 sverma joined #openstack-rally
05:39 shausy joined #openstack-rally
05:42 dratushnyy joined #openstack-rally
05:53 lazy_prince joined #openstack-rally
05:56 e0ne joined #openstack-rally
05:57 openstackgerrit Boris Pavlovic proposed openstack/rally: Reduce default speed of users creation in users context  https://review.openstack.org/299757
05:57 rdas joined #openstack-rally
06:09 ushkalim joined #openstack-rally
06:24 lazy_prince joined #openstack-rally
06:32 winston-d_ joined #openstack-rally
06:36 bkopilov joined #openstack-rally
06:49 tesseract joined #openstack-rally
06:51 akamyshnikova joined #openstack-rally
07:15 e0ne joined #openstack-rally
07:24 abehl joined #openstack-rally
07:30 arxcruz joined #openstack-rally
07:41 pcaruana joined #openstack-rally
07:48 permalac joined #openstack-rally
07:48 openstackgerrit Boris Pavlovic proposed openstack/rally: Calculate duration of load more accurate  https://review.openstack.org/298590
07:55 openstackgerrit Srinivas Sakhamuri proposed openstack/rally: Use glance wrapper for nova and cinder  https://review.openstack.org/298787
07:57 klindgren joined #openstack-rally
07:57 openstackgerrit Merged openstack/rally: Reduce default speed of users creation in users context  https://review.openstack.org/299757
07:59 openstackgerrit Merged openstack/rally: Actually delete result_batch arrays  https://review.openstack.org/298636
08:02 amaretskiy joined #openstack-rally
08:25 permalac joined #openstack-rally
08:27 rpromyshlennikov joined #openstack-rally
08:28 e0ne joined #openstack-rally
08:32 openstackgerrit Merged openstack/rally: [Reports] Various fixes for load profile chart  https://review.openstack.org/298622
08:34 e0ne joined #openstack-rally
08:38 adiantum joined #openstack-rally
08:43 openstackgerrit Rodion Promyshlennikov proposed openstack/rally: Move Credentials initialization to OSClients  https://review.openstack.org/294520
08:43 openstackgerrit Rodion Promyshlennikov proposed openstack/rally: Make Credential class plugin based  https://review.openstack.org/294521
08:44 rdas joined #openstack-rally
08:44 openstackgerrit Rodion Promyshlennikov proposed openstack/rally: Add support for multi-credentials deployment  https://review.openstack.org/297202
08:48 adiantum joined #openstack-rally
09:20 openstackgerrit Andrey Kurilin proposed openstack/rally: DO NOT MERGE: print more debug messages for neutron scenario  https://review.openstack.org/288522
09:30 e0ne joined #openstack-rally
09:40 openstackgerrit Merged openstack/rally: Calculate duration of load more accurate  https://review.openstack.org/298590
09:42 e0ne joined #openstack-rally
09:44 mkrcmari__ joined #openstack-rally
09:45 tosky joined #openstack-rally
09:46 openstackgerrit Boris Pavlovic proposed openstack/rally: [WIP] OMG performance what are you doing? stohp please!!  https://review.openstack.org/298591
09:46 openstackgerrit Boris Pavlovic proposed openstack/rally: [Reports] Fix calc of running iterations in parallel  https://review.openstack.org/299833
09:52 mvk_ joined #openstack-rally
09:55 yanyanhu joined #openstack-rally
10:13 mkrcmari__ joined #openstack-rally
10:17 e0ne joined #openstack-rally
10:20 klindgren joined #openstack-rally
10:21 e0ne joined #openstack-rally
10:23 e0ne joined #openstack-rally
10:28 mvk_ joined #openstack-rally
11:17 e0ne joined #openstack-rally
11:23 e0ne joined #openstack-rally
11:24 bigjools joined #openstack-rally
11:24 bigjools joined #openstack-rally
11:26 sverma joined #openstack-rally
11:33 sverma joined #openstack-rally
11:33 openstackgerrit Merged openstack/rally: devstack: support to be enabled with different plugin name  https://review.openstack.org/296239
11:35 e0ne joined #openstack-rally
11:38 openstackgerrit Piyush Raman Srivastava proposed openstack/rally: Show Missing Argument Error raised by Argparse  https://review.openstack.org/298349
11:40 openstackgerrit Dmitry Ratushnyy proposed openstack/rally: [CLI] Display task errors in human-friendly form  https://review.openstack.org/294727
11:41 openstackgerrit Chris St. Pierre proposed openstack/rally: Add Cinder volume_types context  https://review.openstack.org/295403
11:54 akrzos joined #openstack-rally
11:57 yingjun joined #openstack-rally
11:59 e0ne joined #openstack-rally
12:05 adiantum joined #openstack-rally
12:11 e0ne joined #openstack-rally
12:27 akrzos joined #openstack-rally
12:44 MaxPC joined #openstack-rally
12:45 openstackgerrit Illia Khudoshyn proposed openstack/rally: Refactor DB shema for task results  https://review.openstack.org/297020
