Perl 6 - the future is here, just unevenly distributed

IRC log for #openstack-rally, 2015-04-13

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

All times shown according to UTC.

Time Nick Message
00:05 rook_ joined #openstack-rally
00:06 davideagnello joined #openstack-rally
00:15 prashantS joined #openstack-rally
00:34 dmorita joined #openstack-rally
00:35 prashantS joined #openstack-rally
02:26 panbalag joined #openstack-rally
03:00 prashantS joined #openstack-rally
03:12 subscope_ joined #openstack-rally
03:17 prashantS joined #openstack-rally
03:42 psd joined #openstack-rally
04:15 _kiran_ joined #openstack-rally
04:16 jjmb1 joined #openstack-rally
04:19 jjmb joined #openstack-rally
04:24 jjmb1 joined #openstack-rally
04:26 jjmb joined #openstack-rally
04:34 tfreger joined #openstack-rally
04:39 jjmb joined #openstack-rally
04:40 nkhare joined #openstack-rally
04:41 jjmb joined #openstack-rally
04:45 pradeep joined #openstack-rally
05:15 psd joined #openstack-rally
06:01 rdas joined #openstack-rally
06:08 aswadr joined #openstack-rally
06:15 yfried joined #openstack-rally
06:16 nmagnezi joined #openstack-rally
06:33 pbandzi joined #openstack-rally
06:41 e0ne joined #openstack-rally
07:08 igormarnat joined #openstack-rally
07:11 akuznetsova joined #openstack-rally
07:12 yfried boris-42: hi
07:12 pbandzi joined #openstack-rally
07:33 arxcruz joined #openstack-rally
07:40 fhubik joined #openstack-rally
08:01 karimb joined #openstack-rally
08:08 e0ne joined #openstack-rally
08:11 andreykurilin joined #openstack-rally
08:13 msdubov_ joined #openstack-rally
08:14 e0ne joined #openstack-rally
08:29 openstackgerrit Yair Fried proposed stackforge/rally: [spec] Refactoring scenario utils  https://review.openstack.org/172831
08:30 nkhare joined #openstack-rally
08:41 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
08:43 karimb joined #openstack-rally
08:44 yfried boris-42: are you around?
08:51 openstackgerrit Yair Fried proposed stackforge/rally: [spec] Refactoring scenario utils  https://review.openstack.org/172831
08:51 yfried andreykurilin: https://review.openstack.org/172831
08:51 yfried andreykurilin: could you please take another look?
08:51 * yfried hates rst
08:52 andreykurilin yfried: sure)
09:03 openstackgerrit Yair Fried proposed stackforge/rally: [spec] Refactoring scenario utils  https://review.openstack.org/172831
09:07 yfried andreykurilin: about your question re inheritance
09:08 andreykurilin yfried: ok
09:08 yfried andreykurilin: I think that we could take this opportunity to switch from inheritance to composition
09:08 yfried andreykurilin: instead of NeutronScenario(NeutrionUtils)
09:09 yfried NeutronScenario should have neutrion_utils as an attribute. what do you think?
09:10 boris-42 yfried: yep
09:10 andreykurilin yfried: if you want move to composition, you should mention it in spec:) I agreed that composition is better
09:10 yfried andreykurilin: I just thought of that. of course I will.
09:15 boris-42 andreykurilin: ping
09:15 andreykurilin boris-42: pong
09:15 boris-42 andreykurilin: what about sla stuff ?
09:15 boris-42 andreykurilin: I am almost ready with release notes
09:16 andreykurilin boris-42: heh. I working on it
09:16 boris-42 andreykurilin: do you need some help?
09:17 andreykurilin boris-42: no, thanks
09:18 boris-42 andreykurilin: ok=)
09:19 boris-42 andreykurilin: yfried btw what do you think guys if we start making Meeting Agenda? for each meeting?
09:20 andreykurilin boris-42: do we really need it? for example, nova meeting agenda sometimes is outdated and it is annoying
09:21 boris-42 andreykurilin: I mean now nobody knows what we are going to disucss
09:21 boris-42 andreykurilin: and how to propose changes
09:21 boris-42 in meeting agenda
09:23 yfried boris-42: andreykurilin: +1 on agenda
09:23 yfried I suggest we follow tempest example
09:23 aix joined #openstack-rally
09:24 yfried they have an agenda template and some sort of automated way to log and store each meeting
09:24 boris-42 yfried: no I dislike that approach
09:24 boris-42 yfried: what is the goal of such agenda?