12:48 openstackgerrit Alexander Maretskiy proposed openstack/rally: [Reports] Fix calc of running iterations in parallel  https://review.openstack.org/299833
12:50 e0ne joined #openstack-rally
12:51 openstackgerrit Alexander Maretskiy proposed openstack/rally: [WIP] OMG performance what are you doing? stohp please!!  https://review.openstack.org/298591
13:02 openstackgerrit Merged openstack/rally: [Reports] Refactor processing.plot.plot  https://review.openstack.org/294119
13:02 openstackgerrit Merged openstack/rally: [Reports] Move angular directive to separate file  https://review.openstack.org/298597
13:04 openstackgerrit Andrey Kurilin proposed openstack/rally: Use existing_users in TestTaskSamples  https://review.openstack.org/299919
13:20 yingjun joined #openstack-rally
13:22 baohua joined #openstack-rally
13:24 stpierre joined #openstack-rally
13:29 MaxPC joined #openstack-rally
13:32 e0ne joined #openstack-rally
13:35 diogogmt joined #openstack-rally
13:38 dpaterson joined #openstack-rally
13:40 moravec joined #openstack-rally
13:41 openstackgerrit Yanyan Hu proposed openstack/rally: Add basic support for Senlin service  https://review.openstack.org/298109
13:45 permalac joined #openstack-rally
13:46 klkumar joined #openstack-rally
13:47 arnoldje joined #openstack-rally
13:47 kbaikov joined #openstack-rally
13:54 e0ne joined #openstack-rally
13:54 baohua joined #openstack-rally
14:01 pkarikh joined #openstack-rally
14:02 pkarikh left #openstack-rally
14:05 rook-lappy joined #openstack-rally
14:06 openstackgerrit Andrey Kurilin proposed openstack/rally: Use existing_users in TestTaskSamples  https://review.openstack.org/299919
14:08 baohua joined #openstack-rally
14:17 ushkalim joined #openstack-rally
14:18 baohua_ joined #openstack-rally
14:29 e0ne joined #openstack-rally
14:38 openstackgerrit Roman Vasilets proposed openstack/rally: Add exact error message into runcommand_heat scenario  https://review.openstack.org/299999
14:41 openstackgerrit Roman Vasilets proposed openstack/rally: [Docs] Anniversary patch  https://review.openstack.org/300003
14:42 mvk_ joined #openstack-rally
14:45 e0ne joined #openstack-rally
14:46 openstackgerrit Srinivas Sakhamuri proposed openstack/rally: Make glance web uploads streamable  https://review.openstack.org/297808
14:51 zhenq joined #openstack-rally
15:00 yingjun joined #openstack-rally
15:07 openstackgerrit Andrey Kurilin proposed openstack/rally: Use existing_users in TestTaskSamples  https://review.openstack.org/299919
15:13 ushkalim joined #openstack-rally
15:27 baohua joined #openstack-rally
15:37 e0ne joined #openstack-rally
15:44 e0ne joined #openstack-rally
15:47 e0ne joined #openstack-rally
15:53 klkumar joined #openstack-rally
15:53 harshs joined #openstack-rally
15:55 jaypipes joined #openstack-rally
16:11 e0ne joined #openstack-rally
16:30 davideagnello joined #openstack-rally
16:42 pvaneck joined #openstack-rally
16:47 e0ne joined #openstack-rally
17:01 diogogmt joined #openstack-rally
17:05 tosky joined #openstack-rally
17:28 openstackgerrit Merged openstack/rally: Fix validate_args function in cliutils  https://review.openstack.org/293569
17:34 dratushnyy joined #openstack-rally
17:37 boris-42 @stpierre hi
17:37 boris-42 @stpierre http://logs.openstack.org/33/299833/2/check/gate-rally-dsvm-neutron-rally/c064f03/rally-plot/resources_diff.txt.gz
17:37 boris-42 stpierre: ^ this looks bad
17:40 stpierre i don't think that's a new problem
17:43 stpierre looks like it may have somehow been caused by https://review.openstack.org/#/c/272638
17:45 stpierre yep, loads of tracebacks here: http://logs.openstack.org/38/272638/14/check/gate-rally-dsvm-neutron-rally/40927c3/console.html.gz
17:46 boris-42 stpierre: please be very accurate when you are working on such things like cleanup
17:46 boris-42 it's very easy to mess up everything
17:47 stpierre we need to get the resource checks to influence the success of the job
17:47 stpierre that review was +2'd by four separate cores
17:48 stpierre if five of us could overlook the same thing, that's a problem with the gate IMO
17:49 openstackstatus NOTICE: Job failures for missing traceroute packages are in the process of being fixed now, ETA 30 minutes to effectiveness for new jobs
17:50 boris-42 stpierre: the problem is that core team started thinking that if it pass gates patch is OK and works for any cases
17:50 stpierre we should strive to make that the case
17:50 stpierre people are really good at screwing up, and they always will be. if there's an unexpected increase in resources, that should fail the job.