09:24 boris-42 yfried: every week we are dicussing various stuff
09:25 boris-42 yfried: I mean it's almost equal to not having anything...
09:25 boris-42 yfried: I like TC agenda
09:25 yfried boris-42: I don't understand -
09:25 yfried [12:19] <boris-42> andreykurilin: yfried btw what do you think guys if we start making Meeting Agenda? for each meeting?
09:25 boris-42 yfried: they have just 5-6 points that they want to dicuss
09:25 boris-42 yfried: yep for each meeting
09:25 boris-42 yfried: like TC team
09:26 yfried boris-42: and how is that different from what I've just suggested as an example?
09:26 boris-42 yfried: http://paste.openstack.org/show/203479/
09:27 boris-42 yfried: it's absolutelly differnet
09:27 boris-42 yfried: QA agenda is useless http://paste.openstack.org/show/203480/
09:27 bkopilov joined #openstack-rally
09:27 bkopilov boris-42, Hi
09:27 boris-42 bkopilov: hi
09:28 boris-42 yfried: I mean QA is just template* and TC is what exactly they want to discuss
09:28 yfried boris-42: the QA template is edited each week to have specific agenda for the mtg
09:28 bkopilov boris-42, Quick question  , i want to run benchmark tests , whats the command from command line and how to generate report ?
09:29 bkopilov boris-42, i already have a setup connected to rally (existing cloud)
09:29 boris-42 bkopilov: https://rally.readthedocs.org/en/latest/tutorial/step_1_setting_up_env_and_running_benchmark_from_samples.html
09:29 bkopilov boris-42, cool thanks
09:29 yfried boris-42: but you don't have do define a template. I'm talking about haveing the agenda available for editing for all and logged for history
09:29 rook_ joined #openstack-rally
09:29 boris-42 yfried: ok having template makes sense
09:30 boris-42 yfried: but I was more talking about having Agenda (and not about how we are going to create it)
09:31 yfried boris-42: well having an agenda makes great sense
09:31 boris-42 yfried: are we going to add it like others to openstack wiki?
09:32 boris-42 yfried: and do we need to ping people via mailing list weekly ?
09:32 bkopilov boris-42, the rally task example is for samples , how can i call to python code (including all tests) as we do in tempest
09:33 bkopilov boris-42, i mean something like : rally task tests.py
09:33 boris-42 bkopilov: and what is goal of that?
09:33 boris-42 bkopilov: there is no way to execute random python code
09:33 openstackgerrit Nikita Konovalov proposed stackforge/rally: [Sahara] Improve Image context  https://review.openstack.org/166859
09:34 bkopilov boris-42, i see  , so how we run the gating tests from command line ?  dont we have a bulk of tests to be run
09:35 boris-42 bkopilov: rally is not unit test framework you should specify via yaml what you are going to run
09:35 boris-42 bkopilov: or json(
09:35 boris-42 bkopilov: and there is no such thing like single test that is code
09:36 boris-42 bkopilov: test = scenario + it's args + runner + context + sla
09:36 boris-42 bkopilov:  https://github.com/stackforge/rally/blob/master/rally-jobs/rally.yaml
09:36 boris-42 bkopilov: we have such large yamls that we are running
09:36 boris-42 bkopilov: against all patches in gating
09:37 tosky joined #openstack-rally
09:37 bkopilov boris-42, ok  , thanks
09:38 boris-42 bkopilov: so to get familiar with rally
09:38 boris-42 bkopilov: you can read that step-by-step guide
09:38 boris-42 bkopilov: it's quite short and clear
09:38 bkopilov boris-42, thanks!
09:42 pradeep joined #openstack-rally
09:48 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
09:51 openstackgerrit Yair Fried proposed stackforge/rally: [spec] Refactoring scenario utils  https://review.openstack.org/172831
09:52 nkhare joined #openstack-rally
09:52 yfried boris-42: andreykurilin: ^ pls review new version
09:55 e0ne joined #openstack-rally
09:55 msdubov_ joined #openstack-rally
09:59 openstackgerrit Nikita Konovalov proposed stackforge/rally: Wait for status  https://review.openstack.org/172466
09:59 openstackgerrit Nikita Konovalov proposed stackforge/rally: Wait for status  https://review.openstack.org/172466
10:01 psd joined #openstack-rally
10:05 nithyag_ joined #openstack-rally
10:17 e0ne joined #openstack-rally
10:18 cdent joined #openstack-rally
10:20 fhubik joined #openstack-rally
10:21 yfried boris-42: re wiki - I would like that. it's otherwise hard to follow agenda and go over old mtgs
10:21 yfried boris-42: re ping - what do you mean?