17:50 boris-42 stpierre: it's imposible, there will be always ways to find how to break it without braking gates
17:51 stpierre that doesn't mean we should just give up
17:51 stpierre we should get as close as we can
17:51 boris-42 stpierre: so I agree that there is room for improvement for Ci
17:51 boris-42 stpierre: so the problem here is that WE (as a core team) KNOW that we are NOT testing cleanup very well
17:51 boris-42 in gates
17:51 stpierre well, i'm making a concrete suggestion for how to do it
17:52 stpierre so let's test it better. we've got the data, we just need to make it fail the job
17:52 boris-42 stpierre: yep we can hardocde the amount of possible failuers
17:52 boris-42 stpierre: however it won't be ideal
17:52 stpierre don't let the perfect be the enemy of the good. it'd be more ideal than it is right now
17:53 boris-42 can you try to analyze the reason of failuers?
17:54 stpierre it's pretty obvious
17:54 stpierre and is hopefully a one-line fix
17:55 boris-42 stpierre: ok I will wait for patch
17:55 boris-42 stpierre: btw
17:55 boris-42 https://github.com/openstack/rally/blob/master/tests/ci/osresources.py#L298
17:55 boris-42 stpierre: ^
17:56 boris-42 stpierre: so seems like we should hardcode here len of changes that we are allowing
17:56 boris-42 stpierre: or even do the better job and just fail if there are any changes
17:57 openstackgerrit Chris St. Pierre proposed openstack/rally: Cleanup: Fix getting name from neutron resources  https://review.openstack.org/300121
17:58 stpierre we need to make it a bit smarter, since there are some resources we expect -- 'default' security groups and the _member_ role at a minimum, it looks like
17:58 stpierre but i think we can probably filter those out, and then check for any additional changes
17:58 stpierre that'd be better than just failing on > 3 changes
17:58 stpierre i'll try to get that done this afternoon
18:14 boris-42 stpierre: thanks
18:16 diogogmt_ joined #openstack-rally
18:16 harshs joined #openstack-rally
18:17 shausy joined #openstack-rally
19:04 openstackgerrit joined #openstack-rally
19:04 openstackgerrit Chris St. Pierre proposed openstack/rally: Test: don't stringify cloud resources until printing  https://review.openstack.org/300157
19:04 openstackgerrit Chris St. Pierre proposed openstack/rally: Test: Distinguish expected added resources  https://review.openstack.org/300158
19:04 openstackgerrit Chris St. Pierre proposed openstack/rally: Test: Fail jobs if resources are added or removed  https://review.openstack.org/300159
19:26 openstackgerrit Kevin Benton proposed openstack/rally: Allow max_avg_sla per atomic actions  https://review.openstack.org/298462
19:28 boris-42 stpierre: ping
19:28 stpierre pong
19:28 boris-42 stpierre: I talk to kevinbenton
19:28 boris-42 stpierre: he said that it is known issue that you can't delete security group in neutron
19:28 boris-42 stpierre: so he will try to fix this on neutron side
19:28 stpierre right
19:29 stpierre same is true in nova-net
19:29 boris-42 stpierre: btw you patch works
19:29 boris-42 stpierre: going to merge it
19:29 boris-42 stpierre: so he checked he is saying that we need to delete security groups as a admin
19:29 boris-42 stpierre: so this seems like not a hard to fix stuff
19:30 stpierre ok
19:31 boris-42 stpierre: here is the patch https://review.openstack.org/#/c/296402/2
19:31 stpierre it looks like we set admin_required=True on the base NeutronMixin, but that isn't inherited
19:31 boris-42 stpierre: however it is not properly written
19:32 boris-42 stpierre: so actually it's a bit hard to fix bug
19:32 boris-42 stpierre: because if we are running with existing users and without admin (which will be possible in future) we shouldn't even try to delete default secruity group
19:32 boris-42 stpierre: seems like some refactoring will be required here
19:34 stpierre i dunno, lots of the cleanup resources require admin. it seems reasonable to just set admin_required on NeutronSecurityGroup for now. admittedly it's not perfectly forward-looking, but it should be a one-line fix for an actual bug, and it will be just one of a large list of things that require refactoring later.