10:21 openstackgerrit Marian Krcmarik proposed stackforge/rally: Add to constant and rps runneres limits for maximun Core usage  https://review.openstack.org/164888
10:27 fhubik_afk joined #openstack-rally
10:27 fhubik_afk joined #openstack-rally
10:32 openstackgerrit Yair Fried proposed stackforge/rally: [spec] Refactoring scenario utils  https://review.openstack.org/172831
10:37 e0ne joined #openstack-rally
10:38 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
11:00 boris-42 yfried: so okay
11:00 boris-42 yfried: let's start doing this on wiki
11:27 yfried boris-42: andreykurilin1: https://review.openstack.org/#/c/172831/
11:27 nkhare joined #openstack-rally
11:27 boris-42 yfried: I will review it today a bit later
11:58 andreykurilin joined #openstack-rally
12:00 andreykurilin joined #openstack-rally
12:02 mwagner_lap joined #openstack-rally
12:11 panbalag joined #openstack-rally
12:18 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
12:25 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
12:31 kiran-r Hello! What is project_domain and user_domain?
12:37 boris-42 kiran-r: it's keystone stuff
12:38 kiran-r Could you share significance of it?
12:38 boris-42 kiran-r: ?
12:39 kiran-r Why is the ´default´ loaded here?
12:43 boris-42 kiran-r: hm?
12:45 yfried boris-42: https://review.openstack.org/#/c/172831/ do you understand why this fails Jenkins?
12:45 boris-42 yfried: random bugs
12:45 boris-42 yfried: I believe recheck should help
12:46 yfried boris-42: tnx
12:46 yfried kiran-r: I'm unfamiliar with this. where did you see that?
12:47 boris-42 yfried: I believe you should take a look at keystone docs
12:47 boris-42 yfried: or ask keystone guys
12:48 boris-42 kiran-r: ^
12:48 boris-42 yfried: sry**
12:48 yfried boris-42: np
12:48 boris-42 kiran-r: so actually as far as I know you can create own
12:48 boris-42 kiran-r: domain
12:48 boris-42 kiran-r: and isoloated group of tenants/users inside these domains
12:50 pradeep joined #openstack-rally
13:01 rook_ joined #openstack-rally
13:03 openstackgerrit joined #openstack-rally
13:16 yfried kiran-r: these seem like keystone details. if I understand boris-42's comments (I'm not a keystone guy) seems like you can make the "users" context create users/projects under some kind of "domain"
13:17 yfried kiran-r: I'm guessing. for more details, I ask ://github.com/antonf (did git-blame on these lines)
13:17 boris-42 yfried: it's keystone v3 stuff
13:17 yfried boris-42: kiran-r^
13:17 boris-42 yfried: to isolate group of tenants/users
13:17 yfried boris-42: guessed as much
13:18 yfried boris-42: I usually stop reading after "keystone... stuff" :)
13:19 samueldmq joined #openstack-rally
13:19 yfried pradeep: what do you mean "how to access rally DB"?
13:19 yfried pradeep: are you asking where the code is in the rally tree?
13:24 aix joined #openstack-rally
13:30 andreykurilin joined #openstack-rally
13:33 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
13:34 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
13:35 pradeep yfried: how do i see DB changes while starting rally job
13:37 yfried pradeep: no idea
13:37 * yfried thinking
13:37 yfried pradeep: you could query the DB for changes
13:37 yfried pradeep: rally task start my.json &
13:37 yfried pradeep: rally task details
13:37 yfried pradeep: rally <whatever query you like>
13:38 yfried pradeep: does that help?
13:38 pradeep yfried: yes.
13:38 pradeep let me try
13:43 prashantS joined #openstack-rally
13:44 prashantS joined #openstack-rally
14:00 nkhare joined #openstack-rally
14:00 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
14:01 baker joined #openstack-rally
14:02 pradeep joined #openstack-rally
14:06 nkhare_ joined #openstack-rally
14:11 yfried pradeep: did it work?
14:12 pradeep yfried: nope . thats not what i am looking for
14:12 yfried pradeep: so what are you looking for?
14:13 pradeep yfried: i want to see values updated in variables during transition (while starting any job). For ex: update_at
14:13 yfried pradeep: sorry
14:13 yfried pradeep: you need someone familiar with the rally DB
14:14 yfried pradeep: andreykurilin^
14:20 pradeep yfried: ok. Thanks
14:20 pradeep andreykurilin : ping
14:30 andreykurilin pradeep: pong
14:31 pradeep andreykurilin How do i access rally DB?
14:31 pradeep andreykurilin i want to see values updated in variables during transition (while starting any job)
14:32 andreykurilin pradeep: Do you mean access to rally db in gates?
14:32 pradeep andreykurilin want to see whats being updated in update_at .
14:32 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
14:32 pradeep like this ?
14:33 pradeep is there a place holder or some thing to access rally db
14:33 andreykurilin pradeep: defaul place of rally db: https://github.com/stackforge/rally/blob/master/rally/db/api.py#L50-L51
14:34 pradeep andreykurilin ok
14:34 pradeep Thanks
14:35 andreykurilin also, you can look at "connection = " value in /etc/rally/rally.conf and found correct place of db
14:55 fhubik joined #openstack-rally
14:55 fhubik_afk joined #openstack-rally
15:04 prashantS joined #openstack-rally
15:05 openstackgerrit Andrey Kurilin proposed stackforge/rally: WIP:Add aborted and broken flags to sla checker  https://review.openstack.org/172596
15:05 prashantS joined #openstack-rally
15:06 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
15:08 openstackgerrit Chris St. Pierre proposed stackforge/rally: Add scenario to boot and rebuild server  https://review.openstack.org/172477
15:11 prashantS joined #openstack-rally
15:17 prashantS joined #openstack-rally
15:22 phladky_ joined #openstack-rally
15:38 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
15:46 openstackgerrit Wataru Takase proposed stackforge/rally: [spec] Refactoring Rally cleanup  https://review.openstack.org/172967
16:04 openstackgerrit Andrey Kurilin proposed stackforge/rally: Add aborted and broken flags to sla checker  https://review.openstack.org/172596
16:17 kiran joined #openstack-rally
16:27 kiran-r What are the different types of runners? Serial, constant and what more?
16:28 boris-42 kiran-r: rps
16:40 sputnik13 joined #openstack-rally
16:41 subscope_ joined #openstack-rally
17:18 davideagnello joined #openstack-rally
17:19 kiran joined #openstack-rally
17:20 vkmc joined #openstack-rally
17:20 vkmc joined #openstack-rally
17:37 aix joined #openstack-rally
17:58 prashantS joined #openstack-rally
17:59 prashantS joined #openstack-rally
18:05 baker joined #openstack-rally
18:07 yfried joined #openstack-rally
18:19 sputnik13 joined #openstack-rally
18:27 baker_ joined #openstack-rally
18:32 baker joined #openstack-rally
18:32 rook boris-42: reviewing the vmtask changes : def _boot_server_with_fip(self, image, flavor, is a bit misleading
18:33 boris-42 rook: hm?
18:34 rook https://github.com/stackforge/rally/blob/master/rally/benchmark/scenarios/vm/utils.py#L72
18:34 rook gives the impression the method will boot with a floating ip
18:34 rook IMHO
18:34 rook but it won
18:34 rook however, you could boot w/o fip
18:37 rook just pointing that out.
18:37 rook trying to figure out a different problem right now... not related to Rally -- yet :D
18:39 darkhuy does tox automatically pick up all test_*.py in tests/unit/?
18:40 boris-42 rook: so "f" in fip can mean both
18:40 boris-42 rook: fixed_ip and floating_ip
18:40 rook haha
18:40 rook I like your logic on that one :)
18:40 boris-42 rook: so it seems ok=)
18:40 rook why not just _boot_server then :P
18:43 boris-42 rook: because it is booting with IP
19:04 e0ne joined #openstack-rally
19:17 e0ne joined #openstack-rally
19:48 cdent joined #openstack-rally
19:57 coolsvap joined #openstack-rally
20:05 boris-42 rook: woot https://twitter.com/PavlovicBoris/status/587706418024808449 = 0
20:06 rook Nice! I have been seeing the progress on this. Excellent work!
20:08 boris-42 rook: yep i removed almost all dependencies from it
20:08 boris-42 rook: pbr is the last one
20:08 boris-42 rook: it will be removed soon... I hope
20:09 boris-42 rook: heh it's hard to make lib popular=)
20:10 rook lol yeah, tons of critics out there
20:15 boris-42 rook: lol
20:15 boris-42 rook: nobody cares=)
20:27 rook boris-42: wrt Rally, do we reuse client connections (nova,keystone,etc?)
20:32 boris-42 rook: it depends
20:32 boris-42 rook: if you use cl = Clients(....)