19:35 stpierre i.e., we don't seem to lose anything by fixing it now
19:36 boris-42 stpierre: nope it is not
19:37 boris-42 stpierre: please let not make our life harder and harder
19:37 stpierre that doesn't make it any harder than it already is
19:37 boris-42 stpierre: each hack that we are adding is making rally less maintable
19:37 boris-42 stpierre: it makes it harder to make non-admin mode
19:37 boris-42 stpierre: I have an idea how to fix it
19:37 stpierre it fixes a real actual bug in one line, and there are twelve other places that already ahve admin_required
19:38 boris-42 stpierre: it creates another one
19:38 stpierre we can't just refuse to fix minor, simple bugs any time they conflict with some future idea that isn't even a spec yet
19:38 boris-42 stpierre: actually it creates 2 bugs
19:39 boris-42 1) if you are using existing users, you will delete default security group of it (which is bad idea)
19:39 boris-42 2) scenarios that don't require actually admin will start requiring it
19:39 stpierre that second one isn't a bug, because if they need to clean up security groups then they do actually require admin
19:39 boris-42 stpierre: sorry but I am already working 2 years for making non admin
19:40 boris-42 stpierre: and had to refactor whole Rally to achieve that
19:40 boris-42 stpierre: even plugin base was related to this task
19:40 boris-42 stpierre: and creating rally context*
19:40 stpierre i've gotta say, the amount of time we're spending spinning our wheels and not merging real fixes to actual bugs, or important features that are desperately needed, because of these distant pipe dream features is starting to get quite frustrating.
19:40 boris-42 stpierre: and current cleanups are trying to achieve this
19:41 boris-42 stpierre: so there is way to fix this
19:41 boris-42 stpierre: btw you don't need to be a admin to cleanup security groups
19:41 boris-42 stpierre: just for default security group
19:41 stpierre right, but if we want to do a complete job of cleanup, that's part of it.
19:42 boris-42 stpierre: so
19:42 boris-42 stpierre: there is another way
19:42 stpierre ok
19:42 boris-42 stpierre: btw
19:42 boris-42 stpierre: the code is rally
19:43 boris-42 stpierre: why this doesn't cleanup default security groups?
19:43 boris-42 https://github.com/openstack/rally/blob/master/rally/plugins/openstack/context/keystone/users.py#L124-L139
19:44 stpierre that is an excellent question
19:44 stpierre i suspect that *is* cleaning up most of them, since we're only seeing a single default security group in the osresources output
19:44 stpierre so the question is where that one stray secgroup is coming from
19:44 stpierre which might have a much simpler answer
19:47 boris-42 stpierre: so
19:48 boris-42 stpierre: it's interesting stuff
19:48 stpierre looks like both of the 'default' security groups are for the 'service' tenant
19:48 stpierre which is created by devstack
19:48 stpierre so it should show up in the pre-rally inventory AFAICT
19:55 openstackstatus NOTICE: Any jobs which erroneously failed on missing traceroute packages should be safe to recheck now
19:56 arnoldje joined #openstack-rally
20:08 boris-42 stpierre: btw could you take a look at this patch https://review.openstack.org/#/c/299833/2
20:08 boris-42 stpierre: it's like quite bad bug
20:16 openstackgerrit Merged openstack/rally: Cleanup: Fix getting name from neutron resources  https://review.openstack.org/300121
20:46 openstackgerrit Chris St. Pierre proposed openstack/rally: Test: Distinguish expected added resources  https://review.openstack.org/300158
20:46 openstackgerrit Chris St. Pierre proposed openstack/rally: Test: Fail jobs if resources are added or removed  https://review.openstack.org/300159
20:46 openstackgerrit Chris St. Pierre proposed openstack/rally: Test: don't stringify cloud resources until printing  https://review.openstack.org/300157
21:24 luzC left #openstack-rally
21:44 harshs joined #openstack-rally
22:12 openstackgerrit Merged openstack/rally: Allow max_avg_sla per atomic actions  https://review.openstack.org/298462
23:56 moravec joined #openstack-rally

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