20:32 boris-42 cl.nova()
20:32 boris-42 cl.nova()
20:32 boris-42 will do authentication only once
20:33 rook hm ok
20:35 rook if i change the rally config to have a timeout for the nova-client it will recreate a client if things timeout, correct or will it log that as a error?
20:39 boris-42 rook: what timeouts?
20:39 boris-42 rook: I mean what conf option did you change?
20:40 rook nothing yet.
20:41 e0ne joined #openstack-rally
20:41 boris-42 rook: so timeouts - related to clients means different
20:41 boris-42 rook: it's not reconnect it's how long HTTP reuqest can take
20:42 boris-42 rook: reconnection will be done if token get expired
20:42 boris-42 rook: we merged patch recently
20:42 rook yeah, it seems the default HAProxy configs don't give enough time to Rally, since the Client timeout is None, there really shouldn't be a HAProxy timeout.
20:42 rook so, Rally is getting a BadStatus line from HAProxy.
20:43 rook Hence my questions of timeouts
20:43 rook :)
20:43 rook http://docs.openstack.org/high-availability-guide/content/ha-aa-haproxy.html referencing this config.
20:56 boris-42 rook: by default there are no timeouts
20:56 boris-42 rook: in clients as far as I know
20:56 rook right
20:57 rook so HAProxy is sending back BadStatus to Rally :(
20:57 rook ConnectionError: HTTPConnectionPool(host='20.0.0.41', port=8774): Max retries exceeded with url: /v2/9b0fd278e5aa4bea8e6fca189724be86/servers (Caused by <class 'httplib.BadStatusLine'>: '')
20:57 boris-42 interesting
20:57 boris-42 rook: it sends it always?
20:58 rook oh yeah, 100% reproducible.
20:58 rook once I change the keystone endpoint to remove HAProxy i get much better results
20:58 rook so it is some interaction with HAProxy and Rally right now --- I haven't seen this before.
21:00 rook got to drop
21:00 boris-42 rook: hm
21:00 rook ttyl boris-42!
21:00 rook yeah
21:00 boris-42 rook: maybe you should use explicit ENDPOINT
21:00 boris-42 OS_ENDPOINT
21:00 boris-42 to management netowrk
21:00 rook let me know if you think of something... but this is easily reproducible with upstream HAProxy config.
21:00 rook and remove HAProxy?
21:01 rook I know that works well
21:01 rook but I want to get the HAProxy config right :)
21:01 boris-42 rook: heh
21:09 cdent joined #openstack-rally
21:20 andreykurilin joined #openstack-rally
21:21 aix joined #openstack-rally
21:36 frickler_ joined #openstack-rally
21:47 psd joined #openstack-rally
21:57 jjmb joined #openstack-rally
22:02 prashantS joined #openstack-rally
22:10 baker joined #openstack-rally
22:12 baker joined #openstack-rally
22:20 psd joined #openstack-rally
22:30 rook boris-42: wonder if what I am seeing could be a issue with Rally?
22:31 rook leaving a connection open and not using it.
22:31 rook then attempting to use it.
22:31 rook wonder if I should attempt to set the timeouts to see if that mitigates some of these issues.
22:32 boris-42 rook: I am not sure
22:32 boris-42 that it can mitagate
22:34 rook Have you ever ran against a OpenStack env with HAProxy?
22:35 boris-42 rook: personally I did at the begging=)
22:35 boris-42 rook: but I can say you that it is run already for I think 4-5 months
22:36 boris-42 rook: on daily basis against HA MOS
22:36 boris-42 rook: multple times
22:36 boris-42 rook: like they run 6k times
22:36 rook I will say, i only see this when trying to launch 80 guests concurrently
22:36 boris-42 or even more
22:36 rook anything <80 guests it is just fine.
22:36 rook i am mimicking a bootstorm.
22:36 boris-42 rook: it may be that keystone is dying
22:36 rook nope, keystone is fine.
22:37 rook like I said, when i remove haproxy from the mix, all is good.
22:37 boris-42 are you sure?)
22:37 rook yup
22:37 boris-42 okay
22:37 rook 80 concurent guests with HAProxy, 140 without.
22:38 boris-42 rook: hm very strange
22:38 rook agreed!
22:38 rook frustrating!
22:44 boris-42 rook: hm very strange
22:49 nkhare joined #openstack-rally
22:49 nkhare_ joined #openstack-rally
23:29 baker joined #openstack-rally
23:35 mwagner_lap joined #openstack-rally

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